Social Work 3401 Coursebook

Social Work 3401 Coursebook

AUDREY BEGUN
Social Work 3401 Coursebook by Audrey Begun is licensed under a Creative Commons
Attribution-NonCommercial-NoDerivatives 4.0 International License, except where
otherwise noted.
Contents

Preface & Orientation 1

Module 1

Module 1 Introduction 5
Module 1 Chapter 1: Social Work and "Ways of Knowing" 7
Module 1 Chapter 2: Research and the Professional Ethics 55
Mandate
Module 1 Chapter 3: Ethical Conduct of Research 63
Module 1 Summary 70
Module 1 Key Terms and Definitions 72
Module 1 References 74

Module 2

Module 2 Introduction 83
Module 2 Chapter 1: The Nature of Social Work Research 84
Questions
Module 2 Chapter 2: The Link Between Theory, Research, 116
and Social Justice
Module 2 Chapter 3: What is Empirical Literature & Where 129
can it be Found?
Module 2 Chapter 4: Reviewing Empirical Articles 150
Module 2 Module Summary 161
Module 2 Key Terms and Definitions 162
Module 2 References 166

Module 3

Module 3 Introduction 175

Module 3 Chapter 1: From Research Questions to Research 177
Approaches
Module 3 Chapter 2: Overview of Qualitative Traditions 189
Module 3 Chapter 3: Overview of Quantitative Traditions 206
Module 3 Chapter 4: Overview of Quantitative Study 228
Variables
Module 3 Chapter 5: Overview of Methods for Data 249
Collection and Measurement
Module 3 Chapter 6: Study Participants 293
Module 3 Summary 315
Module 3 Key Terms and Definitions 316
Module 3 References 324

Module 4

Module 4 Introduction 335

Module 4 Chapter 1: Working With Qualitative Data 337
Module 4 Chapter 2: Working With Quantitative 350
Descriptive Data
Module 4 Chapter 3: Univariate Inferential Statistical 387
Analysis
Module 4 Chapter 4: Principles Underlying Inferential 398
Statistical Analysis
Module 4 Chapter 5: Introduction to 5 Statistical Analysis 425
Approaches
Module 4 Summary 457
Module 4 Key Terms and Definitions 459
Module 4 References 466

Module 5

Module 5 Introduction 471

Module 5 Chapter 1: Parts of an Empirical Manuscript or 473
Report
Module 5 Chapter 2: Creating Strong Presentations 487
Module 5 Chapter 3: Creating Figures and Graphs 496
Module 5 Chapter 4: Creating Infographics 517
Module 5 Chapter 5: Course Conclusions 524
Module 5 Key Terms and Definitions 529
Module 5 References 533
Preface & Orientation
Welcome to the online coursebook for SWK 3401 Research & Statistics
for Understanding Social Work Problems and Diverse
Populations course. There are several tools built into these chapters,
designed to enrich and guide your learning.

• Material may be reviewed in three different formats: online

(EPUB), downloaded (PDF), or as Mobilepocket (MOBI) format,
which is better suited to small-screen devices.
• Embedded interactive “Stop and Think” exercises require
internet connectivity, but each can also be downloaded for
offline work. These exercises are presented to help you apply
what you are reading, challenge yourself, prepare for quizzes,
and have a little fun along the way.
• A list of key terms is presented at the end of the module
coursebook to explain text appearing as highlighted in bold
italics throughout the coursebook—in the interactive mode you
can click on a highlighted word to jump to its explanation in the
key terms section. Use the back arrow to return to where you
were reading.
• “Back to Basics” textboxes appear in some chapters of some
modules. These are designed as refreshers of some concepts
that you may have learned in the past and that have applicability
to the present material.
• Green highlighted boxes appearing in some chapters/modules
refer you to learning activities in your online interactive Excel
Workbook. These learning activities relate to content
introduced in your main coursebook. The intent is for you to
complete each Excel Workbook activity as you encounter it,
then return to the coursebook.

Preface & Orientation | 1

Should you find it necessary to cite portions of the coursebook,
here is the recommended citation: Begun, A.L. (2018). SWK 3401
online coursebook: Research & statistics for understanding social work
problems and diverse populations. The Ohio State University, Open
Educational Resources.

2 | Preface & Orientation

MODULE 1

Module 1 | 3
Module 1 Introduction
Our first module allows us to become oriented to the importance of
research and evidence in social work, as well as ethical practices in
social work research. Exploring these topics prepares you to become
engaged in locating, analyzing, and using evidence to inform your
social work practice and to become engaged in social work research
efforts, as well. Additionally, we borrowed a quote from Bannon et al.
(2015) to inform development of the coursebook materials:

“The class operates upon the belief that learning is a product

of performing quantitative analysis, not just reading about the
process” (p. 3).

Thus, throughout our online coursebooks we work with files and

learning activities involving the Excel software program. These are
noted in green boxes throughout the coursebooks; each green box
refers you to a specific location in the online, interactive Excel
Workbook.

Reading Objectives

After engaging with the Module 1 reading materials and learning

resources, you should be able to:

• Recognize and analyze different types of knowledge that inform

social workers’ understanding of social work problems, diverse
populations, and social phenomena;
• Identify principles of critical thinking;
• Distinguish between evidence from science, emerging science,
pseudoscience, and opinion;
• Describethe roles of research in social work and in promoting

Module 1 Introduction | 5
 social justice;
• Identify the philosophical roots of different research
methodologies that inform social workers’ understanding of
social work problems, diverse populations, and social
phenomena;
• Explain important relationships between theory, research, and
practice in understanding diverse populations, social work
problems, and social phenomena;
• Analyze important human subject research principles and the
ethical conduct of research;
• Locate sources of empirical evidence;
• Define key terms related to social work research and research
ethics

6 | Module 1 Introduction
Module 1 Chapter 1: Social
Work and "Ways of Knowing"
This chapter is largely about social work knowledge—how social work
professionals come to “know” what we know, and how that
knowledge can be leveraged to inform practice. Thinking about what
we know and how we come to know it is critically important to
understanding social work values, beliefs, and practices. This means
critically thinking about the sources and types of knowledge on
which social workers rely, and the implications of relying on these
different sources and types of knowledge.

What you will learn from reading this chapter is:

• The philosophical roots of different approaches to scientific

knowledge on which social workers rely (epistemologies);
• Different types and sources of knowledge (“ways of knowing”);
• Strengths and limitations of various “ways of knowing” social
workers encounter and might apply in understanding social
problems and diverse populations;
• Principles of critical thinking;
• Distinctions between science, pseudoscience, and opinions.

Thinking about Knowledge

The study of knowledge and knowing about the world around us

called epistemology, and represents one of the major branches of
philosophy. Throughout much of human history, philosophers have
dedicated a great deal of thought to understanding knowledge and
its role in the human experience. Ancient Greek and western
philosophers like Socrates, Plato, Aristotle, as well as those who came

Module 1 Chapter 1: Social Work and

"Ways of Knowing" | 7
before and after them, made lasting contributions to the way we
think about knowledge and its applications in daily life. A great deal
of this western philosophical activity centered around building
knowledge based on facts and provable “truths,” rather than
spirituality, mythology, and religion. These philosophical efforts
contributed greatly to the evolution of rational thought, science,
theory, and scientific methods that we use in our everyday and
professional lives to make sense of the world around us. Much of
the science on which we often rely to find answers to perplexing
questions is derived from a philosophical position called positivism.
Positivism essentially involves adopting a stance where valid
knowledge, or “truth,” is based on systematic scientific evidence and
proof: in order to be positive about knowing something, that
something must be proven through scientific evidence. Scientific
evidence, developed through a positivism lens, results from a
rational, logically planned process identified as the scientific process.
Similarly, empiricism is about proven knowledge, but contends that
proof also may come through the application of logic or through
direct observable evidence. In our SWK 3401 and SWK 3402
coursework you will be exposed to many of the methods that
investigators apply in scientifically answering questions about social
problems, diverse populations, social phenomena, and social work
interventions. But science is not the only way of knowing, and it
is important for social workers to understand the place science
occupies in an array of ways knowledge is developed and applied.

Much of what is taught in the United States about knowledge and

epistemology is somewhat constrained by traditions of western
philosophy. Challenges to these constraints emerged in the literature
during the 1990s under the headings of naturalistic epistemology,
anthroepistemology, and ethnoepistemology (e.g., Maffie, 1990; 1995).
An important contribution to our understanding of knowledge is an
anthropological appreciation that knowledge is constructed within a

8 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

cultural context. This is quite different from the positivist perspective
concerning single, provable truths that are waiting to be discovered.

“Anthroepistemology views epistemology as an historically and

contingently constituted phenomenon, the nature, aims, and
province of which are to be understood in terms of the life
context in which epistemology is organically rooted and
sustained rather than in terms of divine imperative, rationality
per se, or pre-existing epistemic facts or principles” (Maffie,
1995, p. 223).

In other words, what we know, how we come to know it, and how
we think about knowing all are influenced by the personal, historical,
and cultural contexts surrounding our experiences. For example,
consider what you “know” about deafness as a disability. This
deficiency perspective comes from the cultural context of having
lived in a hearing world. However, members of the Deaf community
offer a different perspective: living within deaf culture and linguistic
structures (e.g., using American Sign Language to communicate)
conveys specific social and cultural implications for human
development, behavior, thinking, and worldview (Jones, 2002). These
implications are viewed the same way other cultures are viewed—as
cultural differences when being compared, not as deficiencies, or
“otherness.” This diversity of experience perspective (rather than a
disability perspective) could be applied to other topics, such as
autism, reflecting human neurodiversity rather than a disorder.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 9

Implications for Social Work.

One implication of this observation might be that social workers

should expect worldviews held by individuals with vastly different
lived experiences to differ, too. Well over 100 years ago, the
philosopher and psychologist William James (1902), discussing the
varieties of religious experience, observed:

“Does it not appear as if one who lived more habitually on one

side of the pain-threshold might need a different sort of religion
from one who habitually lived on the other?” (p. 105).

In broader terms, social workers strive to understand diversity in its

many forms. We appreciate that diverse life contexts, experiences,
opportunities, and biology all interact in complex ways in
contributing to diverse outcomes. These differences include
differences in knowledge and understanding of the
world—differences, not deficiencies. For example, we understand
that the elements of a parenting education program delivered to two-
parent, well-resourced, privileged, reasonably empowered families
of one racial or ethnic background may be grossly inappropriate

10 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

for families existing in a far different reality, facing very different
challenges, very differently resourced or privileged, and responding
to different experiences of discrimination, oppression, exploitation,
threats of violence, and micro-aggression. The pre-existing
knowledge different families bring to their parenting situations and
their parenting knowledge needs differ significantly—even their
parenting goals, approaches, and means of learning and developing
knowledge may differ.

Social workers who adopt an anthropological, ethnoepistemology

perspective are open to considering the beliefs of ordinary people
around the world alongside those of leaders, academics, scientists,
colleagues, and authorities. This line of thought encourages us to
reflect on all epistemologies—wherever in the world they are
practiced and by whomever they are practiced. Western philosophy,
rational logic, and science are situated within this context—as one of
the multitude of epistemologies that exist, not as the first or the most
significant, but simply as one of many. These perspectives contribute
to social work having a rich array of research methodologies available
for understanding social problems, diverse populations, and social
phenomena. A positivism/empiricism perspective contributes to
many of our quantitative research methodologies; an

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 11

anthropological, ethnoepistemology perspective underlies many
qualitative research methods.

This philosophical background is relevant to social work education

for several reasons:

1. Social work professionals engage with individuals from many

diverse backgrounds and social contexts. This means that we
encounter many different ways of knowing and great diversity
of belief among the people with whom we interact. Colleagues,
professionals from other disciplines, clients, community
members, agencies, policy decision makers and others all have
their own understanding about the world we live in, and these
understandings often differ in critically significant ways. We
cannot work effectively with others if we do not have an
appreciation for this diversity of understanding, thinking, and
believing.
2. To be effective in our interactions, social workers also need to
understand and critically evaluate our own personal and
professional epistemologies, and what we “know” about social
work problems, diverse populations, and social phenomena. This
idea fits into the social work practice mandate to “know
yourself” (e.g., see Birkenmaier & Berg-Wegner, 2017).
3. Significant differences exist between the philosophies that
underlie different scientific methodologies. Rational logic
underlies quantitative methodologies and ethnoepistemology
underlies on qualitative methodologies. These differences
contribute to complexity around the qualitative, quantitative,
and mixed methods (integration of qualitative and quantitative)
choices made by investigators whose work helps us understand
social work problems, diverse populations, and social
phenomena.

12 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

Thinking about Different Sources and Types of Knowledge

Humans have many different ways of developing our knowledge

about the world around us. Think about how young children come to
understand the world and all kinds of phenomena they experience.
First, they utilize all five senses to explore the world: vision, hearing,
smell, taste, and touch. For example, a baby might hear food being
prepared, see and smell the food, touch it and taste it.

Sensory evidence provides people of all ages with a great deal of

experiential knowledge about the physical and social world.

Second, people engage in internal mental operations we call

cognition—they engage in thinking and problem solving—to create
meaning from their experiences. To continue our example, it is
through these internal mental experiences that babies develop their
ideas about food. Some of their conclusions are effective, while their
other conclusions are inaccurate; these guesses need to be revised
through further experience and cognition. For example, this toddler
learned (the hard way) to understand that paint is not food. These
cognitions may take a few trials to accurately emerge.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 13

Third, knowledge comes through basic learning principles involving
reinforcement and punishment of behavior, as well as observational
learning of others’ behaviors. Young children often conduct repeated
experiments to develop knowledge concerning the basic principles
by which the physical and social world operate. For example, a young
child might experiment with using “naughty” swear words under
different conditions. The first few times, it might simply be a case of
copying a role model (such as parents, siblings, peers, or television/
movie/music performers). Based on the way the social world
responds to these experimental uses of language, the child may

14 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

continue to experiment with using these words under different
circumstances. This behavior might get one response from siblings
or other children (laughing, giggling, and “Oooohhhhh!” responses), a
different response from a parent or teacher (a corrective message or
scolding response), and yet a different response still from a harried
caregiver (simply ignoring the behavior). We may consider this child
to be an “organic” scientist, naturally developing a complex
understanding about the way the social world works.

People of all ages rely on these multiple ways of knowing about

the world—experiential, cognitive, and experimental. What additional
ways of knowing are important for social workers to understand?
Answers to this question can help us better understand the diverse
people with whom we engage and can help us better understand our
own sources of knowledge as applied in professional practice. By the
way, thinking about our thinking is called metacognition—having an
awareness of your own processes of thinking and knowing.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 15

Ways of Knowing.

Textbooks about the Theory of Knowledge offer various ways

through which human beings derive knowledge (see for example IB,
n.d.; TOK network, n.d). These ways of knowing include:

• Sensory perception
• Memory
• Language
• Reason
• Emotion
• Faith
• Imagination and Intuition

Let’s examine each of these ways of knowing in a little more detail,

and consider their implications for knowledge in social work. It is
helpful to think about how we might consider our use of each source
of knowledge to avoid running “off into a ditch of wrong conclusions”
(Schick & Vaughn, 2010, p. 4).

Sensory perception and selective attention. We have already

considered the ways that individuals across the lifespan come to
know about the world through engaging their five senses. In general,
human beings rely on information acquired through their senses as
being reliable evidence. While this is generally appropriate, there are
caveats to consider. First, knowledge developed through our sensory
experiences are influenced by the degree to which our sensory and
neurological systems are intact and functional. There exists a great
degree of variability in different individuals’ vision, for example. We
can reduce the variability to some degree with adaptive aides like
eye glasses, contact lenses, adaptive screen readers, and tools for
sensing color by persons who are “color blind.” But, it is important

16 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

to recognize that perception is highly individualistic. You may have
played with optical illusions in the past .Or, you may recall the 2015
social media storm created by the “What color is this dress?” dilemma
(#thedress, #whiteandgold, and #blackandblue). Or, for example,
when you look at this image, what do you see?

You might see a duck head looking to the left, a rabbit head looking
to the right, both, or neither. In both the dress and duck/rabbit
examples, different individuals perceive the same stimuli in different
ways.

This raises the question: can we trust our own perceptions to

represent truth? In our everyday lives, we rely heavily on our
perceptions without thinking about them at all. Occasionally, we
might experience the thought: “I can’t believe my own eyes” or “I can’t
believe what I am hearing.” Between these extremes, social workers
often experience the need to confirm our perceptions and checking
with others about theirs.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 17

One reason why our perceptions might not match the perceptions
of another person has to do with what we each pay attention to.
Everyone practices a degree of selective attention . For example, as
you are reading this material, are you aware of the feel of the room
temperature, the sound of the air circulation, or the amount of light
in the room? This information is all available to you, and you might
have paid attention to them when you first sat down to read. But, you
stopped being consciously aware of them as they remained steady
and irrelevant to the task of reading. If any conditions changed, you
might have noticed them again. But, it is human nature to ignore a
great deal of the information collected by our sensory systems as
being irrelevant. A source of individual difference is the degree to
which they attend to different types of information. For example, as
you ride through a neighborhood you might notice the location of
different food or coffee establishments. Someone in recovery from
an alcohol use disorder, taking the very same ride, might attend to
the different liquor advertisements that could trigger a craving and
warrant avoiding. You both experienced the same environment, but
through selective attention you each came to know it differently.

18 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

Memory and memory bias. You might argue that memory is not really
a way of knowing, that it is only a tool in the process of knowing.
You would be partially correct in this argument—it is a tool in the
process. However, it is not ONLY a tool. Memory is an integral and
dynamic part of the human mind’s information processing system.
The diagram below represents steps in human information
processing. The process begins with the sensory systems discussed
earlier. But, look at what happens to those sensory perceptions next:

In terms of the information processing model, once something has

been perceived or an event has been experienced by an individual
(perception), the perception moves to the memory system. A
perception first moves into short-term memory as something the
person is actively aware of or thinking about. The short-term
memory “buffer” is typically limited in space to around 5-7 items at a
time and lasts for only about 18-30 seconds. After that, the memory
either is cleared from the “buffer” zone and forgotten forever, or it is

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 19

moved to longer term memory. If a memory is not converted into a
long-term memory, it cannot be retrieved later.

The long-term memory system has a tremendous storage capacity

and can store lasting memories. However, stored memories are
relatively dynamic, rather than static, in nature. Memories are
remarkably susceptible to change through the incorporation of new
information. Human memory does not work like a video camera that
can be replayed to show exactly what was originally experienced.
Instead, recalling or retrieving memory for an event or experience
happens through a process of recreation. In this process, other
factors and information become woven into the result, making it
a less-than-perfect depiction of the original experience. Consider
research about the ways that additional (mis)information affects what
individuals recalled about a story presented to them in a memory
study (Wagner & Skowronski, 2017). Study participants reported
“remembering” new facts presented later as having been part of the
original story—these later facts became false memories of the original
experience. False or erroneous memories were more likely when
the original facts were scarce (versus robust), the new information
seemed true (rather than questionably reliable), and the person was
highly motivated to retell the story. These conditions provoked the
human mind to fill in gaps with logical information.

The nature of a memory also may be (re)shaped by internally

generated cognition based on our pre-existing understanding of the
world—our thoughts. This process makes memories more relevant to
the individual person but is a significant source of memory bias. We
often use our own thoughts, ideas, cognitions, and other experiences
to fill in knowledge gaps around what we remember from direct
experiences. For example, you may have learned about the poor
reliability of eyewitness testimony—what is remembered by the
eyewitness is influenced by how the situation is understood, making
it susceptible to distortion from beliefs, stereotypes, prejudice, and

20 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

past experiences. This explains why convenience stores often have
rulers painted on the doorframe: if a crime occurs, a shorter witness
might recall a person’s height as being tall, and a taller witness might
recall the person’s height as being short. The yardstick allows more
objective, undistorted recall of the facts.

Memory reshapes itself over time, and sometimes people lose the
connections needed to retrieve a memory. As a result, what we
“think” we know and remember is influenced by other external and
internal information.

Language. Like memory, language is both a product and producer

of knowledge. Language is a created set of signs and symbols with
meaning that exists because of conventional agreement on that
meaning. Language helps describe what we know about the world,
both for ourselves and to communicate with others. But, language
also has the power to shape our understanding, as well. Linguistic

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 21

determinism is about how the structures of language constrain or
place boundaries on human knowledge.

Consider, for example, the limitations imposed on our understanding

of the social world when language was limited to the binary
categories of male and female used in describing gender. Our
language being structured in this binary fashion leads individuals
to think about gender only in this manner. Conventional agreement
is extending to include terms like agender, androgynous, bigender,
binary, cisgender, gender expansive, gender fluid, gender non-
conforming, gender questioning, genderqueer, misgender, non-
binary, passing, queer, transgender, transitioning, transsexual
person, two-spirit (see Adams, 2017). With the addition of multiple
descriptive terms to our language, it becomes easier to have a more
contextual, complex, non-binary understanding of the diverse nature
of gender. Language is one mechanism by which culture has a
powerful influence on knowledge.

Reason and logic. Humans may come to “know” something through

their own internal cognitive processing and introspection about it.
Logic and reasoning represent thought processes that can result in

22 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

knowledge. You may recall the important piece of reasoned logic
offered by Rene Descartes (1637): “I think therefore I am.” This piece
of reasoning was the first assumption underlying his building of an
orderly system of “truth” where one conclusion could be deduced
from other conclusions: deductive reasoning. In deductive reasoning,
we develop a specific conclusion out of assembling a set of general,
truthful principles (see Table 1-1). The specific conclusion is only as
good as the general principles on which it is based—if any of the
underlying premises are false, the conclusion may be incorrect, as
well. The conclusion might also be incorrect if the premises are
assembled incorrectly. On the other hand, inductive
reasoning involves assembling a set of specific observations or
premises into a general conclusion or generalized principle (see Table
1-1). Inductive reasoning is a case of applying cause and effect logic
where a conclusion is based on a series of observations. For example,
estimating future population numbers on the basis of the trend
observed from a series of past population numbers (Kahn Academy
example from www.youtube.com/
watch?time_continue=43&v=GEId0GonOZM). Abductive reasoning is
not entirely inductive or deductive in nature. It generates a
hypothesis based on a set of incomplete observations (inductive
reasoning) and then proceeds to examine that hypothesis through
deductive logic.

Table 1. Examples of deductive, inductive, and abductive reasoning

from Live Science https://www.livescience.com/21569-deduction-
vs-induction.html

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 23

Deductive Reasoning
Inductive Reasoning Example
Example
Spiders have 8 legs.
A black widow is a Pulling a series of coins from a coin purse, 1st,
spider. 2nd, 3rd, and 4th is each a penny.
Conclusion: black Conclusion: the 5th will be a penny, too.
widows have 8 legs

Abductive Reasoning Example

Entering your room, you find that your loose papers are all over the place.
You see that the window is open.
Conclusion: the wind blew the papers around.
But, you consider alternative explanations (another person came in and
threw the papers around) and settle on the most probable/likely
conclusion.

A person does not actually need to witness or experience an event to

draw logically reasoned conclusions about what is likely. For example,
a social worker may be trained in the knowledge that persons with
traumatic brain injury (TBI) in their histories have a greater likelihood
than the general population to experience problems in family
relationships, unemployment, social isolation, homelessness,
incarceration, depression, and suicide risk (Mantell, et al, 2017). In
meeting a client with a TBI history, the social worker might reason
that it is important to screen for the risk of each of these problems.
The social worker does not need to wait for these problems to be
observed; the knowledge gained through reasoning and logic is
sufficient to warrant paying attention to the possibility of their being
present or emerging.

Sometimes a line of reasoning seems rational and logical, but when

carefully analyzed, you see that it is not logically valid—it is fallible,
or capable of being in error. A fallacious argument is an unsound
argument characterized by faulty reasoning. Fallacious arguments
or flawed logic can result in the development of mistaken beliefs,
misconceptions, and erroneous knowledge. For example, in the 1940s

24 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

to 1950s a mistaken theory of autism was offered in the professional
literature (see Rimland, 1964): the “refrigerator mother” theory was
that the low levels of warmth parents were observed directing to an
autistic child caused the child’s autism symptoms. The logic flaw in
this case was a very common error: drawing a false causal conclusion
from a correlational observation. The observed parental behavior
actually was a sensitive response to behavioral cues presented by
the children with autism, not a cause of the autism. Thus, the
“refrigerator mother” misconception was influenced by logical
fallacy. Clearly, the knowledge gained through reasoning is only as
good as the weakest link in the chain of ideas and logic applied.

Emotion and affect. Human feelings have a powerful influence on

what we think and understand about the world around us. Another
word for emotion is affect (not to be confused with effect,
pronounced with an “a as in apple”). For example, the advertising,
entertainment, and news media have long recognized that individuals
more strongly remember information to which they have developed
an emotional response. This explains why we see so many cute
puppies, adorable children, attempts at humor, and dramatic disaster

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 25

footage in advertising and news programming. This also is why we are
more likely to remember the score of a game our favorite team played
than the scores of other games. It is why we remember news events
that “hit close to home” compared to other reported world events.

The strongest emotions we experience as human beings may have

developed as ways to improve our individual chances for survival. For
example, by avoiding experiences that elicit fear we may be avoiding
danger and risk of harm. By avoiding that which we find disgusting,
we may be avoiding that which could make us ill. A lot of questions
surround the emotional world—the origins of emotions and what is
innate versus learned, for example. One thing that research since
the 1970s concerning affective education has demonstrated is that
how we interpret our own and others’ emotions (affect) has a strong
influence on how we behave in different situations (e.g., Baskin &
Hess, 1980). Because the ability to identify and label emotions
(knowledge of your own and those of other people) has such a
significant impact on interpersonal interactions and behaviors, a
great deal of attention has been directed towards trying to
understand individual differences in “emotional intelligence” or “EQ”
(see Goleman, 1995).

26 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

Faith. Religious faith is only one aspect of how faith relates to
knowledge. For some individuals, religious teachings provide them
with knowledge surrounding certain topics. What a person knows
through faith is considered to require no further analysis or
proof—that is the nature of faith. For this reason, we can include other
sources of knowledge derived through faith. This includes relying on
knowledge attained through information shared by a mentor, expert,
or authority. Identifying someone as an authority is intensely
personal and individualized. Just as “beauty is in the eye of the
beholder,” perceptions of another person’s expertise are relative. One
person might have faith in the expert opinion shared by a parent,
grandparent, sibling, or best friend. Another person might place

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 27

greater faith in the opinions expressed by celebrities, sports figures,
news anchors, or political leaders. Social work interns may place faith
in the expertise shared by field instructors and course instructors.
Social work clients often rely on the knowledge shared by others in
their support or treatment group who may have experienced similar
challenges in life.

Knowledge derived from experts and authority is not foolproof,

however. First, there may be differences of opinion expressed by
different experts. For example, legal counsel on different sides of
a court case often engage expert witnesses who have diametrically
opposed expert opinions. Jury members are then faced with the
challenge of evaluating the experts’ qualifications to decide which
opinion is more reliable. Second, the knowledge held by experts and
authorities is subject to the same sources of bias and limitation as any
knowledge. For example, the expertise of social work supervisors was
(at least partially) influenced by their supervisors’ expertise, which
was influenced by their supervisors’ expertise, and so on. What if
the original was wrong from the start? Or, more likely, what if that

28 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

knowledge no longer applies to the current environment? In short,
faith in the knowledge of authority needs to be tempered by
awareness that knowledge from these sources is fallible.

“Arguments from authority carry little weight—‘authorities’

have made mistakes in the past and will do so again” (Sagan,
1987, p. 12).

Imagination and Intuition.Sometimes what a person “knows” is not

based on external events or objects, knowledge might be generated
as new ideas or concepts completely through internal processes.
Imagination, creativity, and intuition are special forms of cognitive
activity, resulting in the creation of unique forms of knowledge. You
have probably been exposed to examples of highly imaginative,
creative art, music, film, dance, or literature. Sometimes the original
insights we gain from these creations make them seem like sheer
genius. We might say the same thing about a person’s apparent
cognitive leaps that we might call intuition: seeming to know
something without being able to trace where the knowledge was
derived; it comes from reasoning that happens at an unconscious
level. These ways of knowing combine elements from other forms
previously discussed, but in different ways: perception, memory,
emotion, language, and reasoning, for example. This is what leads to
that “aha” moment when a new idea is born.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 29

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=22

Culture and Ways of Knowing.

Earlier, you read that language is one way that culture has a powerful

30 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

influence on knowledge. The influence of culture is not only in terms
of what we know, but also in terms of how we learn what we know.
The book Women’s Ways of Knowing: The Development of Self, Voice,
and Mind(Belenky, Clinchy, Goldberger, & Tarule, 1986) describes the
results of a deep qualitative study of how women develop their views
and knowledge. The women’s development of knowledge was heavily
influenced by their cultural experiences with authority, particularly
in learning from family and schools. The authors included an analysis
of five different ways of knowing, outlined below. These may very
well express others’ ways of knowing, based on their own experiences
with majority culture, not only based on gender.

Silence. As a way of knowing, silence is an expression of “blind”

obedience to authority as a means of avoiding trouble or punishment.
The concept of silence refers to an absence of self-expression and
personal voice. The authors indicated that individuals characterized
by silence also found the use of words to be dangerous and that
words were often used as weapons.

Received Knowledge. Carriers of this way of knowing rely on others

for knowledge, even knowledge about themselves. They tend to learn
by listening to others, lacking confidence in their own original ideas
or thoughts. There is a heavy reliance on authority as a source of
knowledge. As a result, thinking is characterized by an inability to
tolerate ambiguity, and is concretely dualistic—everything is
categorized as good or evil, us or them, and black or white with no
shades in between.

Subjective Knowledge. Subjective knowledge is highly individualistic

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 31

because personal experience, “inward listening” and intuition are
major sources of knowledge. A sense of voice exists, expressed as
personal opinion. The source of knowledge is internal, and the
authority listened to is self—and this knowledge is more important
than knowledge from external sources.

Procedural Knowledge. Objective approaches are applied to the

process of acquiring, developing, and communicating procedural
knowledge. In other words, careful observation and critical analysis
are required. Knowledge can come from multiple external sources,
but only after careful internal critical analysis of the arguments
provided by those sources. There exists awareness that others and
one’s self can be wrong, and an effort to remove the impact of feelings
from the process of objective analysis.

Constructed Knowledge. This is an integrative view where both

subjective and objective strategies apply to knowledge. Ambiguity is
tolerated well, as are apparent contradictions in what one knows.
There exists a strong, constantly developing sense of self in relation
to the external world. Constructed knowledge involves belief that
knowledge is relative to context and frame of reference, and a person
is responsible for actively trying to understand by “examining,
questioning, and developing the systems that they will use for
constructing knowledge” (see https://www.colorado.edu/UCB/
AcademicAffairs/ftep/publications/documents/
WomensWaysofKnowing.pdf).

Thinking about Critical Thinking

As we have seen, much of what humans “know” is susceptible to

bias and distortion, often represents only partial or incomplete

32 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

knowledge, and socially constructed knowledge varies as a function
of the different societies constructing it. The concept of critical
thinking describes approaches to knowledge and knowing that rely
on disciplined, logical analysis, and knowledge that is informed by
evidence. Critical thinking extends beyond simple memorization of
facts and information; it involves the analysis and evaluation of
information, leading to reasoned, “thought-out” judgments and
conclusions. You might recognize elements of critical thinking in the
procedural and constructed ways of knowing previously described.
A person engaged in critical thinking expresses curiosity about the
topic or issue under consideration, seeks information and evidence
related to the topic or issue, is open to new ideas, has a “healthy
questioning attitude” about new information, and is humbly able to
admit that a previously held opinion or idea was wrong when faced
with new, contradictory information or evidence (DeLecce, n.d.).

Carl Sagan (1987) argued that critical thinking is, “the ability to
construct, and to understand, a reasoned argument and—especially
important—to recognize a fallacious or fraudulent argument” (p. 12).
To understand what a fallacious argument might be, we can examine
the concepts of logic: a branch of philosophy dedicated to developing
the principles of rational thought. Sagan’s (1987) discussion of critical
thinking indicated that it does not matter how much we might like
the conclusions drawn from a train of reasoning; what matters is
the extent to which the chain of logic “works.” He also contended
that untestable propositions are worthless—if you cannot check out
the assertions, it is not worthwhile considering. However, Sagan also
reminded readers that critical thinking and science itself exist in
a state of constant tension between two apparently contradictory
attitudes:

“an openness to new ideas, no matter how bizarre or

counterintuitive they may be, and the most ruthless skeptical

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 33

 scrutiny of all ideas, old and new. This is how deep truths are
winnowed from deep nonsense” (Sagan, 1987, p. 12).

This quote describes the quality of a critical thinker mentioned

earlier: simultaneously being open-minded and having a reasonable
degree of skepticism about new information or ideas.

In terms of social work and social work education, being able to apply
critical thinking skills, particularly “the principles of logic, scientific
inquiry, and reasoned discernment” has been a curriculum
expectation since the 1990s (Mathias, 2015, p. 457). The
characteristics of critical thinking in social work include rational,
reasoned thinking about complex, “fuzzy” problems that lack at least
some elements of relevant information and readily apparent solutions
(Mathias, 2015; Milner & Wolfer, 2014). Critical thinking in social work
practice involves initially suspending judgment, then engaging in the
process of generating relevant questions, considering assumptions,
seeking information and divergent viewpoints, and applying logical,
creative problem solving.

What is (Scientific) Evidence?

Strong arguments appear in the published literature concerning the

importance of scientific reasoning in social work practice, and its
significance to critical thinking and problem solving (e.g., see
Gambrill, 1997; Gibbs & Gambrill, 1999; Gibbs et al., 1995). Francis
Bacon is credited with presenting an approach to scientific method
that is a foundation for science today (Dick, 1955). The Merriam-
Webster definition of scientific method (https://www.merriam-
webster.com/dictionary/scientific%20method) is:

Principles and procedures for the systematic pursuit of knowledge

involving the recognition and formulation of a problem, the collection

34 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

of data though observation and experiment, and the formulation and
testing of hypotheses.

What you may find interesting is that different authors suggest

different numbers of steps in the process called scientific
method—between four and ten! Let’s look at what these descriptions
have in common:

Step 1. It all begins with an observation about something that arouses

curiosity. For example, a social work investigator and her colleague
observed an emerging population of people who rely on emergency
food pantry support for relatively long periods of time—despite the
programs’ intended period of emergency support being relatively
short-term (Kaiser & Cafer, 2016).

Step 2. The observation leads the curious social worker to develop a

specific research question. In our example, the investigators asked
what might be “the differences, particularly in food security status
and use of federal support programs, between traditional, short-
term emergency pantry users and an emerging population of long-
term users” (Kaiser & Cafer, 2016, p. 46).

Step 3. Out of their questions, the social work scientist develops

hypotheses. This may come from a review of existing knowledge
about the topic—knowledge that is relevant but may not be
sufficient to directly answer the question(s) developed in step 2. In
our food security example, the investigators reviewed information
from literature, community practice, and community sources to
help identify key dimensions to explore. The hypothesis was,

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 35

 essentially, that exploring three factors would be informative:
longevity, regularity, and trends of food pantry use.

Step 4. The social work scientist develops a systematic experiment

or other means of systematically collecting data to answer the
research question(s) or test the research hypotheses from steps
2 and 3. Kaiser and Cafer (2016) described their research
methodology for randomly selecting their sample of study
participants, their research variables and measurement tools, and
their data collection procedures.

Step 5. The social worker will analyze the collected data and draw
conclusions which answer the research question. This analysis not
only includes the results of the study, but also an assessment of
the study’s limitations of the methods used for the knowledge
developed and the implications of the study findings. Our working
example from the food security study described the approaches
to data analysis that were used. They found that almost 67% of
3,691 food pantry users were long-term users, most commonly
female and white. Being a persistent user (versus short-term user)
was predicted by having Social Security, retirement, pension, or
SSI/disability payment income. Also, the probability of being a
persistent pantry user increased with participant age. The authors
also discussed other patterns in use as related to some of the other
study variables and the implications related to food security and
economic vulnerability (Kaiser & Cafer, 2016).

Step 6. Social work scientists share their resulting knowledge with

colleagues and others through professional presentations, training

36 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

 sessions, and professional publications. As noted in the reference
list to this module, the investigators in our study example
published their findings in a professional journal (see Kaiser &
Cafer, 2016).

Step 7. The social work scientist then may ask new, related questions,
and progress through the stages anew. In our food security
example, the authors concluded with a new question: would raising
minimum wage or establishing a living wage be a better means
of supporting families in the precarious position of long-term
dependence on emergency food pantry support?

As you can see, the scientific process potentially incorporates

multiple sources of knowledge that we have discussed in this module
and allows for inclusion of various methods of investigation
depending on the nature of the research questions being asked. The
result of quality science, regardless of methodology, is called
research or scientific evidence.

In an interview where he was asked if religion and science could ever

be reconciled, Stephen Hawking responded:

“There is a fundamental difference between religion, which is

based on authority, [and] science, which is based on observation
and reason. Science will win because it works.”

It may not be necessary to continue pitting these two ways of

knowing against each other. Engaging in science involves multiple
ways of knowing. Scientific activity requires a certain degree of
imagination about which might be true, rational reasoning about how
to demonstrate what might be true, perception and observation of
what is happening, memory about what has happened in the past,

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 37

language to communicate about the process and results, an
emotional investment in what is being explored, (sometimes) a
certain degree of intuition, and faith in the scientific process and
what has come before. Acknowledging the collaborative,
constructive, dynamic nature of scientific knowledge development,
consider a quote Sir Isaac Newton borrowed from an earlier popular
saying:

“If I have seen further, it is by standing on the shoulders of

giants.”

This does not mean always having noncritical faith in the results of
science, however. Scientific results are wrong sometimes. Sometimes
scientific results are inconsistent. Sometimes scientific results are
difficult to interpret. For example, are eating eggs bad for you? In the
early 1970s, a link between cholesterol levels in the body and heart
disease was detected. The advice coming from that observation was
that people should limit their consumption of cholesterol-containing
foods (like eggs) to prevent heart disease. However, this conclusion
was based on faulty logic:eatingcholesterol was not the cause of the
elevated cholesterol levels; cholesterol was being produced by the
body. Further research indicated no increased risk of heart disease
with eating eggs; the elevated cholesterol problem stems from other
unhealthy behavior patterns. This example demonstrates the serial,
sequential, accumulating nature of scientific knowledge where
previous assumptions and conclusions are questioned and tested
in further research. We need to accept a degree of uncertainty as
complex problems are unraveled, because science is imperfect. Again,
quoting Carl Sagan (1987):

“Of course, scientists make mistakes in trying to understand the

world, but there is a built-in error-correcting mechanism: The
collective enterprise of creative thinking and skeptical thinking
together keeps the field on track” (p. 12).

38 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

Critically important is that science be conducted and shared with
integrity. Research integrity concerns conducting all aspects of
research with honesty, fairness, and accuracy. This includes
objectively examining data, being guided by results rather than
preconceived notions, and accurately reporting results and
implications. Research integrity, or “good” science means that a
study’s methods and results can be objectively evaluated by others,
and the results can be replicated when a study is repeated by others.
Science nerds enjoy the playfulness presented in the scientific humor
magazine called The Journal of Irreproducible Results (www.jir.com).

Take a moment to complete the

following activity. For each statement,
decide if it is accurate (True) or not
(False).

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=22

Imperfect Science versus Pseudoscience. Even “good” science has

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 39

certain limitations, imperfections, and uncertainties. For example,
conclusions drawn from statistical analyses are always based on
probability. In later modules you will learn what it means when a
statistical test has p<.05 and about the 95% confidence interval (CI)
of a statistical value. In practical terms, this means that we cannot be
100% certain that conclusions drawn from any one study’s statistical
results are accurate. While accepting that we should not engage in
unquestioning acceptance even of “good” science, it is even more
important to avoid unquestioning acceptance of information offered
through “poor” science. Especially, it is important to recognize what
is being presented as scientific knowledge when it is not science. By
way of introducing the importance of scientific data and statistics,
Professor Michael Starbird (2006, p. 12) introduced a famous quote by
the author Mark Twain:

“There are three kinds of lies: lies, damned lies, and statistics.”

Professor Starbird’s response was:

“It is easy to lie with statistics, but it is easier to lie without

them.”

Pseudoscience is about information that pretends to be based on

science when it is not. This includes practices that are claimed to
be based on scientific evidence when the “evidence” surrounding
their use was gathered outside of appropriate scientific methods. You
may have learned about “snake oil salesmen” from our country’s early
history—people who fraudulently sold products with (usually secret)
ingredients having unverifiable or questionable benefits for curing
health problems—these products often caused more harm than cure!
The pseudoscientific “evidence” used in marketing these products
often included (unverifiable) expert opinions or anecdotal evidence
from “satisfied” customers. See the historical advertisement about
Wizard Oil and all the ills that it will cure: rheumatism (joint pain),

40 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

neuralgia (nerve pain), toothache, headache, diphtheria, sore throat,
lame back, sprains, bruises, corns, cramps, colic, diarrhea, and all
pain and inflammation. The product was 65% alcohol and other
active ingredients included turpentine, ammonia, capsicum,
camphor, gum camphor, fir oil, sassafras oil, cajeput oil, thyme oil,
according to the Smithsonian National Museum of American History.
Fortunately, Wizard Oil was a liniment, applied topically and not
consumed orally.

What does this look like in present times? We still, to this day, have
people relying on unsubstantiated products and treatment
techniques, much of which can be identified as “quackery.” For
example, 9% of out-of-pocket health care spending in the United
States pays for complementary health approaches (e.g., vitamins,

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 41

dietary supplements, natural product supplements and alternative
practitioners): over $30 billion (Nahin, Barnes, & Stussman, 2016).
Many alternative and complementary health approaches lack a
strong scientific evidence base—that does not necessarily mean that
they offer no health benefits, but it may be that they do not help and
sometimes they cause harm, especially with overuse. Proponents of
these approaches adopt what Gambrill and Gibbs (2017) refer to as
a “casual approach” to evidence, where weak evidence is accepted
equally to strong evidence, and an open mind is equated with being
non-critical.

Three reasons why this matters in social work (and other professions)
are:

• There may be harmful “ingredients” in these interventions. For

example, an untested treatment approach to helping people
overcome the effects of traumatic experiences might have the
unintended result of their being re-traumatized or further
traumatized.
• Engaging in unproven interventions, even if not containing
harmful “ingredients,” can delay a person receiving critically
needed help of a type that has evidence supporting its use. For
example, failure with an untested approach to treating a
person’s opioid use disorder extends that person’s vulnerability
to the harmful effects of opioid misuse—including potentially
fatal overdose. This is difficult to condone, particularly in light
of emerging evidence about certain combination approaches,
like medications combined with evidence-supported behavioral
therapies (medication assisted treatment, or MAT), having
significant effectiveness for a range of individuals under a range
of circumstances (Banta-Green, 2015).

Treatment failure with untested, ineffectual interventions can

discourage a person from ever seeking help again, including from

42 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

programs that have a strong track record of success. Engaging in
these ineffectual interventions may also deplete a person’s resources
to the point where evidence-supported interventions may not be
affordable.

Despite limitations, pseudoscience does become applied. For

example, the book about “psychomythology” called 50 Great Myths
of Popular Psychology (Lilienfeld, Lynn, Ruscio, & Beyerstein, 2010)
discusses implications of the myth that most humans use only 10%
of their potential brain power. The authors suggest that this myth is
pervasive, in part because it would be great if true, and because it
is exploitable and lucratively marketable through programs, devices,
and practices available for purchase that will help people tap into the
unused 90% reserve.

Learning to Recognize Pseudoscience. In Science and Pseudoscience

in Clinical Psychology, the authors identified ten signs that raise the
probability that we are looking at pseudoscience rather than science
(Lilienfeld, Lynn, & Lohr, 2015). In their book Science and
Pseudoscience in Social Work, Thyer and Pignotti (2015) help explain
these signs. The ten signs are summarized here.

1. Excuses, excuses, excuses.When a strongly held belief fails to be

supported by rigorous empirical evidence, proponents of that
belief may generate excuses to explain away the failure. These
excuses are often unsupported by evidence. Hundreds of
examples are present in recent political arguments aired on
news talk shows. In science, when an expected result is not
observed through application of scientific method, new
hypotheses are generated to explain the unexpected
results—then these new hypotheses are tested through
scientific methods.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 43

2. Absence of self-correction makes the belief grow stronger. In
pseudoscience, erroneous conclusions go untested, so
corrections of theory and conclusions do not happen.
Proponents of an untested, erroneous belief have no impetus to
make corrections and critical questioning may lead them to “dig
in their heels” more strongly. Scientists admittedly do make
mistakes, but errors are eventually eliminated through
application of scientific methods.
3. Conspiracy theories.Proponents of a pseudoscientific belief may
argue that scientific review of their “evidence” is never going to
be favorable because of bias built into the scientific process and
the science community. The approach is a “blame the system”
strategy, as opposed to recognizing the flaws in their work.
Pseudoscience is not likely to survive peer-review scrutiny
because the methodologies, assumptions, and/or conclusions
are flawed. Of concern in the information technology era are
the many open-source journals appearing globally that claim
peer-review and falsify editorial board membership. These
“predatory journals” do not provide the review services present
with legitimate journals and may be challenging for information
consumers to spot (more on this topic in Chapter 4).
4. & 5. combined: Shifting the burden of proof.A sign of
pseudoscientific beliefs involves demanding that critics prove
them wrong (rather than assuming responsibility for proving the
claim right). In science, the burden of proof is on the one
making the claim. This goes along with the sign where
promoters of a pseudoscientific claim seek only confirmatory
evidence, not any evidence that might disprove the claim.
Scientists, however, “bend over backward” to design studies that
“test and potentially falsify our most cherished notions” (Thyer
& Pignotti, p. 10).
6. Unlike anything seen before.A common claim with any
pseudoscientific beliefs is that it represents a totally new

44 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

 phenomenon, a dramatically different innovation, unlike
anything seen before. In science, innovation is derived from
existing models, theories, or interventions. Often the innovation
involves a unique, novel, new way of thinking about what came
before, but connectivity between past, present, and future can
be demonstrated. With pseudoscience, claims of the
extraordinary are involved, disconnecting the innovation from
anything in the past. “Hence, the well-known saying that
extraordinary claims require extraordinary evidence applies”
(Thyer & Pignotti, 2015, p. 11).
7. The power of testimonials. Promoters of pseudoscientific
innovations rely heavily on testimonial statements and
anecdotal “evidence” provided by those who have been swayed
into belief. While it may be true that an individual providing the
testimonial was helped by the innovation, this alone does not
constitute reliable scientific evidence. It does not attempt to
rule out influences such as social desirability, placebo effects,
and poor retrospective recall on the person’s perceptions. The
proponents of a pseudoscientific claim also do not share any
complaints or reports of negative results in their advertising.
This is at the heart of recent challenges to how social media
product reviews are biased and potentially unreliable.
8. Relying on jargon.Pseudoscientific claims often introduce
technical-sounding jargon that merely obscures the facts.
“Technobabble” may be invented for a specific purpose, it may
involve misapplication of scientific terminology, and/or it may
involve “mashing” together concepts from different fields of
science. Here are some examples presented in an online video
demonstration of pseudoscientific language, Skeptic Presents:
Get Your Guru Going (https://scienceornot.net/2012/05/25/
technobabble-and-tenuous-terminology-the-use-of-pseudo-
scientific-language/). These examples were presented in a “TED
Talks” style and context.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 45

 “Ladies and Gentlemen, consciousness is merely the
potentiality of quantum locality expressing itself beyond
the constraints of four-dimensional space-time. And, that
is the externalization of what we call love.”

“Mindfulness, on the other hand, differentiates itself into

the multi-dimensional expanse of external reality. It really
is that simple. Quantum perception embraces all potential
space-time events exponentially. Cosmic balance is simply
the way the universe transcodes the raw potential of
quantum energy into unbridled happiness.”

9. One size fits everything. In the world of fashion, “one size fits all”
is a false claim—there are always going to be individuals for
whom that item is too large, too small, too short, or too long.
The same is true in social work—the most effective possible
interventions still fail to help some individuals. It is important
for science to identify for whom an intervention might not work
as intended; future scientific work might be able to unravel the
reasons and create interventions targeted specifically to those
individuals. Pseudoscientific claims tend to be overly
inflated—the interventions are often advertised as working for
everyone and/or for everything. A scientific approach requires
evidence that it works for each condition claimed.
10. The Combo Meal. Just as fast-food eateries offer consumers a
combination meal option, rather than ordering individual items,
pseudoscientific approaches are often presented as part of a
package—where the other parts of the package might have
supporting evidence. What makes this pseudoscience is that the
new add-ons have not been scientifically evaluated
independently, and the combination may not have the support
of evidence, either. Science does not stop at demonstrating that
an intervention has an effect on outcomes; a great deal of effort
is directed to understanding the mechanisms of change, how

46 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

 the changes are produced. Not only would each intervention
alone be tested, the mechanisms by which they operate would
be studied, as would their interactions when combined.

When broken down and analyzed this way, you might wonder why
people believe in pseudoscience, especially in the face of challenges
from more critical thinkers. At least two forces are in play at these
times. One is that people sometimes desperately seek simple,
effective solutions to difficult problems. Parents tormented by
watching their child disappear into the altered state of schizophrenia,
individuals haunted by the cravings induced by addiction, anyone
who experiences wrenching grief over the death of someone they
love: these people are particularly susceptible to pseudoscientific
claims.

The second reason people often stick to pseudoscientific beliefs

despite a lack of evidence relates to a psychological phenomenon
known as cognitive dissonance. A significant internal conflict is
created when a person’s deeply held belief or attitude is challenged
by contradictory beliefs, attitudes, or evidence. The discomfort of
this conflicted experience is called cognitive dissonance. The person
is left with few choices: change the original belief/attitude to fit
the new information, remain in limbo with the two in conflict, or
refute the new information and stick to the original belief/attitude.
Remaining in limbo is an extremely uncomfortable state, so the
person is likely to restore balance and comfort by selecting one of the
other two options. Thus, it is not unusual for a person to reject the
new evidence and cling to the original belief.

Consider the example of how you might respond to a homeless

person on the street who asks you for money. Why does this scenario
make you feel uncomfortable? On one hand, you see yourself as a
kindly, generous, giving person who cares about other people. On the
other hand, you do not want to give your hard-earned money away

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 47

to strangers. These two beliefs come into conflict when the person in
need asks for money, a conflict you experience as the uncomfortable
state of cognitive dissonance. (One solution is to resort to
rationalization to resolve the conflict and not give the money: it is
not good to reinforce this behavior, the money might do harm if
the homeless person buys alcohol or drugs with it, someone else
with more money than I have will donate, my taxes go to providing
services to this type of person so I have already donated, letting on
that I have any money will make me vulnerable to being robbed, and
other reasons.)

Emerging Science. Does all of this mean that all unproven techniques
are pointless and should be avoided? No, not really. After all, we
should consider the observation attributed to Sir Isaac Newton:

“What we know is a drop, what we don’t know is an ocean.”

Knowledge and science are always emerging and evolving. It is

possible that what is unheard of today is going to be common
knowledge in the future. In other words, what is today an alternative
or complementary treatment approach may eventually become a
standard practice based on evidence of its effectiveness and safety.
For example, not long ago, meditation practices were considered
to be unconfirmed alternative mental health intervention strategies.
Evidence supports the inclusion of mindfulness-based interventions
(MBIs) and meditation alongside other intervention strategies for
addressing a range of mental health concerns (Edenfield & Saeed,
2012). This includes stress reduction, anxiety, depression, and combat
veterans’ post-traumatic stress disorder (PTSD) (Khusid &
Vythilingam, 2016, Perestelo-Perez, et al., 2017).

A more current conundrum concerns the use of EMDR (Eye

Movement Desensitization and Reprocessing) therapy as a treatment

48 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

approach for addressing PTSD. As scientific study progresses,
elements of this therapy that are similar to other evidence-supported
approaches seem to have significant treatment benefits, but it is
unclear that all of the eye-movement procedures are necessary
elements (Lancaster, Teeters, Gros, & Back, 2016). Investigators are
learning more about EMDR as the mechanisms of change in the
treatment process are being more systematically and scientifically
investigated. This EMDR therapy example fits into that part of the
ocean about which we do not know enough. It serves as a reminder
that we need to be mindful of the tension between what we know is
true, what could be true, and what masquerades as truth.

Distinguishing between Facts and Opinions

We hear hundreds of statements of fact and opinion every day—from

family, friends, the media, and experts in our field. It is important to
decipher which are facts and which are opinions in making decisions
about important matters. A fact is information that can be objectively
proven or demonstrated. An opinion is a personal belief or point of
view, is subjective, and proof is not relevant. Just because a person
expresses an opinion does not mean that person has useful
knowledge to impart.

“There are in fact two things, science and opinion; the former
begets knowledge, the latter ignorance” (Hippocrates, circa 395
BC).

In a 1987 essay,The Fine Art of Baloney Detection: How Not to be

Fooled, Carl Sagan addressed the importance of critical thinking and
a healthy degree of skepticism about the information and knowledge
that we encounter. He stated that:

“Skeptical habits of thought are essential for nothing less than

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 49

 our survival—because baloney, bamboozles, bunk, careless
thinking, flimflam and wishes disguised as facts are not
restricted to parlor magic and ambiguous advice on matters
of the heart. Unfortunately, they ripple through mainstream
political, social, religious and economic issues in every
nation”(p. 13).

It is worth noting that they also ripple through the world of

professional practice, as well. So, how do we tell the difference when
we are provided with practice-related information? Here are some
ways of assessing the information (Surbhi, 2016).

• Is the information based on observation or research, can it be

validated or verified with pieces of evidence (fact)? Or, is it
based on assumptions and personal views, representing a
personal perception, judgment, or belief (opinion)?
• Does it hold true in general (fact) or does it differ individually
(opinion)?
• Is it presented or described in objective, unbiased words (fact)
or is it expressed with subjective, biased words (opinion)?
• Is it a debatable view (opinion) or non-debatable (fact)?
• Facts can change opinions but opinions cannot change facts.

The Problem with Expert Opinion.Over years of practice, social

workers develop a wealth of experience in identifying and addressing
certain problems encountered in routine practice. Under the heading
of “practice makes perfect,” experience contributes to developing
a certain degree of practice wisdom and expert opinion. What is
important in this label is the word “opinion.” Experience does not
change facts, it shapes opinion. Expert opinion is not, in itself, a
bad thing to turn to. The problem lies in non-critical reliance on
expert opinion. Let’s consider an example that a social work intern
experienced in practice.

50 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

 At intake in a mental health center, Mr. R (aged 34) describes
recent scary incidents where he has experienced a loss of
memory for certain events and found himself in dangerous
situations. Most recently, he “came to” standing on the edge of
a river embankment with the potential of falling into rushing
water at night. He has complete amnesia for how he came
to be at the river. In supervision, the intern suggested two
hypotheses: Mr. R has a dissociative disorder and these
incidents might involve suicidality, or Mr. R has alcohol (or
other substance-induced) blackouts. The intern’s social work
supervisor discounted the first hypothesis on the grounds that
dissociative disorders do not occur in men, only women.
Furthermore, the supervisor stated that dissociative disorders
are rare, making it unlikely that the intern would encounter
this in a first semester field placement, since most practitioners
never see such a case in an entire career. The supervisor
indicated that the proper assessment would be a substance use
disorder.

Relatively few facts are present in this example. They include:

• Mr. R was seen by the intern at intake to the mental health

center.
• The intern was in the first semester of field placement.
• The client provided a description of events.
• The intern offered two hypotheses.
• Dissociative disorders are rare.
• The supervisor did not address the potential suicidality risk of
the client’s reported incidents.

The following either fall short of being demonstrable facts or

represent opinion:

• The intern has the client’s description of what happened but

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 51

 does not know for a fact that these things happened.
• Dissociative disorders do not occur in men, only in women (they
are more common among women but do occur among men).
• Because the disorder is rare, this would not show up in a first-
semester internship caseload (a classic logic fallacy).
• The proper assessment is substance use disorder.

Based on this analysis, the social work intern should consider the
“expert opinion” of the supervisor but should also seek more
information to confirm the hypothesis of a substance use disorder.
The intern should also keep an open mind to the other possibilities,
including dissociative disorder (often related to a trauma experience)
and suicidality, and seek information that either supports or refutes
these alternative conclusions.

The Problem with Groupthink. Sometimes an individual makes

decisions or judgments in conformity with a group’s thinking,
decisions that run contrary to what that individual would have
decided alone. The process of group think typically refers to faulty
decision-making resulting in a group’s systematic errors. Groupthink
is characterized by individuals abandoning their own critical analysis,
reality testing, and reasoning—possibly abandoning their own moral
reasoning (Street, 1997). An example discussed by the originator of
the groupthink concept was the space shuttle Challenger disaster:
the commission appointed to determine the probable causes of this
incident concluded, among other things, that “a highly flawed
decision process was an important contributing cause of the disaster
(Janis, 1991, p. 235). A more everyday example follows:

Four friends on vacation at Hawai’i Volcanoes National Park

hiked out to view an active volcano. They arrived at a point
where park rangers had posted signs indicating that no one

52 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

 should go further because of the danger of breaking through
thin lava. The group members were disappointed that they
could only see the active lava flow in the distance. The group’s
informal leader announced they had come this far and it looked
safe enough; he encouraged everyone to follow him out onto
the lava bed to get a closer look. Two other members started to
willingly go along with him. One member decided that she was
not going to cave in to group think, and told the others so. The
two following the leader soon turned back, as well, conceding
that this was ill-advised. The “leader” eventually agreed, and
returned safely, too.

Critics of the original groupthink model proposed by Janis (1991)

argue with the emphasis placed on certain aspects of group cohesion
as a precursor to the observation that group decision-making can
be faulty. However, researchers have observed that “the propensity
for the group to display groupthink symptoms” is greatest when
the group is characterized by a high degree of cohesion based on
the socioemotional dimension and a low degree of task-oriented
cohesion (Street, 2017, p. 78). This difference in groupthink
occurrence is attributed to the importance placed on analytical
(critical) thinking in task-oriented groups. These findings have
important implications for social work practitioners working in
organizations and communities, in particular.

Module 1 Chapter 1: Social Work and "Ways of Knowing" | 53

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=22

54 | Module 1 Chapter 1: Social Work and "Ways of Knowing"

Module 1 Chapter 2: Research
and the Professional Ethics
Mandate
America’s long, complicated, ambivalent, and confusing relationship
with research, evidence, and factual information have profound
implications for how social workers practice our profession. In 1964,
a Pulitzer Prize was awarded to Richard Hofstadter’s book (1963)
analyzing the role played by anti-intellectual passions in a democratic
society—intellect being about critical, creative, and contemplative
thinking. The author’s analysis is relevant today because it addresses
the depreciation of knowledge as a societal counter-force acting
against social change; counter-forces that are especially prevalent in
periods characterized by “complex and bewildering global change”
(Baer, 2017).

What you will learn from reading this chapter is:

• Why research matters to the social work profession and the

nature of the relationship between research and ethical social
work practice
• Principles and standards in the social work code of ethics that
address the relationship between research and practice.

Module 1 Chapter 2: Research and the

Professional Ethics Mandate | 55
 Before you begin reading, please
answer the following question for
yourself—make a note of your answer
using the 5-point scale.

When I think about the importance of

research and social work practice
skills, I think that:

1. practice skills are far more

important
2. practice skills are somewhat more
important
3. research and practice skills are about equally important
4. research skills are somewhat more important
5. research skills are far more important

A Place for Research in The Social Work Profession

One feature of a profession is that members of that profession

participate in the systematic study of its practices and engaging in
knowledge-building processes that shape practice (Kirk & Reid, 2002;
Zlotnik, 2008). Having a body of knowledge and theory, as opposed
to simply having a set of practices or skills, is a distinguishing feature
between an occupation and a profession (Kirk & Reid, 2002).

Research is an important source of knowledge for the profession

of social work. Translated from an original Arabic analysis is the
phrase “knowledge is power” (attributed to Imam Ali). Similar is a
quote from Ralph Waldo Emerson (1862): “…knowledge comes by eyes
always open, and working hands and there is no knowledge that is not
power” (p. 136).

56 | Module 1 Chapter 2: Research and the Professional Ethics Mandate

Consider this description of social work research:

“Social work research is the use of social research methods

for developing, producing, and disseminating knowledge that
is pertinent to policies and practices…It aims to describe and
explain phenomena relevant to social work” (Tripodi &
Lalayants, 2008, p. 512)

And, in an essay about the science of the social work profession,

the authors answer the question “Why social science?” with the
conclusion, “Because social science—and social work research in
particular—is an engine for social progress” (Coffey & Williams, 2017).
This observation leads to an important point: social work is a long-
standing, proud profession that draws on an expanding body of
knowledge, theory, and research—both its own and that generated
through multiple other disciplines. Because social work practice
covers so many broad topics, problems, and populations, and because
social work professionals work in collaboration with so many other
professions, social work scholars are well positioned to contribute to
and draw from knowledge in multiple professions and disciplines.

Module 1 Chapter 2: Research and the Professional Ethics Mandate | 57

The NASW Code of Ethics

By now you may have been introduced to the 36-page National

Association of Social Workers Code of Ethics(NASW, 2017) that
outlines patterns of professional behavior that define ethical social
work practice. In this chapter, we examine several sections of the
professional code of ethics that directly address the relationship
between research/evidence and professional social work practice.
This is not quite the same thing as what we explore in Chapter 4
where we focus more specifically on the ethical conduct of research.

You can access the NASW Code of Ethics at the interactive website
https://www.socialworkers.org/About/Ethics/Code-of-Ethics/
Code-of-Ethics-Englishor as a pdf link in the Carmen course site.

Beginning with The Sixth Principle.The first place where the code
of ethics content directly relates to the relationship between ethical
social work practice and research is in the last of the six broad ethical
principles (NASW, 2017). This principle states:

Social workers practice within their areas of competence and

develop and enhance their professional expertise (p. 6).

This single statement is followed by two additional statements that

elaborate the point:

Social workers continually strive to increase their professional

knowledge and skills and to apply them in practice. Social
workers should aspire to contribute to the knowledge base of the
profession (p. 6).

How can you as a social worker contribute to the knowledge base of

the profession? Engaging in research is a major mechanism for doing
so.

Standard 4.01: Competence.Next, let’s turn to the Standards

58 | Module 1 Chapter 2: Research and the Professional Ethics Mandate

identified in the professional code of ethics for social workers (NASW,
2017). The first place where evidence and research are directly
mentioned is in Standard 4.01, which is about competence. Two
statements are relevant here, the first of which is:

(b) Social workers should strive to become and remain

proficient in professional practice and the performance of
professional functions. Social workers should critically examine
and keep current with emerging knowledge relevant to social
work. Social workers should routinely review the professional
literature and participate in continuing education relevant to
social work practice and social work ethics (p. 25).

What does this mean for you as a social worker?

• First, you need to learn how to identify current relevant

knowledge.
• Second, you need to develop the skills necessary to critically
examine such knowledge.
• Third, you need to develop a strong value and appreciation for
putting forth the effort to maintain your knowledge of the
professional literature—social work education is a lifelong
commitment, it does not end the day you graduate!

The second statement elaborating on the Competence Standard 4.01

is:

(c) Social workers should base practice on recognized

knowledge, including empirically based knowledge, relevant to
social work and social work ethics (p. 25).

For you as a social worker, this means knowing:

• how and where to identify sources of knowledge,

• how to evaluate the information provided through different

Module 1 Chapter 2: Research and the Professional Ethics Mandate | 59

 sources—including the evaluation of research evidence, and
• how research evidence relates to social work practice.

Standard 5.01: Integrity of the Profession. The next place where the
code of ethics directly refers to research and evidence is in Standard
5.01 addressing Integrity of the Profession. Here it is stated:

(b) Social workers should uphold and advance the values, ethics,
knowledge, and mission of the profession. Social workers should
protect, enhance, and improve the integrity of the profession
through appropriate study and research, active discussion, and
responsible criticism of the profession (p. 27).

What this means is that social workers have a responsibility to engage

in critical thinking about what we know and what we do as
professionals.

Research is mentioned again in the next statement as one of the key

activities that promote the profession’s integrity:

(c) Social workers should contribute time and professional

expertise to activities that promote respect for the value,
integrity, and competence of the social work profession. These
activities may include teaching, research, consultation, service,
legislative testimony, presentations in the community, and
participation in their professional organizations (p. 27).

The topic appears again in the last of the Integrity of the Profession
statements:

(d) Social workers should contribute to the knowledge base of

social work and share with colleagues their knowledge related
to practice, research, and ethics. Social workers should seek
to contribute to the profession’s literature and to share their
knowledge at professional meetings and conferences (p. 27).

60 | Module 1 Chapter 2: Research and the Professional Ethics Mandate

Standard 5.02: Evaluation and Research.This standard is the most
obvious and lengthy place where the relationship between research
and social work practice is presented in the NASW ethical code (2017).
Here are the first three points stated in this Evaluation and Research
standard—we will look at the remaining points from this standard
in Chapter 4 where we take a close look at the ethical conduct of
research:

• (a)Social workers should monitor and evaluate policies, the

implementation of programs, and practice interventions (p. 27).
• (b)Social workers should promote and facilitate evaluation and
research to contribute to the development of knowledge (p. 27).
• (c)Social workers should critically examine and keep current
with emerging knowledge relevant to social work and fully use
evaluation and research evidence in their professional practice
(p. 27).

What do these points mean to you as a social work professional?

Taken together, they reinforce some of what was emphasized in
earlier standards—social workers:

• engage in efforts to evaluate their practices and programs

(which includes policy as an intervention strategy),
• promote research that helps build the profession’s knowledge
base,
• keep up with emerging knowledge (especially in our areas of
practice),
• continue to think critically about the emerging knowledge, and
• use evidence in our professional practice.

This last point is the focus of the entire second course in our
sequence, SWK 3402.

Module 1 Chapter 2: Research and the Professional Ethics Mandate | 61

 Before you began reading, you
answered the following question for
yourself using the 5-point scale. Now
compare your answers from before and
after reading the chapter. If your
answer changed after reading the
chapter, what do you think was the
most important information to
stimulate this change? If your answer
remained the same, what do you think
was the most important information
that you read to reinforce your earlier answer?

When I think about the importance of research and social work

practice skills, I think that:

1. practice skills are far more important

2. practice skills are somewhat more important
3. research and practice skills are about equally important
4. research skills are somewhat more important
5. research skills are far more important

62 | Module 1 Chapter 2: Research and the Professional Ethics Mandate

Module 1 Chapter 3: Ethical
Conduct of Research
The NASW professional code of ethics has quite a lot to say about
ensuring that research be conducted in an ethical manner, so here we
revisit the code of ethics to review these research-specific standards.
We begin with an overview of historical events that bring us to where
we are today in the research ethics arena. This content is covered
in greater depth and detail in your CITI (Collaborative Institutional
Training Initiative) program training modules. Then, we return to the
section of NASW’s (2017) Code of Ethics that we initially visited in
Chapter 1.

In reading this chapter, you will learn:

• Key facts about research ethics;

• Statements from the NASW Code of Ethics specific to the
ethical conduct of research

History and Research Ethics

Following the conclusion of World War II and subsequent “Doctors’

Trial” portion of the Nuremberg Trials, the 10-point Nuremberg Code
was developed as a response to numerous examples of unethical,
inhumane “medical” experiments conducted in concentration camps.
The Nuremberg Code was relatively ignored for many years, at least
in the United States. Subsequently, a number of grievously unethical
experiments were conducted on prisoners, institutionalized patients,
and children—disproportionately on poor and racial minority
populations—including exposure to toxins, diseases, radiation, or
torture. The Tuskegee syphilis experiment represents a critical

Module 1 Chapter 3: Ethical Conduct

of Research | 63
turning point in the nation’s tolerance of unethical human research.
In 1932, the U.S. Public Health Service partnered with the Tuskegee
Institute to study the natural course of syphilis among 399 black
men. Over 40 years, the men believed they were being treated for
syphilis, but were in fact receiving no treatment for the disease, even
when penicillin proved to be an effective form of treatment by 1947.
Beginning in 1968, concerns were being raised and condemning news
reports were widely circulated in 1972, leading to the study being
ended. By 1974, the Tuskegee Health Benefit Program was established
by the U.S. government to ensure health and burial benefits to the
study’s remaining survivors, wives, widows, and children.

In response to concerns about unethical research practices, a

national group was formed (in 1974, the National Commission for
the Protection of Human Subjects of Biomedical and Behavioral
Research), and created what is now known as The Belmont Report
(1979). The Belmont Report, in turn, influenced development of
federal policy (1981, revised 2009) concerning protections for human
research subjects—the Common Rule. The Belmont Report presented
a summary of ethical guidelines for engaging in research that involves
human subjects. The guidelines are based on three core principles:

• Respect for persons.This principle is founded on a conviction

that “individuals should be treated as autonomous agents”
(Belmont Report, 1979, p. 4). The report explains this autonomy
in terms of respecting a person’s right to making considered,
informed choices, and places a responsibility on researchers to
ensure that a person has all of the information necessary for
self-determined choices and is free from constraints on making
self-determined choices. The discussion also addresses
situations where a person might not be capable of self-
determination and what protections might be necessary in
these instances.
• Beneficence.This principle is about an obligation for protecting

64 | Module 1 Chapter 3: Ethical Conduct of Research

 participants from harm and “making efforts to secure their well-
being” (Belmont Report, 1979, p. 5). This translates into ensuring
that benefits to participants are maximized and possible harms
or risks are minimized.
• Justice. This principle is about fairness. Justice in this context is
about ensuring a fair and just distribution of both the potential
burdens and the potential benefits of participating in research.
“Another way of conceiving the principle of justice is that equals
ought to be treated equally” (Belmont report, 1979, p. 5).

These three core principles are closely aligned with the core values of
our social work profession. Our code of ethics is based on principles
that include respect for the dignity and worth of the individual, which
includes fostering client self-determination, as well as practicing
with integrity.

The Belmont Report (1979) also contains language about

distinguishing between research and professional practice. The
report defines practice in this way:

“…interventions that are designed solely to enhance the well-

being of an individual patient or client and that have a
reasonable expectation of success” (p. 3).

Research is defined in the Belmont Report as:

“…an activity designed to test an hypothesis, permit

conclusions to be drawn, and thereby to develop or contribute
to generalizable knowledge” (p. 3).

These two types of professional activity are clearly contrasted—their

goals are markedly different. Practice goals relate to benefits for the
individuals being served; research goals are served by the individuals
who participate.

Module 1 Chapter 3: Ethical Conduct of Research | 65

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=199

Revisiting the NASW Code of Ethics

We previously examined what the NASW Code of Ethics had to say

about the role of research and evidence in professional practice and
the role of social work professionals in engaging with research and
evidence. At this point, we turn again to the Code of Ethics to see
what is specified in terms of the ethical conduct of research. These
details are in addition to what was discussed earlier about research
integrity.

Standard 5.02: Evaluation and Research continued. Let’s pick up

where we left off in reviewing the NASW Code of Ethics content
related to engaging in evaluation and research as a social work
professional. The Code of Ethics (p. 27-28) replicates many elements
found in the current federal policy concerning the protection of
human subjects in research and in the 1996 Health Insurance
Portability and Accountability Act (HIPAA). The Code of Ethics states
that:

66 | Module 1 Chapter 3: Ethical Conduct of Research

• (d) Social workers engaged in evaluation or research should
carefully consider possible consequences and should follow
guidelines developed for the protection of evaluation and research
participants. Appropriate institutional review boards should be
consulted

• (e) Social workers engaged in evaluation or research should obtain

voluntary and written informed consent from participants, when
appropriate, without any implied or actual deprivation or penalty
for refusal to participate; without undue inducement to
participate; and with due regard for participants’ well-being,
privacy, and dignity. Informed consent should include
information about the nature, extent, and duration of the
participation requested and disclosure of the risks and benefits of
participation in the research.

• ( f) When using electronic technology to facilitate evaluation or

research, social workers should ensure that participants provide
informed consent for the use of such technology. Social workers
should assess whether participants are able to use the technology
and, when appropriate, offer reasonable alternatives to
participate in the evaluation or research.

• (g) When evaluation or research participants are incapable of

giving informed consent, social workers should provide an
appropriate explanation to the participants, obtain the
participants’ assent to the extent they are able, and obtain written
consent from an appropriate proxy.

• (h) Social workers should never design or conduct evaluation or

research that does not use consent procedures, such as certain
forms of naturalistic observation and archival research, unless
rigorous and responsible review of the research has found it to be

Module 1 Chapter 3: Ethical Conduct of Research | 67

 justified because of its prospective scientific, educational, or
applied value and unless equally effective alternative procedures
that do not involve waiver of consent are not feasible.

• (i) Social workers should inform participants of their right to

withdraw from evaluation and research at any time without
penalty.

• (j) Social workers should take appropriate steps to ensure that

participants in evaluation and research have access to
appropriate supportive services.

• (k) Social workers engaged in evaluation or research should

protect participants from unwarranted physical or mental
distress, harm, danger, or deprivation.

• (l) Social workers engaged in the evaluation of services should

discuss collected information only for professional purposes and
only with people professionally concerned with this information.

• (m)Social workers engaged in evaluation or research should

ensure the anonymity or confidentiality of participants and of the
data obtained from them. Social workers should inform
participants of any limits of confidentiality, the measures that will
be taken to ensure confidentiality, and when any records
containing research data will be destroyed.

• (n) Social workers who report evaluation and research results

should protect participants’ confidentiality by omitting
identifying information unless proper consent has been obtained
authorizing disclosure.

• (o) Social workers should report evaluation and research findings

accurately. They should not fabricate or falsify results and should

68 | Module 1 Chapter 3: Ethical Conduct of Research

 take steps to correct any errors later found in published data
using standard publication methods.

• (p) Social workers engaged in evaluation or research should be

alert to and avoid conflicts of interest and dual relationships with
participants, should inform participants when a real or potential
conflict of interest arises, and should take steps to resolve the
issue in a manner that makes participants’ interests primary.

• (q) Social workers should educate themselves, their students, and

their colleagues about responsible research practices.

Together, these 14 statements reflect a responsibility for social

workers to :

• participate in appropriate Institutional Review Board (IRB)

procedures,
• ensure the safety and protection of participants in social work
research and evaluation studies,
• engage in effective informed consent procedures,
• responsibly utilize information technology in research,
• collect and discuss only information relevant to the study,
• preserve participant privacy and confidentiality (or anonymity),
• engage in research with integrity,
• prevent or responsibly resolve conflict of interest instances, and
• remain informed about responsible research practices.

Module 1 Chapter 3: Ethical Conduct of Research | 69

Module 1 Summary

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=83

In Module 1, we covered a range of topics related to the sources of

knowledge social workers might utilize in practice. One of the major
themes of this module was that there are many different sources of
knowledge available to us as human beings, and additional sources to
consider in relation to ourselves as professional social workers. The
contents of the first chapter reflected different ways of knowing that
might also help us to better understand our clients and colleagues,
as well as ourselves. The Chapter 1 contents set the stage for Chapter
2, creating a context for understanding one of the major ways of
knowing related to professional social work practice: evidence
derived through research. Finally, Chapter 3 addressed issues in
professional responsibility to engage in research that is both ethical
and characterized by integrity. Using this first module as a

70 | Module 1 Summary
springboard, we are now ready to begin looking more closely at the
nature of the evidence that informs social work practice.

Module 1 Summary | 71
Module 1 Key Terms and
Definitions
abductive reasoning: beginning with a set of observations, then
seeking the most likely explanation for what was observed.

Code of Ethics: standards for professional practice, such as the NASW

Code of Ethics for professional social work practice.

cognition: mental processes (thinking, experiencing, sensing)

involved in knowledge and understanding.

cognitive dissonance: uncomfortable experience of inconsistencies in

thoughts, beliefs, or attitudes; often motivates changing one or
more of these cognitions.

critical thinking: forming judgments or conclusions on the basis of

objective analysis and evaluation of information.

deductive reasoning: deriving a specific conclusion based on

application of logical processes, assembling a series of truthful
general premises.

empiricism: belief that knowledge is primarily derived from sensory

experiences and observation.

epistemology: theory of knowledge, its nature, and how it is acquired.

ethnoepistemology: theory of knowledge based on acceptance of all

human experience and epistemology as being equally valid;
strongly based in anthropology.

72 | Module 1 Key Terms and

Definitions
fallacious argument: logic based on faulty reasoning, resulting in
faulty conclusions.

inductive reasoning: deriving a general conclusion or hypothesis

based on a set of specific observations.

linguistic determinism: the limits on knowledge imposed by the

structure and contents of a language.

metacognition: the act of thinking about or being aware of thought

processes and contents (thinking about cognition)

positivism: belief in the possibility of rational/logical or scientific

proof of all justifiable knowledge.

pseudoscience: premises or assertions that have the appearance of

being based on science without their being based on the
application of strong scientific method.

qualitative research: based on a belief that knowledge is socially

constructed, explores individuals’ perceptions, beliefs, attitudes,
experiences, and meaning-making to derive general conclusions or
develop hypotheses.

quantitative research: based on objective measurement for variables

of interest to answer research questions/test hypotheses about
variables or their inter-relationships.

research integrity: a commitment to honesty and assuming

responsibility for research conduct, resulting in replicable, reliable
research results.

Module 1 Key Terms and Definitions | 73

Module 1 References
Adams, C. (2017). The gender identity terms you need to know. CBS
News (March 24). Retrieved from https://www.cbsnews.com/
news/transgender-gender-identity-terms-glossary/

Bacon, F. (1955). Selected writings of Francis Bacon(edited by H.G.

Dick). NY: Modern Library.

Baer, N. (2017). American idiot: Rethinking anti-intellectualism in the

age of Trump. Transformation(August 27). Retrieved from
https://www.opendemocracy.net/transformation/nicholas-baer/
american-idiot-rethinking-anti-intellectualism-in-age-of-trump

Banta-Green, C. (2015). Medication assisted treatment for opioid use

disorders: Overview of the evidence.University of Washington
Alcohol & Drug Abuse Institute brief (June). Retrieved from
http://adai.uw.edu/pubs/infobriefs/MAT.pdf

Baskin, E.J., & Hess, R.D. (1980). Does affective education work? A
review of seven programs. Journal of School Psychology, 18(1), 40-50.

Belenky, M.F., Clinchy, B.M., Goldberger, N.R., & Tarule, J.M. (1986).
Women’s ways of knowing: The development of self, voice, and mind.
NY: Basic Books.

Belmont Report, The. (1979). The Belmont Report: Ethical principles

and guidelines for the protection o human subjects
research.Prepared by The National Commission for the Protection
of Human Subjects of Biomedical and Behavioral Research (April 18).
Washington, DC: Department of Health, Education, and Welfare.
https://www.hhs.gov/ohrp/regulations-and-policy/belmont-
report/read-the-belmont-report/index.html

74 | Module 1 References
Birkenmaier, J., & Berg-Wegner, M. (2017). The practice of generalist
social work, 4thed. NY: Routledge.

Coffey, D.S., & Williams, J.H. (2017).Why social science? Because it

is an engine for social progress.Consortium of Social Science
Associations. Retrieved from www.whysocialscience.com

DeLecce, T. (n.d.). What is critical thinking? Definition, skills &

meaning. Lesson transcript. Retrieved from https://study.com/
academy/lesson/what-is-critical-thinking-definition-skills-
meaning.html

Drisko, J.W., & Grady, M.D. (2012). Evidence-based practice in clinical

social work. NY: Springer.

Edenfield, T.M., & Saeed, S. (2012). An update on mindfulness

meditation as a self-help treatment for anxiety and depression.
Psychology Research and Behavior Management, 5, 131-141.

Emerson, R.W. (1862). Old age. The Atlantic Monthly, 9(51), 134-140
(January). Retrieved from https://www.theatlantic.com/past/
docs/issues/1862jan/emerson.htm

Gambrill, E.D.(1997). Social work practice: A critical thinker’s guide.NY:

Oxford University Press.

Gibbs, L.E., & Gambrill, E.D. (1999). Critical thinking for social workers:
Exercises for the helping professions.Thousand Oaks, CA: Pine Forge
Press.

Gibbs, L.E., Gambrill, E.D., Blakemore, J., Begun, A., Keniston, A.,
Peden, B., & Lefcowitz, J. (1995). A measure of critical thinking about
practice. Research on Social Work Practice, 5(2), 193-204.

Goleman, D. (1995). Emotional intelligence.NY: Bantam Books.

Heussner, K.M. (2010). Stephen Hawking on religion: ‘Science will

Module 1 References | 75
 win.’ ABC News (June 7). Retrieved from https://abcnews.go.com/
WN/Technology/stephen-hawking-religion-science-win/
story?id=10830164

Hofstader, R. (1963). Anti-intellectualism in American life.NY: Knopf.

Holton, G. (1993). Science and anti-science.Cambridge, MA: Harvard

University Press.

International Baccalaureate Organization (IB). (n.d.). Theory of

knowledge guide: Ways of knowing. Retrieved from
https://ibpublishing.ibo.org/exist/rest/app/
tsm.xql?doc=d_0_tok_gui_1304_1_e&part=2&chapter=3

James, W. (1902). The varieties of religious experience: A study in

human nature (Lectures VI and VII). Retrieved from
http://ebooks.adelaide.edu.au/j/james/william/varieties/
complete

Janis, I.L. (1991). Groupthink. Reprinted in E. Griffin, (Eds), A first look

at communication theory,(pp. 235-246). NY: McGraw Hill.

Jones, M.A. (2002). Deafness as culture: A psychosocial perspective.

Disability Studies Quarterly, 22(2), 51-60.

Kaiser, M.L., & Cafer, A.M. (2016). Exploring long-term food pantry
use: Differences between persistent and prolonged typologies of
use. Journal of Hunger & Environmental Nutrition, 12, 46-63.

Khusid, M.A., & Vythilingam, M. (2016). The emerging role of

mindfulness meditation as effective self-management strategy,
Part 1: Clinical implications for depression, post-traumatic stress
disorder, and anxiety. Military medicine, 181(9), 961-968.

Kirk, S.A., & Reid, W.J. (2002). Science and social work: A critical
appraisal. NY: Columbia University Press.

76 | Module 1 References
Lancaster, C.L., Teeters, J.B., Gros, D.F., & Back, S.E. (2016).
Posttraumatic stress disorder: Overview of evidence-based
assessment and treatment. Journal of Clinical Medicine, (5),1-13.
doi: 10.3390/jcm5110105

Lilienfeld, S.O., Lynn, S.J., & Lohr, J.M. (2015). Science and
pseudoscience in clinical psychology: Initial thoughts, reflections,
and considerations. In S.O.Lilienfeld, S.J. Lynn, & J.M. Lohr (Eds.),
Science and pseudoscience in clinical psychology, (pp. 1-16). NY:
Guilford.

Lilienfeld, S.O., Lynn, S.J., Ruscio, J., & Beyerstein, B.L. (2010). 50 great
myths of popular psychology: Shattering widespread misconceptions
about human behavior. West Sussex, UK: Wiley-Blackwell.

Maffie, J. (1990). Recent work on naturalized epistemology. American

Philosophical Quarterly, 27(4), 281-294.

Maffie, J. (1995). Towards an anthropology of epistemology. The

Philosophical Forum, 26(3),218-241.

Mantell, A., Simpson, G.K., Vungkhanching, M., Jones, K.F.,

Strandberg, T., & Simonson, P. (2017). Social work-generated
evidence in traumatic brain injury from 1975 to 2014: A systematic
scoping review. Health and Social Care in the Community, 00(1),
1-16. doi: 10.1111/hsc.12476

Mathias, J. (2015). Thinking like a social worker: Examining the

meaning of critical thinking in social work. Journal of Social Work
Education, 51(3), 457-474.

Milner, M., & Wolfer, T. (2014). The use of decision cases to foster
critical thinking in social work students. Journal of Teaching in
Social Work, 34, 269-284.

Nahin, R.L., Barnes, P.M., & Stussman, B.A. (2016). Expenditures on

Module 1 References | 77
 complementary health approaches: United States, 2012. National
Health Statistics Report, No. 95. Hyattsville, MD: National Center
for Health Statistics. Retrieved from https://www.cdc.gov/nchs/
data/nhsr/nhsr095.pdf

National Association of Social Workers (NASW). (2017). Code of ethics

of the National Association of Social Workers.Washington, DC:
NASW Press. Retrieved from https://www.socialworkers.org/
LinkClick.aspx?fileticket=ms_ArtLqzeI%3D&portalid=0

Perestelo-Perez, L., Barraca, J., Penate, W., Rivero-Santana, A., &

Alvarez-Perez, Y. (2017). Mindfulness-based interventions for the
treatment of depressive rumination: Systematic review and meta-
analysis. International Journal of Clinical and Health Psychology,
17,282-295.

Rimland, B. (1964).Infantile autism: The syndrome and its implications

for a neural theory of behavior. Note: 50th anniversary edition
(2015), edited by W.M. Edelson. Philadelphia, PA: Jessica Kingsley
Publishers.

Russell, J.B. (1991). Inventing the Flat Earth: Columbus and modern
historians. NY: Praeger.

Sagan, C. (1987). The fine art of baloney detection: How not to be

fooled. Parade Magazine, Science on Parade, (February 1), p. 12-13.
Retrieved from https://www.csicop.org/uploads/files/
ParadeFeb11987.pdf

Schick, T., & Vaughn, L. (2010). How to think about weird things:
Critical thinking for a new age, 6thedition.NY: McGraw-Hill.

Soydan, H., & Palinkas, L.A. (2014).Evidence-based practice in social

work: Development of a new professional culture.NY: Routledge/
Taylor & Francis Group.

78 | Module 1 References
Surbhi, S. (2016). Difference between fact and opinion. Retrieved from
Key Differences athttps://keydifferences.com/difference-
between-fact-and-opinion.html

Street, M.D. (1997). Groupthink: An examination of theoretical issues,

implications, and future research suggestions. Small Group
Research, 28(1), 72-93.

Theory of Knowledge Network (TOK). (n.d.). Ways of knowing.

Retrieved from http://www.theoryofknowledge.net/ways-of-
knowing/

Thyer, B.A., & Pignotti, M.G. (2015). Science and pseudoscience in social
work.NY: Springer Publishing Co.

Wagner, M.F., & Skowronski, J.J. (2017). Social influence and mental
routes to the production of authentic false memories and
inauthentic false memories. Consciousness and Cognition, 51,
34-52.

Williams, J.H. (2016). [Editorial] Grand challenges for social work:

Research, practice, and education. Social Work Research, 40(2),
67-70.

Module 1 References | 79
MODULE 2

Module 2 | 81
Module 2 Introduction
Before we can design social work interventions to address specific
problems and help specific populations impacted by the problem,
we first need to understand the nature of these problems for which
populations might be seeking help. In this module, we explore how
social work professionals identify, access, assess, and utilize evidence
in understanding diverse populations, social work problems, and
social phenomena. (The SWK 3402 course involves exploring
research and evidence related to understanding interventions.) In
this Module 2 coursebook, we learn about the implications of
different types of research questions, the role of theory in addressing
research questions, and strategies for identifying and analyzing
existing knowledge related to a research question.

READING OBJECTIVES

After engaging with these reading materials and learning resources,

you should be able to:

• Recognize different types of research, research questions, and

specific aims of a study to address those questions;
• Begin linking theory to research questions about diverse
populations, social work problems, and social phenomena;
• Begin identifying and assessing empirical evidence about
diverse populations, social work problems, and social
phenomena.
• Define key terms related to identifying and assessing evidence.

Module 2 Introduction | 83
Module 2 Chapter 1: The
Nature of Social Work
Research Questions
The search for empirical evidence typically begins with a question
or hypothesis. The nature of the questions asked determine many
features of the studies that lead to answers: the study approach,
design, measurement, participant selection, data collection, data
analysis, and reporting of results. Not just any type of question will
do, however:

“When the question is poorly formulated, the design, analysis,

sample size calculations, and presentation of results may not
be optimal. The gap between research and clinical practice
could be bridged by a clear, complete, and informative research
question” (Mayo, Asano, & Barbic, 2013, 513).

The topic concerning the nature of social work research questions

has two parts: what constitutes a researchquestion, and what makes
it a social workquestion. We begin this chapter by examining a
general model for understanding where different types of questions
fit into the larger picture of knowledge building explored in Module
1. We then look at research questions and social work questions
separately. Finally, we reassemble them to identify strong social work
research questions.

In this chapter, you will learn:

• 4 types of social work research for knowledge building,

• characteristics of research questions,

84 | Module 2 Chapter 1: The Nature

of Social Work Research Questions
 • characteristics of social work research questions.

Translational Science

The concept of translational scienceaddresses the application of

basic science discoveries and knowledge to routine professional
practice. In medicine, the concept is sometimes described as “bench
to trench,” meaning that it takes what is learned at the laboratory
“bench” to practitioners’ work in the real-world, or “in the trenches.”
This way of thinking is about applied science—research aimed at
eventual applications to create or support change. Figure 1-1
assembles the various pieces of the translational science knowledge
building enterprise:

Figure 1-1. Overview of translational science elements

Module 2 Chapter 1: The Nature of Social Work Research Questions | 85

Basic Research. Federal policy defines basic research as systematic
study that is directed toward understanding the fundamental aspects
of phenomena without specific applications in mind (adapted from
32 CFR 272.3). Basic research efforts are those designed to describe
something or answer questions about its nature. Basic research in
social and behavioral science addresses questions of at least two
major types: epidemiology and etiology questions.

Epidemiology questions.Questions about the nature of a population,

problem, or social phenomenon are often answered through
epidemiological methods. Epidemiology is the branch of science
(common in public health) for understanding how a problem or
phenomenon is distributed in a population. Epidemiologists also ask
and address questions related to the nature of relationships between
problems or phenomena—such as the relationship between opioid
misuse and infectious disease epidemics (NAS, 2018). One feature
offered by epidemiological research is a picture of trends over time.
Consider, for example, epidemiology data from the Centers for
Disease Control and Prevention (the CDC) regarding trends in suicide
rates in the state of Ohio over a four-year period (see Figure 1-2,
created from data presented by CDC WONDER database).

Figure 1-2. Graph reflecting Ohio trend in suicide rate, 2012-2016

86 | Module 2 Chapter 1: The Nature of Social Work Research Questions

Since the upward trend is of concern, social workers might pursue
additional questions to examine possible causes of the observed
increases, as well as what the increase might mean to the expanded
need for supportive services to families and friends of these
individuals. The epidemiological data can help tease out some of
these more nuanced answers. For example, epidemiology also tells
us that firearms were the recorded cause in 46.9% of known suicide
deaths among individuals aged 15-24 years across the nation during
2016 (CDC, WONDER database). Not only do we now know the
numbers of suicide deaths in this age group, we know something
about a relevant factor that might be addressed through preventive
intervention and policy responses.

Epidemiology also addresses questions about the size and

characteristics of a population being impacted by a problem or the
scope of a problem. For example, a social worker might have a
question about the “shape” of a problem defined as sexual violence
victimization. Data from the United States’ 2010-2012 National
Intimate Partner and Sexual Violence Survey (NISVS) indicated that

Module 2 Chapter 1: The Nature of Social Work Research Questions | 87

over 36% of woman (1 in 3) and 17% of men (1 in 6) have experienced
sexual violence involving physical contact at some point in their lives;
the numbers vary by state, from 29.5% to 47.5% for women and 10.4%
to 29.3% for men (Smith et al., 2017).

In developing informed responses to a problem, it helps to know

for whom it is a problem. Practitioners, program administrators, and
policy decision makers may not be aware that the problem of sexual
violence is so prevalent, or that men are victimized at worrisome
rates, as well as women. It is also helpful to know how the problem
of interest might interface with other problems. For example, the
interface between perpetrating sexual assault and alcohol use was
examined in a study of college men (Testa & Cleveland, 2017). The
study investigators determined that frequently attending parties and
bars was associated with a greater probability of perpetrating sexual
assault. Thus, epidemiological research helps answer questions about
the scope and magnitude of a problem, as well as how it relates to

88 | Module 2 Chapter 1: The Nature of Social Work Research Questions

other issues or factors, which can then inform next steps in research
to address the problem.

Etiology questions. Etiology research tests theories and hypotheses

about the origins and natural course of a problem or phenomenon.
This includes answering questions about factors that influence the
appearance or course of a problem—these may be factors that
mediate or moderate the phenomenon’s development or progression
(e.g., demographic characteristics, co-occurring problems, or other
environmental processes). To continue with our intimate partner
violence example, multiple theories are presented in the literature
concerning the etiology of intimate partner violence
perpetration—theories also exist concerning the etiology of being
the target of intimate partner violence (Begun, 2003). Perpetration
theories include:

• personality/character traits
• biological/hereditary/genetic predisposition
• social learning/behavior modeling
• social skills
• self-esteem
• cultural norms (Begun, 2003, p. 642).

Evidence supporting each of these theories exists, to some degree;

each theory leads to the development of a different type of
prevention or intervention response. The “best” interventions will be
informed by theories with the strongest evidence or will integrate
elements from multiple evidence-supported theories.

Etiology research is often about understanding the mechanisms

underlying the phenomena of interest. The questions are “how”
questions—how does this happen (or not)? For example, scientists
asked the question: how do opioid medications (used to manage
pain) act on neurons compared to opioids that naturally occur in the

Module 2 Chapter 1: The Nature of Social Work Research Questions | 89

brain (Stoeber et al., 2018)? They discovered that opioid medications
used to treat pain bind to receptors insiden erve cells, which is a
quite different mechanism than the conventional wisdom that they
behave the same way that naturally occurring (endogenous) opioids
do—binding only on the surface of nerve cells. Understanding this
mechanism opens new options for developing pain relievers that are
less- or non-addicting than current opioid medicines like morphine
and oxycodone. Once these mechanisms of change are understood,
interventions can be developed, then tested through intervention
research approaches.

Intervention Research. Interventions are designed around identified

needs: epidemiology research helps to support intervention design
by identify the needs. Epidemiology research also helps identify
theories concerning the causes and factors affecting social work
problems. Intervention development is further supported by later
theory-testing and etiology research. However, developing an
intervention is not sufficient: interventions need to be tested and
evaluated to ensure that they are (1) safe, (2) effective, and (3) cost-
efficient to deliver. This is where intervention research comes into
play. Consider the example of Motivational Interviewing (MI)
approaches to addressing client ambivalence about engaging in a
behavior change effort. Early research concerning MI addressed
questions about its effectiveness. For example, a meta-analytic
review reported that “MI should be considered as a treatment for
adolescent substance abuse” because the evidence demonstrated
small, but significant effect sizes, and that the treatment gains were
retained over time (Jensen et al., 2011). Subsequently, when its safety
and effectiveness were consistently demonstrated through this kind
of evidence, investigators assessed MI as cost-efficient or cost-
effective. For example, MI combined with providing feedback was
demonstrated to be cost-effective in reducing drinking among

90 | Module 2 Chapter 1: The Nature of Social Work Research Questions

college students who engaged in heavy drinking behavior (Cowell et
al., 2012).

Intervention research not only is concerned with the outcomes of

delivering an intervention, but may also address the mechanisms of
change through which an intervention has its effects—not only what
changes happen, but how they happen. For example, investigators are
exploring how psychotherapy works, moving beyond demonstrating
that it works (Ardito & Rabellino, 2011; Kazdin, 2007; Wampold, 2015).
One mechanism that has garnered attention is the role of therapeutic
alliance—the relationships, bonds, and interactions that occur in the
context of treatment—on treatment outcomes.

Therapeutic alliance is one common factor identified across

numerous types of effective psychotherapeutic approaches
(Wampold, 2015). Authors summarizing a number of studies about
therapeutic alliance and its positive relationship to treatment
outcomes concluded that the quality of therapeutic alliance may be

Module 2 Chapter 1: The Nature of Social Work Research Questions | 91

a more powerful predictor of positive outcome than is the nature or
type of intervention delivered (Ardito & Rabellino, 2011). However, it
is important to determine the extent to which (a) therapeutic alliance
enhances clients’ symptom improvement, (b) gradual improvements
in symptoms lead to enhanced therapeutic alliance, or (c) the
relationship between therapeutic alliance and symptom
improvement are iterative—they go back and forth, influencing each
other over time (Kazdin, 2007).

Implementation Science. Social work and other disciplines have

produced a great deal of evidence about “what works” for intervening
around a great number of social work problems. Unfortunately, many
best practices with this kind of evidence support are slow to become
common practices. Implementation science is about understanding
facilitators and barriers to these evidence-supported interventions
becoming adopted into routine practice: characteristics of the
interventions themselves, conditions and processes operating in the
organizations where interventions are implemented, and factors
external to these organizations all influence practitioners’ adoption
of evidence supported interventions.

Even under optimal internal organizational conditions,

implementation can be undermined by changes in organizations’
external environments, such as fluctuations in funding, adjustments
in contracting practices, new technology, new legislation, changes in
clinical practice guidelines and recommendations, or other
environmental shifts” (Birken, et al, 2017).

Research for/about Research. In addition, social work investigators

92 | Module 2 Chapter 1: The Nature of Social Work Research Questions

engage in research that is specifically about scientific methodology.
This is where advances in measurement, participant recruitment and
retention, and data analysis emerge. The results of these kinds of
research studies are used to improve the research in basic,
intervention, and implementation research. Later in the course you
will see some of these products in action as we learn about best
practices in research and evaluation methodology. Here are a few
examples related to measurement methods:

• Concept mapping to assess community needs of sexual minority

youth (Davis, Saltzburg, & Locke, 2010)
• Field methodologies for measuring college student drinking in
natural environments (Clapp et al., 2007)
• Intergenerational contact measurement (Jarrott, Weaver,
Bowen, & Wang, 2018)
• Perceived Social Competence Scale-II (Anderson-Butcher et al.,
2016)
• Safe-At-Home Instrument to measure readiness to change
intimate partner violence behavior (Begun et al., 2003; 2008;
Sielski, Begun, & Hamel, 2015)
• Teamwork Scale for Youth (Lower, Newman, & Anderson-
Butcher, 2016)

And, here are a few examples related to involving participants in

research studies:

• Conducting safe research with at risk populations (Kyriakakis,

Waller, Kagotho, & Edmond, 2015)
• Recruitment strategies for non-treatment samples in addiction
studies (Subbaraman et al., 2015)
• Variations in recruitment results across Internet platforms
(Shao et al., 2015)

Module 2 Chapter 1: The Nature of Social Work Research Questions | 93

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=109

Research Questions

In this section, we take a closer look at research questions and their

relationship to the types of research conducted by investigators.
It may be easier to understand research questions by first ruling
out what are notresearch questions. In that spirit, let’s begin with
examples of questions where applying research methods will not help
to find answers:

• Trauma informed education.The first issue with this example is

obvious: it is not worded as a question. The second is critically
important: this is a general topic, it is not a research question.
This topic is too vague and broad making it impossible to
determine what answers would look like or how to approach
finding answers.
• How is my client feeling about what just happened?This type of
question about an individual is best answered by asking clinical
questions of that individual, within the context of the

94 | Module 2 Chapter 1: The Nature of Social Work Research Questions

 therapeutic relationship, not by consulting research literature
or conducting a systematic research study.
• Will my community come together in protest of a police-involved
shooting incident?This type of question may best be answered
by waiting to see what the future brings. Research might offer a
guess based on data from how other communities behaved in
the past but cannot predict how groups in individual situations
will behave. A better research question might be: What factors
predict community protest in response to police-involved
shooting incidents?
• Should I order salad or soup to go with my sandwich?This type of
question is not of general interest, making it a poor choice as a
research question. The question might be reframed as a general
interest question: Is it healthier to provide salad or soup along
with a sandwich? The answer to that researchable question
might inform a personal decision.
• Why divorce is bad for children.There are two problems with this
example. First, it is a statement, not a question, despite starting
with the word “why.” Second, this question starts out with a
biased assumption—that divorce isbad for children. Research
questions should support unbiased investigation, leading to
evidence and answers representative of what exists rather than
what someone sets out wanting to prove is the case. A better
research question might be: How does divorce affect children?

Module 2 Chapter 1: The Nature of Social Work Research Questions | 95

Tuning back to our first example of what is not a research question,
consider several possible school social work research questions
related to that general topic:

• To what extent do elementary school personnel feel prepared to

engage in trauma informed education with their students?
• What are the barriers and facilitators of integrating trauma
informed education in middle school?
• Does integrating trauma informed education result in lower
rates of suicidal ideation among high school students?
• Is there a relationship between parent satisfaction and the
implementation of trauma informed education in their children’s
schools?
• Does implementing trauma informed education in middle
schools affect the rate of student discipline referrals?

What is the difference between these research questions and the

earlier “not research” questions? First, research questions are
specific. This is an important distinction between identifying a topic
of interest (e.g., trauma informed education) and asking a
researchable question. For example, the question “How does divorce
affect children?” is not a good research question because it remains
too broad. Instead, investigators might focus their research questions
on one or two specific effects of interest, such as emotional or mental
health, academic performance, sibling relationships, aggression,
gender role, or dating relationship outcomes.

96 | Module 2 Chapter 1: The Nature of Social Work Research Questions

Related to a question being “researchable” is its feasibility for study.
Being able to research a question requires that appropriate data can
be collected with integrity. For example, it may not be feasible to
study what would happen if every child was raised by two parents,
because (a) it is impossible to study every child and (2) this reality
cannot ethically be manipulated to systematically explore it. No one
can ethically conduct a study whereby children are randomly
assigned by study investigators to the compared conditions of being
raised by two parents versus being raised by one or no parents.
Instead, we settle for observing what has occurred naturally in
different families.

Second, “good” research questions are relevant to knowledge

building. For this reason, the question about what to eat was not
a good research question—it is not relevant to others’ knowledge
development. Relevance is in the “eye of the beholder,” however. A
social work researcher may not see the relevance of using a 4-item
stimulus array versus a 6-item stimulus array in testing children’s
memory, but this may be an important research question for a

Module 2 Chapter 1: The Nature of Social Work Research Questions | 97

cognitive psychology researcher. It may, eventually, have implications
for assessment measures used in social work practice.

Third, is the issue of bias built into research questions. Remembering

that investigators are a product of their own developmental and
social contexts, what they choose to study and how they choose to
study it are socially constructed. An important aspect at the heart of
social work research relates to a question’s cultural appropriateness
and acceptability. To demonstrate this point, consider an era (during
the 1950s to early 1970s) when research questions were asked about
the negative effects on child development of single-parent, black
family households compared to two-parent, white family households
in America. This “majority comparison” frame of reference is not
culturally appropriate or culturally competent. Today, in social work,
we adopt a strengths perspective, and avoid making comparisons
of groups against a majority model. For example, we might ask
questions like: What are the facilitators and barriers of children’s
positive development as identified by single parents of diverse racial/
ethnic backgrounds? What strengths do African American parents
bring to the experience of single-parenting and how does it shape
their children’s development? What are the similar and different
experiences of single-parenting experienced by families of different
racial/ethnic composition?

98 | Module 2 Chapter 1: The Nature of Social Work Research Questions

Research Questions versus Research Hypotheses. You have now
seen examples of “good” research questions. Take, for example, the
last one we listed about trauma informed education:

Does implementing trauma informed education in middle

schools affect the rate of student discipline referrals?

Based on a review of literature, practice experience, previous

research efforts, and the school’s interests, an investigator may be
prepared to be even more specific about the research question (see
Figure 1-3). Assume that these sources led the investigator to believe
that implementing the trauma informed education approach will have
the effect of reducing the rate of disciplinary referrals. The
investigator may then propose to test the following hypothesis:

Implementing trauma informed education in middle schools

Module 2 Chapter 1: The Nature of Social Work Research Questions | 99

 will result in a reduction in the number of student discipline
referrals.

The research hypothesis is a clear statement that can be tested with

quantitative data and will either be rejected or not, depending on
the evidence. Research hypotheses are predictions about study
results—what the investigator expects the results will show. The
prediction, or hypothesis, is based on theory and/or other evidence.
A study hypothesis is, by definition, quantifiable—the answer lies in
numerical data, which is why we do not generally see hypotheses in
qualitative, descriptive research reports.

Hypotheses are also specific to one question at a time. Thus, an

investigator would need to state and test a second hypothesis to
answer the question:

Is there a relationship between parent satisfaction and the

implementation of trauma informed education in their
children’s schools?

The stated hypothesis might be:

Parent satisfaction is higher in middle schools where trauma

informed education is implemented.

Figure 1-3. Increasing specificity from research topic to question to

hypothesis

100 | Module 2 Chapter 1: The Nature of Social Work Research Questions

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=109

Social Work Questions

It is difficult to find a simple way to characterize social work

research. The National Institutes of Health (NIH) described social
work research in the following way:

Module 2 Chapter 1: The Nature of Social Work Research Questions | 101

 Historically, social work research has focused on studies of
the individual, family, group, community, policy and/or
organizational level, focusing across the lifespan on prevention,
intervention, treatment, aftercare and rehabilitation of acute
and chronic conditions, including the effects of policy on social
work practice (OBSSR, 2003, p. 5).

For all the breadth expressed in this statement, it reflects only how
social work research relates to the health arena—it does not indicate
many other domains and service delivery systems of social work
influence:

• physical, mental, and behavioral health

• substance misuse/addiction and other addictive behaviors
• education
• workplace
• income/poverty
• criminal justice
• child and family welfare
• housing and food security/insecurity
• environmental social work
• intimate partner, family, and community violence
• and others.

In addition to breadth of topic, social work research is characterized

by its biopsychosocial nature. This means that social work
researchers not only pursue questions relating to biological,
psychological, and social context factors, but also questions relating
to their intersections and interactions. Related to this observation is
that social work not only addresses questions related to the multiple
social system levels, social work also addresses the ways multiple
levels intersect and interact (i.e., those levels represented in the NIH
statement about individuals, families, groups, communities,
organizations, and policy).

102 | Module 2 Chapter 1: The Nature of Social Work Research Questions

It is worth noting that research need not be conducted by social
workers to be relevant to social work–many disciplines and
professions contribute to the knowledge base which informs social
work practice (medicine, nursing, education, occupational therapy,
psychology, sociology, criminal justice, political science, economics,
and more). Authors of one social work research textbook summarize
the relevance issue in the following statement:

“To social workers, a relevant research question is one whose

answers will have an impact on policies, theories, or practices
related to the social work profession” (Grinnell & Unrau, 2014,
p. 46).

Social Work Research Questions and Specific Aims

The kinds of questions that help inform social work practice and
policy are relevant to understanding social work problems, diverse
populations, social phenomena, or interventions. Most social work
research questions can be divided into two general categories:
background questions and foreground questions. The major
distinction between these two categories relates to the specific aims
that emerge in relation to the research questions.

Background Questions. This type of question is answerable with a

fact or set of facts. Background questions are generally simple in
structure, and they direct a straightforward search for evidence. This
type of question can usually be formulated using the classic 5
question words: who, what, when, where, or why. Here are a few
examples of social work background questions related to the topic of
fetal alcohol exposure:

• Whois at greatest risk of fetal alcohol exposure?

• Whatare the developmental consequences of fetal alcohol
exposure?

Module 2 Chapter 1: The Nature of Social Work Research Questions | 103

 • Whenin gestation is the risk of fetal alcohol exposure greatest?
• Wheredo women get information about the hazards of drinking
during pregnancy?
• Whyis fetal alcohol exposure (FAE) presented as a spectrum
disorder, different from fetal alcohol syndrome (FAS)?

These kinds of questions direct a social worker to review literature

about human development, human behavior, the distribution of the
problem across populations, and factors that determine the nature of
a specific social work problem like fetal exposure to alcohol. Where
the necessary knowledge is lacking, investigators aim to explore or
describe the phenomenon of interest. Many background questions
can be answered by epidemiology or etiology evidence.

Foreground Questions. This type of question is more complex than

the typical background question. Foreground questions typically are
concerned with making specific choices by comparing or evaluating
options. These types of questions required more specialized evidence
and may lead to searching different types of resources than would be
helpful for answering background questions. Foreground questions
are dealt with in greater detail in our second course, SWK 3402 which

104 | Module 2 Chapter 1: The Nature of Social Work Research Questions

is about understanding social work interventions. A quick foreground
question example related to the fetal exposure to alcohol topic might
be:

Which is the best tool for screening pregnant women for alcohol
use with the aim of reducing fetal exposure, the T-ACE, TWEAK,
or AUDIT?

This type of question leads the social worker to search for evidence
that compares different approaches. These kinds of evidence are
usually found in comparative reviews, or require the practitioner
to conduct a review of literature, locating individual efficacy and
effectiveness studies. Where knowledge is found to be lacking,
investigators aim to experiment with different approaches or
interventions.

Three Question Types and Their Associated Research Aims

Important distinctions exist related to different types of background

questions. Consider three general categories of questions that social
workers might ask about populations, problems, and social
phenomena: exploratory, descriptive, and explanatory. The different
types of questions matter because the nature of the research
questions determines the specific aims and most appropriate
research approaches investigators apply in answering them.

Exploratory Research Questions.Social workers may find themselves

facing a new, emerging problem where there is little previously
developed knowledge available—so little, in fact, that it is premature
to begin asking any more complex questions about causes or
developing testable theories. Exploratory research questionsopen the
door to beginning understanding and are basic; answers would help

Module 2 Chapter 1: The Nature of Social Work Research Questions | 105

build the foundation of knowledge for asking more complex
descriptive and explanatory questions. For example, in the early days
of recognition that HIV/AIDS was emerging as a significant public
health problem, it was premature to jump to questions about how
to treat or prevent the problem. Not enough was known about the
nature and scope of the problem, for whom it was a problem, how the
problem was transmitted, factors associated with risk for exposure,
what factors influenced the transition from HIV exposure to AIDS as
a disease state, and what issues or problems might co-occur along
with either HIV exposure or AIDS. In terms of a knowledge evolution
process, a certain degree of exploration had to occur before
intervention strategies for prevention and treatment could be
developed, tested, and implemented.

In 1981, medical providers, public health officials, and the Centers for
Disease Control and Prevention (CDC) began to circulate and publish
observations about a disproportionate, unexpectedly high incidence
rate of an unusual pneumonia and Kaposi’s sarcoma appearing in
New York City and San Francisco/California among homosexual men
(Curran, & Jaffe, 2011). As a result, a task force was formed and
charged with conducting an epidemiologic investigation of this
outbreak; “Within 6 months, it was clear that a new, highly
concentrated epidemic of life threatening illness was occurring in
the United States” (Curran & Jaffe, 2011, p. 65). The newly recognized
disease was named for its symptoms: acquired immune deficiency
syndrome, or AIDS. Exploratory research into the social networks
of 90 living patients in 10 different cities indicated that 40 had a

106 | Module 2 Chapter 1: The Nature of Social Work Research Questions

sexual contact link with another member of the 90-patient group
(Auerbach, Darrow, Jaffe, & Curran, 1984). Additionally, cases were
identified among persons who had received blood products related
to their having hemophilia, persons engaged in needle sharing during
substance use, women who had sexual contact with a patient, and
infants born to exposed women. Combined, these pieces of
information led to an understanding that the causal infectious factor
(eventually named the human immunodeficiency virus, HIV) was
transmitted by sexual contact, blood, and placental connection. This,
in turn, led to knowledge building activities to develop both
preventive and treatment strategies which could be implemented
and studied. Social justice concerns relate to the slow rate at which
sufficient resources were committed for evolving to the point of
effective solutions for saving lives among those at risk or already
affected by a heavily stigmatized problem.

The exploratory research approaches utilized in the early HIV/AIDS

studies were both qualitative and quantitative in nature. Qualitative
studies included in-depth interviews with identified
patients—anthropological and public health interviews about many
aspects of their living, work, and recreational environments, as well
as many types of behavior. Quantitative studies included
comparisons between homosexually active men with and without
the diseases of concern. In addition, social network study methods
combined qualitative and quantitative approaches. These examples
of early exploratory research supported next steps in knowledge
building to get us to where we are today. “Today, someone diagnosed
with HIV and treated before the disease is far advanced can live
nearly as long as someone who does not have HIV” (hiv.gov). While
HIV infection cannot (yet) be “cured,” it can be controlled and
managed as a chronic condition.

Module 2 Chapter 1: The Nature of Social Work Research Questions | 107

Descriptive Research Questions. Social workers often ask for
descriptions about specific populations, problems, processes, or
phenomena. Descriptive research questions might be expressed in
terms of searching to create a profile of a group or population, create
categories or types (typology) to describe elements of a population,
document facts that confirm or contradict existing beliefs about a
topic or issue, describe a process, or identify steps/stages in a
sequential process (Grinnell & Unrau, 2014). Investigators may elect
to approach the descriptive question using qualitative methods that
result in a rich, deep description of certain individuals’ experiences or
perceptions (Yegidis, Weinbach, & Meyers, 2018). Or, the descriptive
question might lead investigators to apply quantitative methods,
assigning numeric values, measuring variables that describe a
population, process, or situation of interest. In descriptive research,
investigators do not manipulate or experiment with the variables;
investigators seek to describe what naturally occurs (Yegidis,
Weinbach, & Meyers, 2018). As a result of studies answering
descriptive questions, tentative theories and hypotheses may be
generated.

Here are several examples of descriptive questions.

• How do incarcerated women feel about the option of

medication-assisted treatment for substance use disorders?
• What barriers to engaging in substance misuse treatment do
previously incarcerated persons experience during community
reentry?
• How often do emerging adults engage in binge drinking in
different drinking contexts (e.g., bars, parties, sporting events, at
home)?
• What percent of incarcerated adults experience a substance use
disorder?
• What is the magnitude of racial/ethnic disparities in access to
treatment for substance use disorders?

108 | Module 2 Chapter 1: The Nature of Social Work Research Questions

 • Who provides supervision or coordination of services for aging
adults with intellectual or other developmental disabilities?
• What is the nature of the debt load among students in doctoral
social work programs?

An example of descriptive research, derived from a descriptive

question, is represented in an article where investigators addressed
the question: How is the topic of media violence and aggression
reported in print media (Martins et al., 2013)? This question led the
investigators to conduct a qualitative content analysis, resulting in
a description showing a shift in tone where earlier articles (prior to
2000) emphasized the link as a point of concern and later articles
(since 2000) assumed a more neutral stance.

Correlational Research Questions. One important type of descriptive

question asks about relationships that might exist between
variables—looking to see if variable x and variable y are associated
or correlated with each other. This is an example of a correlational

Module 2 Chapter 1: The Nature of Social Work Research Questions | 109

research question; it does not indicate whether “x” causes “y” or
“y” causes “x”, only whether these two are related. Consider again
the topic of exposure to violence in the media and its relationship
to aggression. A descriptive question asked about the existence of
a relationship between exposure to media violence (variable x) and
children’s expression of aggression (variable y). Investigators
reported one study of school-aged children, examining the
relationship between exposure to three types of media violence
(television, video games, and movies/videos) and three types of
aggression (verbal, relational, and physical; Gentile, Coyne, & Walsh,
2011). The study investigators reported that media violence exposure
was, indeed, correlated with all three types of aggressive behavior
(and less prosocial behavior, too).

For a positive correlation (the blue line), as the value of the “x”
variable increases, so does the value of the “y” variable (see Figure
1-4 for a general demonstration). An example might be as age or
grade in school increases (“x”), so does the number of preadolescent,
adolescent, and emerging adults who have used alcohol (“y”). For a
negative correlation (the orange line), as the value of the “x” variable
increases, the value of the “y” variable decreases. An example might
be as the number of weeks individuals are in treatment for depression
symptoms (“x”), the reported depression symptoms decreases (“y”).
The neutral of non-correlation line (grey) means that the two
variables, “x” and “y” do not have an association with each other.
For example, number of years of teachers’ education (“x”) might be
unrelated to the number of students dropping out of high school (“y”).

Figure 1-4. Depicting positive, negative, and neutral correlation lines

110 | Module 2 Chapter 1: The Nature of Social Work Research Questions

Descriptive correlational studies are sometimes called comparison
studies because the descriptive question is answered by comparing
groups that differ on one of the variables (low versus high media
violence exposure) to see how they might differ on the other variable
(aggressive behavior).

Explanatory Research Questions. To inform the design of evidence-

informed interventions, social workers need answers to questions
about the nature of the relationships between potentially influential
factors or variables. An explanatory research question might be
mapped as: Does variable x cause, lead to or prevent changes in
variable y (Grinnell & Unrau, 2014)? These types of questions often
test theory related to etiology.

Comparative research might provide information about a

Module 2 Chapter 1: The Nature of Social Work Research Questions | 111

relationship between variables. For example, the difference in
outcomes between persons experiencing a substance use disorder
and have been incarcerated compared to others with the same
problem but have not been incarcerated may be related to their
employability and ability to generate a living-wage income for
themselves and their families. However, to develop evidence-
informed interventions, social workers need to know that variables
are not only related, but that one variable actually plays a causal
role in relation to the other. Imagine, for example, that evidence
demonstrated a significant relationship between adolescent self-
esteem and school performance. Social workers might spend a great
deal of effort developing interventions to boost self-esteem in hopes
of having a positive impact on school performance. However, what
if self-esteem comes from strong school performance? The self-
esteem intervention efforts will not likely have the desired effect
on school performance. Just because research demonstrates a
significant relationship between two variables does not mean that
the research has demonstrated a causal relationship between those
variables. Investigators need to be cautious about the extent to which
their study designs can support drawing conclusions about causality;
anyone reviewing research reports also needs to be alert to where
causal conclusions are properly and improperly drawn.

112 | Module 2 Chapter 1: The Nature of Social Work Research Questions

The questions that drive intervention and evaluation research studies
are explanatory in nature: does the intervention (x) have a significant
impact on outcomes of interest (y)? Another type of explanatory
question related to intervention research concerns the mechanisms
of change. In other words, not only might social workers be
interested to find out what outcomes or changes can be attributed
to an intervention, they may also be interested to learn how the
intervention causes those changes or outcomes.

Module 2 Chapter 1: The Nature of Social Work Research Questions | 113

Cartoon used with permission, ©Sidney Harris

114 | Module 2 Chapter 1: The Nature of Social Work Research Questions

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=109

Chapter Summary

In this chapter, you learned about different aspects of the knowledge

building process and where different types of research questions
might fit into the big picture. No single research study covers the
entire spectrum; each study contributes a piece of the puzzle as a
whole. Research questions come in many different forms and several
different types. What is important to recall as we move through
the remainder of the course is that the decisions investigators make
about research approaches, designs, and procedures all start with
the nature of the question being asked. And, the questions being
asked are influenced by multiple factors, including what is previously
known and remains unknown, the culture and context of the
questioners, and what theories they have about what is to be studied.
That leads us to the next chapter.

Module 2 Chapter 1: The Nature of Social Work Research Questions | 115

Module 2 Chapter 2: The Link
Between Theory, Research,
and Social Justice
Theory has been mentioned several times in Chapter 1 discussions.
In this chapter, we explore the relationship between theory and
research, paying particular attention to how theory and research
relate to promoting social justice.

In this chapter you will read about:

• Why theory matters to social work

• How theory and research relate to social justice

The Significance of Theory

It is helpful to begin with thinking about what theory is. Theory is

defined as a belief, idea, or set of principles that explain
something—the set of principles may be organized in a complex
manner or may be quite simple in nature. Theories are not facts;
they are conjectures and predictions that need to be tested for their
goodness-of-fit with reality. A scientific theory is an explanation
supported by empirical evidence. Therefore, scientific theory is
based on careful, reasonable examination of facts, and theories are
tested and confirmed for their ability to explain and predict the
phenomena of interest.

Theory is central to the development of social work interventions,

as it determines the nature of our solutions to identified problems.
Consider an example whereby programmatic and social policy

116 | Module 2 Chapter 2: The Link

Between Theory, Research, and Social
Justice
responses might be influenced by the way that a problem like teen
pregnancy is defined and the theories about the problem. In Table 2-1
you can see different definitions or theories of the problem on the
left, and the logical responses on the right. In many cases, the boxes
on the right can also be supplemented with content from other boxes
addressing other problem definitions or theories.

Table 2-1. Analysis of teen pregnancy: How defining a problem

determines responses

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 117
Problem definition/
Policy response option
theory
Teen pregnancy is
No need to respond
not a problem
Keep them otherwise occupied allowing no
Teen girls are having opportunity; keep them separated from potential
sex sex partners (including sexual abuse); teach/
preach "Just Say No!"

Make birth control available, accessible, affordable,

Teen girls are and palatable; keep them separated from boys and
getting pregnant men capable of impregnating them (including
sexual abuse)
Teen girls are having Terminate teen pregnancies; prevent
babies pregnancy—see above

Babies born to teen Provide resources for healthy prenatal

mothers are less development; prevent pregnancy/having
healthy babies—see above

Teen mothers have

Provide health-promoting resources to teen
poorer health than
mothers; prevent pregnancy/having babies—see
other teens, other
above
mothers

Provide resources to promote positive parenting by

Teen mothers are teen mothers (or all parents); place babies with
raising babies other parents (other family members or foster/
inadequately adoptive families); prevent pregnancy/having
babies—see above
Being a teen mother
interrupts her Provide resources to promote positive educational
education and and vocational/income producing outcomes for
future teen mothers; prevent teen mothers raising babies,
employability/ prevent pregnancy/having babies—see above
income potential

Hopefully, through this example, you can see how the way we define
a problem and our theories about its causes determines the types of
solutions we develop. Solutions are also dependent on whether they
are feasible, practical, reasonable, ethical, culturally appropriate, and
politically acceptable.

Theory, Research, and Social Justice

118 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
Theory is integral to research and research is integral to theory.
Theory guides the development of many research questions and
research helps generate new theories, as well as determining
whether support for theories exists. What is important to remember
is that theory is not fact: it is a belief about how things work; theory is
a belief or “best guess” that awaits the support of empirical evidence.

“The best theories are those that have been substantiated or

validated in the real world by research studies” (Dudley, 2011, p.
6).

Countless theories exist, and some are more well-developed and

well-defined than others. More mature theories have had more time
and effort devoted to supporting research, newer theories may be
more tentative as supporting evidence is being developed. Theories
are sometimes disproven and need to be scrapped completely or
heavily overhauled as new research evidence emerges. And,
exploratory research leads to the birth of new theories as new
phenomena and questions arise, or as practitioners discover ways
that existing theory does not fit reality.

Examples of theories and theoretical models with which you may

have become familiar in other coursework are developed, tested,
and applied in research from multiple disciplines, including social
work. You may be familiar with the concepts of multidisciplinary and
interdisciplinary practice, research and theory, but you also might
be interested to learn the concept of transdisciplinary research and
theory. The social work profession engages in all three, as described
in Figure 2-1.

Figure 2-1. Comparison of multidisciplinary, interdisciplinary, and

transdisciplinary as concepts

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 119
An example of transdisciplinarity is demonstrated in motivational
interviewing. The principles and practice of motivational
interviewing skills transcends disciplines—it is relevant and effective
regardless of a practitioner’s discipline and regardless of which
discipline is conducting research that provides supporting evidence.
An example of interdisciplinarity is when social workers, nurses,
pediatricians, occupational therapists, respiratory therapists, and
physical therapists work together to create a system for reducing
infants’ and young children’s environmental exposure to heavy metal
contamination in their home, childcare, recreational, food and water
(e.g., lead and mercury). An example of multidisciplinarity is when
a pediatrician, nurse, occupational therapist, social worker, and
physical therapist each deliver discipline-specific services to children
with intellectual disabilities.

Here are examples of theories that you may have encountered or

eventually will encounter in your career:

• behavioral theory

120 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
 • cognitive theory
• conflict theory
• contact theory (groups)
• critical race theory
• developmental theory (families)
• developmental theory (individuals)
• feminist theory
• health beliefs model
• information processing theory
• learning theory
• lifecourse or lifespan theory
• neurobiology theories
• organizational theory
• psychoanalytic theory
• role theory
• social capital model
• social ecological theory
• social learning theory
• social network theory
• stress vulnerability model
• systems theory
• theory of reasoned action/planned behavior
• transtheoretical model of behavior change

Social work practitioners generate new theories in the field all the
time, but these theories are rarely documented or systematicallyl
tested. Applying systematic research methods to test these practice-
generated theories can help expand the social work knowledge base.
Well-developed theories must be testable through the application
of research methods. Furthermore, they both supplement and
complement other theories that have the support of strong evidence
behind them. In addition to these points, for theories to be relevant
to social work, they need to:

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 121
 • have practice implications at one or more level of
intervention—suggest principles of action;
• be responsive to human diversity;
• contribute to the promotion of social justice (Dudley, 2011).

Social Justice and The Grand Challenges for Social Work. In 2016,
the American Academy of Social Work & Social Welfare (AASWSW)
rolled out a set of 12 initiatives, challenging the social work profession
to develop strategies for having a significant impact on a broad set
of problems challenging the nation: “their introduction truly has the
potential to be a defining moment in the history of our profession”
(Williams, 2016, p. 67). The Grand Challenges directly relate to
content presented in the Preamble to the Code of Ethics of the
National Association of Social Workers (2017):

The primary mission of the social work profession is to enhance

human well-being and help meet the basic human needs of
all people, with particular attention to the needs and
empowerment of people who are vulnerable, oppressed, and
living in poverty. A historic and defining feature of social work
is the profession’s dual focus on individual well-being in a
social context and the well-being of society. Fundamental to
social work is attention to the environmental forces that create,
contribute to, and address problems in living (p. 1).

The ambitious Grand Challenges (http://aaswsw.org/grand-

challenges-initiative/12-challenges/) call for a synthesis of research
and evidence generating endeavors, social work education, and social
work practice (at all levels), with promoting social justice and
transforming society at the forefront of attention. None of the Grand
Challenges can be achieved without collaboration between social
work researchers, practitioners, key stakeholders/constituents,
policy makers, and members of other professions and disciplines
(see Table 2-2). Each of the papers published under the 12 umbrella

122 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
challenges not only reviews evidence related to the challenge, but
also identifies a research and action agenda for promoting change
and achieving the stated challenge goals.

Table 2-2. 12 Grand Challenges for Social Work

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 123
Grand Challenge Topics Under the Challenges

Ensure healthy • Prevention of schizophrenia and severe

development for all mental illness
youths • Unleashing the power of prevention

• Reducing and preventing alcohol misuse and

its consequences
• Health equity: Eradicating health inequalities
Close the health gap
for future generations
• Strengthening health care systems: Better
health across America

• Ending gender-based violence

Stop family violence • Safe children: Reducing severe and fatal
maltreatment

• Increasing productive engagement in later

life
Advance long and
• Increasing productive engagement early in
productive lives
life: Civic and volunteer service as a pathway
to development

Eradicate social • Social isolation presents a grand challenge to

isolation social work

End homelessness • End homelessness

Create social
• Strengthening the social response to the
responses to a
human impacts of environmental change
changing environment

124 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
Grand Challenge Topics Under the Challenges

• Practice innovation through technology in

Harness technology
the digital age
for social good
• Harnessing big data for social good

Promote smart • From mass incarceration to smart

decarceration decarceration

Reduce extreme
• Reversing extreme inequality
economic inequality

Build financial
• Financial capability and asset building for all
capability for all

• Promoting equality by addressing social

stigma
• Achieving equal opportunity and justice: The
Achieve equal
integration of Latina/o immigrants into
opportunity and
American society
justice
• Increasing success for African American
children and youth
• Fair housing and inclusive communities

Theory and its related research is presented in the literature of social

work and other disciplines. Being able to locate the relevant literature
is an important skill for social work professionals and researchers
to master. The next chapter introduces some basic principles about
identifying literature that informs us about theory and the research
that leads to the development of new theories and the testing of
existing theories.

Using Data to Support Social Justice Advocacy

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 125
It is one thing to collect data, statistics, and information about the
dimensions of a social problem; it is another to apply those data,
statistics, and information in action to promote social change around
an identified social justice cause. Advocacy is one tool or role
important to the social work profession since its earliest days. Data
and empirical evidence should routinely support social workers’
social justice advocacy efforts at the micro level—when working with
specific client systems (case advocacy). Social justice advocacy also is
a macro-level practice (cause advocacy), one that often involves the
use of data to raise awareness about a cause, establish change goals,
and evaluate the impact of change efforts. Consider, for example, the
impact of data on social justice advocacy related to opioid misuse
and addiction across the United States. Data regarding the sharp,
upward trend in opioid-related deaths have had a powerful impact
on public awareness, having captivated the attention of mass media
outlets. These are data from the Centers for Disease Control and
Prevention (CDC) for the period 1999 to 2016 (see Figure 2-2). The
death rate from opioid overdose increased remarkably in the most
recent years depicted, including heroin, natural and semisynthetic
(morphine, codeine, hydrocodone, oxycodone), and synthetic opioids
(fentanyl, fentanyl analogs, tramadol); the rate declined somewhat
for deaths related to methadone, a highly controlled prescription
medication used to treat opioid addiction (Hedegaard, Warner, &
Miniño, 2017). In 2016 the overdose death rate was more than triple
the 1999 rate and surpassed all other causes of death (including
homicide and automobile crashes) for persons aged 50 or younger
(Vestal, 2018). The good news is that the rates dropped significantly
across 14 states during 2017 as new policy approaches were
implemented—including more widespread use and access to
naloxone (an opioid overdose reversal drug). The bad news is that the
opioid overdose death rates increased significantly across 5 states
and the District of Columbia as fentanyl and related drugs became
integrated into the illicit drug supply: “Nationally, the death toll is

126 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
still rising” (Vestal, 2018, p. 1). The Ohio Department of Health, with
Ohio being one of the 5 states with more than a 30% increase in
opioid overdose deaths, now recommends that naloxone be used
more widely in overdose situations, whether or not opioids are
known to be involved.

Figure 2-2. Opioid drug overdose death rate trends 1999-2016, by

opioid category

Having data of this sort available to argue the urgency of a cause is

of tremendous value. Imagine the potential impact of data on social
issues such as child maltreatment, intimate partner violence, human
trafficking, and death by suicide. Imagine also the potential impact
of data concerning disparities in health, incarceration, mental and
traumatic stress disorders, and educational or employment
achievement across members of different racial/ethnic, age,

Module 2 Chapter 2: The Link Between Theory, Research, and Social

Justice | 127
economic, and national origin groups. This type of evidence has the
potential to be a powerful element in the practice of social justice
advocacy.

Follow links below to specific AASWSW

Grand Challenges and consider the
ways that research evidence is being
used to advocate for at least one of the
following social justice causes:

• Health Equity: Eradicating Health

Inequalities for Future Generations
• From Mass Incarceration to Smart
Decarceration
• Safe Children: Reducing Severe and
Fatal Maltreatment
• Ending Gender Based Violence: A Grand Challenge for Social
Work
• Prevention of Schizophrenia and Severe Mental Illness
• Increasing Productive Engagement in Later Life
• Strengthening the Social Responses to the Human Impacts of
Environmental Change

128 | Module 2 Chapter 2: The Link Between Theory, Research, and Social
Justice
Module 2 Chapter 3: What is
Empirical Literature & Where
can it be Found?
In Module 1, you read about the problem of pseudoscience. Here, we
revisit the issue in addressing how to locate and assess scientific or
empirical literature. In this chapter you will read about:

• distinguishing between what IS and IS NOT empirical literature

• how and where to locate empirical literature for understanding
diverse populations, social work problems, and social
phenomena.

Probably the most important take-home lesson from this chapter is

that one source is not sufficient to being well-informed on a topic.
It is important to locate multiple sources of information and to
critically appraise the points of convergence and divergence in the
information acquired from different sources. This is especially true
in emerging and poorly understood topics, as well as in answering
complex questions.

What Is Empirical Literature

Social workers often need to locate valid, reliable information

concerning the dimensions of a population group or subgroup, a
social work problem, or social phenomenon. They might also seek
information about the way specific problems or resources are
distributed among the populations encountered in professional
practice. Or, social workers might be interested in finding out about

Module 2 Chapter 3: What is Empirical

Literature & Where can it be
Found? | 129
the way that certain people experience an event or phenomenon.
Empirical literature resources may provide answers to many of these
types of social work questions. In addition, resources containing data
regarding social indicators may also prove helpful. Social indicators
are the “facts and figures” statistics that describe the social,
economic, and psychological factors that have an impact on the well-
being of a community or other population group.The United Nations
(UN) and the World Health Organization (WHO) are examples of
organizations that monitor social indicators at a global level:
dimensions of population trends (size, composition, growth/loss),
health status (physical, mental, behavioral, life expectancy, maternal
and infant mortality, fertility/child-bearing, and diseases like HIV/
AIDS), housing and quality of sanitation (water supply, waste
disposal), education and literacy, and work/income/unemployment/
economics, for example.

Three characteristics stand out in empirical literature compared to

other types of information available on a topic of interest: systematic
observation and methodology, objectivity, and transparency/
replicability/reproducibility. Let’s look a little more closely at these
three features.

130 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
Systematic Observation and Methodology.The hallmark of
empiricism is “repeated or reinforced observation of the facts or
phenomena” (Holosko, 2006, p. 6). In empirical literature, established
research methodologies and procedures are systematically applied to
answer the questions of interest.

Objectivity.Gathering “facts,” whatever they may be, drives the

search for empirical evidence (Holosko, 2006). Authors of empirical
literature are expected to report the facts as observed, whether or
not these facts support the investigators’ original hypotheses.
Research integrity demands that the information be provided in an
objective manner, reducing sources of investigator bias to the
greatest possible extent.

Transparency and Replicability/Reproducibility. Empirical

literature is reported in such a manner that other investigators
understand precisely what was done and what was found in a
particular research study—to the extent that they could replicate
the study to determine whether the findings are reproduced when
repeated. The outcomes of an original and replication study may
differ, but a reader could easily interpret the methods and
procedures leading to each study’s findings.

What is NOT Empirical Literature

By now, it is probably obvious to you that literature based on

“evidence” that is not developed in a systematic, objective,
transparent manner is not empirical literature. On one hand, non-
empirical types of professional literature may have great significance
to social workers. For example, social work scholars may produce
articles that are clearly identified as describing a new intervention or
program without evaluative evidence, critiquing a policy or practice,
or offering a tentative, untested theory about a phenomenon. These

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 131
resources are useful in educating ourselves about possible issues or
concerns. But, even if they are informed by evidence, they are not
empirical literature. Here is a list of several sources of information
that do not meet the standard of being called empirical literature:

• your course instructor’s lectures

• political statements
• advertisements
• newspapers & magazines (journalism)
• television news reports & analyses (journalism)
• many websites, Facebook postings, Twitter tweets, and blog
postings
• the introductory literature review in an empirical article
• textbooks

You may be surprised to see the last two included in this list. Like
the other sources of information listed, these sources also might lead
you to look for evidence. But, they are not themselves sources of
evidence. They may summarize existing evidence, but in the process
of summarizing (like your instructor’s lectures), information is
transformed, modified, reduced, condensed, and otherwise
manipulated in such a manner that you may not see the entire,
objective story. These are called secondary sources, as opposed to
the original, primary source of evidence. In relying solely on
secondary sources, you sacrifice your own critical appraisal and
thinking about the original work—you are “buying” someone else’s
interpretation and opinion about the original work, rather than
developing your own interpretation and opinion. What if they got it
wrong? How would you know if you did not examine the primary
source for yourself? Consider the following as an example of “getting
it wrong” being perpetuated.

132 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
Example: Bullying and School Shootings. One result of the heavily
publicized April 1999 school shooting incident at Columbine High
School (Colorado), was a heavy emphasis placed on bullying as a
causal factor in these incidents (Mears, Moon, & Thielo, 2017),
“creating a powerful master narrative about school shootings”
(Raitanen, Sandberg, & Oksanen, 2017, p. 3). Naturally, with an
identified cause, a great deal of effort was devoted to anti-bullying
campaigns and interventions for enhancing resilience among youth
who experience bullying. However important these strategies might
be for promoting positive mental health, preventing poor mental
health, and possibly preventing suicide among school-aged children
and youth, it is a mistaken belief that this can prevent school
shootings (Mears, Moon, & Thielo, 2017). Many times the accounts
of the perpetrators having been bullied come from potentially
inaccurate third-party accounts, rather than the perpetrators
themselves; bullying was not involved in all instances of school
shooting; a perpetrator’s perception of being bullied/persecuted are
not necessarily accurate; many who experience severe bullying do
not perpetrate these incidents; bullies are the least targeted shooting
victims; perpetrators of the shooting incidents were often bullying
others; and, bullying is only one of many important factors associated
with perpetrating such an incident (Ioannou, Hammond, & Simpson,
2015; Mears, Moon, & Thielo, 2017; Newman &Fox, 2009; Raitanen,
Sandberg, & Oksanen, 2017). While mass media reports deliver
bullying as a means of explaining the inexplicable, the reality is not
so simple: “The connection between bullying and school shootings is
elusive” (Langman, 2014), and “the relationship between bullying and
school shooting is, at best, tenuous” (Mears, Moon, & Thielo, 2017, p.
940). The point is, when a narrative becomes this publicly accepted, it
is difficult to sort out truth and reality without going back to original
sources of information and evidence.

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 133
What May or May Not Be Empirical Literature: Literature Reviews

Investigators typically engage in a review of existing literature as

they develop their own research studies. The review informs them
about where knowledge gaps exist, methods previously employed
by other scholars, limitations of prior work, and previous scholars’
recommendations for directing future research. These reviews may
appear as a published article, without new study data being reported
(see Fields, Anderson, & Dabelko-Schoeny, 2014 for example). Or,
the literature review may appear in the introduction to their own
empirical study report. These literature reviews are not considered
to be empirical evidence sources themselves, although they may be
based on empirical evidence sources. One reason is that the authors
of a literature review may or may not have engaged in a
systematicsearch process, identifying a full, rich, multi-sided pool of
evidence reports.

There is, however, a type of review that applies systematic methods

and is, therefore, considered to be more strongly rooted in evidence:
the systematic review.

134 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
Systematic review of literature. A systematic reviewis a type of
literature report where established methods have been
systematically applied, objectively, in locating and synthesizing a
body of literature. The systematic review report is characterized
by a great deal of transparency about the methods used and the
decisions made in the review process, and are replicable. Thus, it
meets the criteria for empirical literature: systematic observation
and methodology, objectivity, and transparency/reproducibility. We
will work a great deal more with systematic reviews in the second
course, SWK 3402, since they are important tools for understanding
interventions. They are somewhat less common, but not unheard of,
in helping us understand diverse populations, social work problems,
and social phenomena.

Locating Empirical Evidence

Social workers have available a wide array of tools and resources for
locating empirical evidence in the literature. These can be organized
into four general categories.

Journal Articles. A number of professional journals publish articles

where investigators report on the results of their empirical studies.
However, it is important to know how to distinguish between
empirical and non-empirical manuscripts in these journals. A key
indicator, though not the only one, involves a peer review process.
Many professional journals require that manuscripts undergo a
process of peer review before they are accepted for publication. This
means that the authors’ work is shared with scholars who provide
feedback to the journal editor as to the quality of the submitted
manuscript. The editor then makes a decision based on the reviewers’
feedback:

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 135
 • Accept as is
• Accept with minor revisions
• Request that a revision be resubmitted (no assurance of
acceptance)
• Rejection

When a “revise and resubmit” decision is made, the piece will go

back through the review process to determine if it is now acceptable
for publication and that all of the reviewers’ concerns have been
adequately addressed. Editors may also reject a manuscript because
it is a poor fit for the journal, based on its mission and audience,
rather than sending it for review consideration.

Indicators of journal relevance.Various journals are not equally

relevant to every type of question being asked of the literature.
Journals may overlap to a great extent in terms of the topics they
might cover; in other words, a topic might appear in multiple
different journals, depending on how the topic was being addressed.
For example, articles that might help answer a question about the

136 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
relationship between community poverty and violence exposure
might appear in several different journals, some with a focus on
poverty, others with a focus on violence, and still others on
community development or public health. Journal titles are
sometimes a good starting point but may not give a broad enough
picture of what they cover in their contents.

In focusing a literature search, it also helps to review a journal’s

mission and target audience. For example, at least four different
journals focus specifically on poverty:

• Journal of Children & Poverty

• Journal of Poverty
• Journal of Poverty and Social Justice
• Poverty & Public Policy

Let’s look at an example using the Journal of Poverty and Social

Justice. Information about this journal is located on the journal’s
webpage: http://policy.bristoluniversitypress.co.uk/journals/
journal-of-poverty-and-social-justice. In the section headed “About
the Journal” you can see that it is an internationally focused research
journal, and that it addresses social justice issues in addition to
poverty alone. The research articles are peer-reviewed (there appear
to be non-empirical discussions published, as well). These
descriptions about a journal are almost always available, sometimes
listed as “scope” or “mission.” These descriptions also indicate the
sponsorship of the journal—sponsorship may be institutional (a
particular university or agency, such as Smith College Studies in Social
Work), a professional organization, such as the Council on Social
Work Education (CSWE) or the National Association of Social Work
(NASW), or a publishing company (e.g., Taylor & Frances, Wiley, or
Sage).

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 137
Indicators of journal caliber. Despite engaging in a peer review
process, not all journals are equally rigorous. Some journals have
very high rejection rates, meaning that many submitted manuscripts
are rejected; others have fairly high acceptance rates, meaning that
relatively few manuscripts are rejected. This is not necessarily the
best indicator of quality, however, since newer journals may not be
sufficiently familiar to authors with high quality manuscripts and
some journals are very specific in terms of what they publish.
Another index that is sometimes used is the journal’s impact factor.
Impact factor is a quantitative number indicative of how often
articles published in the journal are cited in the reference list of other
journal articles—the statistic is calculated as the number of times
on average each article published in a particular year were cited
divided by the number of articles published (the number that could
be cited). For example, the impact factor for the Journal of Poverty
and Social Justicein our list above was 0.70 in 2017, and for the Journal
of Povertywas 0.30. These are relatively low figures compared to a
journal like the New England Journal of Medicinewith an impact factor
of 59.56! This means that articles published in that journal were, on
average, cited more than 59 times in the next year or two.

Impact factors are not necessarily the best indicator of caliber,

however, since many strong journals are geared toward practitioners
rather than scholars, so they are less likely to be cited by other
scholars but may have a large impact on a large readership. This may
be the case for a journal like the one titled Social Work, the official
journal of the National Association of Social Workers. It is distributed
free to all members: over 120,000 practitioners, educators, and
students of social work world-wide. The journal has a recent impact
factor of.790. The journals with social work relevant content have
impact factors in the range of 1.0 to 3.0 according to Scimago Journal
& Country Rank (SJR), particularly when they are interdisciplinary
journals (for example, Child Development, Journal of Marriage and

138 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
Family, Child Abuse and Neglect, Child Maltreatment,Social Service
Review, and British Journal of Social Work). Once upon a time, a
reader could locate different indexes comparing the “quality” of social
work-related journals. However, the concept of “quality” is difficult
to systematically define. These indexes have mostly been replaced
by impact ratings, which are not necessarily the best, most robust
indicators on which to rely in assessing journal quality. For example,
new journals addressing cutting edge topics have not been around
long enough to have been evaluated using this particular tool, and
it takes a few years for articles to begin to be cited in other, later
publications.

Beware of pseudo-, illegitimate, misleading, deceptive, and suspicious

journals. Another side effect of living in the Age of Information is that
almost anyone can circulate almost anything and call it whatever they
wish. This goes for “journal” publications, as well. With the advent
of open-access publishing in recent years (electronic resources
available without subscription), we have seen an explosion of what
are called predatory or junk journals. These are publications calling
themselves journals, often with titles very similar to legitimate
publications and often with fake editorial boards. These
“publications” lack the integrity of legitimate journals. This caution
is reminiscent of the discussions earlier in the course about
pseudoscience and “snake oil” sales. The predatory nature of many
apparent information dissemination outlets has to do with how
scientists and scholars may be fooled into submitting their work,
often paying to have their work peer-reviewed and published. There
exists a “thriving black-market economy of publishing scams,” and at
least two “journal blacklists” exist to help identify and avoid these
scam journals (Anderson, 2017).

This issue is important to information consumers, because it creates

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 139
a challenge in terms of identifying legitimate sources and
publications. The challenge is particularly important to address when
information from on-line, open-access journals is being considered.
Open-access is not necessarily a poor choice—legitimate scientists
may pay sizeable fees to legitimate publishers to make their work
freely available and accessible as open-access resources. On-line
access is also not necessarily a poor choice—legitimate publishers
often make articles available on-line to provide timely access to the
content, especially when publishing the article in hard copy will be
delayed by months or even a year or more. On the other hand, stating
that a journal engages in a peer-review process is no guarantee of
quality—this claim may or may not be truthful. Pseudo- and junk
journals may engage in some quality control practices, but may lack
attention to important quality control processes, such as managing
conflict of interest, reviewing content for objectivity or quality of
the research conducted, or otherwise failing to adhere to industry
standards (Laine & Winker, 2017).

One resource designed to assist with the process of deciphering

legitimacy is the Directory of Open Access Journals (DOAJ). The DOAJ
is not a comprehensive listing of all possible legitimate open-access
journals, and does not guarantee quality, but it does help identify
legitimate sources of information that are openly accessible and meet
basic legitimacy criteria. It also is about open-access journals, not the
many journals published in hard copy.

An additional caution: Search for article corrections. Despite all of the

careful manuscript review and editing, sometimes an error appears
in a published article. Most journals have a practice of publishing
corrections in future issues. When you locate an article, it is helpful
to also search for updates. Here is an example where data presented

140 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
in an article’s original tables were erroneous, and a correction
appeared in a later issue.

• Marchant, A., Hawton, K., Stewart A., Montgomery, P.,

Singaravelu, V., Lloyd, K., Purdy, N., Daine, K., & John, A. (2017). A
systematic review of the relationship between internet use, self-
harm and suicidal behaviour in young people: The good, the bad
and the unknown. PLoS One, 12(8): e0181722.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5558917/
• Marchant, A., Hawton, K., Stewart A., Montgomery, P.,
Singaravelu, V., Lloyd, K., Purdy, N., Daine, K., & John, A.
(2018).Correction—A systematic review of the relationship
between internet use, self-harm and suicidal behaviour in young
people: The good, the bad and the unknown. PLoS One, 13(3):
e0193937. http://journals.plos.org/plosone/article?id=10.1371/
journal.pone.0193937

Search Tools.In this age of information, it is all too easy to find

items—the problem lies in sifting, sorting, and managing the vast
numbers of items that can be found. For example, a simple Google®
search for the topic “community poverty and violence” resulted in
about 15,600,000 results! As a means of simplifying the process of
searching for journal articles on a specific topic, a variety of helpful
tools have emerged. One type of search tool has previously applied a
filtering process for you: abstracting and indexing databases. These
resources provide the user with the results of a search to which
records have already passed through one or more filters. For
example, PsycINFO is managed by the American Psychological
Association and is devoted to peer-reviewed literature in behavioral
science. It contains almost 4.5 million records and is growing every
month. However, it may not be available to users who are not
affiliated with a university library. Conducting a basic search for our
topic of “community poverty and violence” in PsychINFO returned

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 141
1,119 articles. Still a large number, but far more manageable.
Additional filters can be applied, such as limiting the range in
publication dates, selecting only peer reviewed items, limiting the
language of the published piece (English only, for example), and
specified types of documents (either chapters, dissertations, or
journal articles only, for example). Adding the filters for English, peer-
reviewed journal articles published between 2010 and 2017 resulted
in 346 documents being identified.

Just as was the case with journals, not all abstracting and indexing
databases are equivalent. There may be overlap between them, but
none is guaranteed to identify all relevant pieces of literature. Here
are some examples to consider, depending on the nature of the
questions asked of the literature:

• Academic Search Complete—multidisciplinary index of 9,300

peer-reviewed journals
• AgeLine—multidisciplinary index of aging-related content for
over 600 journals
• Campbell Collaboration—systematic reviews in education, crime
and justice, social welfare, international development
• Google Scholar—broad search tool for scholarly literature across
many disciplines
• MEDLINE/ PubMed—National Library of medicine, access to
over 15 million citations
• Oxford Bibliographies—annotated bibliographies, each is
discipline specific (e.g., psychology, childhood studies,
criminology, social work, sociology)
• PsycINFO/PsycLIT—international literature on material relevant
to psychology and related disciplines
• SocINDEX—publications in sociology
• Social Sciences Abstracts—multiple disciplines
• Social Work Abstracts—many areas of social work are covered
• Web of Science—a “meta” search tool that searches other search

142 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
 tools, multiple disciplines

Placing our search for information about “community violence and

poverty” into the Social Work Abstracts tool with no additional filters
resulted in a manageable 54-item list. Finally, abstracting and
indexing databases are another way to determine journal legitimacy:
if a journal is indexed in a one of these systems, it is likely a legitimate
journal. However, the converse is not necessarily true: if a journal is
not indexed does not mean it is an illegitimate or pseudo-journal.

Government Sources.A great deal of information is gathered,

analyzed, and disseminated by various governmental branches at the
international, national, state, regional, county, and city level.
Searching websites that end in.gov is one way to identify this type of
information, often presented in articles, news briefs, and statistical
reports. These government sources gather information in two ways:
they fund external investigations through grants and contracts and
they conduct research internally, through their own investigators.
Here are some examples to consider, depending on the nature of the
topic for which information is sought:

• Agency for Healthcare Research and Quality (AHRQ) at

https://www.ahrq.gov/
• Bureau of Justice Statistics (BJS) at https://www.bjs.gov/
• Census Bureau at https://www.census.gov
• Centers for Disease Control and Prevention (CDC) at
https://www.cdc.gov

◦ Morbidity and Mortality Weekly Report of the CDC

(MMWR-CDC) at https://www.cdc.gov/mmwr/index.html
• Child Welfare Information Gateway at
https://www.childwelfare.gov
• Children’s Bureau/Administration for Children & Families at
https://www.acf.hhs.gov

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 143
 • Forum on Child and Family Statistics at
https://www.childstats.gov
• National Institutes of Health (NIH) at https://www.nih.gov,
including (not limited to):
• National Institute on Aging (NIA at https://www.nia.nih.gov
• National Institute on Alcohol Abuse and Alcoholism (NIAAA) at
https://www.niaaa.nih.gov
• National Institute of Child Health and Human Development
(NICHD) at https://www.nichd.nih.gov
• National Institute on Drug Abuse (NIDA) at
https://www.nida.nih.gov
• National Institute of Environmental Health Sciences at
https://www.niehs.nih.gov
• National Institute of Mental Health (NIMH) at
https://www.nimh.nih.gov
• National Institute on Minority Health and Health Disparities at
https://www.nimhd.nih.gov
• National Institute of Justice (NIJ) at https://www.nij.gov
• Substance Abuse and Mental Health Services Administration
(SAMHSA) at https://www.samhsa.gov/
• United States Agency for International Development at
https://usaid.gov

Each state and many counties or cities have similar data sources
and analysis reports available, such as Ohio Department of Health
at https://www.odh.ohio.gov/healthstats/dataandstats.aspxand
Franklin County at https://statisticalatlas.com/county/Ohio/
Franklin-County/Overview. Data are available from international/
global resources (e.g., United Nations and World Health
Organization), as well.

Other Sources.The Health and Medicine Division (HMD) of the

144 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
National Academies—previously the Institute of Medicine (IOM)—is a
nonprofit institution that aims to provide government and private
sector policy and other decision makers with objective analysis and
advice for making informed health decisions. For example, in 2018
they produced reports on topics in substance use and mental health
concerning the intersection of opioid use disorder and infectious
disease, the legal implications of emerging neurotechnologies, and
a global agenda concerning the identification and prevention of
violence (see http://www.nationalacademies.org/hmd/Global/
Topics/Substance-Abuse-Mental-Health.aspx). The exciting aspect
of this resource is that it addresses many topics that are current
concerns because they are hoping to help inform emerging policy.
The caution to consider with this resource is the evidence is often
still emerging, as well.

Numerous “think tank” organizations exist, each with a specific

mission. For example, the Rand Corporation is a nonprofit
organization offering research and analysis to address global issues
since 1948. The institution’s mission is to help improve policy and
decision making “to help individuals, families, and communities
throughout the world be safer and more secure, healthier and more
prosperous,” addressing issues of energy, education, health care,
justice, the environment, international affairs, and national security
(https://www.rand.org/about/history.html). And, for example, the
Robert Woods Johnson Foundation is a philanthropic organization
supporting research and research dissemination concerning health
issues facing the United States. The foundation works to build a
culture of health across systems of care (not only medical care) and
communities (https://www.rwjf.org).

While many of these have a great deal of helpful evidence to share,

they also may have a strong political bias. Objectivity is often lacking
in what information these organizations provide: they provide
evidence to support certain points of view. That is their purpose—to

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 145
provide ideas on specific problems, many of which have a political
component. Think tanks “are constantly researching solutions to a
variety of the world’s problems, and arguing, advocating, and
lobbying for policy changes at local, state, and federal levels” (quoted
from https://thebestschools.org/features/most-influential-think-
tanks/). Helpful information about what this one source identified
as the 50 most influential U.S. think tanks includes identifying each
think tank’s political orientation. For example, The Heritage
Foundation is identified as conservative, whereas Human Rights
Watch is identified as liberal.

While not the same as think tanks, many mission-driven

organizations also sponsor or report on research, as well. For
example, the National Association for Children of Alcoholics (NACOA)
in the United States is a registered nonprofit organization. Its
mission, along with other partnering organizations, private-sector
groups, and federal agencies, is to promote policy and program
development in research, prevention and treatment to provide
information to, for, and about children of alcoholics (of all ages).
Based on this mission, the organization supports knowledge
development and information gathering on the topic and
disseminates information that serves the needs of this population.
While this is a worthwhile mission, there is no guarantee that the
information meets the criteria for evidence with which we have been
working. Evidence reported by think tank and mission-driven
sources must be utilized with a great deal of caution and critical
analysis!

In many instances an empirical report has not appeared in the

published literature, but in the form of a technical or final report
to the agency or program providing the funding for the research
that was conducted. One such example is presented by a team of
investigators funded by the National Institute of Justice to evaluate a
program for training professionals to collect strong forensic evidence

146 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
in instances of sexual assault (Patterson, Resko, Pierce-Weeks, &
Campbell, 2014): https://www.ncjrs.gov/pdffiles1/nij/grants/
247081.pdf. Investigators may serve in the capacity of consultant to
agencies, programs, or institutions, and provide empirical evidence
to inform activities and planning. One such example is presented
by Maguire-Jack (2014) as a report to a state’s child maltreatment
prevention board: https://preventionboard.wi.gov/Documents/
InvestmentInPreventionPrograming_Final.pdf.

When Direct Answers to Questions Cannot Be Found.Sometimes

social workers are interested in finding answers to complex
questions or questions related to an emerging, not-yet-understood
topic. This does not mean giving up on empirical literature. Instead,
it requires a bit of creativity in approaching the literature. A Venn
diagram might help explain this process. Consider a scenario where
a social worker wishes to locate literature to answer a question
concerning issues of intersectionality. Intersectionality is a social
justice term applied to situations where multiple categorizations or
classifications come together to create overlapping, interconnected,
or multiplied disadvantage. For example, women with a substance
use disorder and who have been incarcerated face a triple threat
in terms of successful treatment for a substance use disorder:
intersectionality exists between being a woman, having a substance
use disorder, and having been in jail or prison. After searching the
literature, little or no empirical evidence might have been located on
this specific triple-threat topic. Instead, the social worker will need
to seek literature on each of the threats individually, and possibly will
find literature on pairs of topics (see Figure 3-1). There exists some
literature about women’s outcomes for treatment of a substance use
disorder (a), some literature about women during and following
incarceration (b), and some literature about substance use disorders
and incarceration (c). Despite not having a direct line on the center

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 147
of the intersecting spheres of literature (d), the social worker can
develop at least a partial picture based on the overlapping literatures.

Figure 3-1. Venn diagram of intersecting literature sets.

148 | Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found?
 Take a moment to complete the
following activity. For each statement
about empirical literature, decide if it is
true or false.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=144

Module 2 Chapter 3: What is Empirical Literature & Where can it be

Found? | 149
Module 2 Chapter 4:
Reviewing Empirical Articles
After carefully reviewing the source you have located, it is time to
critically review the piece itself. In this chapter, you will read about:

• Steps in reviewing different sections of an empirical article

• The importance of maintaining a critical perspective on what
you are reading—being an “active reader”

The seven steps considered in this chapter relate to the structure of

typical journal articles published in social work and allied discipline
journals. The structure is familiar to anyone who works with the
American Psychological Association (APA) guide to how journal
articles are written and structured, thePublication Manual of the
American Psychological Association(APA, 2009 for the sixth edition).

Step 1. What is in a Name? Reviewing the title

Authors vary tremendously in their approach to titling their work,

much as parents differ markedly in their approach to naming their
babies. Ideally, a title is sufficiently precise and specific to tell a
reader what the article is about. Sometimes titles have attention
catching phrases added at the front or back end. Ideally, a title is also
not overly elaborate and lengthy.

For example, the following article titles clearly and succinctly

communicates what each article is about.

• “Race and ethnic differences in early childhood maltreatment in

150 | Module 2 Chapter 4: Reviewing

Empirical Articles
 the United States” (Lanier, Maguire-Jack, Walsh, & Hubel, 2014).
• “Tracking the when, where, and with whom of alcohol use:
Integrating ecological momentary assessment and geospatial
data to examine risk for alcohol-related problems” (Freisthler,
Lipperman-Kreda, Bersamin, & Gruenewald, 2014).
• “Meeting them where they are: An exploration of technology use
and help seeking behaviors among adolescents and young
adults” (Cash & Bridge, 2012).
• “A systematic review of the relationship between internet use,
self-harm and suicidal behaviour in young people: The good, the
bad and the unknown” (Marchant et al, 2017). [Note the word
“behavior” is spelled “behaviour” in the United Kingdom, but
without the “u” in the United States.]

Examples of (hypothetical) article titles that are non-communicative

or miscommunicate include:

• “The problem of drugs in America.” This title is not sufficiently

specific for an empirical article. First, a reader does not know if
this is about how the pharmacology industry manufactures
drugs (quality control), an epidemiology report about the scope
of the problem, an etiology report about theories related to the
causes of the problem, a test of a sociological, psychological, or
biological theory. It is more likely a title about an opinion piece,
a general book chapter, or even an entire book.
• “The effectiveness of PFI and MI in AOD treatment.” This title is
swimming in jargon, making it difficult to interpret without
looking into what each acronym means, and also makes it
difficult to locate in the literature. (PFI is Personal Feedback
Intervention, MI is motivational interviewing, and AOD is about
alcohol and other drugs.) In addition, there exists a problem of
ambiguity with acronyms: BPD could refer to either
bronchopulmonary dysplasia, a lung complication common

Module 2 Chapter 4: Reviewing Empirical Articles | 151

 among prematurely born infants, borderline personality
disorder, or bipolar disorder (both are mental disorders).
• “The effects of self-esteem on high school student retention
and drop-out.” This title is only good if the study design and
methods actually allow for a causal inference. If the study design
only allows for conclusions about the existence of a relationship
between these two variables, the title is a poor choice—the word
“effects” implies causality and misrepresents the study.

Step 2. What is it about? Reviewing the abstract

Authors present a summary of their manuscript in a brief abstract

that appears at the start of a published article. Abstracts are also
published in a number of indexing and abstracting resources, making
them relatively easy to access. Journals limit the length of abstracts,
usually to somewhere between 150-250 words depending on the
journal. This makes it challenging to explain the important aspects of
the manuscript, with enough detail to be clear, but without the luxury
of unlimited space to present nuances. An abstract should address
the following points:

• Study’s purpose, research aims, questions, and/or hypotheses

• Study approach
• Study design and methods (including study participants and
measures)
• Data analysis and key results
• Key implications of the study results.

As a reader, the abstract should provide enough information for you

to determine whether it is relevant for you to pursue the full article.
An article abstractis not sufficient information for you to evaluate
the evidence, even the evidence that appears in the description of
results! To evaluate the evidence, you need to acquire and review the
full article.

152 | Module 2 Chapter 4: Reviewing Empirical Articles

This point is so important that it warrants repeating:

An article’s abstract is not sufficient information for you to

evaluate the evidence, even the evidence that appears in the
description of results! To evaluate the evidence, you need to
acquire and review the full article.

A Note about research abstracts. An excellent search of

literature will often turn up published research abstracts. In this case,
there will not be a full article to locate. The abstract describes a
conference presentation, and these are the precursors to publishing
an article about a research study. Many professional organizations
publish these abstracts in a special journal issue, sometimes a
“supplement” to the journal. The next step for your search, if the title
and abstract seem relevant, will be to determine whether a paper was
ever published based on the study described in the abstract. This is
a place where you would search by author name(s) rather than by
subject or topic alone.

Step 3. What is the rationale and background knowledge?

Reviewing the introduction

Once you have acquired an article that seems interesting and

relevant (based on its title and abstract), you will next encounter its
introduction. The purpose of an introduction is to provide the reader
with a background orientation to the study that was conducted. This
might include background information regarding the scope of the
problem being addressed by the research. It should certainly provide
the reader with a review of literature related to the topic and
research questions. This might include an overview of the theory or
theories related to the research that was conducted. In the end, the
reader should understand the following:

Module 2 Chapter 4: Reviewing Empirical Articles | 153

 1. Why was the study was undertaken, why was it important, why
does it matter?
2. What was known from the literature that informed the study’s
development?
3. What knowledge gap or gaps did the study aim to fill, or what
did the study aim to contribute to the body of knowledge?
4. What research questions did the investigators address in their
study?

The introduction often also informs readers about the study’s

approach (e.g., qualitative, quantitative, mixed methods) and type
of study that was implemented (e.g., exploratory, descriptive,
experimental). After reviewing the introduction, you should have an
even better idea of whether the article is relevant for your purposes.

Step 4. What happened? Reviewing the methods

If it was not made clear in the introduction, the study approach

and type of study should be made explicit in the methods section
of the article. The methods section, at a minimum, needs to explain
who participated in the study and how data were collected. In a
quantitative study, the study design is also described in the methods
section; in a qualitative study, the type of study is described. There
are three basic sub-sections in the methods section of an empirical
article: study participants, study measures, and study procedures.

Study participants. A methods sub-section describes who actually

participated in the study, including numbers and characteristics of
the study participants, as well as the pool from which these
participants were drawn. The purpose of this sub-section in
describing a quantitative study is to inform readers about

154 | Module 2 Chapter 4: Reviewing Empirical Articles

generalizability of the study’s results and inform other investigators
about what they would need to do to replicate the study to determine
if they achieve similar results. Authors may present some of the
description material in the form of tables with information about
numbers and proportions reflecting categorical variables (like gender
or race/ethnicity) and distribution on scale/continuous variables
(like age). The method of selecting these participants should be clear
along with any inclusion or exclusion criteria that were applied. The
participant response rate might also be calculated as the number of
participants enrolled in the study divided by the number of persons
eligible to be enrolled (the “pool”), multiplied by 100%. Very low
response rates make a study vulnerable to selection bias—the few
persons who elected to participate might not represent the general
population. In a qualitative study, the study participants section again
describes how the individuals were selected for participation, and
details describing these individuals are provided. Generalizability is
not a goal in qualitative studies, but information about study
participants should provide an indication to a reader of how robust
the results might be. Robust descriptions come from participants
who exhibit a range of defining and/or experiential characteristics.
Finally, regardless of study approach, authors typically make evident
that the study was reviewed by an Institutional Review Board for the
inclusion of human participants.

Study Measures.Another methods sub-section explains how the data

were collected. For quantitative studies, the data collection
instruments used to measure each study variable are described. If the
tools used for data collection were previously published, authors cite
the sources of those tools and published literature about them—their
reliability and validity, for example. The authors may also summarize
literature concerning how the measures are known to perform with
the specific type of study participants involved in the study—for

Module 2 Chapter 4: Reviewing Empirical Articles | 155

example, different ages, diagnoses, races/ethnicities, or other
characteristics. For qualitative studies the interview protocols or
questions asked of participants are described in detail. In
observational data collection studies, the approach to recording and
scoring/coding observed behavior are described. In any case, the
approach to data collection or measurement is described in sufficient
detail for a reader to critically appraise the adequacy of the data
collection approach and conclusions that can be drawn from the data
collection process, and for other investigators to be able to replicate
the study should they wish to confirm the results.

Study procedures.Sometimes study procedures is a separate

methods sub-section and sometimes this content is incorporated
into the participants and measures sub-sections. This sub-section
includes information about activities in which the study participants
engaged during the study. In a quantitative, experimental study, the
methods utilized to assign study participants to different
experimental conditions might be described here (i.e., the
randomization approach used). Additionally, procedures used in
handling data are usually described. In a quantitative study,
investigators may report how they scored certain measures and what
evidence from the literature informs their scoring approach. In a
qualitative study, details about how data were coded are reported
here. Procedures for ensuring inter-observer or inter-rater reliability
and agreements will also be reported for either type of study.
Regardless of the study’s research approach, a reader should come
away with a detailed understanding of how the study was executed.
As a result, the reader should be sufficiently informed about the
study’s execution to be able to critically analyze the strength of the
evidence developed from the methods that were applied.

156 | Module 2 Chapter 4: Reviewing Empirical Articles

Step 5. What was found? Reviewing the results

The results section is where investigators describe the data they

collected, how it was analyzed, and what was observed in the
data.The structure and format of the results section varies markedly
for different research approaches.

Qualitative methods results description.The nature of qualitative

research questions and methods leads to data that are richly
descriptive. The results derived from the data are, therefore,
generally descriptive in nature. There may be a great deal of direct
quotes, presenting information in participants’ own words.
Descriptions may include thematic or concept maps constructed
by the study investigators as a means of “sense making” from the
data. If statistics are included, they tend to be of a descriptive
nature—perhaps demonstrating the frequency with which certain
results were observed in the data. The results may include tables or
figures representing results. Ideally, a reader can identify the way
that reported results relate to the research aims or questions
originally asked by the study.

Quantitative methods results description. The nature of

quantitative research questions and methods leads to numeric data
that can be summarized using various forms of statistical analyses.
The results section of a quantitative study report will describe which
statistical analyses were utilized, and should indicate the rationale for
selecting those analyses, as well as discussing how well the data were
suited to those analyses. Analyses that involve hypothesis testing
will indicate the statistical support for conclusions drawn from the
data (i.e., descriptive statistics, test statistics, significance levels, and
confidence intervals). The results may be presented in a combination

Module 2 Chapter 4: Reviewing Empirical Articles | 157

of text descriptions, tables, and figures. An informed reader should be
able to determine the appropriateness of the statistical approaches
used in the analyses and the conclusions drawn from those analyses.
Ideally, study results are presented in association with each study
question or hypothesis as it is answered. Problems encountered with
any specific analyses are also reported here, such as when data were
not suitably distributed, sample sizes were inadequate, or
assumptions underlying specific types of analytic approaches were
violated.

Mixed methods results description. As mixed methods approaches

continue evolving, so are creative ways of presenting results of
studies that integrate qualitative and quantitative approaches.
Results for a study employing mixed methods are often presented
question-by-question. Where qualitative questions were addressed,
the descriptive results will be presented as outlined above. Where
quantitative questions were addressed, the numeric and statistical
results will be presented as outlined above.

Regardless of study approach, a reader should have a clear

understanding of the way data were analyzed and results of those
analyses. This is not a place where authors have drawn conclusions
about the implications of those results—that belongs in the article’s
discussion section.

Step 6. What was concluded? Reviewing the Discussion

In the end, the authors will offer their interpretation of the evidence
described in their Results section. This discussion should include
several elements:

• A brief overview summary of the key results.

• Discussion of how each key result relates to the study aims,

158 | Module 2 Chapter 4: Reviewing Empirical Articles

 questions, and/or hypotheses.
• Discussion of how the observed results relate to the previous
existing literature (are they mutually confirming or
contradicting), if the study results were completely new
contributions, or if they were ambiguous and no conclusions
could be drawn.
• Discussion of the study’s methodological or analysis/results
limitations.
• Practical implications of the study results for practice and for
future research.

It is important to remember that the discussion is the authors’ own

interpretation of the results. This, again, is a place where readers
must apply their own critical analysis to the study implications. For
example, sometimes authors get a bit carried away with their
interpretation and make suggestions that are not supported by the
evidence in their studies. Or, they may not have gone far enough, and
you see potential implications that they did not.

Step 7. Where are other relevant pieces? Reviewing the reference

list

As you search for relevant literature, you might want to review the
reference list to an article that you found to be relevant. Sometimes
your own search methods and search terms might have missed some
important items that the article’s authors were able to identify. This
review will provide you with titles to consider, and possibly you will
recognize the names of key scholars in the topic area. You can then
pursue these background resources as part of your own search.

Module 2 Chapter 4: Reviewing Empirical Articles | 159

 Interactive Excel Workbook Activities

Complete the following Workbook Activities:

• Workbook Introduction and Steps to Install Microsoft

Office 365
• SWK 3401.2-4.1 Getting Started: Excel and Data
Analysis ToolPak Access

160 | Module 2 Chapter 4: Reviewing Empirical Articles

Module 2 Module Summary
The chapters in this module covered a wide range of topics. You read
about the nature of social work research questions and how these
relate to different types of research. The different types included
those involved in a translational science process, as well as those
addressing epidemiology, etiology, or mechanisms of change
questions. The framework of exploratory, descriptive, and
explanatory research was introduced, and this framework will be
revisited in future modules as we examine different research
methodologies. The role of theory in relation to research and the
role of theory and research in promoting social justice were also
examined in this module. The importance of recognizing how social
work interfaces with other disciplines was emphasized, as well.
Finally, you were presented with a great deal of information about
empirical literature—what it is, where to find it, and how to review
it. All of this information is designed to assist you as a social work
professional as you strive to develop your understanding of diverse
populations, social work problems, and social phenomena, and
engage in efforts to develop culturally responsive intervention
strategies. This information serves as a foundation, relevant in our
next course about understanding interventions.

Module 2 Module Summary | 161

Module 2 Key Terms and
Definitions
abstracting and indexing databases: provide article abstracts and
information about accessing full articles of interest; may not be in the
public domain; access depends on the library use.

article abstract: a summary of an article; helps to determine if the full

article will be of interest/relevant.

background question: answerable with facts, direct a straightforward

search for evidence; typically get at the epidemiology or etiology
of a social work problem or social phenomenon (as compared to
foreground question).

basic research: fundamental or foundational research, usually to build

or test theory or improve prediction; tends to impose rigid control
over variables (compared to applied research).

biopsychosocial: a perspective that integrates biological,

psychological, and social system factors and forces in explaining a
phenomenon; also, may consider the interactions among these
factors and forces.

correlational research question: seeking descriptions about

relationships that might exist between variables; questions of
association.

descriptive research question: directs research around developing

descriptions about populations, problems, processes, or phenomena.

empirical literature: literature providing information based on

systematic observation and applying systematic methodology,

162 | Module 2 Key Terms and

Definitions
objectively developed, characterized by transparency, and
reproducible (meeting definition of research).

epidemiology: science of understanding the distribution of problems

or phenomena in the population; also, may include understanding
relationships between problems/phenomena, and the root causes;
common approach used in public health.

etiology: science testing theories and hypotheses about the origins

and natural course of a problem or phenomenon

explanatory research question: directs research around testing

theories and hypotheses about causality and causal relationships
between variables; usually leads to experimental study approaches.

exploratory research question: directs research around developing a

basic, beginning understanding of a newly recognized phenomenon
of interest where not much is yet known or understood, to the point
where knowing what questions about theory and causes are
premature.

foreground question: answers allow comparison of options, usually

practice decisions; often complex (as compared to background
questions).

impact factor: a statistic used to evaluate how much articles

published in a particular journal have been cited by others, as a gauge
of the journal’s impact on the field.

implementation science: the study of how interventions are

disseminated and adopted into practice.

interdisciplinary: where different disciplines work together, creating

an integrated, unified whole.

intervention research: science of the impact of specific interventions

Module 2 Key Terms and Definitions | 163

on outcomes of interest; in addition to what outcomes are observed,
may also include studies of mechanisms by which interventions have
their impact on outcomes (mechanisms of change).

mechanisms of change: the specific factors, variables, and processes

by which changes in outcome are produced; how an intervention has
its effects.

multidisciplinary: where different disciplines contribute separate

parts, individually, that maintain their distinct, unique characteristics
in the whole.

open-access journals: electronic publications available to anyone

without a subscription, designed to make information as widely
accessible to public, global audiences as possible; note cautions
regarding discerning between legitimate and predatory/junk open-
access resources.

peer review process: publication (and grant award) process by which

submissions are evaluated for adequacy and accuracy by other
scholars familiar with the topic/field/area.

predatory or junk journals: publications lacking integrity of legitimate

journals, proliferating in the open environment of the internet/web
access.

research abstracts: published summaries of a research report; may or

may not be accompanied by a full manuscript.

research hypothesis: a clear, testable statement predicting an

outcome or results, and that will either be rejected or not, depending
on the evidence; the prediction or hypothesis is based on theory
and/or other evidence.

social indicators: statistics that describe the social, economic, and

164 | Module 2 Key Terms and Definitions

psychological factors that have an impact on the well-being of a
community or other population group.

theory: explanatory belief(s), idea(s), or principles; may be simple or

from complex system of principles.

think tank: formal organizations that provide ideas on specific

problems, many of which have a political component and advocate/
lobby for policy change based on their “research.”

transdisciplinary: where disciplinary differences become

insignificant; the “whole” is equally relevant across disciplines.

translational science: a system whereby basic research informs

intervention research, and implementation science addresses
adoption of evidence-supported interventions.

Module 2 Key Terms and Definitions | 165

Module 2 References
American Psychological Association (APA). (2009). Publication Manual
of the American Psychological Association, sixth edition.
Washington, DC: American Psychological Association.

Anderson, R. (2017). Cabell’s new predatory journal blacklist: A review.

Blog posting of Society for Scholarly Publishing, The Scholarly
Kitchen (July 25).

Anderson-Butcher, D., Amorose, A.J., Lowers, L., Gibson, A., Riley, A.,
& Ruch, D. (2016). The case for the Perceived Social Competence
Scale-II. Research on Social Work Practice. 26(4), 419-428.

Ardito, R.B., & Rabellino, D. (2011). Therapeutic alliance and outcome

of psychotherapy: Historical excursus, measurements and
prospects for research. Frontiers in Psychology, 2, 270-281.

Begun, A.L., Brondino, M.J., Bolt, D., Weinstein, B., Strodthoff, T., &
Shelley, G. (2008). The Revised Safe At Home Instrument for
assessing readiness to change intimate partner violence. Violence
and Victims, 23(4), 508-524.

Begun, A.L., Murphy, C., Bolt, D., Weinstein, B., Strodthoff, T., Short,
L., & Shelley, G. (2003). Characteristics of the Safe At Home
Instrument for assessing readiness to change intimate partner
violence. Research on Social Work Practice, 13(1), 80-107.

Birken, S.A., Bunger, A.C., Powell, B.J., Turner, K., Clary, A.S., Klaman,
S.L., Yu, Y., Whitaker, D.J., Self, Sh.R., Rostad, W.L., Chatham, J.R.S.,
Kirk, M.A., Shea, C.M., Haines, E., & Weiner, B.J. (2017).
Organizational theory for dissemination and implementation
research. Implementation Science, 12(62).

166 | Module 2 References

Centers for Disease Control and Prevention (CDC). Multiple causes
of death 1999-2016, CDC WONDER database, National Center for
Health Statistics (https://wonder.cdc.gov).

Clapp, J.D., Holmes, M.R., Reed, M.B., Shillington, A.M., Freisthler,

B. & Lange, J.E. (2007). Measuring college students’ alcohol
consumption in natural drinking environments: Field
methodologies for bars and parties. Evaluation Review, 31, 469-489.

Cowell, A.J., Brown, J.M., Mills, M.J., Bender, R.H., & Wedehase, B.J.
(2012). Cost-effectiveness analysis of motivational interviewing
with feedback to reduce drinking among a sample of college
students. Journal of Studies in Alcohol and Drugs, 73(2), 226-237.

Davis, T. S., Saltzburg, S., & Locke, C. R. (2010). Assessing community

needs of sexual minority youth: Modeling Concept Mapping for
service planning. Journal of Gay & Lesbian Social Services, 22(3),
226- 249.

Dudley, J.R. (2011). Research methods in social work: Being producers

and consumers of research, 2ndedition/updated edition. Boston:
Allyn & Bacon.

Fields, N.L., Anderson, K.A. & Dabelko-Schoeny, H. (2014). The

effectiveness of adult day services for older adults: A review of the
literature from 2000 to 2011. Journal of Applied Gerontology, 33(2),
130-163.

Freisthler, B., Lipperman-Kreda, S., Bersamin, M., & Gruenewald, P.J.

(2014). Tracking the when, where, and with whom of alcohol use:
Integrating ecological momentary assessment and geospatial data
to examine risk for alcohol-related problems. Alcohol Research:
Current Reviews, 36, 29-38.

Grinnell, R.M., & Unrau, Y.A. (2014). Social work research and

Module 2 References | 167

 evaluation: Foundations of evidence-based practice, 10thedition. NY:
Oxford University Press.

Hedegaard, H., Warner, M., & Miniño, A.M. (2017). Drug overdose
deaths in the United States, 1999-2016. NCHS Data Brief, no 294.
Hyattsville, MD: National Center for Health Statistics.

Ioannou, M., Hammond, L., & Simpson, O. (2015). A model for

differentiating school shooters’ characteristics. Journal of Criminal
Psychology, 5(3), 188-200.

Jarrott, S.E., Weaver, R.H., Bowen, N.K., & Wang, N. (2018). Measuring
dimensions of intergenerational contact: Factor analysis of the
Queen’s University Scale. Aging & Mental Health, 22(4), 568-573.

Jensen, C.D., Cushing, C.C., Aylward, B.S., Craig, J.T., Sorell, D.M.,
& Steele, R.G. (2011). Effectiveness of motivational interviewing
interventions for adolescent substance use behavior change: A
meta-analytic review. Journal of Consulting and Clinical Psychology,
79(4), 433-440.

Kazdin, A.E. (2007). Mediators and mechanisms of change in

psychotherapy research. Annual Review of Clinical Psychology, 3,
1-27.

Kyriakakis, S., Waller, B., Kagotho, N., & Edmond, T. (2015).

Conducting safe research with at risk populations: Design
strategies from a study with unauthorized immigrant women
experiencing intimate abuse. Qualitative Social Work, 14(2) 259–274.

Laine, C., & Winker, M.A. (2017). Identifying predatory or pseudo-

journals. Biochemia Medica, 27(2), 285-291.

Lanier, P., Maguire-Jack, K., Walsh, T., & Hubel, G. (2014). “Race and
ethnic differences in early childhood maltreatment in the United

168 | Module 2 References

 States.” Journal of Behavioral and Developmental Pediatrics, 35(7),
419-426.

Langman, P. (2014). Statistics on bullying and school shootings.

Retrieved from www.schoolshooters.info

Lower, L.M., Anderson-Butcher, D., & Newman, T.J. (2017). Validity and
reliability of the Teamwork Scale for Youth, Research on Social Work
Practice, 27(6), 716-725.

Maguire-Jack, K. (2104). Child maltreatment prevention programs—A

worthwhile investment. What it Will Take: Investing in Wisconsin’s
future by keeping kids safe today, Madison, WI: Wisconsin
Children’s Trust Fund. Retrieved from
https://preventionboard.wi.gov/Documents/
InvestmentInPreventionPrograming_Final.pdf

Marchant, A., Hawton, K., Stewart A., Montgomery, P., Singaravelu, V.,
Lloyd, K., Purdy, N., Daine, K., & John, A. (2017). A systematic review
of the relationship between internet use, self-harm and suicidal
behaviour in young people: The good, the bad and the unknown.
PLoS One, 12(8): e0181722.

Marchant, A., Hawton, K., Stewart A., Montgomery, P., Singaravelu,

V., Lloyd, K., Purdy, N., Daine, K., & John, A. (2018). Correction—A
systematic review of the relationship between internet use, self-
harm and suicidal behaviour in young people: The good, the bad
and the unknown. PLoS One, 13(3): e0193937.

Mears, D.P., Moon, M.M., & Thielo, A.J. (2017). Columbine revisited:
Myths and realities about the bullying-school shootings
connection. Victims & Offenders, 12(6), 939-955.

Mayo, N.E., Asano, M., & Barbic, S.P. (2013). When is a research
question not a research question? Journal of Rehabilitation
Medicine, 45(6), 513-518.

Module 2 References | 169

National Academies of Sciences, Engineering, and Medicine (NAS).
(2018). Integrating responses at the intersection of opioid use
disorder and infectious disease epidemics: Proceedings of a
workshop.Washington, DC: The National Academies Press.

Newman, K.S., & Fox, C. (2009). Repeat tragedy: Rampage shootings

in American high school and college settings, 2002-2008. American
Behavioral Scientist, 52(9), 1286-1308.

Office of Behavioral and Social Sciences Research (OBSSR). (2003).

NIH Plan for social work research. Retrieved from

https://obssr.od.nih.gov/wp-content/uploads/2016/07/
SWR_Report.pdf.

Patterson, D., Resko, S., Pierce-Weeks, J., & Campbell, R. (2014).

Delivery and evaluation of Sexual Assault Forensic (SAFE) training
programs. Final Report. Submitted to National Institute for Justice.
Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/
247081.pdf.

Raitanen, J., Sandberg, S., & Oksanen, A. (2017). The bullying-school

shooting nexus: Bridging master narratives of mass violence with
personal narratives of social exclusion. Deviant Behavior,
online,1-14.

Shao, W., Guan, W., Clark, M.A., Liu, T., Santelices, C.C., Cortes, D.E., &
Merchant, R.C. (2015). Variations in recruitment yield, costs, speed
and participant diversity across internet platforms in a global study
examining the efficacy of an HIV/AIDS and HIV testing animated
and live-action video among English- or Spanish-speaking Internet
or social media users. Digital Culture and Education, 7(1), 40-86.

Sielski, C.L., Begun, A.L., & Hamel, J. (2015). Expanding knowledge

concerning the Safe at Home Instrument for assessing readiness

170 | Module 2 References

 to change among individuals in batterer treatment. Partner Abuse,
6(3), 255-272.

Smith, S.G., Chen, J., Basile, K.C., Gilbert, L.K., Merrick, M.T., Patel,
N., Walling, M., & Jain, A. (2017). The National Intimate Partner
and Sexual Violence Survey (NISVS): 2010-2012 State Report.Atlanta,
GA: National Center for Injury Prevention and Control, Centers for
Disease Control and Prevention.

Stoeber, M., Jullié, D., Lobingier, B.T., Laeremans, T., Steyaert, J.,
Schiller, P.W., Manglik, A., & von Zastrow, M. (2018). A genetically
encoded biosensor reveals location bias of opioid drug
action.Neuron, 98, 1-14.

Subbaraman, M.S., Ludet, A.B., Ritter, L.A., Stunz, A., & Kaskutas, L.A.
(2015). Multisource recruitment strategies for advancing addiction
recovery research beyond treated samples. Journal of Community
Psychology, 43,560-575.

Testa, M., & Cleveland, M. (2017). Does alcohol contribute to college

men’s sexual assault perpetration? Between- and within-person
effects over five semesters. Journal of Studies on Alcohol and Drugs,
78, 5-13.

Vestal, C. (2018). Overdose deaths fall in 14 states. PEW Trusts,

retrieved from http://www.pewtrusts.org/en/research-and-
analysis/blogs/stateline/2018/02/22/overdose-deaths-fall-
in-14-states.

Wampold, B.E. (2015). How important are the common factors in

psychotherapy? An update. World Psychiatry, 14(3), 270-277.

Williams, J.H. (2016). [Editorial] Grand challenges for social work:

Research, practice, and education. Social Work Research, 40(2),
67-70.

Module 2 References | 171

MODULE 3

Module 3 | 173
Module 3 Introduction
In this module, we build on what we learned in earlier modules about
the nature of our research questions concerning diverse populations
and the problems that social workers address. This module affords us
the opportunity to become familiar with several different approaches
for answering questions about populations and social work problems.
We explore the philosophies and assumptions underlying
quantitative, qualitative, and mixed methods approaches, and we
analyze various advantages/disadvantages associated with their use.
We describe and analyze different types of study designs used to
answer population and social work problem questions (survey, group
comparison/experimental, grounded theory, narrative/ethnography,
and others). We explore:

• Qualitative, quantitative, and mixed methods approaches to

understanding diverse populations,social work problems, and
social phenomena;
• The types of approaches used for drawing samples for different
study design;
• Measurement strategies for different types of variables
(independent, dependent, categorical, continuous);
• Varied approaches to data collection (e.g., observation,
interview, self-administered survey, online surveys, focus
groups, and others), and cultural competence in measurement

READING OBJECTIVES

After engaging with these reading materials and learning resources,

you should be able to:

• Identify the similarities and differences underlying qualitative,

Module 3 Introduction | 175

 quantitative, and mixed methods study approaches, and how
they relate to different types of research questions;
• Recognize different qualitative research traditions;
• Describe longitudinal and cross-sectional approaches to
designing research studies and how they relate to different
types of research questions;
• Recognize different types of quantitative variables and how they
relate to our research questions;
• Identify advantages and disadvantages of various measurement
and data collection strategies;
• Critique measurement and data collection approaches on the
dimension of cultural competence;
• Explain basic principles associated with designing a plan for
who might be engaged as participants in research studies about
diverse populations and social work problems;
• Define key terms related to social work research approaches
and methods.

176 | Module 3 Introduction

Module 3 Chapter 1: From
Research Questions to
Research Approaches
The approaches that social work investigators adopt in their research
studies are directly related to the nature of the research questions
being addressed.In Module 2 you learned about exploratory,
descriptive, and explanatory research questions. Let’s consider
different approaches to finding answers to each type of question.

In this chapter we build on what was learned in Module 2 about

research questions, examining how investigators’ approaches to
research are determined by the nature of those questions. The
approaches we explore are all systematic, scientific approaches, and
when properly conducted and reported, they all contribute empirical
evidence to build knowledge. In this chapter you will read about:

• qualitative research approaches for understanding diverse

populations, social problems, and social phenomena,
• quantitative research approaches for understanding diverse
populations, social problems, and social phenomena,
• mixed methods research approaches for understanding diverse
populations, social problems, and social phenomena.

Overview of Qualitative Approaches

Questions of a descriptive or exploratory nature are often asked

and addressed through qualitative research. The specific aim in these
studies is to understand diverse populations, social work problems,
or social phenomena as they naturally occur, situated in their natural

Module 3 Chapter 1: From Research

Questions to Research
Approaches | 177
environments, providing rich, in-depth, participant-centered
descriptions of the phenomena being studied. Qualitative research
approaches have been described as “humanistic” in aiming to study
the world from the perspective of those who are experiencing it
themselves; this also contributes to a social justice commitment in
that the approaches give “voice” to the individuals who are
experiencing the phenomena of interest (Denzen & Lincoln, 2011).
As such, qualitative research approaches are also credited with being
sensitive and responsive to diversity—embracing feminist, ethnic,
class, critical race, queer, and ability/disability theory and lenses.

In qualitative research, the investigator is engaged as an observer and

interpreter, being acutely aware of the subjectivity of the resulting
observations and interpretations.

“At this level, qualitative research involves an interpretive,

naturalistic approach to the world” (Denzin & Lincoln, 2011, p.
3)

Because the data are rich and deep, a lot of information is collected
by involving relatively few participants; otherwise, the investigator
would be overwhelmed by a tremendous volume of information to
collect, sift through, process, interpret, and analyze. Thus, a single
qualitative study has a relatively low level of generalizability to the
population as a whole because of its methodology, but that is not the
aim or goal of this approach.

In addition, because the aim is to develop understanding of the

participating individuals’ lived experiences, the investigator in a
qualitative study seldom imposes structure with standardized
measurement tools. The investigator may not even start with
preconceived theory and hypotheses. Instead, the methodologies
involve a great deal of open-ended triggers, questions, or stimuli to
be interpreted by the persons providing insight:

178 | Module 3 Chapter 1: From Research Questions to Research Approaches

 “Qualitative research’s express purpose is to produce
descriptive data in an individual’s own written or spoken
words and/or observable behavior” (Holosko, 2006, p. 12).

Furthermore, investigators often become a part of the qualitative

research process: they maintain awareness of their own influences
on the data being collected and on the impact of their own
experiences and processes in interpreting the data provided by
participants. In some qualitative methodologies, the investigator
actually enters into/becomes immersed in the events or phenomena
being studied, to both live and observe the experiences first-hand.

Qualitative data and interpretations are recognized as being

subjective in nature—that is the purpose—rather than assuming
objectivity. Qualitative research is based on experientially derived
data and is interpretive, meaning it is “concerned with understanding
the meaning of human experience from the subject’s own frame of
reference” (Holosko, 2006, p. 13). In this approach, conclusions about
the nature of reality are specific to each individual study participant,
following his or her own interpretation of that reality. These
approaches are considered to flow from an inductive reasoning
process where specific themes or patterns are derived from general
data (Creswell & Poth, 2018).

Several purposes of qualitative approaches in social work include:

• describing and exploring the nature of phenomena, events, or

relationships at any system level (individual to global)
• generating theory
• initially test ideas or assumptions (in theory or about practices)
• evaluate participants’ lived experiences with practices,
programs, policies, or participation in a research study,
particularly with diverse participants
• explore “fit” of quantitative research conclusions with

Module 3 Chapter 1: From Research Questions to Research Approaches | 179

 participants’ lived experiences, particularly with diverse
participants
• inform the development of clinical or research assessment/
measurement tools, particularly with diverse participants.

Overview of Quantitative Approaches

Questions of the exploratory, descriptive, or explanatory type are

often asked and addressed through quantitative
research approaches, particularly questions that have a numeric
component. Exploratory and descriptive quantitative studies rely on
objective measures for data collection which is a major difference
from qualitative studies which are aimed at understanding subjective
perspectives and experiences. Explanatory quantitative studies often
begin with theory and hypotheses, and proceed to empirically test
the hypotheses that investigators generated. By their quantitative
(numeric) nature, statistical hypothesis testing is possible in many
types of quantitative studies.

Quantitative research studies utilize methodologies that enhance

generalizability of results to the greatest extent possible—individual
differences are de-emphasized, similarities across individuals are
emphasized. These studies can be quite large in terms of participant
numbers, and the study samples need to be developed in such a
manner as to support generalization to the larger populations of
interest.

The process is generally described as following a deductive logical

system where specific data points are combined to lead to developing
a generalizable conclusion. The philosophical roots (epistemology)
underlying quantitative approaches is positivism, involving the
seeking of empirical “facts or causes of social phenomena based on
experimentallyderived evidence and/or valid observations” (Holosko,
2006, p. 13). The empirical orientation is objective in that

180 | Module 3 Chapter 1: From Research Questions to Research Approaches

investigators attempt to be detached from the collection and
interpretation of data in order to minimize their own influences
and biases. Furthermore, investigators utilize objective measurement
tools to the greatest extent possible in the process of collecting
quantitative study data.

Several purposes of quantitative approaches in social work include:

• describing and exploring the dimensions of diverse populations,

phenomena, events, or relationships at any system level
(individual to global)—how much, how many, how large, how
often, etc. (including epidemiology questions and methods)
• testing theory (including etiology questions)
• experimentally determining the existence of relationships
between factors that might influence phenomena or
relationships at any system level (including epidemiology and
etiology questions)
• testing causal pathways between factors that might influence
phenomena or relationships at any system level (including
etiology questions)
• evaluate quantifiable outcomes of practices, programs, or
policies
• assess the reliability and validity of clinical or research
assessment/measurement tools.

Overview of Mixed-Method Approaches

Important dimensions distinguish between qualitative and

quantitative approaches. First, qualitative approaches rely on
“insider” perspectives, whereas quantitative approaches are directed
by “outsiders” in the role of investigator (Padgett, 2008). Second,
qualitative results are presented holistically, whereas quantitative
approaches present results in terms of specific variables dissected
from the whole for close examination; qualitative studies emphasize

Module 3 Chapter 1: From Research Questions to Research Approaches | 181

the context of individuals’ experiences, whereas quantitative studies
tend to decontextualize the phenomena under study (Padgett, 2008).
Third, quantitative research approaches tend to follow a positivist
philosophy, seeking objectivity and representation of what actually
exists; qualitative research approaches follow from a post-positivist
philosophy, recognizing that observation is always shaped by the
observer, therefore is always subjective in nature and this should be
acknowledged and embraced. In post-positivist qualitative research
traditions, realities are perceived as being socially constructed,
whereas in positivist quantitative research, a single reality exists,
waiting to be discovered or understood. The quantitative perspective
on reality has a long tradition in the physical and natural sciences
(physics, chemistry, anatomy, physiology, astronomy, and others).
The social construction perspective has a strong hold in social
science and understanding social phenomena. But what if an
investigator’s questions are relevant to both qualitative and
quantitative approaches?

Given the fundamental philosophical and practical differences, some

scholars argue that there can be no mixing of the approaches, that
the underlying paradigms are too different. However, mixed-methods
research has also been described as a new paradigm (since the 1980s)
for social science:

“Like the mythology of the phoenix, mixed methods research

has arisen out of the ashes of the paradigm wars to become
the third methodological movement. The fields of applied social
science and evaluation are among those which have shown
the greatest popularity and uptake of mixed methods research
designs” (Cameron & Miller, 2007, p. 3).

182 | Module 3 Chapter 1: From Research Questions to Research Approaches

Mixed-methods research approaches are used to address in a single
study the acknowledged limitations of both quantitative and
qualitative approaches. Mixed methods research combines elements
of both qualitative and quantitative approaches for the purpose of
achieving both depth and breadth of understanding, along with
corroboration of results (Johnson, Onwuegbuzie, & Turner, 2007, p.
123). One mixed-methods strategy is related to the concept
of triangulation: understanding an event or phenomenon from the
use of varied data sources and methods all applied to understanding
the same phenomenon (Denzin & Lincoln, 2011; see Figure 1-1).

Module 3 Chapter 1: From Research Questions to Research Approaches | 183

Figure 1-1. Depiction of triangulation as synthesis of different data
sources

For example, in a survey research study of student debt load

experienced by social work doctoral students, the investigators
gathered quantitative data concerning demographics, dollar amounts
of debt and resources, and other numeric data from students and
programs (Begun & Carter, 2017). In addition, they collected
qualitative data about the experience of incurring and managing debt
load, how debt shaped students’ career path decisions, practices
around mentoring doctoral students about student debt load, and
ideas for addressing the problem. Triangulation came into play in
two ways: first, collecting data from students and programs about
the topics, and second, a sub-sample of the original surveyed
participants engaged in qualitative interviews concerning the “fit”
or validity of conclusions drawn from the prior qualitative and
quantitative data.

Three different types of mixed methods approaches are used:

• Convergent designs involve the simultaneous collection of both

qualitative and quantitative data, followed by analysis of both

184 | Module 3 Chapter 1: From Research Questions to Research Approaches

 data sets, and merging the two sets of results in a comparative
manner.
• Explanatory sequential designs use quantitative methods first,
and then apply qualitative methods to help explain and further
interpret the quantitative results.
• Exploratory sequential designs first explore a problem or
phenomenon through qualitative methods, especially if the
topic is previously unknown or the population is understudied
and unfamiliar. These qualitative findings are then used to build
the quantitative phase of a project (Creswell, 2014, p. 6).

Mixed methods approaches are useful in developing and testing new

research or clinical measurement tools. For example, this is done
in an exploratory sequential process whereby detail-rich qualitative
data inform the creation of a quantitative instrument. The
quantitative instrument is then tested in both quantitative and
qualitative ways to confirm that it is adequate for its intended use.
This iterative process is depicted in Figure 1-2.

Figure 1-2. Iterative qualitative and quantitative process of

instrument development

Module 3 Chapter 1: From Research Questions to Research Approaches | 185

One example of how this mixed-methods approach was utilized was
in development of the Safe-At-Home instrument for assessing
individuals’ subjective readiness to change their intimate partner
violence behavior (Begun et al., 2003; 2008). The transtheoretical
model of behavior change (TMBC) underlies the instrument’s
development: identifying stages in readiness to change one’s behavior
and matching these stages to the most appropriate type of
intervention strategy (Begun et al., 2001). The first step in developing
the intimate partner violence Safe-At-Home instrument for assessing
readiness to change was to qualitatively generate a list of statements
that could be used in a quantitative rating scale. Providers of
treatment services to individuals arrested for domestic or
relationship violence were engaged in mutual teaching/learning with
the investigators concerning the TMBC as it might relate to the
perpetration of intimate partner violence. They independently
generated lists of the kinds of statements they heard from individuals
in their treatment programs, statements they believed were
demonstrative of what they understood as the different stages in the

186 | Module 3 Chapter 1: From Research Questions to Research Approaches

change process. The investigators then worked with them to reduce
the amassed list of statements into stage-representative categories,
eliminating duplicates and ambiguous statements, and retaining the
original words and phrases they heard to the greatest extent possible.
The second phase was both quantitative and qualitative in nature:
testing the instrument with a small sample of men engaged in batters’
treatment programs and interviewing the men about the experience
of using the instrument. Based on the results and their feedback, the
instrument was revised. This process was followed through several
iterations. The next phases were quantitative: determining the
psychometric characteristics of the instrument and using it to
quantitatively evaluate batterer treatment programs—the extent to
which individuals were helped to move forward in stages of the
change cycle.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.3-1.1 Getting Started: Understanding an

Excel Data File

Chapter Summary

In this chapter you were introduced to three general approaches

for moving from research question to research method. You were
provided with a brief overview of the philosophical underpinnings
and uses of qualitative, quantitative, and mixed-methods approaches.

Module 3 Chapter 1: From Research Questions to Research Approaches | 187

Next, you are provided with more detailed descriptions of qualitative
and quantitative traditions and their associated methodologies.

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=211

188 | Module 3 Chapter 1: From Research Questions to Research Approaches

Module 3 Chapter 2: Overview
of Qualitative Traditions
In Chapter 1 you read about three different approaches to social work
research: qualitative, quantitative, and mixed-methods approaches.
In this chapter we examine different traditions and methods
commonly utilized in qualitative approaches to research. The
qualitative research literature presents a wide array of approaches:
ethnography, life histories, symbolic interactionism, grounded
theory, phenomenology, hermeneutics/heuristic research,
interpretivism, collaborative social research, social or cognitive
anthropology, case studies, narrative inquiry, critical inquiry,
participatory action research, content analysis, study of artifacts, and
participant observation (Creswell & Poth, 2018). Here, we examine
four of the five approaches identified by Creswell and Poth (2018), as
well as two other approaches. You will read about:

• Narrative research
• Phenomenological research
• Grounded theory research
• Ethnographic research
• Participatory action research
• Content/artifact analysis research

Qualitative Research Approaches

Qualitative research approaches are most appropriately used when

the nature of the research questions and study goals meet certain
criteria (Padgett, 2008). These include:

• Relatively little is known or understood about the topic of

Module 3 Chapter 2: Overview of

Qualitative Traditions | 189
 interest
• The topic involves stigma and/or taboos, thereby lending itself
to analysis with emotional depth and sensitivity (rather than
quantifiable facts)
• The goal is to study and create meaning from the perspectives
of individuals who have “lived the experiences” of study interest
• The goal is to understand processes, particularly complex
processes, in social work problems and practices/interventions
• The goal is to interpret quantitative research results,
particularly unexpected results
• Engaging in action or participatory research, involving social
action or advocacy in combination with conducting the
research.

Common across qualitative approaches is employing the

“researcher-as-instrument” (Padgett, 2008, p. 17). Qualitative
research results are heavily influenced by the experiences,
perceptions, and personal qualities of the investigator. Thus, a critical
component involves the investigator being reflective about the data,
participants, analysis process, and results.

Narrative Research. Narrative research “begins with the experiences

as expressed in lived and told stories of individuals” (Creswell & Poth,
2018). It is about both the nature of the expressed narratives and
exploration of the contexts in which the narratives and narrators are
situated. The method of gathering and interpreting these stories is
called narrative inquiry. This is typically performed with one or a
very few individuals being involved with any one study. The purpose
is to learn the individual stories and their meaning to individuals.

190 | Module 3 Chapter 2: Overview of Qualitative Traditions

Of great interest in narrative inquiry is how a story is presented
by the individual: it is seldom presented in a simple, chronological
manner. Instead, stories are often conveyed in terms of critical
transitions or nodal events where “turning points” occur in a person’s
lived experience. Narrative examples are present in biographies,
autobiographies (autoethnography), and oral histories. For example,
the movie documentary Whitneypresents singer Whitney Houston’s
stories and lived experiences as a performer, celebrity, and family
member, including her relationship with alcohol and other drugs.
Parts of the documentary are presented as others’ descriptions, parts
are presented in her own words. The point is to better understand
this person’s life.

Module 3 Chapter 2: Overview of Qualitative Traditions | 191

A social work example of narrative inquiry is reported as one woman’s
journey of recovery from mental illness (Thomas & Rickwood, 2016).
Case interviews were conducted with a woman on three different
occasions when she was (re)admitted to a residential mental health
program. The woman’s narratives demonstrated changes over time in
hope, self-identity, a meaningful life, and assuming responsibility for
one’s own recovery. Furthermore, the investigators learned from the
woman’s perspective what elements of the services provided were
helpful in her road to recovery and that repeated admissions were
less indicative of repeated failures and more about progress toward
recovery.

Phenomenological Research. Like narrative research,

phenomenological research involves recording the stories of
individuals who experience a particular phenomenon or event of
interest. Unlike narrative inquiry, however, phenomenology is about
deriving or constructing a shared meaning about those events among
a group of participants: the goal is “meaning making” (Oktay, 2012).
The emphasis in this approach is placed on identifying what the
participants and their shared stories have in common. The emerging
phenomenological description is a composite description of the
phenomenon.

192 | Module 3 Chapter 2: Overview of Qualitative Traditions

Ideally, the investigator engages with a heterogeneous group of
individuals, all of whom have experienced the phenomenon of
interest. For example, if the phenomenon of interest is about the
experiences of siblings to persons with intellectual disabilities, the
investigator would want to engage with families of different ages,
social or economic standing, religious and cultural background,
sibling composition (numbers, genders, and spacing of siblings), and
individuals with different types or severity of disability.

One example of this type of research is a qualitative study concerned

with how trafficking affects the health and health care access among
15 women engaged in sex work in the red-light district (Kamathipura)
of Mumbai, India (Karandikar, Gezinski, & Kaloga, 2016). The
investigators worked with transcripts of interviews with the women,
preserving the voice and language of the women who participated.
They identified themes and relationships between themes in their
coding of the data. Three themes were discussed in their manuscript:

Module 3 Chapter 2: Overview of Qualitative Traditions | 193

how participants first entered into sex work (all but two reported
deception by family members, friends, or acquaintances as the
pathway to prostitution); health problems experienced by the women
(physical violence, sexually transmitted infection, alcohol addiction,
tuberculosis, HIV infection, miscarriage, abortion, cough/cold, and
malaria were most frequently reported as notable events); and,
disparities in access to health care resulting from being captively
held in isolation or seclusion to prevent escape from trafficking and
stigma attached to sex work.

Grounded Theory Research. Grounded theory research develops

theoretical explanations about a phenomenon or process being
studied—in other words, moving beyond description to generating
or discovering a theory (Creswell & Poth, 2018). The theory that
emerges can then be tested using other approaches and methods
to determine its generalizability. The name of this approach comes
from the theory being “grounded” in the data, rather than the data
being used to test a theory that has been presented “off the shelf”
(Creswell & Poth, 2018, p. 82). This approach, first described as a
unified approach by Glaser and Strauss (1967), is particularly relevant
to social work:

Because grounded theory creates theories that are derived

directly from real-world settings, it has the potential to produce
theories that can be used by social workers to guide practice.
This theory can also be used to develop theoretically based
interventions that can be tested in practice settings. In fact,
when grounded theory was developed, the authors envisioned a
collaborative venture between researchers and practitioners to
test and adapt the theory in practice settings” (Oktay, 2012, p.
5).

194 | Module 3 Chapter 2: Overview of Qualitative Traditions

Grounded theory as an approach utilizes abductive reasoning,
moving back and forth in an iterative manner between data gathering
and analysis processes, including inductive and deductive logic in
generating and cross-checking, testing the emerging theory (Oktay,
2012). Thus, theories grow from the grounded base of real-world
data. The approach is heavily influenced by the discipline of
sociology, but is used to inform other professions like social work
and nursing. Study participants in grounded theory research are
purposively selected because of their ability to provide relevant data.

An example of the grounded theory approach in use was presented

Module 3 Chapter 2: Overview of Qualitative Traditions | 195

in a study of professional decision-making processes undertaken by
social workers in child protection roles (Kettle, 2018). Based on 22
interviews with social workers in Scotland, the investigator explored
social workers’ own perceptions of negotiating delicate balances
between the interests of children and adults; the past, present, and
future; functioning as an investigator versus building a relationship;
getting too close versus not close enough; and, having power over
versus power together. Also studied was the nature of their
transactions and exchanges with other professionals. Emerging
theory emphasized social workers’ internal mental processes over
rational-technical solutions for practice decisions, the role of
perceived responsibility in degree of cooperation between
professionals, and social workers’ perceptions of different family
strategies for managing closeness of and power in their relationships
with the social worker.

Ethnographic Research. Like the qualitative approaches that we have

explored so far, ethnographic research aims to understand
phenomena from the perspectives of those experiencing it. However,
there exists a major difference: the unit of analysis in ethnographic
research is a collective group where the individuals interact with one
another and have shared, common experiences. As a result, the
individuals become a “culture-sharing group” by virtue of the “shared
patterns of behavior, beliefs and language” that they develop
(Creswell & Poth, 2018, p. 90). Ethnographic research focuses on
cultural systems, “operating on several levels simultaneously to infer
the tacit rules of the culture or subculture from the myriad of actions
and interactions being witnessed” (Padgett, 2008, p. 32). The
approach is strongly influenced by the discipline of anthropology,
but is critically important as an option for social work to understand
diverse populations, social work problems, and social phenomena.

196 | Module 3 Chapter 2: Overview of Qualitative Traditions

Helm (2016) described an ethnographic study of how practice
decisions are made within and across a team of social workers when
the judgments are made under conditions of subjectivity and
uncertainty, and when the available information is incomplete,
inconclusive, or possibly contested. The aim of the study conducted
in Scotland was to learn about “sense-making” under these
circumstances, with attention directed to how this becomes a shared
activity, rather than an individual activity. The study participants
included a single team manager, four team leaders acting as
supervisors, and six social workers from the larger team of
professionals working at the study location. Data were collected
through the investigator’s direct,non-participatory observation of

Module 3 Chapter 2: Overview of Qualitative Traditions | 197

team member behavior, including what was said in the normal course
of practice (naturalistic observation). The observation notes were
coded for themes, the two strongest being “making sense through
‘framing’ the situation” and “constructing responsibility” (Helm, 2016,
p. 29). Social work practitioners were observed to discuss cases with
others, usually by opening the discussion with a framing statement,
which then guided the discussion; for example, framing it in terms of
the worker’s own feelings about a case situation. Marked differences
were noted between how social workers framed their discussions
with each other, in supervision, and with clients. In addition, the
social workers’ sense-making occurred within the context of
reference to their perceived professional roles and responsibilities.
These perceptions were observed to influence decisions and actions
by the social workers. The ethnographic observations help the
profession better understand how practice decisions are made in
complex, under-defined situations: in sense-making, individual
practitioners rely on multiple sources of information, not the least of
which is provided by their understanding of the professional context
(roles and responsibilities) and input from co-workers and
supervisors.

Ethnography might be an important approach for social workers

to better understanding diverse populations in the ability/disability
arena. Consider, for example, the meaning of a deaf culture. For
hearing persons, not being able to hear and communicate through
spoken language might seem like an overwhelming disability. For
persons who have never been part of the dominant hearing culture,
living and thriving in a deaf community is important and meaningful.
Deaf culture is experienced as communicating and experiencing the
world differently compared to the hearing community. Members of
deaf culture do not view themselves as disabled and needing to be
“fixed.” Similar discussions have emerged among individuals who
experience the world differently than the mainstream by virtue of
their neurodiversity experience. These are individuals, many

198 | Module 3 Chapter 2: Overview of Qualitative Traditions

diagnosed with autism spectrum or other neurological disorders,
who advocate for recognizing and appreciating variations in
functioning as just that, not as mental disorders or disabilities
needing to be cured. Ethnographic research helps social workers
better understand their lived experiences from a strengths
perspective.

Participatory Action Research. Sometimes social workers are

members of a group attempting to solve a problem or influence
change at the mezzo or macro level; sometimes members of such
a group engage social workers to help them promote the problem
solving and change-making processes. Either way, the key
stakeholders in problem solving or advocating for change may decide
to integrate research methodologies into their process. The result
might be a participatory action research (PAR) scenario. This is a type
of research activity where investigators are engaged in dual roles,
both as researchers and change agents in the group or process under
study.

PAR is conducted in the context of a non-hierarchical, democratic,

engaged partnership of decision-making at every point in the
process. Together, stakeholders and social workers develop the aims,
questions, methods, analyses, results, and reporting of the change
process. The outcomes of PAR are locally relevant solutions,
developed by key stakeholders empowered to act as local experts
and co-researchers supported by relevant theory and research
knowledge shared by social work practitioners in a collaborative,
problem-solving process. The results often include enhanced
research skills among the key stakeholder partners as a form of
further empowerment. What is learned through the PAR
collaboration may have generalizable knowledge applicable to other
communities, but the generalizability is carefully interpreted since

Module 3 Chapter 2: Overview of Qualitative Traditions | 199

it is, essentially, a single case study at the community level. Lawson
(2015) described the relevance of PAR in community development
efforts:

“Homogenization is rooted in part in mainstream researchers’

claims about the wholesale generalizability of the knowledge
gained from their investigations, and it is facilitated by the
worldwide movement toward evidence-based policy and
practice. Granting the importance, indeed the strengths,
associated with this international development, there are
manifest risks and dangers when local voices, choices, and
knowledge are neglected, ignored, and discounted. Under these
circumstances, rigorous research knowledge has the potential
to serve as an instrument for domination, marginalization, and
oppression. And when these conditions prevail, research-based
knowledge has the potential to cause harm in the name of doing
good work. PAR’s expressly local knowledge for locally tailored
solutions thus provides an important safeguard and a quality
assurance mechanism for policy and practice” (Lawson, 2015, p.
ix-x).

Community-based participatory research (CBPR) is a type of

participatory research involving a collaborative inquiry process. As
above, local experts are “people other than formally trained
researchers” engaged to share their good ideas about every aspect
of the research being conducted: the questions asked, study design
features, how results are interpreted, and how findings should be
used (Lawson, 2015, p. xv). University-community collaborations
often provide the context through which CBPR occurs.

200 | Module 3 Chapter 2: Overview of Qualitative Traditions

An example of CBPR was a needs assessment concerning parenting
support programs for fathers in Detroit, Michigan (Lee, Hoffman,
& Harris, 2016). The key stakeholders engaged in the effort with
trained social work investigators from a university’s school of social
work included formal service delivery system providers, a community
advisory board, other key informants, and fathers. The team
proceeded through a five-stage community-engaged research
process, and discussed the importance of operating as a “learning
community” where team members function as both teachers and
learners throughout the collaborative process.

Stage 1. Identification and assessment was centered around locating

and interviewing formal service providers in the local community. As
part of the interview process, additional providers were identified by
the participants.

Stage 2. The team engaged in mapping nontraditional settings, as

well. This was directed to identifying places where fathers might

Module 3 Chapter 2: Overview of Qualitative Traditions | 201

seek or receive parenting support outside of formal service delivery
systems: barber shops, for example. They developed a map of the
community served by formal and informal/nontraditional systems.

Stage 3. This phase was labeled “engagement & relationship building”

(Lee, Hoffman, & Harris, 2016, p. 78). It involved creating a project
community advisory board made up of key stakeholders, and
developing working connections and relationships with researchers
involved with this particular community.

Stage 4. At this point, key stakeholder and university partners

collaborated to develop strategies and events to build the service
providers’ capacity to provide parenting support as needed and
desired by fathers in the community.

Stage 5. The final, evaluation stage involved trained investigators

coding data, then stakeholders involved in checking/confirming the
themes identified in the data (called “member checking” of results).
This final phase also included dissemination of the study findings.

In the end, this CBPR project identified multiple themes and at least
15 recommendations. Service providers identified service needs of
the community’s fathers and strategies for engaging fathers. The
fathers identified themes around engaging with children and
interacting with service providers. The recommendations are specific
with regards to what can change to provide support for the parenting
efforts of fathers in this community, and many are relevant to
consider adopting in other communities.

Content or Artifact Analysis Research. Content analysis is not so

much a qualitative approach as it is a means of analyzing qualitative
data. The important distinction lies in the sources of data compared
to studies where interviews are conducted. In content or artifact

202 | Module 3 Chapter 2: Overview of Qualitative Traditions

analysis, the investigator’s source of information is records or
artifacts previously produced during or about the phenomenon of
interest. Content analysis is typically conducted on artifacts that
were created for naturally occurring reasons, rather than for
purposes of the study being conducted. Historically, content analysis
was frequently conducted with personal diaries and letters written
to family members or close associates. Examples include studies of
first-hand report artifacts concerning how military personnel
experienced military combat or their experiences of being separated
from family members.

Given the important role of media on socialization, contents of media

Module 3 Chapter 2: Overview of Qualitative Traditions | 203

to which individuals are exposed might be analyzed for common
themes and messages that audiences might be receiving. For
example, a social worker might be concerned with learning about
the contents of books commonly read to young children in families
expecting a new baby. Parents read these books as a form of
“bibliotherapy” to inform their children about what to expect and
socialize them about how to behave in relation to the expected or
new family member. A content analysis of these books divulged a
common structure and theme: the new baby’s arrival is initially met
with a great deal of negative affect (jealousy, feeling neglected by
parents, being annoyed by the baby’s behavior), then in the end the
child has an epiphany whereby the baby is fun and becomes loved
(Begun, unpublished). Unfortunately, this plot structure and theme
approach is adult-centric—young children learn concepts not by the
logical sequencing of events but by what is most salient, affect-laden,
or recent. Therefore, the message to these children could be “don’t
like the baby” as much as it is “like the baby” when exposed to these
books.

Recent literature includes studies relying on content presented in

204 | Module 3 Chapter 2: Overview of Qualitative Traditions

social media postings as the studied artifacts (Facebook, twitter,
Snapchat, MySpace, and others). For example, investigators reviewed
the contents of more than 1,000 MySpace comments posted by
adolescents and emerging adults (aged 13-24 years) for the presence
of potential suicide statements (Cash et al, 2013). The content analysis
resulted in identifying 64 messages where a theme of potential
suicidality was present; these statements were associated with
subthemes about relationships, mental health concerns, substance
misuse, and suicide methods. The lesson learned from this study is
that young people are expressing themselves about suicide-related
thoughts and behaviors to their social networks using social
networking sites. One implication is that individuals using social
network sites might need to be educated about how to recognize,
interpret, and respond to these types of messages to become part of
a suicide prevention network.

Chapter Summary

In this chapter you read about different traditions of qualitative

research. You were introduced to narrative, phenomenological,
grounded theory, ethnographic, participatory action, and content/
artifact analysis research traditions. In future chapters you will learn
about the specific methods involved in data collection and engaging
participants in qualitative research studies.

Module 3 Chapter 2: Overview of Qualitative Traditions | 205

Module 3 Chapter 3: Overview
of Quantitative Traditions
Just as there were multiple traditions under the umbrella heading
of qualitative research approaches, different types of quantitative
research approaches are used in social work, social, and behavioral
sciences. In this chapter you are introduced to:

• cross-sectional research
• longitudinal research
• descriptive research
• correlational research
• experimental research.

Cross-Sectional and Longitudinal Research

Research for understanding diverse populations, social work

problems, and social phenomena might be conducted either cross-
sectionally or longitudinally. As you will see, the decision about using
a cross-sectional research or longitudinal research design has a lot
to do with the nature of the research questions being addressed,
the type of data available to investigators, and a weighing of the
advantages and disadvantages of either approach.

Cross-Sectional Research Design. A cross-sectional study involves

data collection at just one point in time. It lends itself to research
questions that do not relate to change over time but are descriptive,
exploratory, or explanatory about a specific point in time. In a study
addressing an epidemiological or other descriptive question, for

206 | Module 3 Chapter 3: Overview

of Quantitative Traditions
example, the cross-sectional design allows investigators to describe
a phenomenon at a single time point. An example might be a study
of the prevalence of individuals’ exposure to gun violence in their
communities and ratings of collective efficacy for the community
(Riley et al., 2017). Survey data were collected from participants at one
point in time (2014) in two communities with high rates of violent
crime. Collective efficacy was conceptualized as each individual’s
perception about formal and informal cohesion in their
neighborhood, the strength of social bonds among neighbors, and
the ability of community members to achieve public order through
informal mechanisms. The team hypothesized an inverse relationship
between these two factors: that higher rates of gun violence
exposure would correlate with lower scores for perceived collective
efficacy. Responses on 153 surveys were analyzed, demonstrating
that exposure to gun violence was very high (95% had heard
gunshots, 21% heard gunshots at least weekly, over 1/3 had been
present when someone was shot). The hypothesized relationship
between individuals’ higher exposure to gun violence and their lower
perceptions of community’s collective efficacy was observed.

Module 3 Chapter 3: Overview of Quantitative Traditions | 207

Cross-sectional research designs are also used in comparison studies
where a phenomenon’s expression is compared for two or more
groups—groups differing in race or ethnicity, sex or gender identity,
geographical location, or other social demographic factors, for
example. These studies might also compare groups of individuals
who received a certain intervention and those who did not. Or, this
design might compare groups of individuals who differ in severity
or other characteristics of social work or behavioral health problem.
A cross-sectional design was used to answer a question concerning
how bipolar disorder might differ from major depressive disorder
outside of major mood episodes (Nilsson, Straarup, & Halvorsen,
2015). Investigators compared 49 individuals with bipolar disorder
to 30 participants with major depressive disorder at a point in time
when each was in remission. Participants were compared on 15 “early
maladaptive schemas” which included their self-ratings of
abandonment, mistrust, social isolation, defectiveness, failure to
achieve, dependence, emotional inhibition, insufficient self-control,
and others. The investigators observed statistically different scores
between the two groups on 7 of the early maladaptive schemas, with
the bipolar diagnosis being associated with less favorable ratings.
This evidence contributes to a clinical understanding of differences
between these two diagnosed mood disorders, differences with
potential implications for intervention, especially with those
diagnosed with bipolar disorder.

208 | Module 3 Chapter 3: Overview of Quantitative Traditions

In other studies, the groups being compared using a cross-sectional
research design might also represent different age groups. The
conclusions drawn from such a study could be intended to help
develop or test theory concerning the developmental course of a
phenomenon or changes in it over time.

A cross-sectional study example explored the potential impact of

young children having been exposed to intimate partner violence
(Bowen, 2017). The investigator utilized one data point (when the
children were four years old) to assess the presence of conduct
disorder symptoms among children whose mothers reported their
partner having been emotionally cruel and/or having physically hurt
them during the child’s first 33 months. Among 7,743 children, the
mothers of 18.4% of girls and 17% of boys reported experiencing
a measured form of intimate partner violence during that
developmental period. Conduct problems were significantly more
common among these children compared to children whose mothers

Module 3 Chapter 3: Overview of Quantitative Traditions | 209

reported no exposure to intimate partner violence: 2.12 times more
common among exposed boys and 2.85 times more common among
exposed girls. The source of data for this study was longitudinal:
one cohort of families participated in data collection at multiple time
points during the course of the child’s development; however, the
study results reported by Bowen (2017) were based on one time point
alone, making it a cross-sectional study.

The cross-sectional study presents a single snapshot description of

the group’s aggregated ratings. A longitudinal study design is more
sensitive to fluctuations over time and to variation patterns
experienced by individual couples, presenting a more video-like
description. When the effect of time (or age) is of interest,
conclusions from cross-sectional studies must be drawn with caution
since many alternative explanations for any observed group
differences exist, most of which have little to do with the passage of
time or developmental changes. That is where longitudinal designs
are more helpful.

Longitudinal Research Design. A longitudinal study involves data

collection from the same participants at two or more points in time.
The beauty of this approach is that it allows investigators to directly
address questions involving change over time. Longitudinal studies
are a common means of studying or evaluating social work
interventions, for example. This design strategy is also common in
studies about human developmental processes. For example, a
longitudinal study design was used to study behavior and cognitive
functioning in 189 young adolescents who had experienced prenatal
cocaine exposure and 183 who had not (Minnes et al., 2016). Data were
collected when these individuals were 12 years old, and again when
they were aged 15. At age 12, girls with prenatal cocaine exposure
had more problems regulating their behavior and with metacognition

210 | Module 3 Chapter 3: Overview of Quantitative Traditions

than did girls without prenatal cocaine exposure; there was no
observed difference between the groups for boys. At this point, the
study conclusions could have been derived from a cross-sectional
study comparing the two groups at age 12 alone. However, the
investigators also reported that both behavioral and metacognition
problems improved from age 12 to 15 for the prenatally cocaine
exposed girls at a greater rate than the other groups. While some
differences did exist at age 15, it is important to note that the
conclusions that might have been drawn from the cross-sectional
element of the study (girls at age 12) are different from conclusions
drawn from the longitudinal report comparing ages 12 and 15.

The group of individuals being followed in a longitudinal study are

called a study cohort. In the above example, the cohort was a group
of babies born 12 years before the analyzed data were first collected.
This group continued to be followed at age 15. In addition to studying
a birth cohort like this, investigators might study other than birth
cohorts. For example, an incoming class of BSW students might be
followed each semester until graduation from the program. Or,
investigators might longitudinally follow a cohort of veterans and
their families as they return from deployment at about the same time.
Or, investigators might follow a group of survivors of a traumatic
event such as the 9/11 collapse of the twin towers in New York City or
the Washington, DC attack on the Pentagon. The concept of “cohort”
indicates that the study participants share a common experience or
event.

Comparing Cross-Sectional and Longitudinal Designs. These

two study designs result in somewhat different perspectives even if
the same phenomenon is being examined; much as an orange sliced
across its equator looks different from one sliced from top to bottom.

Module 3 Chapter 3: Overview of Quantitative Traditions | 211

Each design approach is associated with advantages and
disadvantages that come under consideration by investigators as they
plan their studies. Let’s take a closer look at what these
considerations might include.

Individual variation. Longitudinal studies, compared to cross-

sectional studies, are more sensitive to the unique, diverse patterns
of behavior observed for individuals over time. In cross-sectional
studies, individual differences are blended together for each group
being compared. In longitudinal studies, behavior at one point in time
is maintained within the context of behavior at prior and later points
in time. This provides useful information about how individuals
change.

Adaptability. One limitation of longitudinal studies is that

investigators are “stuck” with continuing to use the same (or very
similar) measurement approach throughout the life of a study.
Measurement consistency ensures that differences at different time
points are a function of time, rather than different measurement
approaches. Unfortunately, the science of measurement may

212 | Module 3 Chapter 3: Overview of Quantitative Traditions

progress over the course of a longitudinal study so that by the end of
the study, a better way to measure the variables of interest may have
emerged. A cross-sectional study can capitalize on the best, latest
measurement strategy available at the time the study is conducted.
The longitudinal study must continue to use what was the best,
latest measurement strategy at the start of the study, regardless of
progress in measurement science that has occurred by the end of the
study period.

Concurrent versus retrospective measurement.If investigators are

addressing questions about change, data collected concurrently at
each point in time may be more reliable than data collected at one
point asking about previous points in time. This latter retrospective
data may be contaminated by memory or other sources of bias,
making inferences about change less accurate.

Time. Imagine that investigators wanted to know how driving after

drinking excessive amount of alcohol might change as people age. A
study conducted in 2008 demonstrated that the percent of persons
who did so during the prior 12 months was highest among persons
aged 16-20 years (57%), and declined with each subsequent age group
to those aged 65 and older (see Figure 3-1). This cross-sectional
study, covering an age span of more than 50 years was conducted
in about four months. If the same question were to be studied
longitudinally, following the 16-year-old participants until they were
aged 65, and asking the question at intervals along the way, it would
take every bit of those 50 years to complete. Sometimes longitudinal
studies are simply impractical to undertake because of the time
involved.

Figure 3-1. Percent driving during past 12 months when thought to be

over the legal limit for alcohol consumption (N=1466 drivers who also
consume alcohol, adapted from NHTSA, 2010).

Module 3 Chapter 3: Overview of Quantitative Traditions | 213

Attrition. Investigators conducting cross-sectional studies, because
they only involve one measurement point for each study participant,
are not concerned with losing participants over time. Participant
attrition (drop-out) from longitudinal studies is a very serious
concern, and a great deal of effort needs to be directed toward
retaining each participant longitudinally over time. Participants
disappear for a number of reasons: losing interest in study
participation, the condition or phenomenon being studied changes,
moving out of the area, becoming incarcerated or otherwise being
lost to study continuation/follow-up. Furthermore, there exists
some degree of investigator turn-over/attrition possible, as well.

Cost. Longitudinal studies may be more costly than cross-sectional

studies to implement because of the need to maintain participants’
involvement over time (retention).

Statistical analyses. Different types of analyses need to be conducted

214 | Module 3 Chapter 3: Overview of Quantitative Traditions

on data that are longitudinal compared to cross-sectional data. This
is neither an advantage nor disadvantage, just a difference worthy of
note.

Descriptive, Correlational, and Experimental Research

Back in Module 2, you were introduced to research questions of an

exploratory, descriptive, or explanatory nature. Here we examine the
traditions in quantitative research that help address these different
types of research questions: descriptive, correlational, and
experimental studies.

Descriptive Research. The aim of a descriptive research study is

either to create a profile or typology of a population, or to describe a
phenomenon or naturally occurring process (Grinnell & Unrau, 2014;
Yegidis, Weinbach, & Meyers, 2018). Descriptive research contributes
important information for understanding diverse populations, social
work problems, and social phenomena. An example of descriptive
research was a study aimed at identifying different substance use
patterns among pregnant adolescents (Salas-Wright, Vaughn, &
Ugalde, 2016). The study investigators started with an understanding
from previous research that there exists an overall pattern of
elevated substance use levels prior to conception compared to non-
pregnant peers, and of reduced use during pregnancy. However,
these investigators were concerned that this general, aggregate
picture of pregnant adolescents might mask different patterns
(heterogeneity) having important social work practice implications.
The team analyzed substance use data for 810 pregnant adolescents,
aged 12-17. They were able to discern and describe four different
pattern types: abstaining from any substance use, using only alcohol,
using both alcohol and cannabis, and polydrug use. Those who only
drank alcohol prior to conception tended to cut back or eliminate

Module 3 Chapter 3: Overview of Quantitative Traditions | 215

drinking during pregnancy. Those who used alcohol and cannabis or
engaged in polydrug use were fewer in number than those who only
used alcohol, but also were less likely to stop using substances during
pregnancy. Those who engaged in polydrug use were the youngest
group and most likely to meet criteria for a substance use disorder.
The different substance use types were not evenly distributed by
demographics (ethnicity, age, family income) or pregnancy stage
across study participants. The results of this descriptive study have
social work and public health implications related to intervening in
different ways with adolescents of different types to prevent or
reduce prenatal exposure to substances.

Correlational Research. The aim in some studies is to evaluate the

216 | Module 3 Chapter 3: Overview of Quantitative Traditions

existence and nature ofrelationships that might exist between
variables—looking to see if variable xand variable yare associated
or correlated with each other. These are called correlational
research studies. As mentioned in reference to descriptive studies,
the investigator does not manipulate variables experimentally:
naturally occurring relationships among variables is the focus of
study. An interesting approach to understanding social work
problems and social phenomena involves the use of geospatial data.
You may have heard the term GIS—short for geographic information
system—which is about data that have a geographical, spatial
component. This type of data is useful for social workers to
understand how a problem or phenomenon might be experienced
in the environmental contexts where people live, work, recreate, or
otherwise function.

A correlational study using geospatial data examined relationships

between use of and physical access to marijuana with child
maltreatment (Freisthler, Gruenewald, & Price Wolf, 2015). First, the
investigators demonstrated that parental use of marijuana was
positively correlated with the frequency of child physical abuse,
negatively correlated with physical neglect, and not correlated with
supervisory neglect. Parents reporting marijuana use reported
engaging in physical abuse three times more frequently than parents
who did not use marijuana, with boys more likely than girls to be
physically abused, and older children more likely to be physically
abused than younger children. The data for these results came from
telephone interviews with 3,023 persons in a state with “legalized”
medical marijuana dispensaries and delivery services. The geospatial
data came into play in analyses indicating that greater geographical
density of medical marijuana dispensaries and delivery services was
positively related to greater frequency of physical child abuse. These
findings have important implications for social work practice and
social policy, including that marijuana use may be problematic even

Module 3 Chapter 3: Overview of Quantitative Traditions | 217

if a marijuana use disorder is not diagnosed in the child welfare
context, and that there may exist differences in the impact of
marijuana use in communities where distribution is legitimized,
relegated to illicit “street-level dealers,” or not present.

Correlation versus Causality. The major limitation of correlational

research is that the tradition is unable to definitively determine
causality in observed relationships between variables of interest.
When a correlation is observed between two variables, xandy, at least
four possibilities exist:

• x influences y

• y influences x

218 | Module 3 Chapter 3: Overview of Quantitative Traditions

 • x and y influence each other in a more complex, circular,
iterative manner, where x influences y which influences x which
influences y and so forth

• some third variable, z, causes both x and y but these two do not
directly influence each other

To logically draw conclusions about causality, at least three

conditions need to be met.

1. Evidence exists that the two variables are related—as values for the
“cause” change, values for the “effect” also change systematically in
relation to those changes in the “cause.” This is what the correlation
statistic can demonstrate: changes in xrelate to changes in y. In
addition, that the two are related must be consistently
demonstrated—that changing the “cause” consistently (not

Module 3 Chapter 3: Overview of Quantitative Traditions | 219

arbitrarily) results in “effect” changes (Cozby, 2007). Consider the
example presented earlier about exposure to gun violence and
perception of community efficacy (Riley et al., 2017). The investigators
observed that as ratings of exposure to gun violence increased (the
xor hypothesized “cause” variable), ratings of perceived collective
efficacy decreased (the y or hypothesized “effect” variable). This kind
of relationship between the variables is called a negative or inverse
correlation—as one goes up the other goes down. Depending on the
actual values of the relationship, it might look something like Figure
3-2 where more exposure to gun violence (moving from left to right)
is associated with lower community efficacy.

Figure 3-2. Depiction of negative (inverse) correlation using

hypothetical values

2. The “cause” and “effect” must be sequenced in time, with the

“cause” preceding (coming before) the “effect.”For example, consider
the earlier-mentioned Bowen (2017) study concerning conduct
disorder among four-year-old children who have been exposed to a
mother’s experience of intimate partner violence. Before concluding
that the exposure caused the conduct disorder symptoms, it is

220 | Module 3 Chapter 3: Overview of Quantitative Traditions

important to demonstrate that the children’s conduct disorder
symptoms did not predate the exposure to intimate partner violence.
If the conduct disorder came first, it is quite possible that the stress
of managing a child with these behaviors influenced the intimate
partner violence (y caused x) instead.

3. There is not another, third factor that “causes” both other factors.
A demonstrative example comes from a purposefully absurd graph
attributed on the internet to Bobby Henderson, used to demonstrate
that correlation does not imply causation. The erroneous conclusion
drawn from the data is that global warming (y variable) is a direct
effect of the shrinking number of pirates (variable x). As you can see
from Figure 3-3, between the years 1860 and 2000 the number of
Caribbean pirates declined from 45,000 to 17, and in the same period,
the average global temperature climbed from 14.25 degrees Celsius
to 16 degrees Celsius (“data” retrieved from https://en.wikipedia.org/
wiki/Flying_Spaghetti_Monster).

Module 3 Chapter 3: Overview of Quantitative Traditions | 221

While the “data” indicate an inverse association between the two
variables (notice that the approximate number of pirates is declining
as year is increasing on the x axis), it is absurd to conclude that the
reduction in pirate numbers caused global warming. A third variable
might be causing both: for example, the mechanism of change in both
might be related to the global spread of industrialization.

In summary, correlational research studies allow investigators to

identify possible associations between factors, helping to develop
better understanding about diverse populations, social work
problems, and social phenomena. However, additional research is
necessary to determine if there are causal relationships involved. This
is important if we are to use evidence to inform the interventions
we develop and test. If we do not understand the causal nature of
the relationships and the mechanisms by which observed effects
are caused, then we cannot determine the best places and ways
to intervene. This understanding is enhanced through experimental
research.

Experimental Research

The aim of experimental research is to answer explanatory questions:

to test hypotheses or theory about phenomena and processes. This
includes testing theory and social work interventions. Our next
course, SWK 3402, emphasizes research and evaluation to
understand interventions. Here we focus on experimental research
to understand diverse populations, social work problems, and social
phenomena. Experiments are designed to help develop our
understanding of causal relationships between factors, the etiology
of social work problems and social phenomena, and the mechanisms
by which change might occur. Investigators develop and discover
this kind of information through systematic manipulation of specific

222 | Module 3 Chapter 3: Overview of Quantitative Traditions

factors and observing the impact of those manipulations on the
outcomes of interest. The study designs and methods applied in
experimental research are selected to eliminate as many alternative
explanations for the results as possible, thereby increasing
confidence that the tested mechanism or explanation is accurately
accepted: “Explanatory research depends on our ability to rule out
other explanations for our findings” (Engel & Schutt, 2013, p. 19). This
observation leads naturally to a discussion of internal validity and
external validity in experimental research.

Internal Validity. The concept of internal validity refers to the degree

of confidence that can be applied to the results of an experimental
study. Remember, the purpose of an experiment is to develop an
understanding of causal relationships between variables—in other
words, to test explanatory hypotheses. Strong internal validity means
greater confidence in drawing causal or explanatory conclusions
from the data; this is sometimes referred to as study integrity.
Experimental design is sometimes referred to as “the gold standard”
against which all others are compared. This, however, would only
be true if the aim of those being compared is explanatory—many
other designs are important, depending on the nature of the research
questions and aims. If, however, the aim is explanatory, designing a
study that allows for investigator control over the greatest number of
possible explanatory factors, ruling out alternative explanations for
the observed experimental results, is most desirable—it means the
study has strong internal validity.

External Validity. External validity is about the degree to which a

study’s findings can be generalized to the population represented
by that sample. External validity is again about confidence in the
results, but this time it is about how well the results from a particular

Module 3 Chapter 3: Overview of Quantitative Traditions | 223

study represent a general reality for other similar people and other
similar settings. For example, on attitudes about social issues, studies
conducted with college student samples from private universities
may poorly represent the nation’s population of college-aged adults,
or even the population of the nation’s college students. Similarly,
studies about a social work problem conducted with people in
treatment or receiving services might not be representative of people
experiencing the same problem but who are not in treatment and
may never have sought treatment for the problem. This is called the
clinical sample problem related to external validity concerns. As you
will see in Chapter 6, external validity is powerfully influenced by
the procedures involved in developing and retaining a study sample,
including (but not limited to) problems of small sample size.

The Monitoring the Future study represent an example of strong

cross-sectional survey methodology, contributing to the study’s high
external validity. The Monitoring the Future study has been
conducted with 12thgrade students every year since 1975, and with
8th, 10th, and 12thgraders since 1991—you may very well have
participated in the study at some point in your life. Some young
adults are being surveyed in more recent years, as well. The purpose
is to study trends in U.S. adolescents’ beliefs, attitudes, and behavior.
To enhance generalizability of the annual findings, about 50,000
students are surveyed each year. These students are from about 420
public and private high schools across the nation, and include about
18,000 8thgraders from about 150 schools, about 17,000 10thgraders,
and about 16,000 12thgraders from about 133 schools. A randomly
selected subset of participants has been followed longitudinally every
two years since 12thgrade, as well. The geographic regions are
randomly selected each year, with random selection of the schools
in the selected areas, and random selection of the classes to be
included within each school. Random selection is a strong external
validity contributor. You may be interested to see some of the trend

224 | Module 3 Chapter 3: Overview of Quantitative Traditions

results reported by the study investigators, located at
http://www.monitoringthefuture.org/.

If you have enjoyed television shows like Bill Nye the Science Guyand
Mythbusters, then you have witnessed systematic experiments being
implemented. Their demonstrations typically include elements of
scientific process designed to enhance rigor of the experiments
under way. A great deal of social science, behavioral science, and
social work research follows scientific methods and logic developed
in physical and natural sciences. You might wonder why all social
work research is not conducted under these scientific, experimental
procedures. The simple answer is that a great deal of social work
research is designed to address other types of questions:
experimental research is designed to answer explanatory questions,
not the other types that we have discussed.

Module 3 Chapter 3: Overview of Quantitative Traditions | 225

A more complex answer includes that many possible experiments
cannot be practically or ethically conducted. For example, we cannot
ethically expose people to conditions that might cause them harm. To
understand the impact of disrupted parent-child attachment bonds
we cannot experimentally manipulate this condition to see what
happens; instead, we need to rely on observing the effects through
what happened “naturally,” such as when policy led to children and
immigrant/refugee parents being separated at U.S. borders while
parents were detained for having entered the country illegally.
Similarly, we could not ethically conduct controlled experiments
where children are randomly assigned to being raised by parents
either with or without substance misuse problems, where families
are randomly assigned to living in homes with or without lead
contamination, or where communities are randomly assigned to
having or not having access to affordable, healthy food. Abuse and
ethical violations in experiments conducted with unknowing, under-
informed, unempowered, and otherwise vulnerable persons led to
the development of policies, guidelines, and practices discussed in
your CITI training about research involving human
subjects—including the Belmont Report and Institutional Review
Boards.

226 | Module 3 Chapter 3: Overview of Quantitative Traditions

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=213

Chapter Summary

This chapter introduced the traditions of quantitative research. You

read about the distinctions, advantages, and disadvantages of cross-
sectional and longitudinal study designs. You also read about
descriptive, correlational, and experimental studies, and were
reminded that correlation does not imply causation. Finally, the
topics of internal and external validity were examined. The next
chapter further develops a foundation for understanding the wide
range of available options and decisions to be made by investigators
planning experimental research. The focus includes understanding
the nature of variables involved in quantitative studies, whether they
are descriptive, exploratory, or experimental in nature.

Module 3 Chapter 3: Overview of Quantitative Traditions | 227

Module 3 Chapter 4: Overview
of Quantitative Study
Variables
The first thing we need to understand is the nature of variables
and how variables are used in a study’s design to answer the study
questions. In this chapter you will learn:

• different types of variables in quantitative studies,

• issues surrounding the unit of analysis question.

Understanding Quantitative Variables

The root of the word variable is related to the word “vary,” which
should help us understand what variables might be. Variables are
elements, entities, or factors that can change (vary); for example,
the outdoor temperature, the cost of gasoline per gallon, a person’s
weight, and the mood of persons in your extended family are all
variables. In other words, they can have different values under
different conditions or for different people.

We use variables to describe features or factors of interest. Examples

might include the number of members in different households, the
distance to healthful food sources in different neighborhoods, the
ratio of social work faculty to students in a BSW or MSW program,
the proportion of persons from different racial/ethnic groups
incarcerated, the cost of transportation to receive services from a
social work program, or the rate of infant mortality in different
counties. In social work intervention research, variables might

228 | Module 3 Chapter 4: Overview

of Quantitative Study Variables
include characteristics of the intervention (intensity, frequency,
duration) and outcomes associated with the intervention.

Demographic Variables. Social workers are often interested in what

we call demographic variables. Demographic variables are used to
describe characteristics of a population, group, or sample of the
population. Examples of frequently applied demographic variables
are

• age,
• ethnicity,
• national origin,
• religious affiliation,
• gender,
• sexual orientation,
• marital/relationship status,
• employment status,
• political affiliation,
• geographical location,
• education level, and
• income.

At a more macro level, the demographics of a community or

organization often includes its size; organizations are often measured
in terms of their overall budget.

Independent and Dependent Variables. A way that investigators

think about study variables has important implications for a study
design. Investigators make decisions about having them serve as
either independent variables or as dependent variables. This
distinction is not something inherent to a variable, it is based on

Module 3 Chapter 4: Overview of Quantitative Study Variables | 229

how the investigator chooses to define each variable. Independent
variables are the ones you might think of as the manipulated “input”
variables, while the dependent variables are the ones where the
impact or “output” of that input variation would be observed.

Intentional manipulation of the “input” (independent) variable is not

always involved. Consider the example of a study conducted in
Sweden examining the relationship between having been the victim
of child maltreatment and later absenteeism from high school: no
one intentionally manipulated whether the children would be victims
of child maltreatment (Hagborg, Berglund, & Fahlke, 2017). The
investigators hypothesized that naturally occurring differences in the
input variable (child maltreatment history) would be associated with
systematic variation in a specific outcome variable (school
absenteeism). In this case, the independent variable was a history of
being the victim of child maltreatment, and the dependent variable
was the school absenteeism outcome. In other words, the
independent variable is hypothesized by the investigator to cause
variation or change in the dependent variable. This is what it might
look like in a diagram where “x” is the independent variable and “y”
is the dependent variable (note: you saw this designation earlier, in
Chapter 3, when we discussed cause and effect logic):

For another example, consider research indicating that being the

victim of child maltreatment is associated with a higher risk of
substance use during adolescence (Yoon, Kobulsky, Yoon, & Kim,

230 | Module 3 Chapter 4: Overview of Quantitative Study Variables

2017). The independent variable in this model would be having a
history of child maltreatment. The dependent variable would be risk
of substance use during adolescence. This example is even more
elaborate because it specifies the pathway by which the independent
variable (child maltreatment) might impose its effects on the
dependent variable (adolescent substance use). The authors of the
study demonstrated that post-traumatic stress (PTS) was a link
between childhood abuse (physical and sexual) and substance use
during adolescence.

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=222

Types of Quantitative Variables

There are other meaningful ways to think about variables of interest,

as well. Let’s consider different features of variables used in
quantitative research studies. Here we explore quantitative variables
as being categorical, ordinal, or interval in nature. These features
have implications for both measurement and data analysis.

Module 3 Chapter 4: Overview of Quantitative Study Variables | 231

Categorical Variables. Some variables can take on values that vary,
but not in a meaningful numerical way. Instead, they might be
defined in terms of the categories which are possible. Logically, these
are called categorical variables. Statistical software and textbooks
sometimes refer to variables with categories as nominal variables.
Nominal can be thought of in terms of the Latin root “nom” which
means “name,” and should not be confused with number. Nominal
means the same thing as categorical in describing variables. In other
words, categorical or nominal variables are identified by the names
or labels of the represented categories. For example, the color of the
last car you rode in would be a categorical variable: blue, black, silver,
white, red, green, yellow, or other are categories of the variable we
might call car color.

What is important with categorical variables is that these categories

have no relevant numeric sequence or order. There is no numeric
difference between the different car colors, or difference between
“yes” or “no” as the categories in answering if you rode in a blue car.
There is no implied order or hierarchy to the categories “Hispanic
or Latino” and “Not Hispanic or Latino” in an ethnicity variable; nor
is there any relevant order to categories of variables like gender, the
state or geographical region where a person resides, or whether a
person’s residence is owned or rented.

If a researcher decided to use numbers as symbols related to

categories in such a variable, the numbers are arbitrary—each
number is essentially just a different, shorter name for each category.
For example, the variable gender could be coded in the following
ways, and it would make no difference, as long as the code was
consistently applied.

232 | Module 3 Chapter 4: Overview of Quantitative Study Variables

Coding Option A Variable Categories Coding Option B
1 male 2
2 female 1
3 other than male or female alone 4
4 prefer not to answer 3

Race and ethnicity.One of the most commonly explored categorical

variables in social work and social science research is the
demographic referring to a person’s racial and/or ethnic
background. Many studies utilize the categories specified in past U.S.
Census Bureau reports. Here is what the U.S. Census Bureau has to
say about the two distinct demographic variables, race and ethnicity
(https://www.census.gov/mso/www/training/pdf/race-ethnicity-
onepager.pdf):

What is race? The Census Bureau defines race as a person’s

self-identification with one or more social groups. An
individual can report as White, Black or African American,
Asian, American Indian and Alaska Native, Native Hawaiian
and Other Pacific Islander, or some other race. Survey
respondents may report multiple races.

What is ethnicity? Ethnicity determines whether a person is of

Hispanic origin or not. For this reason, ethnicity is broken out
into two categories, Hispanic or Latino and Not Hispanic or
Latino. Hispanics may report as any race.

In other words, the Census Bureau defines two categories for the
variable called ethnicity (Hispanic or Latino and Not Hispanic or
Latino), and seven categories for the variable called race. While these
variables and categories are often applied in social science and social
work research, they are not without criticism.

Module 3 Chapter 4: Overview of Quantitative Study Variables | 233

Based on these categories, here is what is estimated to be true of the
U.S. population in 2016:

Dichotomous variables.There exists a special category of categorical

234 | Module 3 Chapter 4: Overview of Quantitative Study Variables

variable with implications for certain statistical analyses. Categorical
variables comprised of exactly two options, no more and no fewer
are called dichotomous variables. One example was the U.S. Census
Bureau dichotomy of Hispanic/Latino and Non-Hispanic/Non-
Latino ethnicity. For another example, investigators might wish to
compare people who complete treatment with those who drop out
before completing treatment. With the two categories, completed
or not completed, this treatment completion variable is not only
categorical, it is dichotomous. Variables where individuals respond
“yes” or “no” are also dichotomous in nature.

The past tradition of treating gender as either male or female is

another example of a dichotomous variable. However, very strong
arguments exist for no longer treating gender in this dichotomous
manner: a greater variety of gender identities are demonstrably
relevant in social work for persons whose identity does not align
with the dichotomous (also called binary) categories of man/woman
or male/female. These include categories such as agender,
androgynous, bigender, cisgender, gender expansive, gender fluid,
gender questioning, queer, transgender, and others.

Ordinal Variables. Unlike these categorical variables, sometimes a

variable’s categories do have a logical numerical sequence or order.
Ordinal, by definition, refers to a position in a series. Variables with
numerically relevant categories are called ordinal variables. For
example, there is an implied order of categories from least-to-most
with the variable called educational attainment. The U.S. Census data
categories for this ordinal variable are:

• none
• 1st-4thgrade
• 5th-6thgrade

Module 3 Chapter 4: Overview of Quantitative Study Variables | 235

 • 7th-8thgrade
• 9thgrade
• 10thgrade
• 11thgrade
• high school graduate
• some college, no degree
• associate’s degree, occupational
• associate’s degree academic
• bachelor’s degree
• master’s degree
• professional degree
• doctoral degree

In looking at the 2016 Census Bureau estimate data for this variable,
we can see that females outnumbered males in the category of having
attained a bachelor’s degree: of the 47,718,000 persons in this
category, 22,485,000 were male and 25,234,000 were female. While
this gendered pattern held for those receiving master’s degrees, the
pattern was reversed for receiving doctoral degrees: more males than
females obtained this highest level of education. It is also interesting
to note that females outnumbered males at the low end of the
spectrum: 441,000 females reported no education compared to
374,000 males.

Here is another example of using ordinal variables in social work

research: when individuals seek treatment for a problem with alcohol
misuse, social workers may wish to know if this is their first, second,
third, or whatever numbered serious attempt to change their
drinking behavior. Participants enrolled in a study comparing
treatment approaches for alcohol use disorders reported that the
intervention study was anywhere from their first to eleventh
significant change attempt (Begun, Berger, Salm-Ward, 2011). This
change attempt variable has implications for how social workers

236 | Module 3 Chapter 4: Overview of Quantitative Study Variables

might interpret data evaluating an intervention that was not the first
try for everyone involved.

Rating scales. Consider a different but commonly used type of ordinal

variable: rating scales. Social, behavioral, and social work
investigators often ask study participants to apply a rating scale to
describe their knowledge, attitudes, beliefs, opinions, skills, or
behavior. Because the categories on such a scale are sequenced (most
to least or least to most), we call these ordinal variables.

Examples include having participants rate:

• how much they agree or disagree with certain statements (not

at all to extremely much);
• how often they engage in certain behaviors (never to always);
• how often they engage in certain behaviors (hourly, daily,
weekly, monthly, annually, or less often);
• the quality of someone’s performance (poor to excellent);
• how satisfied they were with their treatment (very dissatisfied
to very satisfied)
• their level of confidence (very low to very high).

Interval Variables. Still other variables take on values that vary in

a meaningful numerical fashion. From our list of demographic

Module 3 Chapter 4: Overview of Quantitative Study Variables | 237

variables, age is a common example. The numeric value assigned to
an individual person indicates the number of years since a person
was born (in the case of infants, the numeric value may indicate
days, weeks, or months since birth). Here the possible values for the
variable are ordered, like the ordinal variables, but a big difference is
introduced: the nature of the intervals between possible values. With
interval variables the “distance” between adjacent possible values are
equal. Some statistical software packages and textbooks use the
term scale variable: this is exactly the same thing as what we call an
interval variable.

For example, in the graph below, the 1 ounce difference between this
person consuming 1 ounce or 2 ounces of alcohol (Monday, Tuesday)
is exactly the same as the 1 ounce difference between consuming
4 ounces or 5 ounces (Friday, Saturday). If we were to diagram the
possible points on the scale, they would all be equidistant; the
interval between any two points is measured in standard units
(ounces, in this example).

238 | Module 3 Chapter 4: Overview of Quantitative Study Variables

With ordinal variables, such as a rating scales, no one can say for
certain that the “distance” between the response options of “never”
and “sometimes” is the same as the “distance” between “sometimes”
and “often,” even if we used numbers to sequence these response
options. Thus, the rating scale remains ordinal, not interval.

What might become a tad confusing is that certain statistical

software programs, like SPSS, refer to an interval variable as a “scale”
variable. Many variables used in social work research are both
ordered and have equal distances between points. Consider for
example, the variable of birth order. This variable is interval because:

• the possible values are ordered (e.g., the third-born child came
after the first- and second-born and before the fourth-born),
and
• the “distances” or intervals are measured in equivalent one-
person units.

Continuous variables. There exists a special type of numeric interval

variable that we call continuous variables. A variable like age might
be treated as a continuous variable. Age is ordinal in nature, since
higher numbers mean something in relation to smaller numbers. Age

Module 3 Chapter 4: Overview of Quantitative Study Variables | 239

also meets our criteria for being an interval variable if we measure it
in years (or months or weeks or days) because it is ordinal and there
is the same “distance” between being 15 and 30 years old as there is
between being 40 and 55 years old (15 calendar years). What makes
this a continuous variable is that there are also possible, meaningful
“fraction” points between any two intervals. For example, a person
can be 20½ (20.5) or 20¼ (20.25) or 20¾ (20.75) years old; we are
not limited to just the whole numbers for age. By contrast, when
we looked at birth order, we cannot have a meaningful fraction of a
person between two positions on the scale.

The Special Case of Income. One of the most abused variables in

social science and social work research is the variable related to
income. Consider an example about household income (regardless
of how many people are in the household). This variable could be
categorical (nominal), ordinal, or interval (scale) depending on how it
is handled.

Categorical Example: Depending on the nature of the research

questions, an investigator might simply choose to use the
dichotomous categories of “sufficiently resourced” and “insufficiently
resourced” for classifying households, based on some standard
calculation method. These might be called “poor” and “not poor”
if a poverty line threshold is used to categorize households. These
distinct income variable categories are not meaningfully sequenced
in a numerical fashion, so it is a categorical variable.

Ordinal Example: Categories for classifying households might be

ordered from low to high. For example, these categories for annual
income are common in market research:

• Less than $25,000.

• $25,000 to $34,999.

240 | Module 3 Chapter 4: Overview of Quantitative Study Variables

 • $35,000 to $49,999.
• $50,000 to $74,999.
• $75,000 to $99,999.
• $100,000 to $149,999.
• $150,000 to $199,999.
• $200,000 or more.

Notice that the categories are not equally sized—the “distance”

between pairs of categories are not always the same. They start out
in about $10,000 increments, move to $25,000 increments, and end
up in about $50,000 increments.

Interval Example. If an investigator asked study participants to report

an actual dollar amount for household income, we would see an
interval variable. The possible values are ordered and the interval
between any possible adjacent units is $1 (as long as dollar fractions
or cents are not used). Thus, an income of $10,452 is the same
distance on a continuum from $9,452 and $11,452—$1,000 either way.

The Special Case of Age. Like income, “age” can mean different things
in different studies. Age is usually an indicator of “time since birth.”
We can calculate a person’s age by subtracting a date of birth variable
from the date of measurement (today’s date minus date of birth). For
adults, ages are typically measured in years where adjacent possible
values are distanced in 1-year units: 18, 19, 20, 21, 22, and so forth.
Thus, the age variable could be a continuous type of interval variable.

Module 3 Chapter 4: Overview of Quantitative Study Variables | 241

However, an investigator might wish to collapse age data into ordered
categories or age groups. These still would be ordinal, but might
no longer be interval if the increments between possible values are
not equivalent units. For example, if we are more interested in age
representing specific human development periods, the age intervals
might not be equal in span between age criteria. Possibly they might
be:

• Infancy (birth to 18 months)

• Toddlerhood (18 months to 2 ½ years)
• Preschool (2 ½ to 5 years)
• School age (6 to 11 years)
• Adolescence (12 to 17 years)
• Emerging Adulthood (18 to 25 years)
• Adulthood (26 to 45 years)
• Middle Adulthood (46 to 60 years)
• Young-Old Adulthood (60 to 74 years)
• Middle-Old Adulthood (75 to 84 years)
• Old-Old Adulthood (85 or more years)

Age might even be treated as a strictly categorical (non-ordinal)

variable. For example, if the variable of interest is whether someone
is of legal drinking age (21 years or older), or not. We have two
categories—meets or does not meet legal drinking age criteria in the
United States—and either one could be coded with a “1” and the other
as either a “0” or “2” with no difference in meaning.

What is the “right” answer as to how to measure age (or income)?

The answer is “it depends.” What it depends on is the nature of
the research question: which conceptualization of age (or income) is
most relevant for the study being designed.

242 | Module 3 Chapter 4: Overview of Quantitative Study Variables

Alphanumeric Variables. Finally, there are data which do not fit into
any of these classifications. Sometimes the information we know is
in the form of an address or telephone number, a first or last name,
zipcode, or other phrases. These kinds of information are sometimes
called alphanumeric variables. Consider the variable “address” for
example: a person’s address might be made up of numeric characters
(the house number) and letter characters (spelling out the street, city,
and state names), such as 1600 Pennsylvania Ave. NW, Washington,
DC, 20500.

Actually, we have several variables present in this address example:

• the street address: 1600 Pennsylvania Ave.

• the city (and “state”): Washington, DC
• the zipcode: 20500.

This type of information does not represent specific quantitative

categories or values with systematic meaning in the data. These are
also sometimes called “string” variables in certain software packages
because they are made up of a string of symbols. To be useful for an
investigator, such a variable would have to be converted or recoded
into meaningful values.

Module 3 Chapter 4: Overview of Quantitative Study Variables | 243

A Note about Unit of Analysis

An important thing to keep in mind in thinking about variables is

that data may be collected at many different levels of observation.
The elements studied might be individual cells, organ systems, or
persons. Or, the level of observation might be pairs of individuals,
such as couples, brothers and sisters, or parent-child dyads. In this
case, the investigator may collect information about the pair from
each individual, but is looking at each pair’s data. Thus, we would
say that the unit of analysis is the pair or dyad, not each individual
person. The unit of analysis could be a larger group, too: for example,
data could be collected from each of the students in entire
classrooms where the unit of analysis is classrooms in a school or
school system. Or, the unit of analysis might be at the level of
neighborhoods, programs, organizations, counties, states, or even
nations. For example, many of the variables used as indicators of
food security at the level of communities, such as affordability and
accessibility, are based on data collected from individual households
(Kaiser, 2017). The unit of analysis in studies using these indicators
would be the communities being compared. This distinction has
important measurement and data analysis implications.

A Reminder about Variables versus Variable Levels

A study might be described in terms of the number of variable

categories, or levels, that are being compared. For example, you
might see a study described as a 2 X 2 design—pronounced as a two
by two design. This means that there are 2 possible categories for the
first variable and 2 possible categories for the other variable—they are
both dichotomous variables. A study comparing 2 categories of the
variable “alcohol use disorder” (categories for meets criteria, yes or

244 | Module 3 Chapter 4: Overview of Quantitative Study Variables

no) with 2 categories of the variable “illicit substance use disorder”
(categories for meets criteria, yes or no) would have 4 possible
outcomes (mathematically, 2 x 2=4) and might be diagrammed like
this (data based on proportions from the 2016 NSDUH survey,
presented in SAMHSA, 2017):

Illicit Substance Use Disorder (SUD)

No Yes
No 500 10
Alcohol Use Disorder (AUD)
Yes 26 4

Reading the 4 cells in this 2 X 2 table tells us that in this (hypothetical)

survey of 540 individuals, 500 did not meet criteria for either an
alcohol or illicit substance use disorder (No, No); 26 met criteria for
an alcohol use disorder only (Yes, No); 10 met criteria for an illicit
substance use disorder only (No, Yes), and 4 met criteria for both an
alcohol and illicit substance use disorder (Yes, Yes). In addition, with
a little math applied, we can see that a total of 30 had an alcohol use
disorder (26 + 4) and 14 had an illicit substance use disorder (10 + 4).
And, we can see that 40 had some sort of substance use disorder (26
+ 10 + 4).

Module 3 Chapter 4: Overview of Quantitative Study Variables | 245

 To make this distinction
between variables and variable levels or categories crystal clear, let’s
consider one more example: a 2 X 3 study design. First, doing the
math, we should see 6 possible outcomes (cells). Second, we know
that the first variable (age group) has 2 categories (under age 30, 30
or older) and the other variable (employment status) has 3 categories
(fully employed, partially employed, unemployed). This time the 6
cells of our design are empty because we are waiting for the data.

Employment Status
Fully Employed Partially Employed Unemployed

<30
Age Group
≥30

Thus, when you see a study design description that looks like two

246 | Module 3 Chapter 4: Overview of Quantitative Study Variables

numbers being multiplied, that is essentially telling you how many
categories or levels of each variable there are and leads you to
understand how many cells or possible outcomes exist. A 3 X 3 design
has 9 cells, a 3 X 4 design has 12 cells, and so forth. This issue becomes
important once again when we discuss sample size in Chapter 6.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.3-4.1 Beginning Data Entry

Chapter Summary

In summary, investigators design many of their quantitative studies

to test hypotheses about the relationships between variables.
Understanding the nature of the variables involved helps in
understanding and evaluating the research conducted.
Understanding the distinctions between different types of variables,
as well as between variables and categories, has important
implications for study design, measurement, and samples. Among
other topics, the next chapter explores the intersection between
the nature of variables studied in quantitative research and how
investigators set about measuring those variables.

Module 3 Chapter 4: Overview of Quantitative Study Variables | 247

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=222

248 | Module 3 Chapter 4: Overview of Quantitative Study Variables

Module 3 Chapter 5: Overview
of Methods for Data
Collection and Measurement
Remember that one defining characteristic of scientific inquiry is
that it involves observation. This chapter is concerned with the many
different approaches to observation that are commonly utilized in
social, behavioral, and social work research. Each has its advantages
and disadvantages, and investigators are faced with significant
challenges in planning the measurement/observation aspects of
their studies. In this chapter you will read about:

• validity and reliability principles related to quantitative

measurement
• qualitative and quantitative measurement and data collection
approaches

Measurement Validity & Reliability in Quantitative Research

The point of quantitative measurement is to objectively measure

characteristics, features, or variables of interest. While this may seem
obvious, it is important not to confuse variables and
measures—measurement tools provide information about variables,
they are not variables themselves. This measurement topic is focused
on quantitative studies because it is not relevant to the aims of
qualitative research. Two measurement traits are important in
achieving the quantitative measurement task: measurement
validity and measurement reliability. While neither trait can be
realized 100%, the aim is to maximize both to the greatest extent

Module 3 Chapter 5: Overview of

Methods for Data Collection and
Measurement | 249
possible. Let’s look more closely at each of these traits and what they
mean in quantitative research.

Measurement Validity. Measurement validity is about the extent

to which a measurement tool or instrument adequately measures
the concept, construct, or variable of interest (Perron & Gillespie,
2015). In other words, does it really measure what it purports to be
measuring (Grinnell & Unrau, 2014). It may help to note that this use
of the term “validity” is getting at a very different point than did the
terms internal and external validity that your read about earlier in
terms of experimental study designs. There we were concerned with
either how confident we were in the results of the study (internal
validity) or the conclusions of the study being generalizable (external
validity). Here, we are concerned with the measurement procedures
employed in generating those results. Measurement validity plays
a role in both internal validity and external validity of a study, but
operates at a more specific level of the research process.

Our general definition of measurement validity includes aspects or

components that you might read about: construct, content, and face
validity, for example. While these are not all the same thing, they all
contribute to a general assessment of measurement validity. Validity
is sometimes assessed by convergent validity analysis, as well. In
convergent validity, measurement scores or values are compared
with those obtained on other measures of the same construct or
variable, measured at the same time. For example, the URICA
(University of Rhode Island Change Assessment) is a 32-item tool
used by practitioners to assess clients’ motivational readiness to
change (McConnaughy, Prochaska, & Velicer, 1983). If investigators
or clinicians desired a shorter, briefer version or a version for self-
administered use on mobile devices, they would first need to

250 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
demonstrate that the new version or format led to the same results
as the original form.

Additionally, it is important that investigators recognize limitations of

instruments and measures for work with different target populations.
For example, the need to know how validly an instrument or measure
translates into other languages. The problem with direct translation
into another language is that nuances relevant to different cultures
may play a role in how the measure is experienced. For example,
translating an instrument developed in English into an instrument
for Spanish-speaking persons is not simply a matter of using an
English-Spanish dictionary and substituting words. There are many
differences between the Spanish spoken and used in daily life among
persons from Spain, Mexico, Chile, Cuba, and other areas, and there
are many colloquialisms and slang terms that do not translate word-
for-word. For example, the word “dope” has many meanings even in
English: someone who does something stupid, marijuana or heroin
(depending on region and cohort), insider information, or something
that is awesome. A recommended practice is to have a native speaker
of the language conduct a translation, then have another native
speaker reverse translate it back into English and compare the
original to the double-translated version to see where meaning was
altered or lost.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 251
Furthermore, it is important to be aware that language interpretation
is not the same thing as cultural interpretation. In other words, just
because the words are literally translated correctly does not mean
that they have the same cultural meaning across cultures. Careful
attention must be directed toward creating or selecting culturally
relevant or culturally grounded approaches to measurement,
instruments and measures that accurately capture the lived
experiences of the studied population.

Latent variables. One important difference between the physical and

natural sciences compared to social, behavioral, and social work
science is related to whether the phenomena of interest are directly
or only indirectly observable. Much of the science directed towards
understanding diverse populations, social work problems, and social
phenomena involves studying internal mental processes, attitudes,
beliefs, and interpretations.A variable that is not directly observed

252 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
but is mathematically or statistically inferred from other variables
that are directly measuredis called a latent variable. For example, in
creating the Safe At Home instrument for measuring readiness to
change battering behavior (Begun et al., 2003; 2008), investigators
inferred the variables for Precontemplation, Contemplation, and
Preparation/Action from the study participants’ ratings of their
agreement with 35 readiness statements. They created these 3 latent
variables through a statistical approach called exploratory factor
analysis (EFA)—see Figure 5-1. Validity of these sorts of inferred
variables is an important issue to consider, and evaluating a research
study may involve evaluating the evolution of inferred, latent
variables that may have been used.

Figure 5-1. Items loaded in Contemplation (latent variable) through

EFA (constructed from results presented in Begun et al., 2003)

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 253
Proxy variables. A commonly encountered validity issue arises with
the use of proxy variables. The word proxy means something (or
someone) is representing something (or someone) else. A proxy voter
is authorized to vote on behalf of someone else. A proxy variable is a
variable that stands for or represents something else. A proxy variable
is not directly relevant in itself, but serves as a close approximation
of the variable of interest. Proxy variables are used when the variable
of interest is difficult (or impossible) to measure; they are easily
measured variables that can be used in place of more difficult-to-
measure variables.

For example, birth order can be either a variable of interest used

on social research, or it can act as a proxy variable for something
more difficult to measure, such as parental attention or parenting
experience. When we see a correlation between birth order and
a trait (like intelligence or personality), we do not necessarily
understand that relationship. Could it be the result of a biological
phenomenon related to health of the prenatal, uterine environment
which may change with each subsequent birth? Or, is it a function
of how much time, undivided attention, and social/intellectual
attention each subsequent child receives from parents? This is the
problem with proxy variables—we do not know how validly they fit the
actual construct for which they are standing in. Factors other than
actual birth order number (1st, 2nd, 3rd and so forth) are associated
with how much parental time, attention, and stimulation is received,
making birth order a less than ideal proxy variable for the
hypothesized mechanism of effect on child intelligence. Investigators
generally do better to find direct measures of the constructs in which
they are interested than to rely on proxy variables. Individuals
reviewing a research report need to critically analyze the degree
to which they believe a proxy variable is a good representative for
the variable of interest. By the way, you may be interested to learn
the results of a large-scale study about birth order (Rohrer, Egloff,
& Schmukle, 2015): the authors concluded that “birth order does

254 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
not have a lasting effect on broad personality traits outside of the
intellectual domain” (p. 14224).

Also consider the problem of using age or income as a proxy variable

for something else. For example, age is a relatively unreliable proxy
for physical condition or health. Consider research that indicates
20% or more of U.S. children and adolescents have abnormally high
cholesterol levels, a condition associated with poor cardiovascular
health among adults (Nguyen, Kit, & Carroll, 2015). Or, consider older
adults whose condition is similar to average adults in much younger
age groups—marathons have age groupings of 70 and over or 80
and over—while many younger adults could not meet the physical
challenges involved. Age is not a great proxy for fitness. Similarly,
income might not be a good proxy for families consuming nutritious

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 255
food. Some persons reside in “food deserts” where nutritious food is
difficult to obtain within a reasonable geographic distance regardless
of their available economic resources. Income is not entirely
predictive of this experience since some families with low incomes
are able to grow much of their own healthful food, while others
cannot, and some families with higher incomes spend their resources
on poor quality “fast” or “junk” foods instead of healthful foods.

Proxy variables are risky to apply to populations where there exists

a great deal of heterogeneity. Unfortunately, investigators are
sometimes forced to rely on proxy variables when measuring the true
variable of interest is not possible. This problem is often encountered
when working with administrative or secondary data.

Measurement Reliability. As a complement to measurement validity,

measurement reliability indicates the degree of accuracy or precision
in measuring the construct or variable of interest. This accuracy
indicator is concerned with stability in measurement, for one thing:
if you use the tool or instrument to measure the same thing multiple
times, you would hope to get the same result. Imagine trying to
measure your cereal box with a rubber band—you want to know how
many stretched rubber bands big it is. If you measured it 5 times in
a row, the chances are that you would not get the same answer all 5

256 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
times because the rubber band is such a flexible tool and you cannot
be sure you applied the same amount of pressure each time .

Reliability problems arise in real-world research, as well. Imagine,

for example, that your goal was to measure aggressive behaviors
exhibited by a toddler. It might be easy to count kicking, biting,
shoving, and hitting behaviors. But how would you count instances
where the toddler yells or spits at someone, raises a hand to strike
but doesn’t follow through, or kicks the person’s chair rather than
the person? Would everyone observing that behavior reliably count
it the same way? This gets at the issue of precision. An important
dimension of measurement reliability is called inter-observer
reliability. This might also be called inter-rater or inter-coder
reliability, depending on the nature of the measurement data. In
studies where qualitative, open-ended questions, or observational
data are being evaluated for analysis, strong research methodology
involves applying methods to assess and ensure that different
observers, raters, or coders would get the same result when
evaluating the same data. For example, the team might spend time
up front training a group of observers to rate or code a set of sample
data. Once everyone has developed a strong consensus on the
practice data (at least 90% agreement is one standard), pairs of
observers might rate or code a percentage of the actual data, with
their evaluations being compared. Again, the expectation is that their
level of agreement is very high, and that it remains high over time
(not just when the training was still fresh in their minds).

There exists a tremendous need to ensure reliability in measurement,

particularly in social work research where variables might be difficult
to directly observe and quantify. For example, social work
practitioners and researchers might be concerned with reliably
measuring anxiety. Asking different individuals to rate how anxious
they feel right now on a scale of 0-10 might not generate very reliable
answers. While zero might be clear to everyone (no anxiety at all),

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 257
for a person who often experiences very high levels of anxiety, a “5”
might mean something very different than for a person who seldom
experiences anxiety or whose anxiety is mild when experienced.

The goal with measurement reliability is to minimize measurement

error. Measurement error is a recognized problem in social work,
social science, and behavioral science research. It concerns the
difference between value obtained through measurement of a
variable and the variable’s true (unknown) value. Investigators do
their best to be accurate in measurement, but a certain amount
of “noise” is always a concern in the data collected. Measurement
error leads investigators to either over- or under- estimate reality.
Turning back to our anxiety measurement example, measurement
error or bias may be introduced by the context in which anxiety is
measured. Measuring immediately before or after a quiz might lead
to different results than measuring first thing in the morning after a
good night’s sleep. Or, women being interviewed about experiences
of sexual harassment might respond differently to male versus female
interviewers. Or, data collected on Mondays might be either more
accurate after participants have had a restful weekend or might be
less accurate after their partying weekends, compared to data
collected on other days of the week.

Cultural Competence. An issue integral to both measurement validity

and reliability concerns cultural competence in
measurement procedures. You may be familiar with the long-
standing controversy concerning cultural bias in standardized
testing, particularly intelligence tests. A demonstrative example
comes from an intelligence test purportedly administered to Koko,
the gorilla who learned to communicate with humans using American
Sign Language. When asked where you should go if it starts to rain,
Koko “failed” the item by responding “up a tree” instead of “inside

258 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
a house.” The test is clearly “human-centric” and biased against
gorillas. Similarly, there are many examples of ethnic, cultural,
geographic, gender, and other bias built into testing materials. For
example, the T-ACE (Tolerance, Annoyance, Cut Down, and Eye
Opener) and TWEAK (Tolerance, Worry, Eyeopener, Amnesia, Kut
Down) are two alcohol screening tests developed for use with women
when it was determined that existing screening measures were more
appropriate for use only with men (Russell, 1994).

Cultural equivalence/nonequivalence and avoiding ethnocentrism

are important consideration when evaluating or planning
measurement in quantitative research. The issue is that concepts
or constructs may have very different meaning to persons living in
different cultural contexts. Consider, for example, the meaning of
marriage to individuals in same- versus different-sex relationships,
particularly in light of the unstable history of marriage policy across
the nation and the globe. Also consider that equivalents for a
phenomenon that exists in one cultural group may not exist in
another. An example might be the sibling-like cousin relationships
that exist among cultural groups characterized by high degrees of
cooperation and community compared to the nuclear family
conception of siblings.

One type of research question concerns measurement science:

validating research measurement and clinical assessment tools for
use with diverse populations. This may begin with the initial
development of a measurement approach/instrument, or involve
recommended modifications of existing tools based on evidence. An
important “how to” resource is a book called Developing cross-
cultural measures in social work research and evaluation, 2ndedition
(Tran, Nguyen, & Chan, 2016).

A concern beyond the actual measurement tools is the measurement

procedures involved in a study. Of great concern is the way that

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 259
the measurement process might put study participants in jeopardy.
One example is measurement that asks about illegal behaviors, such
as underage drinking, illicit use of drugs at any age, or substance
use during pregnancy. Asking for honest, truthful answers to these
kinds of questions can place participants in jeopardy with legal or
child welfare systems if the confidentiality is breached. One tool
for managing this risk is the Certificate of Confidentiality that an
investigator might request from the National Institutes of Health
(NIH) in connection with securing local IRB approval to conduct this
type of research. The Certificate of Confidentiality is designed to
protect investigators from being forced by court order to disclose
individually identifiable data.

Another scenario to consider is discussed in an article about safety

in research involving unauthorized immigrant women experiencing
intimate partner violence:

“Social science research with vulnerable populations is

necessary in order to address social and health problems among
those most in need. Yet, this research is fraught with inherent
risks to its participants” (Kyriakakis, Waller, Kagotho, &
Edmond, 2015, p. 260).

The authors expressed concern over the potential physical,

emotional, psychological, economic, and legal harms to which these
women could be exposed by virtue of the research process. The risk
to the 29 Mexican immigrant women was related to their recent
history of exposure to intimate partner violence, and for 26 of the
women the risk was compounded by their unauthorized immigrant
status. The authors described various strategies used to minimize
the women’s risks of participation, including cautious recruitment,
screening, informed consent, interviewing, data management/
storage, resource referral, and data reporting procedures.

260 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Ways of Measuring and Collecting Data

Whether a scientific study adopts a qualitative or quantitative

approach, investigators need to determine a strategy for collecting
data about the phenomenon being observed. Options vary along
several dimensions, one of which is the degree to which strategies
are obtrusive. Think about your family’s home videos and the pictures
you have on your phone. There is a great deal of difference between
the candid and posed images—the candid images capture more
natural behavior than do the posed images.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 261
The same is true of scientific measurement: behavior observed under
natural conditions is more representative than is behavior observed
under laboratory conditions or conditions where those being
observed are aware that their behavior is being studied. This is the
first dimension we explore with regards to ways of measuring and
collecting data.

Naturalistic Observation. Both qualitative and quantitative studies

can utilize naturalistic observation techniques in data collection. This
means that information about behavior is collected in normally
experienced contexts. If an investigator wishes to learn about
aggression between preschool-aged children, the behavior might be
observed in naturally occurring situations and natural
environments—daycare, playgrounds, at home, or riding in the car.
The advantage of this approach is that the data are representative
of natural behavior. The disadvantages of this approach include lack
of investigator control over circumstances or variables that might
interfere with the behavior being observed. For example, home

262 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
observations might be interrupted by the doorbell or telephone
ringing and the dog barking, which might alter the behavior of those
being observed. Laboratory conditions can minimize these
confounding factors, but observations of behavior expressed in the
laboratory may not accurately reflect or generalize to natural
behavior conditions. Another factor which investigators take into
account is the possibility that the behavior of interest might not
occur often enough to actually be observed under natural conditions.
Investigators may feel the need to create laboratory conditions to
“trigger” the behavior of interest so that it can be observed. For
example, imagine that investigators are interested in observing how
families negotiate the distribution of resources across members and
for the family as a whole. This may be difficult to observe during
a conveniently scheduled period of observation. The investigators
might, instead, provide the family members with a specific
distribution task and observe how they negotiate a solution—perhaps
providing them with a scratch-off State lottery ticket and asking
them to decide what to do if were a $1,000 winner. While this is
not a typical, routine resource distribution problem that families
encounter, the task is designed to “trigger” resource distribution
behavior from the family members being observed. The limitation
being that family members will need to advise the investigators as
to how typical their behavior was compared to naturally occurring
situations.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 263
The type of data collected from naturalistic observation differs
between qualitative and quantitative studies. Qualitative data might
include rich, detailed descriptions of who does/says what, when they
do/say it, how they do/say it, and the sequential context of what
happens–the precursors and consequences of each behavior. These
data may later be coded for themes or other qualitative analyses.
Quantitative data might include numeric counts of the frequency of
certain behaviors or the duration (in seconds or minutes) of specific
behaviors of interest. To accomplish this, the investigators need to
have predetermined what behaviors will be recorded as data, being
very precise as to the definition and criteria for those behaviors.
In both cases, qualitative and quantitative, the ability to review the
action is helpful to ensure data accuracy and completeness. Thus,
recording the observed events is common so that the events can
repeatedly be replayed. Furthermore, working with recordings allows
investigators to cross-check data decisions made by individuals
coding, scoring, or rating the observed behaviors.

Direct observational data is particularly useful in studies where

participants are not able to provide information about themselves
or their usual behavior. This is generally true of infants and young
children, for example. As their language/communication skills
develop and mature, individuals’ ability to answer questions about
their internal thought processes and behavior improves, at least
somewhat. Observation may be a strong option in studying persons
who have cognitive or intellectual barriers to consistently providing
information—persons with dementia, for example.

Photovoice. An approach to qualitative data collection that differs

from observing individuals in their natural environments involves
having individuals themselves actively engage in observing their
natural environments. Photovoice is a powerful participatory action

264 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
strategy for data collection (Wang, 1999). The approach begins with
having individuals take photographs of what is meaningful to them
in their daily lives. It is particularly important as a method used
with traditionally marginalized groups or communities (Fisher-Borne
& Brown, 2018). The second phase of photovoice methodology is
equally, if not more, important—recording the participants’
explanations of the images they chose to capture or create. The use
of photovoice methodology was presented as a strategy for engaging
young, Black men as researchers in community-based participatory
research (CBPR) with the goal of shaping an HIV prevention and care
project, ensuring that they are empowered to engage and have their
voices heard (Fisher-Borne & Brown, 2018). In this study, participants
documented engaged in action-oriented, problem-solving group
discussion sessions about their photos. This feature distinguished the
study from documentary photovoice projects—the emphasis of the
discussions was on helping the men identify issues and action steps
toward personal and societal change. Each week of the project had a
theme around which the men took pictures (e.g., stereotypes, Black
masculinity, sexuality, self-esteem, emotions, multiple “identities”).
The pictures were discussed in group sessions, with each
photographer selecting their own discussion photo, providing a name
or label for it, and leading discussion about it in relation to the theme.
The analyses which included the study participants resulted in 4
themes: identity, stereotypes reinforce oppression, expectations, and
breaking free/embracing self. The authors argued in favor of using
this approach as a qualitative, action-oriented research method:

“Embedded within the Photovoice technique is the belief that

people can participate and define for themselves issues that
will shape their lives. While Photovoice is a vehicle for personal
expression, its structure provides for the vocalization of
community and public issues. As such, the visual images and

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 265
 participatory process serve as a tool for advocacy and collective
action” (Fisher-Borne & Brown, 2018, p. 171).

Artifact Analysis. Another form of unobtrusive observation involves

analyzing the remains or artifacts that remain after naturally

266 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
occurring behavior. For example, alcohol researchers may wish to
record data concerning the alcohol purchases made by study
participants. Participants could be asked to save all receipts (paper
and electronic) for all purchases made in a month. These receipts
can then be content analyzed for purchases involving alcohol. This
artifact analysis approach may be more representative of actual
alcohol purchasing behavior than asking about the
behavior—individuals may be unaware of or mistaken about their
actual purchasing behavior, either under- or over-estimating reality.
Other types of artifacts that might be analyzed include personal
diaries and social media postings, as well as data recordings from
devices such as fitness and GPS trackers. As in the case of Photovoice
methodology, other arts-based approaches might be appropriate:
eliciting and analyzing dance, poetry, drawings/painting, sculpture,
and other artifacts.

Administrative and Secondary Data Sources. Client records and

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 267
administrative data are forms of artifacts that might usefully be
analyzed to answer some research questions, as well. The
information may not have originally been collected for research
purposes, but may be informative in research all the same.
Administrative data includes information routinely recorded in
programs, agencies, and institutions as part of normal operations.
For example, investigators analyzed data routinely collected at the
state level as child abuse and neglect (CA/N) reports (Lanier at al.,
2014). The investigators calculated disproportionality ratios for each
state by comparing the state’s child maltreatment rate for black and
white groups, and for Hispanic and white groups. They then related
these state-by-state data to statistics about each state’s poverty,
teenaged mother, and single mother rates. The results of this study
using administrative data indicated that disproportionality in child
maltreatment rates was correlated with disproportionality in poverty
rates, and this was observed for both black and Hispanic groups.
Furthermore, a significant amount of the child neglect
disproportionality between black and white groups (66%) was
accounted for by disproportionality in teen motherhood rates. This
kind of study and the observed disproportionate poverty-
maltreatment links have profound social justice implications for
social work practice and policy.

Administrative data differs from what you may have heard called
secondary data analysis. Secondary analysis concerns information
collected for a research purpose, but analyzed differently than in
the primary study for which the information was originally collected.
The original study involved primary data collection, the later study
involves secondary data analysis. For example, numerous
investigators have secondarily analyzed data from the original Fragile
Families and Child Wellbeing study. The original is an ongoing, multi-
year, longitudinal study of child development outcomes in relation
to numerous parent, family, and neighborhood factors. Barnhart and

268 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Maguire-Jack (2016) reported on their analyses concerning the role
of parenting stress and depression in the relationship between single
mothers’ perception of community social cohesion and child
maltreatment. They analyzed data from a subset of 1,158 single
mothers from the full original study. The investigators reported that
single mothers’ perception of community social cohesion has an
effect on child maltreatment, and the nature of this effect is
influenced by maternal depression: greater social cohesion was
associated with lower depression rates, and lower depression was
associated with lower child maltreatment rates. These findings
demonstrate the importance of social workers attending to both
individual and community level factors when developing
interventions to prevent or reduce child maltreatment; intervening
around community social cohesion might make a difference in child
maltreatment because it can influence maternal depression. This is
also important because social workers not only care about child well-
being, we care about the well-being of mothers, as well.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 269
The two examples (Lanier, et al., 2014 and Barnhart & Maguire-Jack,
2016) demonstrate the greatest advantage associated with
administrative and secondary data. The investigators addressed their
research questions without having to expend tremendous time and
resources in generating new primary data. They avoided duplicating
the data gathering efforts that have already been expended. That
does not mean that administrative and secondary data analysis
studies are easier or less effort-intensive than primary data studies:
the effort and time is distributed differently. These studies still need
to undergo IRB review for the protection of human subjects to ensure
that the risks are minimal or less. These two studies also demonstrate
one of the most significant limitations involved with administrative
and secondary data. The primary data collection may not have
included the best, direct measures of the variables in which the
investigators were interested. There may exist a need for faith in the
leap of logic that extends between the available data and the variables
that the investigators wish to analyze.

Furthermore, using administrative or secondary data leaves the new

study vulnerable to any limitations in methodology associated with
the primary study or the administrative data. For example, in a study
of prisoner visitation patterns by family members, investigators were
able to access data concerning every visit made to every state
prisoner in Ohio over the preceding five years (Begun, Hodge, &
Early, 2017). The study was limited, however, by imprecise and high
variability in how family relationships were reported in the
administrative data: mother, mother-in-law, step-mother, and parent
were all used differently by different recorders and at different
institutions. Similarly, someone might be described as a sibling,
brother, brother-in-law, step-brother, half-brother, step-sibling, or
half-sibling. Thus, the investigators were limited in the degree of
detail they could apply to the variable of family relationship—the
possible categories had to be combined into more general, less
specific categories for the sake of accuracy.

270 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Geographic Information Systems (GIS). You were introduced to
geographic information systems (GIS) research in the earlier
discussion about correlational studies, and the example about
marijuana dispensary density in relation to child maltreatment rates.
This is considered a non-intrusive form of data collection, since it is
not directly experienced by the individuals whose environments are
mapped. The approach can be combined with other, more intrusive
approaches, to generate answers to important questions about
diverse populations, social work problems, and social phenomena.
For example, investigators examined the nature of parents’
engagement with “activity spaces” using both qualitative interview
data and GIS methods (Price Wolf, Freisthler, Kepple, & Chavez, 2017).
The aim of the study was to better understand how physical and
social environments shape parenting behaviors, which in turn,
influence child development and health outcomes. They found that
parenting had a powerful influence on the places where parents
go (child-centric places with their children, “adults-only” places to
escape from parenting demands). In addition, day of the week,
children’s ages, role as primary or other caregiver, and income all
influenced parents’ use of activity spaces by imposing boundaries
that limited options. This information provides social work with
insights concerning where community-based interventions for
parents might best be situated to be accessible for these families, and
when different spaces might best be targeted.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 271
Key Informants. One method of collecting data is to have
knowledgeable others provide descriptions rather than collecting
information directly from individuals. Sometimes this approach is
combined with data collection from study participants, as well. Key
informants are helpful when studying individuals who cannot
effectively share reliable information about themselves—young
children, for example, or persons experiencing significant short- or
long-term cognitive/intellectual impairments. Child development
studies often involve assessment tools completed by parents and
teachers. This approach is reasonable when the studied behaviors
are observable, but not with studies of internal mental or emotional
processes. Great care must also be taken to protect the rights of the
individuals being studied.

Surveys. A great deal of social research is conducted using survey

tools for data collection. It is important to distinguish between
surveys as tools for data collection and survey methodology. Survey
methodology refers to how survey study samples are drawn from

272 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
the population (more about this in Chapter 6) and the use of survey
instruments. Surveys are administered through a variety of
mechanism: hard copy, paper and pencil surveys; telephone or in-
person interview surveys; and, on-line, e-mail or web-based surveys.
Most likely, you have participated in surveys from product marketing,
course or instructor evaluations, public opinion research, and others.

A great deal has been written about constructing “good” survey

instruments and questions. First, it helps to consider the type of
information to be provided for a survey question: objective, factual
data or information about a person’s subjective opinions, beliefs,
attitudes, perceptions, ideas, or states (Fowler & Cosenza, 2009).
Asking effective factual questions is not always as simple and
straightforward as it may seem. Imagine a study where investigators
want to know about a person’s alcohol consumption. They plan to ask
study participants:

How often and how much do you drink?

You have probably already identified one problem with this question:
drink what? It is very important to clearly define terms in survey
questions. The revised question might be:

How often and how much do you drink beverages containing

alcohol?

While this question is better defined than the first draft, it still suffers
from another common problem: it is a double-barreled question. A
double-barreled question is presented as a single question, but
actually asks two different questions needing two distinct answers.
In this case, one question is about frequency of alcohol consumption,
the other is about amount. Recognizing this, the investigators might
change their survey to ask:

How often do you drink beverages containing alcohol?

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 273
 When you drink beverages containing alcohol, how much do
you drink?

There remains a significant problem with the second of these two

questions—a lack of precision about how amounts might be reported.
One person might respond “a lot,” another might respond “it
depends.” The investigators need to make clear that they are asking
for a specific, concrete, objective number here. Still, if a person
responds “3” in reference to the number of drinks, the investigators
still do not know how much alcohol was consumed because different
types of alcoholic beverages contain different amounts of alcohol,
and it is unclear how large the 3 drinks might be. Figure 4-1 shows
how questions about drinking amounts could be asked.

Figure 4-1. Using the chart* below, on a day that you drink alcohol,
how many standard drinks do you consume?

Retrieved from https://www.niaaa.nih.gov/alcohol-health/

overview-alcohol-consumption/what-standard-drink

274 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Asking questions about a person’s subjective internal states can also
be complicated. As we saw with asking factual questions, defining
terms and avoiding double-barreled question both remain important.
It is important that the wording of a question not be biased at the
outset. Bias can be introduced unintentionally by using specific
labels—consider the difference between the labels “victim” and
“survivor” of intimate partner violence, for example. Another
potential source of bias comes from the phenomenon recognized
as social desirability responding. Sometimes the answers people
provide reflect what they believe is the most favorable or “correct”
response rather than a response that truly reflects their own
reality—aiming to be “PC” (“politically correct” rather than accurate.
The literature contains several measures of social desirability,
questions that can be integrated into a survey, to determine the
extent to which an individual’s answers might be suffering from social
desirability bias.

In addition, the options provided in the survey for responding to

questions need to be carefully considered. Response options include
(but are not limited to):

• Open ended or fill-in-the-blank. This option provides the

greatest degree of latitude and potential variability in
responding. Depending on the nature of the questions asked,
the responses can be qualitatively (rather than quantitatively)
analyzed. If the goal of the study is quantitative, these types of
response are more difficult to work with. Problems include
difficulty with interpreting handwriting and spelling errors, or
non-standard uses of words or terms. Responses may also vary
greatly in terms of how much information and detail different
participants provide—some may give a word or two while others
provide great amounts of detail. This response variability can be
difficult to analyze.
• Yes/no. This option offers only two possibilities, meaning that it

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 275
 is “forced choice” question. This is sometimes used as a
screening question to determine the appropriateness of asking
further questions on the topic. For example, our investigators
asking about alcohol consumption amounts might first ask:
Have you consumed any beverages containing alcohol in the
past 12 months? Then, they would as the frequency and amount
questions only of individuals who responded “yes” to the
screening questions; those responding “no” would skip the
frequency and amount questions. One problem with force
choice questions is that study participants may become
frustrated, especially when the answer is not quite so cut and
dried. For example, a screening question that asks, “Do you
drink beverages containing alcohol?” would be difficult to
answer with either “yes” or “no” by a woman who drank until she
became pregnant and now does not, but will again after the
baby is born.
• Rating scales:Survey questions often ask participants to apply a
continuum or rating scale in answering a subjective question.
This might be in the form of how much they agree with specific
statements, their evaluation of something (degree of goodness,
fairness, attractiveness, or other qualities), or estimations about
how often something occurs. Two general types of scales exist.
One is a single, unidimensional scale, such as rating happiness
from low to high. The other is a two-dimensional scale, such as
rating affect from unhappy to happy with a neutral value in the
middle. It is important when deciding which type of scale to
apply that both ends of the two-dimensional scale cannot occur
at the same time for the same person. For example, a person
might be unhappy about some things and happy about others at
the same time. In this case, it would be more informative to have
two separate unidimensional scales (one for happiness and one
for unhappiness) instead of relying on a single two-dimensional
scale, since feeling neutral is very different from feeling both

276 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
 unhappy and happy.

It might be wise, to offer participants options outside of the main

choices. Response options like “unsure,” “don’t know,” or “prefer not
to answer” might keep someone from either skipping the question
or providing an inaccurate response. On the other hand, it may be
that a forced choice is desired by the investigators and an alternative,
unscaled option is not offered.

Another consideration with rating scales is how many points on the

scale are offered. Too few options (e.g., a 3-point scale being 0, 1,
2) may lead to there being very little variation in the responses
provided—it would be too easy for everyone to pick the same value,
perhaps the middle of the range (ratings of 1). On the other hand,
having too many options makes it difficult for people to decide on
an appropriate rating—for example, when rating something on a
100-point scale, would there be a meaningful difference between
ratings of 70 and 71? It would be easier for someone to use a 10-point
scale (where the response might be 7). A common solution is the
5-point scale with two extremes, a neutral midpoint, and two
modified options between the midpoint and extremes. For example:

Notice in this example that each possible numeric response has an

associated description. In survey terms, these ratings are “anchored”
by the words provided. Sometimes a scale is presented with some
anchors being unspecified. For example:

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 277
This leaves the meaning of the ratings between 2 to 6 open to
individual interpretation. Additionally, the anchors do not have to be
words or phrases. For example, symbols or emoticons might be used,
especially in working with children. The pain scale is an example that
might be usefully modified for surveys (retrieved from www.disabled-
world.com/health/pain/scale.php):

A note about Likert scales. A common error in describing survey scales

revolves around use of the term “Likert scale.” By definition, a Likert
scale is a set of individual items that, in combination, form a single
index (or scale). A Likert scale is made up of Likert items. Likert items
generally have five or seven options on the rating continuum, but
the midpoint is always neutral. Just because items have five rating
options arrayed on a continuum does not make it a Likert item—a
unidimensional item has no neutral midpoint, so it is not a Likert
item. A Likert scale score is calculated by mathematically combining
the ratings for individual Likert items, usually as either a total (sum)
or as an average (mean) score. In some cases, the developer of a scale
might provide alternative scoring instructions (such as multiplying
some item ratings to provide weighted scores if some items are
more important than others to the meaning of the final score). For

278 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
example, ratings on the following four Likert items about readiness
to succeed in a social work course about research and statistics are
simply added together to create a composite Likert scale score for
each study participant—one participant’s hypothetical responses are
scored in this example, receiving a score of 14 on a possible range of
4 to 20 (see Table 4-1).

Table 4-1. Likert items and Likert scale scoring example

How much do you agree or disagree with each statement

below?

1=completely disagree

2=somewhat disagree Item

Rating
3=neither disagree nor agree

4= somewhat agree

5=completely agree
Research evidence is important for social work practice 4
Math is relatively easy for me 2
Intellectual challenges and puzzles are enjoyable to work on 3
I am confident that I can master the content in this course 5

Scale Score 14

Reverse scoring scale items. In order to compute a single, composite

score by adding ratings together requires all of the items to run in
the same direction with regard to the construct being measured.
Consider what would happen if the second item (B) had been phrased
differently: Math is relatively difficult for me. In this case, the item
would be negative with respect to the construct of readiness to
succeed in the course. Thus, the rating scale would need to be

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 279
reversed before the rating values could be added together. In other
words, ratings of 1 and 5 would be flipped so that ratings of 1 become
5, and ratings of 5 become 1; similarly, ratings of 2 and 4 would be
flipped so that ratings of 2 become 4, and ratings of 4 become 2.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.3-5.1 Summing Scores

Screening and Assessment Tools. Very often, investigators utilize

existing screening and assessment tools for measuring the
phenomena or constructs of interest in a research study. The
literature is rich with measures designed to screen for or assess many
different behaviors, conditions, or social work problems. It is usually
preferable to use pre-existing measurement instruments rather than
creating new ones for a particular study, particularly if the literature
supports their reliability and validity for the job (more about this
later in this chapter). Electronic search terms that help investigators
and practitioners locate pre-existing measures on a specific topic
include: tests, measures, measurements, assessment(s), scale(s), and
screening.

Locating measures that can be applied in research studies is

facilitated by various catalogues, compilations, guides, and
handbooks that review them. For example, the Mental Measurements
Yearbook(Carlson, Geisinger, & Jonson, 2017) is a frequently updated
published resource with information to facilitate the selection of

280 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
measurement tools by practitioners and investigators. An
investigator interested in measures concerning attention deficit with
hyperactivity disorder (ADHD) would find at least 4 options reviewed
in the series (from 1985 to the present). A listing of all reviewed
measures is presented on the publication website
(https://buros.org/tests-reviewed-mental-measurements-
yearbook-series). Many of the measures are commercially available,
meaning that there is potentially a fee associated with their use—they
may not be available in the public domain—and many require
specialized training or certification.

Additional guides to research measures, including other than clinical

behaviors, are:

• Handbook of clinical rating scales and assessment in psychiatry

and mental health(Baer & Blais, 2010)
• Handbook of family measurement techniques, v. 1-3(Touliatos,
Perlmutter, Straus, & Holden, 2001)
• Handbook of psychiatric measures, 2nded. (Rush, First, & Blacker,
2008)
• Handbook of research design and social measurement,
6thed.(Miller & Salkind, 2002)
• Measures for clinical practice and research, 5thed, v.
1&2(Corcoran & Fischer, 2013)
• Positive psychological assessment: A handbook of models and
measures(Lopez & Snyder, 2003)
• Rating scales in mental health(Sajatovic & Ramirez, 2012)

Important cautions revolve around the use of screening and

assessment results as research measures. First, investigators must
develop procedures for providing test results to study participants,
especially those whose results indicate a clinical problem or high
risk of developing one. Second, the investigator needs to understand
the measurement tool sufficiently well to determine whether actual

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 281
scores can be used as a scale variable or if the test results can only be
relied on for the yes/no dichotomous categories related to whether
or not the person meets criteria for the condition being studied. For
example, a measure of depression might help determine if a person
is “over” or “under” the threshold for meeting depression criteria,
but actual raw depression scores cannot be treated as a scale: the
difference between a score of 38 and 40 is not the same as the
difference between a score of 48 and 50. The AUDIT (Alcohol Use
Disorder Identification Test) is an example of a screening instrument
that has been used both ways (Babor, Higgins-Biddle, Saunders, &
Monteiro, 2010). This screening measure is in the public domain and
has 10 items, each of which is rated from 0 to 4 points based on
the rubric presented for each item. An AUDIT score of 8 or higher
for men, 7 or higher for women, is considered a “positive” screening
result, indicating that an alcohol use disorder possibly exists. On the
other hand, higher AUDIT scores are indicative of greater risk, and
the actual scores can be used as a scale (ranging from 0 to 40). Or, the
scores can be broken into 4 categories:

• Zone 1 scores 0-7 intervene with alcohol education alone,

• Zone 2 scores 8-15 intervene with simple advice about alcohol
misuse,
• Zone 3 scores 16-19 intervene with simple advice plus brief
counseling and continued monitoring,
• Zone 4 scores 20-40 intervene with referral to specialist for
diagnostic evaluation and treatment (Babor, Higgins-Biddle,
Saunders, & Monteiro, 2010, p. 22).

Interviews. Whether they are engaged in a qualitative, quantitative,

or mixed-methods study, investigators may rely on interviews to
gather data. Interview protocols vary in terms of the extent to which
they are pre-scripted and structured. Qualitative studies in the
grounded theory tradition typically are relatively unstructured since

282 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
the purpose is to learn about the individual’s own perspectives and
interpretations about events, processes, or phenomena. The “flow”
of such an interview follows from what the study participant shares
rather than from a script predetermined by the investigator. At the
other end of the continuum, quantitative study interviews typically
are quite structured and scripted so as to elicit similar types of
information (though varying in content) from each study participant.
The semi-structured interview format is situated between the
unstructured and highly structure interview formats. The questions
in a semi-structured interview are general and open-ended and the
“script” may include specific prompts to help guide the process. The
research interview requires an investigator to employ many of the
same skills used in social work clinical practice.

An important characteristic defining interview questions is whether

they are open- or close-ended questions. Questions that can be
answered with simple statements of fact or with yes/no responses
are close-ended questions. An example might be:

How many brothers and sisters lived with you 50% of the time
or more when you were growing up?

The answer to this question is likely to be short and factual

(objective). On the other hand, an open-ended question requires the
person to elaborate on the information being shared. For example:

If you had to pick the one brother or sister to whom you feel
emotionally closest, who would it be and why?

A common error that interviewers make is failing to frame questions

in open-ended format when they are seeking rich, descriptive
responses. Even with open-ended questions, prompts to keep the
story expanding may be necessary:

Can you tell me more about that?

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 283
 What else should I know about it?

I am not sure I understand. Can you explain it to me

differently?

Interview data might be recorded in notes kept by the interviewer,

audio recordings, video recordings, or notes kept by a third-party
observer. Ideally, more than one strategy is utilized in order to ensure
the most complete data—any of these approaches can fail or data
from one approach might be unintelligible/unreadable later. It is
helpful to have multiple forms for cross-checking and completion
purposes. Furthermore, they complement each other. For example,
tone of voice and meaningful pauses may be detected in an audio
recording but not be evident in written notes or transcripts of the
interview dialogue. Video recordings may pick up nonverbal cues that
are missed in audio-only or written transcript records.

Focus Groups. A focus group is essentially a group interview. There

are two main reasons for using focus group procedures rather than
individual interviews. The first is that it might be more practical,
feasible, or advisable to meet with small groups rather than individual
study participants. For example, information might be gathered from
6 individuals in a 90-minute group session more quickly than in 6
separate 30- or 60- minute interviews. The second (and perhaps
more important) reason is that the focus group participants may
build off each other’s comments in such a way that a more rich, in-
depth picture emerges from the whole group than would emerge
from assembling information provided by 6 separate individuals. A
third reason is that individuals may feel more comfortable discussing
certain topics in a group context than in individual interviews with
an unfamiliar investigator. Of course, the opposite may also be true:

284 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
some may feel more comfortable discussing the topic individually
rather than in a group.

Focus group methodology was used in a study of the importance

of HIV/AIDS prevention education among older adults (Altschuler &
Katz, 2015). The investigators hosted five 90-minute focus groups
with adults aged 50 to 82 years. The study was prompted by the
observation that increasing numbers of older adults are living with
HIV/AIDS, prevention education efforts are seldom targeted/
delivered to older adults, and programs serving older adults seldom
address HIV/AIDS prevention education. They wanted to learn from
study participants what they recommend, what would motivate them
to participate in HIV/AIDS prevention education or to seek such
information. This information was shared in responding to 5 general
questions in a semi-structured group interview context. Transcripts
of the recorded focus group sessions were thematically coded by the
investigators. Analyses led to the identification of 4 themes:

• Wanting to learn about the topic was fostered by knowing

someone who has HIV/AIDS or a general sense of compassion
for these individuals.
• Recommending that HIV/AIDS prevention education be

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 285
 delivered in settings where older adults work or volunteer.
• Reminding others that despite stereotypes, older adults remain
sexually active “both within and outside of monogamous
relationships,” and therefore remain at risk for exposure
(Altschuler & Katz, 2015, p. 694).
• Ambivalence about discussing the topic with their medical care
providers and discomfort with how providers responded to their
questions about this topic.

The investigators concluded that HIV/AIDS prevention education is

appropriate and needed among older adults, and the study
participants provided insight as to how this might best be
undertaken.

Journaling and Ecological Momentary Assessment (EMA). One

286 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
problem with relying on surveys, questionnaires, clinical instruments,
and interviews is that much of the information individuals are asked
to provide is retrospective prospective in nature. In other words,
people are asked to describe what has happened in the past or what
they believe will happen in the future. Unless the questions are about
the “here and now,” people are not providing information about what
is happening in real time, as it happens. We learned back in Module
1 that recalled information is only as good as the person’s memory
and how memories are constantly being revised by experiences and
interpretation processes.

To improve the quality of information provided by study participants,

investigators might ask them to maintain a log, diary, or journal on a
frequent, routine, automatic basis, or as the phenomenon of interest
occurs. For example, if the study is about anxiety, the person might
write down answers to a set of questions about each anxiety episode
as it is happening or as soon after as possible. Not only would this
provide investigators with the ability to count the frequency of
anxiety episodes, they would also have information about episode
intensity, what preceded each episode, and what might have helped
the person resolve each episode. The data can be analyzed more
contextually than would be possible without these event records,
using more global, retrospective accounts.

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 287
Another format for collecting event data is called ecological
momentary assessment (EMA). This approach minimizes recall and
memory bias by sampling an individual’s behavior or experiences
at periodic intervals (Shiffman, Stone, & Hufford, 2008). A study
participant may be signaled by the investigators at either regular or
random times throughout the day or week to immediately record
information about what is happening. This approach has been used in
a number of alcohol-related studies to help investigators understand
how alcohol use is influenced by situational, contextual
circumstances—cravings, stress, affect, motivation to drink,
motivation to abstain from drinking, concurrent other substance use,
social and place factors (Wray, Merrill, & Monti, 2014). The approach
is also presented as a possible technology assist to treating alcohol
use problems by providing real-time support rather than supporting
individuals’ change efforts on a clinic or office schedule
(Morgenstern, Kuerbis, & Muench, 2014). EMA data are collected from
individuals as they function in their natural environments. The data
are sensitive to naturally occurring fluctuations and patterns in the
behavior of interest, rather than relying on a single, retrospective
summary of the behavior. This is particularly important when what
is being studied is dynamic—frequent changes in mood, behaviors
influenced by environmental context, transitory thoughts or feelings,
for example. EMA is facilitated by technology—cell phone access, text
messaging, and web-based survey tools that can be accessed from
most locations. The method is sometimes limited by certain contexts,
such as schools, jails, bathrooms, locker rooms, driving, and other
situations/settings where use of these technologies is discouraged
or banned.

288 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Concept Map. An individual or group process approach to
understanding social phenomena is called a concept map. The result
of this process is a diagram depicting how ideas about a topic are
related, possibly as a web of ideas. While concept mapping is useful
in social work practice with individuals, families, small groups,
communities, and organizations, it also can be useful in conducting
exploratory or descriptive research studies. The underlying research
philosophy and methodology are essentially qualitative in nature,
involving a structured step-by-step approach to achieve the end
goals; the approach also involves both qualitative grouping of
concepts and the use of statistical methods to analyze participant
input in generating the resulting concept map.

“Concept mapping is a structured process, focused on a topic

or construct of interest, involving input from one or more
participants, that produces an interpretable pictorial view
(concept map) of their ideas and concepts and how these are
interrelated” (Trochim & Donnelly, 2007, p.27).

A team of investigators utilized concept mapping as a method for

exploring types of support needs identified by a group of diverse
sexual minority youth (Davis, Saltzburg, & Locke, 2010). The 20 study
participants were served by a GLBTQ youth-focused. The

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 289
participating youth generated 58 statements of need in the “idea
generation” phase of the process. Next, each individual participant
sorted the collection of statements in clusters that were conceptually
meaningful to them, and provided a label for each of their item
clusters. This step concluded with each participant rating each
statement on a 7-point scale of importance (not important to
extremely important) in meeting their emotional or psychological
needs and again in meeting their social needs as a GLBT youth. The
investigators used these data in multidimensional scaling statistical
analysis, and generated concept maps.

“Concept maps were produced, providing graphic

representations of relationships and relevance of the ideas. The
maps illustrated how the youths’ ideas clustered together and
the rating data reflected the value given to each conceptual idea
and cluster” (Davis, Saltzburg, & Locke, 2010, p. 232).

The study participants next engaged with the investigators in group

discussions about how to interpret the preliminary concept mapping
results. A combination of additional qualitative and statistical
methods led to the final results. The youth’s statements fit into 5
clusters of needs (educating the public, policies, community support
& involvement, individual responsibilities, and GLBTQ acceptance
& individuality). In terms of importance, the highest ratings were
assigned to “policies” statements, lowest to “community support &
involvement” statements. Results from studies like this have
important implications for planning culturally and developmentally
relevant services in programs, organizations, communities, and
service delivery systems—both in terms of process and products.

Social networks. Another method that allows graphical depiction of

relationships comes from asking questions about social networks and

290 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
using research methods designed to gather social network data.
Social networks are the patterns of formal and informal connections
that exist between individual units in complex social systems. Social
network maps depict the strength and/or frequency of connections
between individual units (individual persons or groups). A great deal
of such research is conducted using social media interaction data,
for example. Questions about how ideas, rumors, or innovations are
disseminated or diffused can be answered in this manner. Social
network analysis allows investigators to identify key “nodes” or
“nodal points” where different networks connect, as well. Questions
about how people use their social networks might rely on other forms
of data collection (survey or interview questions, for example). This
type of research helps us understand how social relationships and
social contexts influence human behavior.

Social network data collection and analysis was used in a study to

understand the phenomenon of suicide diffusion among youth social
networks—how suicidality might spread among adolescents (Mueller
& Abrutyn, 2015 ). These investigators used data previously collected
in the National Longitudinal Study of Adolescent to Adult Health
to map how suicidal ideation and suicide attempts by individual
adolescents (“ego” individuals) might be influenced by suicide
attempts of their friends (“alter ego” individuals). Ego individuals who
knew about an alter ego’s attempted suicide (disclosed suicide
attempt) during the past year experienced higher levels of emotional
distress and suicidality themselves compared to egos with an alter
ego whose suicide attempt was undisclosed (ego did not know about
it) or whose alter egos had no previous suicide attempts. The authors
did find evidence of “the social contagion of negative emotions” but
that emotional contagion alone is not relevant to the social contagion
of suicide (Mueller & Abrutyn, 2015, p. 144).

Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement | 291
 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=223

Chapter Summary

In summary, there exist many options for collecting qualitative and

quantitative data in social work research. The issues of measurement
validity and reliability are important concerns for quantitative
research. Ensuring the appropriateness and cultural competence of
measurement approaches and procedures is also critically important
in any type of research involving people as participants.

292 | Module 3 Chapter 5: Overview of Methods for Data Collection and

Measurement
Module 3 Chapter 6: Study
Participants
Research informing our understanding of diverse populations, social
work problems, and social phenomena is entirely dependent on the
participants who provide the data and observations for analysis.
Guidelines for engaging participants vary by research approach:
qualitative, quantitative, or mixed-methods, as well as by study
purpose (descriptive, exploratory, or explanatory questions). In this
chapter you will learn about:

• Sample size implications

• Strategies for establishing study samples
• Participant recruitment
• Participant retention
• IRB concerns for research involving human participants

What Is a Sample?

First things first: understanding this chapter requires an

understanding of what a sample is. In research, a sample is a group
drawn from a population for the purposes of observation or
measurement. If we measured an entire population, we would not
be talking about a sample. It is usually unreasonable to attempt this
feat of measuring an entire population to answer a research
question—although it is sometimes possible in evaluating a small-
sized program. Ideally, the sample drawn is a good representation of
the population. This allows an investigator to offer generalizations
based on the study results derived from the sample. Many of the
quantitative statistics used in research are based on certain

Module 3 Chapter 6: Study

Participants | 293
assumptions about the relationship of the sample to the population.
If sampling error is small, then statistical values observed for the
sample will be a good representation of the statistical values we
would see for the entire population if we could really measure them
all.

A note about populations is warranted here: a research population

is not the same as the entire population. A research population is
defined by certain variables of interest in the study being conducted.
For example, consider a study of service needs experienced by adults
following release from incarceration (Begun, Early, Hodge, 2016). The
population relevant to this study was defined as men and women
preparing for release from jail or (state) prison—it was not all people,
or even all incarcerated people (not juveniles, not persons in federal
prison, not individuals who are still serving sentences). The results of
the study are generalizable only to this specific population, and since
the data were collected only in the state of Ohio, generalizations to
other states are made with caution (see Figure 6-1).

294 | Module 3 Chapter 6: Study Participants

Figure 6-1. Sphere of generalizability to study population

Matters of Sample Size and Diversity

The question of how many study participants are needed for a

specific study is complicated to answer. First, it depends on the study
approach and aims. Most qualitative research is not intended to be
generalizable to a larger population, but most quantitative research
is. If the aim of a qualitative study is to develop a deep, rich
understanding of a construct, event, experience, or other
phenomenon, then the number of study participants needed might
be relatively few. For example, in grounded theory studies, the need
is to get to a point where all or most facets of the topic are addressed
and where enough responses cluster together to have a picture of

Module 3 Chapter 6: Study Participants | 295

common themes emerge from the data. This ideal point of saturation
is achieved when new ideas, not expressed by previous participants
are no longer being presented by subsequent participants. The
recommended number is in the 20-30 participant range (Creswell
& Poth, 2018). Similarly, in ethnography, the number of participants
depends on how many observations are required to achieve a clear
picture of the culture-sharing group (Creswell & Poth, 2018). In
narrative studies the aim is to develop the story of a single individual,
or a few individuals involved in a single event. In focus group studies,
the emphasis is on group process, so the recommendation is 5-10
persons per group—the number of groups varies from 1 to several.

The aim of quantitative research lies in sharp contrast: the aim is

to generate information that is generalizable from the sample to
the population. Therefore, the numbers needed are usually greater
for studies designed around quantitative than around qualitative
approaches. Table 6-1 presents an example generated from 12 social
work students using samples of M&Ms® candies to estimate the
proportion of green M&Ms® in the population of M&Ms® candies
(adapted from Begun, Berger, & Otto-Salaj, 2018, p. 8).

296 | Module 3 Chapter 6: Study Participants

a This estimate is per https://www.exeter.edu/documents/mandm.pdf which
also states that the Mars Candy Company no longer publishes the proportions
of colors in the M&M’s® product line.

As you can see, when the individual students’ sample sizes were
between 18 to 37, the range of values for proportion (%) of green
candies was wide—as low as 6.5% and as high as 33.3%. As students
combined their samples (teams of 3), the sample sizes from 106 to
119 had much less variability in the observed proportions: 10.9% to
20.8%. Finally, when the teams’ samples were combined into one
single class sample of 447 candies, the sample’s observed proportion
of 15% was close to the actual proportion for the population reported
by the company (16%). (Yes, they ate the candy after completing the
demonstration.)

Module 3 Chapter 6: Study Participants | 297

Second, the diversity of study participants is at least as important as
numbers in qualitative research studies. To ensure rich coverage of
a topic, investigators seek participants with differences likely to be
important to that topic—an approach that maximizes variation among
study participants so that different perspectives will arise in the data.
The approach to selecting participants is purposeful. For example, in
a grounded theory qualitative study of benefits experienced by older
adults in adult day services programs, investigators interviewed 28
service recipients (Dabelko-Schoeny & King, 2010). More important
than the numbers, however, is that the investigators intentionally,
strategically engaged participants from 4 different programs. The
group of study participants ended up being Caucasian and African
American women and men, differing in marital status, ranging in
age around a mean of 78 years, experiencing different types and
numbers of medical conditions, and ranging in income ($0-$88,000
annual income). The investigators’ conclusions were that five themes
expressed by the study participants fit into two general categories:
program experiences and perceived impact of experiences. The
program experience themes included social connections with other

298 | Module 3 Chapter 6: Study Participants

participants, empowering relationships with staff, and participation/
enjoyment of activities and services. The perceived impact of
experiences themes included perceived improvement in psychosocial
well-being and perceived decrease in dependence and burden on
their primary caregiver. These results contribute to an understanding
of the perceived benefits experienced by persons who participate in
adult day services programs.

Just as sample diversity is important in qualitative research, the issue

of sample heterogeneity is an important consideration in quantitative
research. The challenge played out in quantitative research concerns
the sample having diversity that is representative of the population
to which results will be generalized. In contrast to the purposeful

Module 3 Chapter 6: Study Participants | 299

selection process used in qualitative research, strategies for
randomly selecting participants into quantitative studies are more
common. More about this topic is presented in the next two sections.

Filling a Quantitative Study Design

Let’s revisit a study design example from Chapter 4—the 2 X 3 design

with empty cells. The original example was about waiting to see how
individuals sorted themselves out in terms of the two variables of
interest. This time, we are going to use the diagram to help figure
out how many people of each type investigators need to include in
our sample for a quantitative study. Imagine that a planned study is
about participants’ ratings of an outcome variable—maybe likelihood
of engaging in additional college education—and the investigators
wish to determine how age group and employment status relate
to this outcome variable. The statistical analyses planned by the
investigators require a minimum of 10 persons in each cell (more
about this in Module 4).

The result would be the need for ending up with a minimum of 60

study participants, recruited from each of the 6 types (10 x 6=60).

Employment Status
Fully Employed Partially Employed Unemployed
<30 10 10 10
Age Group
≥30 10 10 10

Now imagine that the investigators, based on their review of

literature, have decided that they really need 3 age groups: 18-25,
26-35, and over 35. Let’s see what this does to the design and number
of participants needed. As you can see, we now have a 3 X 3 design,

300 | Module 3 Chapter 6: Study Participants

which makes 9 cells. With 10 participants needed per cell, the
minimum sample size is now increased to 90.

Employment Status
Fully Employed Partially Employed Unemployed
18-25 10 10 10
Age Group 26-35 10 10 10
>35 10 10 10

Thus, one answer to the question of how many study participants are
needed in a quantitative study is that it depends on a combination of
the study design and the intended statistical analysis requirements.
In our next course, SWK 3402, you will also learn about matters
of statistical power and effect size in determining how many
participants are needed in an experimental study about
interventions. For now, it is sufficient to understand that the more
complex a study design, the greater the number of study participants
are required.

Sampling Strategies: Probability and Nonprobability

Earlier, when discussing the strategy used for generating the sample
in a qualitative study we used the word “purposeful.” This means that
selection of study participants was not left to chance. In quantitative
studies, however, investigators need to minimize the extent to which
bias enters into the participant selection process if the sample drawn
from the population is to be reflective of that population. They need
every individual in the population to have the same chance, or
probability, of being selected. This means using random selection
strategies.

Module 3 Chapter 6: Study Participants | 301

Random Selection. Random selection refers to strategies used to
recruit participants into a quantitative study with the aim of
maximizing external validity (the extent to which results from the
sample can be generalized to the larger population). Random
selection is considered a probability sampling strategy since the
probability for each and every member of a population to be included
in the sample is the same. This use of the word “random” should
not be confused with the concept of random assignment in an
experimental study. Random assignment refers to dividing the
randomly selected sample into assigned experimental groups—like
who receives the experimental, new intervention being tested and
who receives the usual form of intervention (a lot more about this in
our next course, SWK 3402).

Random selection is never perfect—there is always some error in the

degree to which a sample represents a population. The aim is to
minimize this error, especially to minimize systematic error or bias
in the sample. Simple random sampling begins with a known pool of
possible participants: they are the sampling frame. The investigator
is going to select a specific number of those possible participants, at
random, so that each possible participant has the same probability of
being selected as any other participant. For example, imagine that a
social worker wants to use the residential zipcodes for a randomly
selected group of individuals in a program for men and women court-
ordered to receive treatment to stop their intimate partner violence
behaviors to determine the best locations to place a new branch of
the program. The social worker decided that 20% of the zipcodes
should be utilized. The sampling frame, or available pool, is all those
served over the past year—for this example, let’s say it was 1,200
persons. This means that the need is to select 240 participants for
the study (1,200 x .20=240). The social worker in this situation could
assign each “case” in the program a number from 1 to 1200. Then,
a statistical program could be used to randomly select 240 of those

302 | Module 3 Chapter 6: Study Participants

numbers. An example is the free random number generator called
Stat Trek (https://stattrek.com). The social worker enters 240 in the
box for how many random numbers are desired, 1 for the minimum
value, 1200 for the maximum value, and does not allow duplicate
entries (“false” means that once a number has been selected it is
removed from the pool so every number generated is unique.) The
resulting table tells the social worker to use the zipcodes from “cases”
numbered 186, 251, 58, 635 and so forth to the last one, 246.

Essentially, this is like drawing numbers from a hat or generating

lottery numbers on television. Unfortunately, sometimes the size of
the starting pool is unknown or changes. For example, a social worker
may want to obtain client satisfaction information from 10% of
families served by an agency. There is no pre-existing pool to draw
from, the need is for 10% of whoever comes into the program moving
forward in time. In this case, investigators need to rely on systematic
random sampling strategies—retaining an element of random
selection, but more vulnerable to systematic bias. For example, an

Module 3 Chapter 6: Study Participants | 303

investigator might decide to collect mental health and substance use
screening data from every 10thperson who comes to the emergency
department for care. Assuming that there exists no systematic bias
in the pattern by which people come to the emergency department,
this plan is effective. Where bias enters the picture is if the sample
selection is conducted only during “office hours” (M-F, 8-5). The
people who come in on weekends and late night may be very different
than those who are sampled.

As a different example, for a study about family recreational time,

taking every 10th child enrolled in after school programs is fraught
with potential bias problems. First, families with more than one child
in the programs have a greater chance than single-child families
of being selected. Second, this type of list is already systematically
arranged from the sequencing of when children are enrolled—there
may be some systematic differences between those who enroll early
and those who enroll later. Working from alphabetical listings has a
similar problem: certain last names are more common than others
in the U.S.—Smith, Johnson, Williams, and Brown are the top 4. This
alters the chances of the persons with names in that part of the

304 | Module 3 Chapter 6: Study Participants

alphabet. Sampling bias is potentially introduced when the pool from
which selection is made is systematically arranged.

Convenience Sampling. Probability sampling through random

selection is not always essential because sometimes the research
aims do not include generalizability back to the population. An
example of this scenario is a pilot or feasibility study being conducted
in preparation for a larger scale study with external validity being
addressed. Another example is a study where it simply is not feasible
to draw a representative sample from the general population of
interest. Investigators might rely on selection from among those
most easily accessed—called convenience sampling. In these
instances, potential bias from non-probability sampling is recognized
as a generalizability limitation in interpreting the results.

For example, consider the comparative, exploratory study conducted

with social work students in 4 countries: United States, Greece,
Cyprus, and Jordan (Kokaliari, Roy, Panagiotopoulos, & Al-
Makhamreh, 2017). The purpose of the study was to extend our
understanding of social workers’ perceptions concerning non-
suicidal self-injury. The selection of these countries was intended
to maximize heterogeneity (diversity) in terms of cultural values and
religious belief systems: inclusion of students with Christian, Greek
Orthodox, and Muslim religious backgrounds, and nations differing
in social/political orientation toward individual and women’s rights.
The students participating in the survey study were those easily
accessible to the investigators who were associated with social work
programs in those countries. The investigators identified differences
and similarities in the students’ responses which are suggestive of
the need to better address the problem of non-suicidal self-injury
in social work education; however, they also identified the potential
limitations of the study related to the convenience sampling

Module 3 Chapter 6: Study Participants | 305

approach. Student surveys may not be a good representation of social
work practitioners, either. The observed differences in this study
were that Jordanian students viewed the problem behaviors as being
associated with weak/absent ties to religion, while U.S. and Greek/
Cypriot students attributed the problem to mental illness and
socioeconomic factors. Students from each nation were equally
unprepared to understand treatment of the problem. One concern
with how to better educate social workers about non-suicidal self-
injury is that the behaviors may differ in the varied cultures/
countries, and marked differences in stigma exist, as well.

Snowball Sampling. Social work investigators are often asking

research questions related to difficult-to-locate
populations—questions about uncommon problems, stigmatized
phenomena, and intersectionality issues. Quite possibly, individuals
who meet the study criteria know others like themselves. In these
cases, investigators may rely on snowball sampling strategies: the
sample builds itself from a few core individuals who identify other
participants, and they identify still others. The name comes from how
you might build the base of a snowman: start with a small snowball,
roll it in the snow and as more snow adheres to it, the snowball grows
larger (assuming it is good packing snow). This is another form of
non-probability sampling.

306 | Module 3 Chapter 6: Study Participants

A more complex version of snowball sampling is called respondent
driven sampling. Not only does this approach involve the study
participants helping to identify other potential participants, it
involves some analysis of the social networks that are tapped into
throughout the snowball process. For example, investigators
interested in addressing questions related to the experience of
human trafficking might begin with identifying a small number of
individuals willing to participate in their study (the sample nucleus).
These individuals would each identify one or more others who could
be invited to participate. Those who accept the invitation and
become participants would each identify one or more others. The
hard work for study investigators begins with identifying a
sufficiently diverse initial core: if they identify too many within the
same social network, these members will all be identifying each other
or the same few and the snowball will cease to grow. The hard work
continues in terms of ensuring protection of privacy rights for those
nominating others and for those being nominated for invitation.
Then, the investigators must take into consideration the potential

Module 3 Chapter 6: Study Participants | 307

bias built into the snowball approach—these people are sufficiently
similar to know each other. Entire networks of diverse others may
be missed in the snowball process. This limitation is based on the
principle of homophily: the observation that people tend to associate
with others similar to themselves in ways that are meaningful to them
(i.e., race, ethnicity, social class, gender identity, sexual orientation,
political or religious belief systems, athletic team loyalty, drinking
habits, and many more). The saying “birds of a feather flock together”
is true of human social networks, as well.

Participant Recruitment

308 | Module 3 Chapter 6: Study Participants

Successfully engaging sufficient numbers and diversity of
participants in a study is critically important to the study’s success.
Participant recruitmentcan be conceptualized as a 3-step process
(adapted from Begun, Berger, & Otto-Salaj, 2018):

• Generating initial contacts is about identifying and soliciting

potential participants. This step is accomplished through a
variety of advertising strategies: media advertisements
(including social media outlets), posters and flyers, and mass
mailings (postal and e-mail). Some success is achieved through
telephone calls, as well. In recent years, the rate at which people
respond to many of these approaches has declined
sharply—people are finding themselves saturated with claims for
their attention, and are ignoring many of these messages as
“spam” or “junk” mail. Investigators also need to critically
consider the potential sources of bias related to different
possibilities. For example, individuals’ responsiveness different
types of on-line social media is influenced by age group
(Snapchat, twitter, and Facebook, for example, are used
differently by adolescents, emerging adults, and older adults);
reading newspapers is heavily influenced by geography and
social class; receiving church newsletter postings is dependent
on affiliation; and, reading bus advertisements is dependent on
who rides buses and which routes they frequent. In-person
recruitment might be effective in specialized locations (e.g.,
clinics) but may not be effective with the general public (e.g.,
stores and shopping malls). These, too, are subject to bias in
terms of who will agree to listen to the recruitment message
and who will not.

Like any advertising effort, the nature of the message is important

to consider. The direct contact tool needs to be crafted with
consideration of why someone would WANT to engage with the

Module 3 Chapter 6: Study Participants | 309

study. The message needs to capture attention first, then needs to
capture interest next. It also needs to be easy for someone to
respond—a 24/7 phone line or e-mail address or website is more
accessible than a phone number with an answering machine.
Furthermore, cultural competence of the message needs to be taken
into consideration: images and language need to convey a sense that
“this study is about people like me.”

• Screening is part of the recruitment process, as well. If strong

effort is placed on the first step (generating contacts), there is
strong likelihood that individuals who do not meet the study
criteria will be contacted. The study investigators need to have
clearly defined criteria for study inclusion/exclusion. This
process can be as simple or as complex as the criteria for
participation. If the study is limited to adults, then the screening
can simply be a question: Are you at least 18 years of age?
Anyone who answers “no” to this question fails to meet the
inclusion criteria and is excluded from participation. Consider
the example of a study concerning the relationship of social
work education and self-esteem on social work students’ social
discrimination of persons with disabilities (Bean, & Hedgpeth,
2014). The direct contact method used in this survey study was
to distribute the survey only to individuals who met the study
criteria: being in the last semester of their social work program.
On the other hand, had they distributed the survey to the entire
school, they would have needed to impose a screening strategy
to ensure that other students did not become enrolled in the
study.

Four difficulties arise in regard to screening. First, investigators need

to identify reliable, valid screening measures for each of the study
criteria. Second, the screening information cannot be treated as data
in the study unless consent was secured for screening and for the

310 | Module 3 Chapter 6: Study Participants

main study. Third, investigators need to have an appropriate
response for individuals who become excluded. Consider, for
example, individuals with a substance use problem who are
interested in participating in a treatment study, but who fail to meet
the study criteria. Ethically, investigators cannot simply turn them
away without offering strong alternatives. Fourth, screening is an
important part of the participant-investigator relationship. If the
screening experience is tedious or otherwise unpleasant, the chances
of retaining participants in the actual study diminish.

• Consenting study participants is a process, not an outcome. A

signed consent form simply provides documentation that the
process was engaged—the process is the goal. This means that
investigators studies need to engage in interactions with
potential study participants that ensures they are fully informed
about the study and what should be expected—including any
potential risk and benefits, as well as steps the investigators
take to minimize risk. Making sure of potential participants’
comprehension of the consent information is a responsibility of
study investigators. The IRB review process helps ensure that
the proposed procedures meet expectations for protecting
potential participants rights. In addition to providing
participants with sufficient information to make an informed
choice about engaging with the study, investigators need to
ensure that participants are making the decision without
influence of coercion. Some potential coercion scenarios are
more obvious than others. For example, instructors who wish to
engage their students in a study should not be in the position of
knowing who does or does not consent to participate, since
students may believe that this knowledge could affect their
grade in the course. This is a problem to be addressed in social
work intervention research: clients may feel obligated to
participate in a study conducted by the person delivering them

Module 3 Chapter 6: Study Participants | 311

 services, out of concern that the decision might affect the
services they receive or out of a desire to help the person who
has been helping them. Problems might arise in relation to
incentives offered to study participants, too. The amount
offered might be considered coercive: $25 may not be coercive
to people in some circumstances, but is a lot of money to people
in other circumstances, potentially affecting their decision
about participation. Legally, children and adolescents cannot
provide consent—they have not achieved the age of majority.
Thus, persons responsible for ensuring their welfare need to
provide consent (parents, legal guardians). At the same time,
investigators have a responsibility to secure the child’s/
adolescent’s assent to participate. This means that the child or
adolescent agrees to the conditions of participation.

Exceptions do exist: sometimes the consent process or consent

documentation is waived by an IRB. Sometimes it is not reasonable to
secure assent. Consider, for example, studies conducted with infants
who would not understand an investigator’s description of what is
expected to happen. The investigators need to engage parents/
guardians in consent and then observe the infants for cues as to their
willingness to continue—for example, crying is a clear sign that the
baby is done participating, assent has ended.

Participant Retention

In longitudinal research, it is critical that study participants remain

engaged with the study through all phases of data collection. While
participant recruitment is about people initially enrolling, participant
retention is about them staying involved over time. Obviously,
retention is not an issue in cross-sectional research studies. An
important point for investigators to keep in mind is that retention

312 | Module 3 Chapter 6: Study Participants

begins with recruitment—retention only happens when study
participants decide that the experience continues to be worthwhile.

The opposite of participant retention is calledparticipant

attrition—people dropping out of a study before completion. The
obvious problem with attrition is that the number of participants
may drop below the level needed for a strong study design and
generalizability from the smaller sample to the population. A less
obvious problem is that attrition is seldom a random process. This
means that sampling bias may become introduced into the study
despite all the care and effort that investigators may have applied
in their initial probability random selection. For example, men or
persons of color may drop out at a higher rate than women or white
participants. The sample at the end of the study is what matters,
so minimizing attrition is an important issue for investigators to
address. It is costly to replace missing participants—starting over
with recruitment and all the data collection efforts that need to be
repeated (not to mention paying incentives again). Investigators need
to constantly address the question: why would participants WANT to
remain involved with the study?

Researchers have demonstrated that the rate of participant attrition

tends to be greatest at the start of a study: if someone is going to drop
out, this is most probable early on (Begun, Berger, & Otto-Salaj, 2018).
That does not mean they will certainly be retained once they are
involved in middle and later phases of a study; the dropout rate tends
to slow down to more of a trickle as time passes. This means that
investigators might engage in their most strenuous, costly retention
efforts early on. In addition to making the study protocol interesting
and a positive experience for participants, developing good rapport
and effective systems for tracking and maintaining contact are
critical.

Module 3 Chapter 6: Study Participants | 313

Chapter Summary

This is the final chapter of Module 3. Here you learned about

sampling in quantitative research studies, and about participant
numbers and heterogeneity (diversity) as important issues in both
qualitative and quantitative research. This chapter also discussed
aspects of participant recruitment and retention that are important
in both qualitative and quantitative studies.

While it is not quite the same as

recruiting you to participate in a
research study, let’s think about how
the content of this chapter relates to
your participation in end-of-semester
course/instructor evaluations. For
many courses, the response rate is 30%
or less—meaning that 2/3 of your
classmates do not participate.

What is the greatest potential bias risk

to the program’s evaluation results of
failing to recruit participants like you (and many of your peers)?

In a class of 25 students, how many do you think would be a good

number to recruit for a strong sample?

What are 3-5 things the program could do to increase your

participation in the evaluation process—what makes you WANT to
participate?

Does reading this chapter and thinking about this topic make the
likelihood of your participation in evaluating this course greater,
less, or about the same? Why?

314 | Module 3 Chapter 6: Study Participants

Module 3 Summary
This module was concerned with the many research methodology
choices faced by social work, social science, and behavioral science
investigators. You read about the ways in which research questions
lead to research approaches: specifically, about qualitative,
quantitative, and mixed-methods approaches. Next, you learned a
great deal about different qualitative and quantitative research
traditions and the many available options: each has different
advantages, disadvantages, and goodness-of-fit with study aims to
consider. In Chapter 4, you read about variables and their role in
quantitative research. Next you learned about many different
strategies that investigators use to collect data, and again learned
about their relative advantages and disadvantages. Finally, you were
presented with issues related to participants in social work, social
science, and behavioral science research. This included issues of
sample size and diversity, IRB concerns with involving humans as
study participants, and issues in participant recruitment and
retention (for longitudinal studies). This Module 3 content has great
relevance to the next course module, Module 4, where we look at data
analytic approaches.

Module 3 Summary | 315

Module 3 Key Terms and
Definitions
administrative data: information routinely recorded in programs,
agencies, and institutions as part of normal operations, not with a
research intent.

alphanumeric variables: data in words or numbers that have no

mathematical meaning (for example, telephone numbers, addresses,
names or descriptions).

categorical variables: variables with categories of a non-numeric

type, characterized by their names/labels rather than mathematical
numbers (for example, race, religion, or national origin categories)

cohort: a group of individuals with a shared, common history or

experience (for example a birth cohort like “baby boomers,” “Gen X,”
“Gen Y,” or “Gen Z.”

community-based participatory research (CBPR): a collaborative

inquiry process involving an egalitarian, mutual learning partnership
between community partners and researchers, both as experts in
their own ways, sharing that expertise toward a common research-
related goal.

concept map: the diagram resulting from a systematic research

process where concepts, ideas, or constructs are identified and
organized in terms of their mutual relationships.

content/artifact analysis: research analysis of information, records,

recordings, or objects created in the normal course of events/living,
outside of a research intent.

316 | Module 3 Key Terms and

Definitions
continuous variables: numeric variables on a continuum of values
where possible values are equally distant (see interval variables).

convenience sampling: a type of non-probability sampling, subject

to sampling bias, that draws a group of participants (sample) in a
systematic, non-random manner based on ease of access to recruit
them (for example, college students, clinic populations, club
members).

correlational research: research designed to evaluate the existence

and nature of relationships between variables (whether they are
statistically correlated, and if so, the strength and direction of the
correlation).

cross-sectional research: research designs involving only one data

collection time per individual element or unit being measured,
providing a single-time “snapshot” of data.

cultural competence in measurement: using measures and

measurement procedures that are appropriate for the group being
measured, as free from bias as possible; “cultural” is broadly defined
to include gender, race, ethnicity, language, religion, national origin,
ability/disability, and other factors relevant to a study population.

demographic variables: variables used to describe specific

characteristics of a population, group, or sample (for example, age,
gender, race, ethnicity, or income level)

dependent variable: the variable in a study design presumed or

hypothesized to vary as a function of variation in the independent
variable; its value is presumed to depend on the independent variable
values; the “outcome” or “output” variable in a study design (i.e.,
usually denoted as the “y” variable, where “x” denotes the
independent variable).

Module 3 Key Terms and Definitions | 317

descriptive research: research with the purpose of creating a profile
or typology (description) of a population, phenomenon, or process.

dichotomous variables: a special type of categorical variable with only

two possible categories or values (for example, yes/no, meets/does
not meet clinical criteria, completed/did not complete treatment).

double-barreled question: a single question that asks for information

representing two different questions, may not be accurately
answered in a single response; often uses the word “and” to connect
the two different questions being asked.

ecological momentary assessment (EMA): a form of real-time data

collection, often using technology to record events as they are
happening (rather than retrospective accounts)

ethnographic research (ethnography): one of the major qualitative

research traditions with the aim of understanding a population, social
work problem, social phenomenon, or experience from the collective
perspective of a group (rather than discrete individual perspectives).

experimental research: research designed to answer explanatory

questions, testing hypotheses or theory about a phenomenon or
process.

external validity: the degree to which quantitative research results

based on a study sample can accurately be generalized to the larger
population or other settings.

focus group: a research tradition involving interview data collected in

a group setting (rather than from individual participants).

geographic information systems (GIS): a form of data representing

geospatial relationships between places where events, experiences,
or behaviors occur and the nature of those events, experiences, or

318 | Module 3 Key Terms and Definitions

behaviors (for example, locations of state prisons and prisoner
visitation patterns).

generalizability: the extent to which findings based on sample data

can be expected to reflect what is true of the population from which
the sample was drawn (see external validity).

grounded theory research: one of the major qualitative research

traditions, aimed at developing theoretical explanations about a
phenomenon or process from data provided by intentionally/
purposively selected individuals.

heterogeneity: differences, diversity, variability in a group or

population.

homophily: observation that people tend to associate with others

similar to themselves.

independent variable: the variable in a study design presumed or

hypothesized to vary on its own, but that determines variation
observed in the dependent (outcome or output) variable; the “input”
variable in a study design (i.e., usually denoted as the “x” variable,
where “y” denotes the dependent variable).

internal validity: the degree of confidence that can be applied to the

results of an experimental, explanatory study; study integrity; ability
to rule out alternative explanations for the experimental results.

inter-observer reliability: different observers, raters, or coders

getting the same result when evaluating the same qualitative, open-
ended question, or observational data; also called inter-rater or
inter-coder reliability.

interval variable: continuous variable (on a continuum) where the

distance between possible values (intervals) are equivalent (for

Module 3 Key Terms and Definitions | 319

example, age where the 1-point interval between 40 and 41 is the
same as the 1-point interval between 49 and 50, 1 year).

interview protocols: data are systematically collected by investigators

interact with participants to provide information in a “live” dialogue
(may be in-person, by telephone, or other interactive media).

key informants: persons identified as close and knowledgeable about

someone for whom data are being recorded (for example, parents
of young children, spouse or caregiver to a person with special care
needs)

latent variable: a variable not directly observed but mathematically or

statistically inferred from other variables that are directly measured.

longitudinal research: study design where data are collected

repeatedly from the same individuals/elements/units over time.

measurement reliability: indicates the degree of accuracy or

precision in measuring the construct or variable of interest.

measurement validity: the extent to which a measurement tool or

instrument adequately measures the concept, construct, or variable
of interest, comprised of several types of validity (construct,
concurrent, predictive, and others).

mixed-methods research: a study approach that integrates aims,

questions, and methods in both qualitative and quantitative
traditions.

narrative inquiry: one of the major qualitative traditions, eliciting

and analyzing the “stories” and their contexts as experienced by one
or a few individuals.

naturalistic observation: a form of data collection concerning

320 | Module 3 Key Terms and Definitions

individuals’ behavior in their naturally occurring environments; the
observer is a non-participant (objective observer).

nominal variables: another name for categorical variables (categories

have names, are not numeric)

non-probability sampling: forms of participant selection that are not

random; participants have unequal chances of being selected from
the population.

ordinal variables: categories of a variable have numeric relationships,

can be sequenced in numeric order but are not continuous interval
variables (for example, ratings of frequency from never to always)

participant attrition: the opposite of participant retention, refers to

participants dropping out of a (longitudinal) study before completion

participant recruitment: process of initially engaging potential

participants in a research study

participant retention: the opposite of attrition, refers to maintaining

participation for the duration of a (longitudinal) study

participatory action research (PAR): investigators are active both as

researchers and change agents in the group/process under study.

phenomenological research (phenomenology): one of the major

qualitative traditions, involving the constructing of shared, common
meaning about an event, experience, or phenomenon out of the
experiences and perceptions of a group of study participants.

photovoice: a data collection approach involving a combination of

participants recording images (photographing) aspects of their
experiences and environments along with qualitative data about their
explanations of those images/photographs.

Module 3 Key Terms and Definitions | 321

primary data: data collected with the intention of meeting aims of
the specific research study for which the data are collected.

probability sampling: selecting research participants using methods

that ensure every member of the original population has an equal
chance (probability) of being selected; (see random selection).

proxy variables: variables that represent or stand for something not

directly measured (for example, birth order representing amount of
parental attention experienced).

qualitative research: research with the aim of describing or exploring

populations and phenomena as they naturally occur; approaches flow
from inductive reasoning process; multiple qualitative traditions
exist.

quantitative research: research with descriptive, exploratory, or

explanatory aims where the data collected are in quantifiable terms;
approaches flow from deductive reasoning process; usually involves
some form of statistical analysis or reporting; built from a

random assignment: refers to dividing the randomly selected sample

into assigned experimental groups

random selection: refers to strategies used to recruit participants into

a quantitative study with the aim of maximizing external validity (the
extent to which results from the sample can be generalized to the
larger population)

sample: a group drawn from a population for the purposes of

observation or measurement.

scale variable: another name for an interval variable.

screening and assessment tools: instruments or tools used in

322 | Module 3 Key Terms and Definitions

professional practice to evaluate clients, but which may also be used
in research to collect data.

secondary data: research using data originally collected for a

different research purpose (a secondary use of the data).

snowball sampling: building a study sample from having enrolled

participants identify others who may be recruited as participants,
too.

social desirability: a person’s tendency to provide answers that seem

to be more acceptable/correct than would be their true answers; a
potential type of bias in data.

social networks: the formal and informal connections that exist

between individual units in complex social systems.

survey methodology: research approach involving probability

sampling and data collection by use of survey tools (whether by in-
person interview, telephone, paper-and-pencil, on-line, or otherwise
collected).

triangulation: strategy of applying multiple methods to study a single

phenomenon, generating a unified understanding by combining the
results.

unit of analysis: data may be collected from individuals, but the unit
of analysis might be the individuals themselves, or higher order units
to which they belongdyads, couples, families, classrooms,
neighborhoods, and others.

variable: elements, entities, or factors that can change (vary) in a

research study.

Module 3 Key Terms and Definitions | 323

Module 3 References
Altschuler, J., & Katz, A. (2015). “Of course it’s relevant!”: A focus
group study of older adults’ perceived importance of HIV/AIDS
prevention education. Qualitative Social Work, 14(5), 687-701.

Auerbach, D.M., Darrow, W.W., Jaffe, H.W., & Curran, J.W. (1984). A
cluster of cases of the acquired immune deficiency syndrome:
Patients linked by sexual contact. American Journal of Medicine, 76,
4878-492.

Babor, T.F., Higgins-Biddle, J.C., Saunders, J.B., & Monteiro, M.G.

(2010). The Alcohol Use Disorders Identification Test (AUDIT):
Guidelines for use in primary care, 2nded. Geneva: The World
Health Organization (WHO), Department of Mental Health and
Substance Dependence.

Baer, L., & Blais, M.A. (2010). Handbook of clinical rating scales and
assessment in psychiatry and mental health.NY: Humana Press.

Barnhart, S. & Maguire-Jack, K. (2016). Single mothers in their

communities: The mediating role of parenting stress and
depression between social cohesion, social control and child
maltreatment. Children and Youth Services Review, 70, 37-45.

Begun, A.L. (unpublished master’s thesis). Content analysis of sibling

relationships as presented in preschool children’s books. University
of Michigan, Department of Psychology (1980).

Begun, A.L., Berger, L.K., & Salm Ward, T.C. (2011). Using a lifecourse
context for exploring alcohol change attempts and treatment
efforts among individuals with alcohol dependency. Journal of
Social Work Practice in the Addictions, 11(2), 101-123.

324 | Module 3 References

Begun, A. L., Brondino, M. J., Bolt, D., Weinstein, B., Strodthoff, T.,
& Shelley, G. (2008). The revised Safe At Home Instrument for
assessing readiness to change intimate partner violence. Violence
and Victims, 23(4), 508-524.

Begun, A.L., & Carter, J. (2017). Career implications of doctoral social

work student debt load. Journal of Social Work Education, 53(2),
161-173.

Begun, A.L., Early, T.J., & Hodge, A. (2016). Mental health and
substance abuse service engagement by men and women during
community reentry following incarceration. Administration and
Policy in Mental health and Mental Health Services Research, 43(2),
207-218.

Begun, A.L., Hodge, A.I., & Early, T.(2017).A family systems perspective
in prisoner reentry. In S. Stojkovic (Ed.), Prisoner reentry: Critical
issues and policy directions, (pp.85-144).NY: Palgrave Publishing.

Begun, A. L., Murphy, C., Bolt, D., Weinstein, B., Strodthoff, T., Short,
L., & Shelley, G. (2003). Characteristics of the Safe At Home
Instrument for Assessing Readiness to Change Intimate Partner
Violence,Research on Social Work Practice, 13(1), 80-107.

Begun, A. L., Shelley, G., Strodthoff, T., & Short, L. (2001). Adopting
a stages of change approach for individuals who are violent with
their intimate partners. Journal of Aggression, Maltreatment &
Trauma, 5(2), 105-127.

Begun, S. (2015). A mixed-methods examination of family planning

and social work: Implications for research, education, and the role
of personal beliefs. Contraception, 92(4), 389-390.

Bowen, E. (2017). Conduct disorder symptoms in pre-school children

exposed to intimate partner violence: Gender differences in risk
and resilience. Journal of Child & Adolescent Trauma, 10, 97-107.

Module 3 References | 325

Cameron, R., & Miller, P. (2007). Mixed methods research: Phoenix of
the paradigm wars. Paper presented at 21stannual Australian & New
Zealand Academy of Management (ANZAM) conference, Sydney,
Australia (December). Retrieved from
https://www.researchgate.net/publication/
37357300_Mixed_method_research_phoenix_of_the_paradigm
_wars

Carlson, J.F.. Geisinger, K.F., & Jonson, J.L. (Eds.). (2017). The twentieth
mental measurements yearbook.Lincoln, NE: Buros Center for
Testing.

Cash, S.J., Thelwall, M., Peck, S.N., Ferrell, J.Z., & Bridge, A. (2013).
Adolescent suicide statements on MySpace. Cyberpsychology,
Behavior and Social Networking, 16(3), 166-174.

Corcoran, K., & Fischer, J. (2013). Measures for clinical practice and
research, 5thed., v. 1-2. NY: Oxford University Press.

Cozby, P.C. (2007). Methods in behavioral research, ninth ed.Boston:

McGraw-Hill.

Creswell, J.W. (2014). A concise introduction to mixed methods

research. Thousand Oaks, CA: Sage.

Creswell, J.W., & Poth, C.N. (2018).Qualitative inquiry & research

design: Choosing among five approaches, 4thed. Thousand Oaks, CA:
Sage.

Curran, J.W., & Jaffe, H.W. (2011). AIDS: The early years and CDC’s
response. Morbidity and Mortality Weekly Report (MMWR)
Supplements, 60(4), 64-69 (October 7).

Dabelko-Schoeny, H., & King, S. (2010). In their own words:

Participants’ perceptions the impact of adult day services.Journal of
Gerontological Social Work, 53(2), 176-192.

326 | Module 3 References

Davis, T. S., Saltzburg, S., & Locke, C. R. (2010). Assessing community
needs of sexual minority youth: Modeling Concept Mapping for
service planning. Journal of Gay & Lesbian Social Services, 22(3),
226- 249.

Denzin, N.K., & Lincoln, Y., Eds. (2011). The SAGE handbook of
qualitative research, 4thed. Thousand Oaks, CA: Sage.

Engel, R.J., & Schutt, R.K. (2013). The practice of research in social
work, third ed. Thousand Oaks, CA: Sage.

Fisher-Borne, M., & Brown, A. (2018). A case study using photovoice

to explore racial and social identity among young Black men:
Implications for social work research and practice. Journal of
Ethnic & Cultural Diversity in Social Work, 27(2), 157-175.

Fowler, F.J., & Cosenza, C. (2009). Design and evaluation of survey

questions. In L. Bickman & D.J. Rog, Eds.,The Sage handbook of
applied social research methods, 2nded.,(pp. 375-412). Thousand
Oaks, CA: Sage.

Freisthler, B., Gruenewald, P.J., & Price Wolf, J. (2015). Examining the
relationship between marijuana use, medical marijuana
dispensaries, and abusive and neglectful parenting. Journal of Child
Abuse & Neglect, 48,170-178.

Glaser, B.G., & Strauss, A.L. (1967). The discovery of grounded theory:
Strategies for qualitative research.Chicago: Aldine Publishing.

Grinnell, R.M., & Unrau, Y.A. (2014). Social work research and
evaluation: Foundations of evidence-based practice, 10thed. NY:
Oxford University Press.

Hagborg, J.M., Berglund, K., & Fahlke, C. (2018). Evidence for a

relationship between child maltreatment and absenteeism among
high-school students in Sweden. Child Abuse & Neglect, 75, 41-49.

Module 3 References | 327

Helm, D. (2016). Sense-making in a social work office: An
ethnographic study of safeguarding judgements.Child & Family
Social Work, 21(1), 26-35.

Holosko, M.J. (2006). Primer for critiquing social research: A student

guide. Belmont, CA: Brooks/Cole.

Jirek, S.L. (2017). Narrative reconstruction and post-traumatic growth

among trauma survivors: The importance of narrative in social
work research and practice. Qualitative Social Work, 16(2), 166-188.

Johnson, R. B., Onwuegbuzie, A. J., & Turner, L. A. (2007). Toward a

definition of mixed methods research. Journal of mixed methods
research, 1(2), 112-133.

Kaiser, M. (2017). Redefining food security in a community context:

An exploration of community food security indicators and social
worker roles in community food strategies. Journal of Community
Practice, 25(2), 213-234.

Kettle, M. (2018). A balancing act: A grounded theory study of the

professional judgement of child protection social workers. Journal
of Social Work Practice: Psychotherapeutic Approaches in Health,
Welfare, and the Community, 32(2), 219-231.

Kokaliari, E.D., Roy, A.W., Panagiotopoulos, C., & Al-Makhamreh, S.

(2017). An exploratory comparative study of perspectives on non-
suicidal self-injurious behaviors among social work students in the
United States, Greece, Cyprus, and Jordan: Implications for social
work practice and education. International Social Work, 60(4),
1015-1027.

Kyriakakis, S., Waller, B., Kagotho, N., & Edmond, T. (2015).

Conducting safe research with at-risk populations: Design
strategies from a study with unauthorized immigrant women

328 | Module 3 References

 experiencing intimate abuse. Qualitative Social Work, 14(2),
259-274.

Lanier, P., Maguire-Jack, K., Walsh, T., Drake, B., & Hubel, G. (2014).
Race and ethnic differences in early childhood maltreatment in
the United States. Journal of Developmental & Behavioral Pediatrics,
35(7), 419-326.

Lawson, H.A. (2015). Introduction. In H.A. Lawson, J.C. Caringi, L.

Pyles, J.M. Jurkowski, & C.T. Bozlak, Eds. (2015). Participatory action
research, (pp. ix-xxvii).NY: Oxford University Press, Pocket Guides
to Social Work Research.

Lee, S.J., Hoffman, g., & Harris, D. (2016). Community-based

participatory research (CBPR) needs assessment of parenting
support programs for fathers.Children and Youth Services Review,
66, 76-84.

Lopez, S.J., & Snyder, C.R. (2003). Positive psychological assessment:

A handbook of models and measures: A handbook of models and
measures.Washington, DC: American Psychological Association.

McConnaughy, E.A.., Prochaska, J.O., & Velicer, W.F. (1983). Stages

of change in psychotherapy: Measurement and sample profiles.
Psychotherapy: Theory, Research & Practice, 20(3), 368-375.

Miller, D.C., & Salkind, N.J. (2002). Handbook of research design &
social measurement, 6thed.Thousand Oaks, CA: Sage.

Minnes, S., Min, M.O., Short, E.J., Wu, M., Lang, A., Yoon, S., & Singer,
L.T. (2106). Executive functioning in children with prenatal cocaine
exposure (12-15 years). Neurotoxicology and teratology, 57, 79-86.

Morgenstern, J., Kuerbis, A., & Muench, F. (2014). Ecological

momentary assessment and alcohol use disorder treatment,Alcohol
Research: Current Reviews, 36(1), 101-109.

Module 3 References | 329

Mueller, A.S., & Abrutyn, S. (2015). Suicidal disclosures among friends:
Using social network data to understand suicide contagion. Journal
of Health and Social Behavior, 56(1), 131-148.

National Highway Transportation Safety Administration (NHTSA).

(2010). National survey of drinking and driving attitudes and
behaviors: 2008. Volume I: Summary report.Washington, DC: U.S.
Department of Transportation. Retrieved from
https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/811342.pdf

Nguyen, D., Kit, B., & Carroll, M. (2015). Abnormal cholesterol among
children and adolescents in the United States, 2011-2014. NCHS data
brief, no 228. Hyattsville, MD: National Center for Health Statistics.

Nilsson, K.K., Straarup, K.N., & Halvorsen, M. (2015). Early maladaptive

schemas: A comparison between bipolar disorder and major
depressive disorder. Clinical Psychology and Psychotherapy, 22,
387-391.

Oktay, J.S. (2012). Grounded theory. NY: Oxford University Press,

Pocket Guides to Social Work Research series.

Olate, R., Salas-Wright, C., & Vaughn, M.G. (2012). Predictors of

violence and delinquency among high risk youth and youth gang
members in San Salvador, El Salvador. International Social Work,
55(3), 383-401.

Padgett, D.K. (2008). Qualitative methods in social work research,

2nded.Thousand Oaks, CA: Sage.

Perron, B.E., & Gillespie, D.F. (2015). Key concepts in measurement. NY:
Oxford University Press. Pocket Guides to Social Work Research.

Price Wolf, J., Freisthler, B., Kepple, N.J., & Chavez, R. (2017). The
places parents go: Understanding the breadth, scope, and
experiences of activity spaces for parents. GeoJournal, 82,355-368.

330 | Module 3 References

Riley, C., Roy, B., Harari, N., Vashi, A., Violano, P., Greene, A., …..Wang,
E.A. (2017). Preparing for disaster: A cross-sectional study of social
connection and gun violence. Journal of Urban Health, 94, 619-628.

Rohrer, J.M., Egloff, B., & Schmukle, S.C. (2015). Examining the effects
of birth order on personality. Proceedings of the National Academy
of Sciences, 112(46), 14224-14229.

Rush, A.J., First, M.B., & Blacker, D. (2008). Handbook of psychiatric

measures, 2nded.Washington, DC: American Psychiatric
Publications.

Russell, M. (1994). New assessment tools for risk drinking during

pregnancy: T-ACE, TWEAK, and others. Alcohol Health & Research
World,18(1), 55-61.

Sajatovoc, M., & Ramirez, L.F. (2012). Rating scales in mental

health,3rded. Baltimore, MD: Johns Hopkins University Press.

Salas-Wright, C.P., Vaughn, M.G., & Ugalde, J. (2016). A typology of

substance use among pregnant teens in the United States.
Maternal & Child Health Journal, 20(646-654).

Substance Abuse and Mental Health Services Administration

(SAMHSA). (2017). Key substance use and mental health indicators in
the United States: Results from the 2016 National Survey on Drug Use
and Health(HHS Publication No. SMA 17-5044, NSDUH Series H-52).
Rockville, MD: Center for Behavioral Health Statistics and Quality.
Retrieved from https://www.samhsa.gov/data/

Shiffman, S., Stone, A.A., & Hufford, M.R. (2008). Ecological

momentary assessment.Annual Review of Clinical Psychology, 4,
1-32.

Thomas, K.A., & Rickwood, D.J. (2016). One woman’s journey of

Module 3 References | 331

 recovery from mental illness—Hopes, back-up plans, rebuilding self
and service support. Qualitative Social Work, 15(4), 501-517.

Touliatos, J., Perlmutter, B.F., Straus, M.A., & Holden, G.W. (2001).
Handbook of family measurement techniques, vols. 1, 2 & 3.Thousand
Oaks: Sage.

Tran, T., Nguyen, T., & Chan, K. (2016). Developing cross-cultural

measures in social work research and evaluation, 2ndedition. NY:
Oxford University Press, Pocket Guides to Social Work Research
series.

Trochim, W.M.K., & Donnelly, J.P. (2007). The research methods

knowledge base, 3rded. Mason, OH: Thomson Custom Solutions.

Vaughn, M.G., Perron, B.E., Abdon, A., Olate, R., Groom, R., & Wu, L.T.
(2012). Correlates of handgun carrying among adolescents in the
United States. Journal of Interpersonal Violence, 27(10), 2003-2021.

Wang, C. (1999). Photovoice: A participatory action research strategy

applied to women’s health. Journal of Women’s Health, 8(2), 185-192.

Wray, T.B., Merrill, J.E., & Monti, P.E. (2014). Using ecological
momentary assessment (EMA) to assess situation-level predictors
of alcohol use and alcohol-related consequences. Alcohol Research:
Current Reviews, 36(1), 19-27.

Yoon, S., Kobulsky, J.M., Yoon, D., & Kim, W. (2017). Children and Youth
Services Review, 82(1), 271-279.

332 | Module 3 References

MODULE 4

Module 4 | 333
Module 4 Introduction
This module builds on earlier modules about understanding and
working with data concerning diverse populations, social
phenomena, and social work problems. First, we learn basics about
coding and analyzing qualitative data. Then, we learn basics about
analyzing quantitative data; these basics are further developed
through completing several linked Excel workbook learning activities.
In this module, we learn about conducting and interpreting basic
univariate and bivariate statistical analyses related to proportion,
frequency, probability, percentage, mean, median, mode, variance,
normal distribution, standard deviation, bar graphs, pie charts,
histograms, and scatterplots. We also learn about comparing
different demographic groups using 5 statistical analytic approaches
to test hypotheses (null and alternative hypothesis testing): single
sample t-tests, independent sample t-tests, one-way analysis of
variance, chi-square, and correlation. We become familiar with Type
I and Type II errors, and with basic nonparametric principles, as well.

Reading Objectives

After engaging with these reading materials and learning resources,

you should be able to:

• Describe how different types of qualitative data might be

interpreted or analyzed;
• Recognize and practice univariate data analysis for different
types of quantitative variables (categorical, dichotomous
categorical, ordinal, and interval/continuous variable types);
• Explain the logic underlying inferential statistics and null
hypothesis testing, as well as Type I and Type II errors;
• Identify underlying principles and practice 5 inferential

Module 4 Introduction | 335

 statistics approaches with quantitative data (univariate and
bivariate);
• Interpret results of one-sample t-test, independent samples t-
test, one-way analysis of variance (Anova), chi-square, and
correlation analyses;
• Recognize basic principles of non-parametric data analysis;
• Identify and practice steps in data analysis conducted using
Excel software;
• Define key terms related to working with qualitative and
quantitative data.

336 | Module 4 Introduction

Module 4 Chapter 1: Working
With Qualitative Data
Prior modules introduced qualitative approaches and methods for
study design and data collection. You learned that exploratory and
descriptive research questions are often addressed using qualitative
methodologies—naturalistic observation, interview, focus group,
social network, GIS, or open-ended survey data, to name a few.
Qualitative studies do not necessarily test hypotheses about the data
(although they can test hypotheses generated by prior theory (Glaser
& Struss, 1967). Qualitative studies often use data to develop an
understanding of social work problems, social phenomena, or diverse
populations. The focus of this chapter is on what investigators do
with the collected qualitative data to begin answering their research
questions.

In this chapter you will learn about:

• moving from qualitative data collection to data preparation;

• coding qualitative data.

Qualitative Data Preparation

Investigators engaged in qualitative research are careful to collect

data in a manner that preserves, to the closest possible degree, the
specific wording and context or what their study participants share.
Technology tools can assist in capturing participants’ statements
verbatim. For example, digital audio or video recording is possible
with small, portable recorders, cell phones, or dictation devices.
Digital recording is preferable to recordings based on “tapes” because
they are more durable and less vulnerable to accidental destruction.

Module 4 Chapter 1: Working With

Qualitative Data | 337
Furthermore, digital software exists to help transcribe these
recordings into written text (more about this later). The entire
process of data collection and preparation (recording and
transcription) needs to be approved by the institutional review board
(IRB) for human subject participation and consented to by the study
participants.

However, seasoned investigators do not rely on technology alone.

There exist many painful stories concerning data lost through
technology failures: not picking up the narrative in the first place
because microphones or cameras were not sufficiently sensitive or
were improperly placed, batteries running out, or damaged recording
devices; and, including accidentally erased recordings or other epic
failures. Audio and video recordings are usually supplemented with
field notes—these are written (or typed) either by the person
conducting the observation, interview, or focus group, or by a
collaborating observer/recorder. It is difficult to effectively play both
roles, interviewer and recorder, and maintain strong interview
rapport with participants; engaging a collaborator as recorder/note-
taker is well-advised.

338 | Module 4 Chapter 1: Working With Qualitative Data

There exists another, important reason for these field notes: they
contribute rich descriptive detail about the context of statements
made, supplementing the recorded and transcribed participant
statements, infusing the record with greater meaning. For example,
field notes can clarify who was the speaker when recorded voices
sound similar. And, the notes can describe changes in body language,
long pauses, facial expressions, making or losing eye contact, or other
events that can help investigators interpret meaning from the
context of what is said.

To demonstrate, consider an early study of children’s emerging

sibling relationships (Nadelman & Begun, 1982). The investigators
engaged firstborn preschool aged children (2 ½ to 5 years old) from
families expecting their second baby in projective doll play, a
structured format for observational data. Each of the children, in
their own homes, were introduced to the standardized, portable doll
house and doll family (parents, child, and baby matching their own
race and family composition). What the child said was audio-
recorded throughout the doll play session. The investigator also kept
field notes describing each child’s behavior with the dolls with
comments made every 30 seconds. These field notes were
particularly helpful in understanding child utterances, since the
children’s language skills were emerging. The notes also explained
long silences that occurred when children, upon discovering the toy
toilet, ran off to the potty themselves—potty training was an active
part of the children’s reality at this age, and a reminder to go was
often triggered by play with the toy toilet. Field notes also helped
when the doll play sessions were repeated after the birth of the
younger sibling. The children often ran to the baby when playing
with the doll baby, and the investigators had notes characterizing
a child’s touching, eye contact, and other interactions with the real
baby brother or sister. Combining these different types of qualitative
data allowed investigators to develop a rich description of emerging

Module 4 Chapter 1: Working With Qualitative Data | 339

sibling relationships that extended far beyond the preexisting ‘sibling
rivalry’ paradigm.

Qualitative Data Transcription. Investigators may choose to work

with observational data in real time—coding observed behaviors as
they occur. For the sake of reliable and verifiable data, however, they
often choose to work from recordings. One challenge with recordings
is the necessity to replay them, over and over, to analyze the data.
Instead, investigators often choose to work from transcripts of the
recorded interview, focus group, or observation sessions; reading the
transcript is often faster than repeatedly rewinding and replaying
content for coding purposes. As previously noted, however, some
of the rich context may be lost in translation from audio or video
recording to written transcription.

Data transcription is time consuming. Transcribing a one-hour

340 | Module 4 Chapter 1: Working With Qualitative Data

interview could require four or more hours of transcription
time—assuming the recording is clear and easy to interpret and there
is only one person speaking. It could require upwards of ten hours
in real time to transcribe a one-hour group interview (a family with
multiple members or a focus group, for example). Digital
transcription software can help (e.g., Dragon Speak®) but often
introduces inaccuracies, necessitating a great deal of time re-
checking the transcription to be certain of its accuracy. These
software packages MAY have contractual agreements that violate
research standards for confidentiality and data security—reading the
fine print before clicking on “ACCEPT” is critical, especially for “free”
transcription software packages. Investigators need to consider the
relative cost of their own time, the time and experience of study
team members, or professional transcription (e.g., ranges $1 to $5 per
minute across several websites visited in August, 2018). Furthermore,
transcription services need to be approved by the institutional review
board (IRB) when a study involving human participants is proposed,
and participant consent is required, as well.

Module 4 Chapter 1: Working With Qualitative Data | 341

 Try this out for yourself: watch a brief Youtube video, such as
“Charlie Bit My Finger—Again.” While listening to the video but not
watching the screen, write down everything you hear the children
vocalize. Now replay the video, listening while reviewing your
transcription.

• How accurately did you record the children’s vocalizations?

Replay the video again, this time both watching and listening.
Compare your audio transcription to what you see and hear–the
seeing being a context for what you hear.

• How accurately did your audio transcription reflect what

happened—the children’s vocalizations and the context?
• How much meaning do you think you lost without tone of voice
and body language/gestures?
• What do you think would improve the quality of the data you
transcribed, moving from the video to the written word?

Qualitative Data Coding

The process of coding qualitative data is systematic and should be

replicable—that being one hallmark of empirical evidence. This does

342 | Module 4 Chapter 1: Working With Qualitative Data

not mean that different investigators would draw the same
conclusions about the data—it means that others would be able to
repeat the process, following in a systematic fashion, use of the
same raw data. This represents a major distinction between
qualitative and quantitative research: in qualitative studies, the
investigator is recognized and accepted as part of the data
interpretation process; in quantitative studies, investigator impact is
minimized to enhance internal validity. Either way, however, the
procedures used are clearly identified and replicable, even if the
results in qualitative coding differ with different investigators.
Qualitative data coding and analysis is about grouping terms,
concepts, ideas, images, or other elements together into themes or
categories. The themes and categories provide a means of
organizing participants’ data in meaningful ways. Qualitative
analysis:

“aims to make sense of and give meaning to the data

collected. In general, the process of qualitative data analysis
involves the identification and organization of themes or
patterns from the words, text, and narratives obtained in the
data collection” (Corcoran & Secret, 2013, p. 166).

Coding themes, patterns, or categories derives from one of two

sources, depending on the qualitative study approach adopted and
the nature of the qualitative data collected. Contents may be
analyzed based on predetermined categories (a priori coding) or the
categories may emerge from the contents being analyzed.
Predetermined categories come from hypotheses based on theory or
literature. Categories emerging from the data is a process referred
to as open coding, which involves creating the categories or
groupings, confirming that the contents of each have points of
similarity or overlap, and providing the categories or groupings with
meaningful labels or names.
Tools for helping interpret qualitative data may be as simple as

Module 4 Chapter 1: Working With Qualitative Data | 343

colored highlighters applied to a printed transcription, statements
copied onto index cards which can be shuffled and re-arranged, or
digital highlighting applied to a Word® or .pdf document. A number
of sophisticated analysis-assistance software tools also exist: four
often reported in the literature are Atlas.ti®, NVivo®, MAXQDA®,
and Dedoose®. Several qualitative analysis assistance software
packages are free to download, at least in their “lite” form (e.g.,
Provalis QDA Miner®).
When selecting qualitative data analysis programs, investigators
need to consider several factors:

• These programs do not DO the analysis, they support the

investigator doing analyses—remember, the investigator is the
“tool” for determining themes and codes.
• Some programs only assist with text data, others assist with
analysis of images and other forms of data, as well (e.g.,
Dedoose®).
• Cloud-based programs and some free packages may have
practices that violate data confidentiality or security
requirements established by an institutional review board (IRB).

344 | Module 4 Chapter 1: Working With Qualitative Data

Coding and Coding Confirmation

As previously noted, an investigator may approach qualitative data

with a pre-established list of categories that are applied to the
data—deductive or a priori coding. The investigator applies the
predetermined coding categories with the data in terms of whether
each them appears or, possibly, how frequently the theme appears
(depending on the research question). For example, investigators
were interested in exploring the nature and extent of personal
information shared by adolescents in the MySpace social networking
site (Hinduja & Patchin, 2008). The study addressed a widespread
concern that individuals were providing information that left them
vulnerable to sexual predators. The team began with an a priori
coding scheme for the type of information youth publicly post: they
conducted a content analysis with a randomly sampled set of 9,282
MySpace profile pages. The coding included first name, full name,
birth date, telephone number, address or city/state of residence,
school attended, email address, instant messaging screen name,
pictures, pictures in swimsuit or underwear, and evidence of
alcohol/other substance use. In terms of identification, over 38%
of pages sampled provided the adolescent’s first name, almost 9%
provided a full name, over 81% provided a city of residence, and
almost 28% referenced their school. Furthermore, almost 57% of
sampled profiles included a photograph, and over 15% had a friend’s
photograph in a swimsuit or underwear, while slightly over 5% had
a photograph of themselves in swimsuit/underwear; over 18%
presented evidence of alcohol use, 7.5% tobacco use, and 1.7%
marijuana use. About 40% of youth set access to “private,” meaning
that the other 60% of profiles were potentially viewable by anyone.
While single pieces of information alone might not lead to identifying
an individual, in combination (especially with pictures included)
information shared makes identification and/or personal contact
possible. Since this research was completed, public education about

Module 4 Chapter 1: Working With Qualitative Data | 345

safer Internet use may have had an impact on individuals’ posting
patterns, and organizations have introduced security measures that
may reduce vulnerability. However, concern remains that individuals,
including adolescents, are vulnerable to predators or exploitation
based on what is shared across public domains of the Internet.

Open Coding. In contrast, when investigators approach qualitative

data without predetermined or preconceived ideas about categories
and themes present in participants’ responses, they engage in open
coding (an inductive process). This is the foundation of the qualitative
approach called grounded theory. Open coding was defined by Rubin
and Babbie (2013) as:

A qualitative data processing method in which, instead of

starting out with a list of code categories derived from theory,
one develops code categories through close examination of
qualitative data” (p. 337).

Accompanying the investigators’ description of the resulting coding

categories is a set of memos, notes, or journaling entries that depict
what the investigator was thinking in making these coding decisions.
This information helps guide others in how the results emerged to
provide transparency and replicability. While qualitative research is

346 | Module 4 Chapter 1: Working With Qualitative Data

science and results in empirical evidence, performing it well can be
compared to the art of dance (Engel & Schutt, 2018). This is because
it requires the investigator to maintain a state of openness in
interpreting the variety of ideas shared by participants while
concurrently maintaining objectivity in applying the methods and a
subjective awareness and reflection about the dynamic processes by
which themes and categories emerge.

For example, investigators reported on a study that utilized grounded

theory methods to explore the ways that transgender persons are
depicted in the media (McInroy & Craig, 2015). They conducted in-
depth, semi-structured interviews of several hours length with 19
young adults who self-identify as LGBTQ persons. The interviews
were audio-recorded and transcribed; coding of each interview was
conducted by three separate coders. The investigators found, first of
all, that participants described their experiences with offline media
(e.g., television and movies) differently from their experiences with
online media (e.g., websites and social media). Transphobic
representations (negative reactions or opinions concerning
transgender persons) emerged from the thematic coding as an issue
with offline media more significantly than issues of homophobia
(negative reactions of opinions concerning gay, lesbian, or bisexual
persons). The participants also expressed that offline media exhibits
very little in the way of positive representations of transgender
people; this was experienced as a contrast with more positive
representations of LGBTQ persons. The participants also
experienced transphobia to a greater extent in the online
environment, possibly because of the anonymity allowed in this
environment. On the other hand, the online environment offered
greater numbers and range of supportive, helpful options for
transgender youth. The study authors shared a large number of
verbatim quotes from their study participants to demonstrate the
coding categories that emerged from the data.

Module 4 Chapter 1: Working With Qualitative Data | 347

Cross-Checking Coding Decisions

In the example just presented, notice that the authors reported

three individuals coded each interview. This is an important aspect
in preserving the integrity of qualitative analysis—that one person’s
coding decisions be confirmed by others’ independent decisions.
Furthermore, many studies take their findings back to a subsample
of the original participants or a new sample to learn if their
conclusions are a good fit with the participants’ lived experiences.
These activities are part of the “assessing interpretations” step in the
qualitative “data analysis spiral” (adapted from Creswell & Poth, 2018,
p. 186):

• data collection
• managing and organizing the data
• reading and making memo notes about emergent ideas (initial
analysis)
• describing and classifying codes into themes (data reduction)
• developing interpretations (including how themes relate to one

348 | Module 4 Chapter 1: Working With Qualitative Data

 another, how they are distinct)
• assessing interpretations (how themes/interpretations hold up
in other examples, how well other investigators and study
participants agree with interpretations)
• representing and visualizing the data
• presenting an account of the findings.

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=428

Chapter Summary

In this chapter, you were introduced to general concepts and issues

involved in managing and analyzing qualitative data. There exist
excellent textbooks and online demonstrations for learning to master
these skills. This content also applies to mixed methods research
since many aspects of mixed methods data are qualitative.

Module 4 Chapter 1: Working With Qualitative Data | 349

Module 4 Chapter 2: Working
With Quantitative Descriptive
Data
Prior modules introduced quantitative approaches and methods,
including study design and data collection methods. This chapter
examines what investigators do with collected quantitative data to
begin answering their research questions. As you work your way
through the contents of this chapter, you will find it helpful to have
a basic calculator handy (even one on your smart phone will do).
In addition, you are presented with links to the Excel workbook for
learning activities that need to be completed using a computer (data
analyses do not work on most tablets or phones).

In this chapter you will learn:

• the distinction between univariate and bivariate analyses,

• descriptive statistics with categorical and numeric variables,
• computing and interpreting basic inferential statistics
• understanding graphs, figures, and tables with basic statistics.

Univariate and Bivariate Analyses

The way we think about quantitative variables has to do with how

they are operationalized in a particular study. First, let us consider
the distinction between univariate and bivariate work with
quantitative variables, then explore distinctions between
independent and dependent variables. When the goal is to describe a
social work problem, diverse population, or social phenomenon along

350 | Module 4 Chapter 2: Working

With Quantitative Descriptive Data
a particular dimension, investigators engage in univariate analyses.
The prefix “uni” means one (like a unicycle having one wheel).

Univariate Analysis. Univariate analysis means that one variable is

analyzed at a time; if multiple variables are of interest, univariate
analyses are repeated for each. Univariate analysis has the purpose of
describing the data variable-by-variable, and are often referred to as
descriptive statistics, or simply as “descriptives.” See Table 2-1 for an
example of how this might appear in a research report. The example
is based on a subset of data about the Safe At Home instrument
developed to assess readiness to change among men entering a
treatment program to address their intimate partner violence
behavior (Begun et al., 2003). The study participants are described in
terms of several variables, with the descriptive statistics reported for
each variable, one at a time.

Table 2-1. Univariate descriptive data example for N=520* men

entering treatment

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 351

 Mean
N % Range sd
(M)
Referral Type
nonvoluntary 424 81.5% --- --- ---
voluntary 96 18.5% --- --- ---

Race / Ethnicity
African American 215 41.8% --- --- ---
White 205 39.9% --- --- ---
Latino/Hispanic 71 13.8% --- --- ---
Other 23 4.5% --- --- ---

Relationship Status
Married/living with
234 45.6% --- --- ---
partner
Single in a relationship 101 19.7% --- --- ---
Not in a relationship 178 34.8% --- --- ---
Age --- --- 18-72 33.2 8.8668

0-520 40 56.4120
Time Since IPV Incident --- ---
weeks weeks weeks
Readiness Score --- --- .14-7.62 4.60 1.2113

*Note: totals may not equal 520 due to missing data or may not total
100% due to rounding.

Bivariate Analysis. When investigators

examine the relationship between two
Back to Basics: variables, they are conducting bivariate
Rounding Numbers analysis (“bi” meaning two, as in a
bicycle having two wheels). This is
contrasted to the descriptive work
using univariate analysis (one variable
at a time).

352 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 Rounding numbers:
The note at the
bottom of Table 1
mentions
“rounding” as a
reason why totals
might not equal
For example, investigators might be 100%. Rounding is
interested to know if two variables often used when
related to each other (correlated or there are more
associated). Turning back to the Safe at digits than
Home example (Begun et al., 2003), the necessary to
variable indicating men’s readiness to convey the
change their intimate partner violence important message.
behavior was positively correlated with For example, the
the variable about their assuming fraction 2/3 is the
responsibility for the behavior: men same as 0.666666
with higher readiness to change scores with the “6s”
were more likely to assume repeating
responsibility for their behavior than indefinitely when 2
were men with lower readiness scores. is divided by 3. In
Looking at this pair of variables social work, social
together, in relation to each other, is an science, and
example of bivariate analysis. Other behavioral
examples of research questions leading research, it is
to bivariate analysis include: having a seldom necessary
history of child maltreatment to go beyond two or
victimization associated with school three decimal
absenteeism in a Swedish study
(Hagborg, Berglund, & Fahlke, 2018);

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 353

 low empathy being associated with
gang membership (as were a number of
other variables) in an El Salvador study
places. We “round
(Olate, Salas-Wright, & Vaughn, 2012);
up” when the next
and mothers’ smoking during
digit is 5 or higher
pregnancy being associated with
(5-9) and “round
children’s behavioral regulation
down” when the
problems (and other variables) at the
next digit is less
age of 12 years (Minnes, eAdd Newt al.,
than 5 (0-4). In our
2016).
example, 0.666666
would be rounded How Univariate and Bivariate
up to 0.67. The Analyses Are Used: In earlier modules
decimal version of you were introduced to the types of
1/3, 0.333333 with research questions that lend
the “3s” repeating themselves to quantitative approaches
indefinitely, would (Module 2) and to different kinds of
be rounded down quantitative variables (Module 3).
to 0.33. Thus, with Likewise, the type of data analyses and
slight rounding reports that investigators generate are
error being directly related to the nature of the
introduced, we questions the data answer. Univariate
might end up with a and bivariate analyses are used
total slightly lower descriptively. In other words, when the
or higher than question to be answered calls for a
100%. description of a population, social work
problem, or social phenomenon, these
are useful analyses to conduct and
present. Univariate and bivariate
analyses help develop a picture of how the numeric values for each
quantitative variable are distributed among a sample of study
participants. Bivariate analyses are also used to test hypotheses
about the relationships between variables—the subject of Chapters 3,

354 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

4, and 5 in this module. The next section goes into greater detail as to
the nature of descriptive analyses.

Univariate Descriptive Analysis

Here we explore ways that investigators might analyze specific types

of variables in order to help answer their research questions. First,
our focus is on descriptive statistics: statistics that help describe
populations and groups, or the dimensions of social work problems
and social phenomena. As we work through these Module 4 materials,
you may find that the question “Statistics or sadistics?” asked in the
book Statistics for people who (think they) hate statistics (Salkind, 2017)
changes to the statement “Statistics, not sadistics!”

Descriptive Statistics

Descriptive statistics help develop the picture of a situation. Since

the point of variables is that their values vary across individual cases,
investigators need to understand the way those values are distributed
across a population or in a study sample. The three main features
about distribution are of great interest to investigators:

• frequency,
• central tendency, and
• variance.

These three features are reported for one variable at a time

(univariate analysis). The descriptive statistics reported about a
variable depend on the type of variable: categorical, ordinal, and
interval variables that you learned about in Module 3. In Module 3
you also learned about descriptive research: quantitative descriptive
research reports the results of descriptive statistical analyses.
Descriptive analyses are reported for variables used in quantitative

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 355

exploratory and explanatory studies as background for
understanding the additional (bivariate and other) statistical analyses
reported.

Frequency Analysis: Categorical Variables

The first descriptive analyses are concerned with frequency.

Frequency is a count of how much, how often, or how many,
depending on what the variable was measuring.

Frequency Count. Counting “how many” is among the simplest

descriptive statistics. In the Safe At Home example above (Table 2-1),
investigators wanted to know how many men participating in the
study entered batterer treatment voluntarily and how many were
mandated to enter treatment by court order (nonvoluntary clients).
As you can see in Table 2-2, the total number of participating clients
was 520 (this would be written as N=520, where the capital N reflects
total sample size). In addition, you can see that the study included 96
“voluntary” clients and 424 “nonvoluntary” clients (this is sometimes
written as n=96 and n=424, where the lower case “n” is about a
group within the larger, full set where we used “N”). Not only do we
see that more clients entered nonvoluntarily, we can calculate how
many more nonvoluntary clients there were: 424 – 96 = 328 more
nonvoluntary than voluntary clients participated in the study.

Table 2-2. Frequency report for voluntary and nonvoluntary

participants.

N %
nonvoluntary 424 81.5%
voluntary 96 18.5%

Total 520 100%

356 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.4-2.1 Frequency Counts

Pie-Chart. The pie-chart is a useful tool for graphically presenting

frequency data. It is a quick, easily-interpreted visual device. Here is
a sample pie chart depicting the nonvoluntary and voluntary client
frequency (and proportion) data described in Table 2-2.

Frequency, Proportion, or Percentage. The investigators in our

example not only wished to know how many voluntary and

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 357

nonvoluntary clients were involved in their study, they also wanted to
consider their relative proportions. You might wonder why it matters
since they already know how many men were in each group, and
that there were 328 more involuntary than voluntary clients in the
study. It matters because the investigators wish to evaluate how
meaningful that difference might be. A difference of 328 clients is
proportionately large in a sample of 520, but the same 328 would be
proportionately small among a very large sample such as 5,200.

A proportion is simply a fraction, or

Back to Basics:
ratio, calculated by dividing the
Converting Fractions
number in each group (fraction
to Decimal Units
numerator) by the total number in the
whole sample (fraction denominator).
To simplify interpretation, the fraction
is often converted to percent—the
Mathematical
proportion of 100% represented in the
operations are
fraction.
simpler when
fractions are
• For nonvoluntary clients, this would be
computed as: (424/520) x 100%, which
comes out to be (.815) x 100%, or

358 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 81.5%.
• For voluntary clients, this would
be computed as: (96/520) x 100%,
expressed in
which comes out to be (.185) x
decimals. To make
100%, or 18.5%.
the conversion,
think about what
The investigators provided these
the fraction stands
percentages in Table 2-2; the
for: a top number
percentages total to 100% (81.5% +
(numerator) over a
18.5% = 100%). You can see that the two
bottom number
groups were not equivalent in actual
(denominator)
size, relative proportion, or
means to divide the
percentages of the total sample. In
numerator by the
Chapters 4 and 5 you will learn about
denominator.
statistical analysis approaches that
Example: 3/4
allow investigators to test if the size of
means divide 3 by 4,
the observed difference is statistically
so 3/4 = .75
significant.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.4-2.2 Proportion and Percentage

Univariate Central Tendency Analysis: Numeric Variables

Knowing the frequency with which certain categories or levels of a

categorical variable are reported in the data is useful information, as

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 359

witnessed above. Frequency counts are less useful in understanding
data for a numeric variable, however. For example, in the Safe At
Home study, it would be difficult to interpret information about the
participants’ ages: 3 were age 18, 7 were age 19, 17 were age 20, 8 were
age 21, 14 were age 22, and so forth for all 520 individuals between 18
to 72 years. A more useful way of looking at these numeric (interval
type) data is to consider how the study participants’ values for age
were distributed and how any individual’s values for age might differ
from the rest of the group. Ideally, we would like to see a meaningful
summary of the data.

The concept of central tendency refers to a single value that

summarizes numeric data by characterizing the “center” values of
a data set—a single value around which individuals’ values tend to
cluster. Central tendency helps make sense of the variability
observed among participants on a numeric variable. The most
commonly used central tendency ways to summarize data are:

• mean,
• median, and
• mode.

For example, within a classroom of public school students we might

see many different values for the variable “number of days absent”
during a school year. Most states in the U.S. require 180 student
instruction days in the school year. Let’s consider a hypothetical
example (see Table 2-3) where hypothetically:

• we have a class of 28 students

• the number of days absent for each student ranges from 0 to 49
out of the year’s 180 days
• the total number of absent days for the class members
combined was 194.

360 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

Table 2-3: Number of days absent for each student in (hypothetical)
class.

student # days student # days

1 2 16 6
2 4 17 4
3 0 18 49
4 0 19 0

5 26 20 3
6 0 21 0
7 15 22 4
8 3 23 10
9 9 24 1

10 12 25 2
11 1 26 0
12 1 27 2
13 3 28 12
14 19 --- ---
15 6 total 194

Our goal is to find a single central

tendency value that helps represent
the group on this days absent variable; Back to Basics:
summarizing the variable without Computing and
having to rely on all 28 values at once. Average or Mean Value
Let’s look at each of the common
central tendency options.

Mean. The first central tendency

indicator we consider is the mean. You
might recognize this as numeric

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 361

 average; average and mean for a set of
values are the same thing. The mean is
computed by adding up all the values,
Step 1. Add up the
then dividing by the actual number of
values for each
values. In our example, we would first
member of the
find the sum of the number of days
group on the
absent across all the students together
variable of interest
(194), then divide that total by the
Step 2. Divide that
number of students contributing to
total by the number
that total (N=28). This would be the
of values there
mean value for our students on the
were in the group
variable for number of days absent: 194
(how many
divided by 28 (194/28) = 6.93 with
individual members
rounding. We would report this as
contributed to that
M=6.93.
total).
Step 3. The result is
the “average” value
Interactive Excel Workbook Activities
for the group, also
called the “mean”
value (written as a
capital, italicized
M). Complete the following Workbook
Activity:

• SWK 3401.4-2.3 Computing a Mean

Median. The median is the next common central tendency feature,

and it can be thought of as the half-way point in a distribution of
values. In other words, it is the point where half of the values are
higher and half of the values are lower. Another name for the median
is the 50th percentile—since 50% is halfway to 100%. In other words,
the 50th percentile is the value where 50% (half) of the scores are

362 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

lower and the other 50% (other half) are higher. In our classroom
absenteeism example, the median value is 3 days: three is the point
where half of the students had more and half had fewer days absent.
We can easily compute that half of the 28 students is 14, so we identify
the median point as where the lowest 14 values are included and all
of the higher values are excluded—14 students had scores of 3 or less.
We would report this information as: Mdn=3.

Mode. The last common measure of central tendency to consider is

the mode. Sometimes it is useful to know the most common value
for a variable. This is simply a frequency count—how many times did
each value appear in our data. For our student absenteeism data, the
most common value was 0—six of our students did not miss any days
of school (see Table 2-4). This would be written as Mode=0.

Table 2-4. Frequency of each value for days absent example.

# days # of students with # days # of students with

absent that value absent that value
0 6 10 1
1 3 12 2
2 3 15 1
3 3 19 1

4 3 26 1
6 2 49 1
9 1 total 28

Interactive Excel Workbook Activities

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 363

 Complete the following Workbook Activity:

• SWK 3401.4-2.4 Computing Descriptive Statistics

(Mean, Median, Standard Deviation)

Comparing Mean, Median, and Mode. Why wouldn’t we always use

the mean to describe the variable of interest? The answer is
demonstrated in our student absenteeism example: even a few
extreme values (sometimes called outliers) can seriously distort or
skew the picture. The most common value (mode) was 0 but we had
a couple of students with some really high values compared to the
others: 26 and 49. When we add those two extreme values into the
total number of absent days to compute the group mean, they have a
big impact on the total (194): the mean for the 28 students was 6.93
days absent. Without those two extreme values, the total would have
been 119, making the mean for the 26 remaining students (28 – 2 =
26) only 4.6 days (119/26=4.6). This mean of 4.6 days is considerably
less than the mean of 6.93 days we observed with these two extreme
students included. Thus, you can see that the two extreme values had
a powerful skewing impact on the mean.

As a real-world example, consider data about the 2017 incomes

(salary plus bonuses) for 1,999 NFL professional football players. The
range was $27,353 to $23,943,600—a difference between 5-figure and
8-figure incomes! In 2017:

• mean player income was $1,489,042 (we would report this as

M=$1,489,042),
• median income was $687,500 (Mdn=$687,500), and
• mode income was $465,000.

364 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

The mean and median incomes are very different, and the mean
differs markedly from the mode, as well. This NFL player income
example demonstrates what data look like when great disparities
exist between individuals across a group or population—because the
mean is highly sensitive to extreme values, there exist large
differences between the mean, the median, and the mode. This is
an important social work and social justice issue in poverty, health,
incarceration, and other rates where large disparities between
population groups exist.

In a relatively small sample (like our 28-student classroom), one or

two extreme values made a big difference between the mean and
median. Think about the mean income of people in a room where
everyone is a social work student learning about social justice issues.
Then, think about what would happen to that mean if Bill Gates
arrived as a guest speaker invited to talk about social justice on
a global scale. Just that one person would drastically change the
summary picture based on mean income. However, the median would
not really shift much; Bill Gates would be only one person out of
the total number of people in the room; the halfway point is not
affected much by one person being extreme, since the median is
a frequency count of people on either side of the central midpoint
value regardless of how far out the values are spread. The mode
also is not affected since Bill Gates’ one extreme value is certainly
not going to be the most common value (otherwise it would not
be extreme). This all leads to a discussion of the importance in
understanding the nature of variance as it relates to distribution of
values for a variable.

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 365

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=429

Univariate Data Distribution: Numeric Variables

Central tendency is about how the set of values for a specific variable
are similar, how they cluster around a mean, median, or mode value.
However, investigators are also interested in how individual values
differ. This is a curiosity about how those values are distributed or
spread across a population or sample: the range and variance of the
values. These elements are explained below, along with the more
commonly reported standard deviation.

Range. Turning once again to our school absenteeism example, we

can see that there exist very great differences in the numbers of
days absent among the students in this class. The minimum value
observed was 0 and the maximum value observed was 49. The range
of values in a set of data is computed as the highest value minus the
lowest value. In our example, the range would be 49 – 0 = 49 days
absent. It is just a coincidence that the range in our example equals
the highest value: this is only because our lowest value was 0 (if the

366 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

lowest value was different from zero, the range would be different
from the highest value).

The range has its greatest meaning when we understand the context
in which it exists: a 49-point range is large for the possible 180
days but would not be if the context for possible values was 1,800
instead. Therefore, it becomes helpful to consider the range in ratio
or percent absent days, giving the range greater meaning. This is
calculated as the observed value divided by the possible value for the
ratio, then multiplying by 100% to determine percent.

• Missing 0 of 180 days possible is a division (ratio) problem: (0/

180) x 100% = 0 x 100% = 0% days absent;
• Missing 49 of 180 days possible is computed as (49/180) x 100%
= 27.2 x 100% = 27.2% days absent.
• Thus, the range is 0% to 27.2% days absent.

One statistical approach to assessing this variability in days absent is

our next topic: understanding variance.

Variance. One way of thinking about

variability in the data is to compare
Back to
each individual’s score to the mean for
Basics: Subtracting
the group—how far from that central, Negative Numbers
summary value each individual’s value
falls. When individuals’ scores are
clustered very close to the mean, the
variance is small; when individuals’
Computing the
scores are spread far from the mean,
range can be a little
the variance is large.
tricky if the lowest
This sequence of steps describes how value is a negative
variance is computed in a sample.

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 367

 • Step 1. Compute the mean of all the
values together (overall mean).
• Step 2. Compute the distance of each
number and the
individual value from that mean—this
highest value is a
becomes a subtraction problem where
positive
the mean is subtracted from each
number–unless you
individual score, showing the
remember that
“difference” between the individual
subtracting a
score and the group mean.
negative number is
• Step 3. Multiply each of those distance
the same as adding
values by itself (square the difference
it as a positive
value). Why bother with this step, you
number (the two
might wonder? This is done to resolve
negative signs
potential problems with negative
cancel each other
numbers from our earlier subtraction
out). For example: if
step. Some values were smaller than
the lowest value is
the mean and some were larger than
-4 and the highest
the mean—when we subtract the mean
value is 26, then the
from those smaller values, we end up
range would be
with a negative number. In our next
computed as: (26) –
step (step 4) we are going to add the
(-4) = (26) + (4) = 30
computed distances together. If we
add in negative numbers, we are
essentially subtracting values from the
total which does not give an accurate picture of the total
distances from the mean. To solve this negative number
problem, we capitalize on the fact that multiplying two negative
numbers together gives a positive value. Therefore, if we
multiply a negative value by itself (square a negative difference
score) we get a positive distance value (the distance squared).
• Step 4. This step is where we add up the squared distance values
to get a single total (sum) of the squared distances.
• Step 5. Divide this sum by the number of values in the data set

368 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

minus 1—one less than the number of contributions to the
variance calculation. Why is this not just divided by the number
of values instead, you might wonder? It has to do with the fact
that we are computing variance of a sample, not the whole
population. Without this adjustment, the computed sample
variance would not be quite the same as the population
variance—more about this distinction between sample and
population parameters later, when we talk about statistical
analyses and the conclusions drawn from the statistical
calculations.

Back to Basics:
Multiplying With
Negative Numbers

Remember that
multiplying two
numbers with the
same sign results
in a positive sign
number, whether
those two signs
are both positive
or both negative.
So, when
multiplying two
negative
numbers, the

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 369

 Here is what it looks like with our
school absenteeism example.

negative signs
• Step 1. We know that the mean is
cancel each other
6.93 from our earlier discussion
out.
about means.
positive x positive • Step 2. Our first student missed 2
= positive days, so the distance or difference
positive x score is:
negative =
◦ (2 – 6.93) = -4.93 (a negative
negative
number).
negative x
• Step 3. Now we compute those
negative =
differences for every one of our 28
positive
students. Once we have done this,
we compute the square of each
difference by multiplying it by itself.
For our first student, this is:

◦ (-4.93)2= (-4.93) x (-4.93) = 24.3 (rounding)

• Step 4. After computing each student’s squared difference, we
add them up (sum or total).
• Step 5. We take that total value and divide it by one less than the
number of students contributing data: (28-1) = 27. The value we
end up with for variance in the class is 109.25 (see Table 2-5).

Table 2-5. Computing variance in student absenteeism example:

(2949.86)/(27)=109.25

370 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 distance distance
# 2 # 2
student from (difference) student from (difference)
days days
mean mean
1 2 -4.93 24.30 16 5 -.93 .86
2 4 -2.93 8.58 17 4 -2.93 8.58
3 0 -6.93 48.02 18 49 42.07 1769.88

4 0 -6.93 48.02 19 0 -6.93 48.02

5 26 19.07 363.66 20 3 -3.93 15.44
6 0 -6.93 48.02 21 0 -6.93 48.02
7 15 6.07 65.12 22 4 -2.93 8.58
8 3 -3.93 15.44 23 10 3.07 9.42

9 9 2.07 4.28 24 1 -5.93 35.16

10 12 5.07 25.70 25 2 -4.93 24.30
11 1 -5.93 35.16 26 0 -6.93 48.02
12 1 -5.93 35.16 27 2 -4.93 24.30
13 3 -3.93 15.44 28 12 5.07 25.70

14 19 12.07 145.68 ------ ------ ------ ------

15 6 -.93 .86 total 194 ------ 2949.86

Standard Deviation

Working with variances can be unwieldy, for two reasons. A practical

reason is that the numbers can become very large, making them
difficult to work with. Another practical reason is that variance is a
bit difficult to interpret. Remember when we calculated the square of
the differences in days absent minus the mean? We started with the
unit of analysis being days absent. By squaring the values, our units
are now the square of the difference in days absent. That just is not
very intuitive. Instead, we find it easier to convert the figure back to
the unit of days of absent. To do this is straightforward math: we take
the square root of the variance of the squared distances. This gets

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 371

both the value and the units back to a more interpretable, practical
place.

• Square root of 109.254 is 10.452 (using the square root calculator

on www.math.com)
• Cross-check this yourself: multiply 10.452 x 10.452 and see what
result you get—it is slightly different in the third decimal place
because of rounding.
• Thus, the standard deviation in our example is reported as:
sd=10.45 days absent.

The way to think about standard deviation is that the greater a

standard of deviation is, the more variable the values are—the more
they spread rather than cluster around the mean. This is an
important aspect of understanding the distribution of values on the
variable of interest. A graph is also very helpful.

Graphing Numeric Variables

The way the numeric values of a variable are distributed can be

visualized using what is called a histogram. The histogram is a graph
where the x-axis represents the set of observed values for a specific
variable (a univariate analysis), and the y-axis represents how many
times each of those values was observed in the data (see Figure 2-1).
You can see from the height of the bars that the frequencies begin to
taper off after about 4 days absent and how far from the rest of the
class those few extreme students are situated.

Figure 2-1. Histogram depicting distribution of days absent data,

mean, and median.

372 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

Another clue about the degree of spread or variability is the relative
size of the sample’s standard deviation compared to the sample mean.
In our example, the sd=10.45 is greater than the M=6.93. This
indicates that a great deal of spread exists in the data. This
observation relates to understanding the characteristics of normal
distribution.

NORMAL DISTRIBUTION

You have probably been exposed to at least the general concept of the
bell curve or normal curve at some point in your education. A normal
distribution curve is sometimes called a “bell” curve because of its
shape: high in the middle, low in the “tails,” and even on both sides
of center. Bell-shaped curves do not necessarily meet all criteria of a
normal distribution curve—you will see examples in this chapter. We
focus on the true normal curve at this point, a fundamental concept
related to many of the statistics used in social work, social science,
and behavioral research.

Our in-depth exploration of its importance in understanding data

begins with the picture of a normal distribution curve (Figure 2-2).
The type of variable mapped onto a normal distribution curve is an
interval or continuous variable—the units are equal intervals across a
continuum.

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 373

Figure 2-2. Generic normal distribution curve (histogram).

In a normal distribution curve, the values are arranged around the

mean in very specific ways.

• The mean is the center of the curve.

• The observed data values are distributed equally on either side
of the mean—the right and left sides are mirror images. This
goes along with the mean being in the center; the right and left
ends are called “tails” of the distribution curve.
• The median is equal to the mean. This also goes along with the
mean being in the center of the distribution.
• The set of observed data values (the frequency each value is
observed) are distributed in the following way, with regard to
the standard deviation value (see figure 2-3):

◦ 68% are within one standard deviation of the mean (34% on

each side of the mean)—see the grid-marked area under the
curve in Figure 2-3.
◦ Combining all values observed for 2 standard deviations
either side of the mean will include a total of 95% of the

374 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 values (all of those within 1 standard deviation plus those
from 1 to 2 standard deviations from the mean).
◦ Combining all values within 3 standard deviations of the
mean accounts for 99% of the values.

Figure 2-3. Proportion of values under the normal curve by standard

deviations from the mean.

Percentile Scores.You may wonder how this relates to information

such as someone having a 95th percentile score on some measure.
Look at the right-side tail, out at 2 standard deviations from the
mean. You can see that very few scores are way out in the tail beyond
that point—in other words, it is rare to be more than 2 standard
deviations away from the mean, and exceedingly rare to be more than
3 standard deviations away.

We saw that 95% of the scores fall between -2 and +2 standard

deviations. This, however, is NOT the 95th percentile score. The
reason it is not has to do with the definition of the 95th percentile:
it is the point with 95% of the observed scores falling below and only

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 375

5% of the observed scores falling above. In this case, at 2 standard
deviation above the mean (+2) we have 95% of scores in the -2 to +2
standard deviation range, which means 5% of scores are outside the
range (100% – 95% = 5%). But, these are not all on one tail; there
are those scores in the -2 to -3 standard deviation range to contend
with. Those scores are also below the point we identified for being
2 standard deviations above the mean. So, using the +2-standard
deviation value as our decision criterion, we would actually have half
of the left-over 5% (5%/2 = 2.5%), plus the original 95%. This means 2
standard deviations above the mean is the 97.5 percentile (95% + 2.5%
= 97.5%). The trick is to keep in mind which standard deviation range
is relevant to the question being asked—is it about a range around the
mean or is it about a range up to a standard deviation criterion value
(see Figure 2-4)?

Figure 2-4. Computation of percentile for +2 SD.

Here is another example to consider. Imagine that policy makers

want to provide services to individuals whose IQ scores are equal
to or less than 1 standard deviation below the mean (this would

376 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

be written as: ≤ -1 SD). This decision concerns one tail of the
distribution—those at the far left (bottom) of the range, not the mirror
tail on the right (top) of the range. If you wanted to determine what
percentile would be covered, you would start with figuring out the
total percent that fall outside of the -1 to +1 standard deviation range.
That would be 100% minus the 68% that ARE in the range: 100% –
68% = 32% outside of the range. We also do not want to include those
individuals in the range greater than 1 standard deviation above the
mean (> +1 SD). They account for half of the number who are outside
the -1 to +1 standard deviation range, which was 32%. So, we are going
to subtract half of 32% (32%/2 is 16%) and we are left with only 16%
of individuals falling below 1 standard deviation being eligible for the
services (see Figure 2-5).

Figure 2-5. Computation of percentile for -1 SD.

In fact, this scenario became an issue in one state during the 1970s.
IQ scores are presumed to distribute on a normal curve with a mean
of 100 across the population, and each standard deviation defined
by 26 points on the test. As a cost-saving plan, the state determined

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 377

that services to persons with intellectual developmental disabilities
would be offered to individuals whose IQ scores were 68 or below.
Previously, the criterion value had been 70. Rather than having
“cured” individuals whose scores were between 68 and 70, the state
simply ceased providing services to them—several thousand
individuals no longer met criteria.

Next, let’s consider the real-world example of NFL player salaries

mentioned earlier. Looking at the histogram, it does not seem to be
very normally distributed. First, the mean does not seem to be in
the middle, the bars to the right and left of the mean are not mirror
images, and the median is very much lower than the mean—they are
not even close to equal (see Figure 2-6). These are ways in which
distribution on a variable fails to be normal.

Figure 2-6. Histogram of NFL player salaries.

Bimodal Curve. Another way in which distributions are non-normal

is when we see a bimodal curve. The normal, or “bell” curve, has only
one peak or hump in the data. Sometimes, we get a curve that has
two humps. If there had been more players in the 400,000 to 600,000

378 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

range we might have seen a second “hump” after the first one at
around 500,000 to 600,000.

This scenario is “scary” in statistical terms, because it means that

our assumption of normal distribution, the basis of many statistical
analyses, has been violated. If you ever read The Little Prince (de
Saint-Exupéry, 1943), you might understand the scary picture better.
The Little Prince shows his scary picture (Drawing Number one) to
the adults, who say a hat is not scary at all.
© de
Saint-Exupér
y

But, the picture is scary because it is not a picture of a hat, it is a

picture of a snake who ate an elephant.

© de Saint-Exupéry

To statisticians, a bimodal curve is a scary picture—it has two

“humps.” In other words, there exists a second peak in the data and
this needs to be explored. As an example for explaining the bimodal

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 379

curve in The Little Prince’s Drawing Number One, perhaps the curve
is showing salaries where the left peak (lower salaries) is for women
and the right peak (higher salaries) is for men; or, it could be about
salary differences for two different raceial groups; or, it could depict
an inequity or disparity in the frequency for some other variable.

Skew and Kurtosis. Two other ways that a variable might be

distributed in a non-normal way is that it is skewed to one side or the
other in relation to the mean, and/or the peak of the bell is either too
tall or too flat to represent those probabilities we mentioned earlier
(e.g., 68% being within one standard deviation of the mean). The first
of these is called skew and the second is called kurtosis. A curve with
a lot of skew fails to look normal because the curve is no longer
symmetrical on both sides of the mean. This is how skewed curves
might look:

Univariate statistics often report a value for skewness. While the

computer can do the calculations, the trick lies in knowing how to
interpret the values. Here are some rules of thumb for interpreting
skewness values:

380 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 • If the value for skewness is less than -1.0 or greater than +1.0, far
from zero, the distribution is highly skewed (asymmetrical or
non-normally distributed).
• If the value for skewness is between -1.0 and -0.5, or it is
between +.5 and +1.0, then the distribution is considered to be
moderately skewed (asymmetrical).
• If the value for skewness is between -0.5 and +0.5 (including
zero), the distribution is considered relatively symmetrical.

This diagram might help you visualize what these words mean:

Kurtosis, on the other hand, has to do with how much of the

distribution is sitting out in the tail or tails instead of being in the
center “peak” area compared to a normal distribution. On one hand,
there may be too few cases in the tails (clustered too much in the
“peak”); on the other hand, there may be too many cases in the tails
(spread away from the “peak”). This is how curves with kurtosis might
look:

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 381

Kurtosis values are a bit more complicated to interpret than
skewness values. When we learn about statistical tests of significance
we can revisit the issue. In general, kurtosis is considered to be
problematic when the value falls outside of the range from -7.0 to
+7.0 (for social work, social science, and behavioral research). Some
scholars argue for applying criteria with a tighter range, from -2.0 to
+2.0.

The website called Seeking Alpha (α, https://seekingalpha.com/

article/2100123-invest-like-taleb-why-skewness-matters) explains
the importance of skew and kurtosis by using an example from
Nassim Taleb (2012):

Suppose you placed your grandmother in a room with an

average temperature of 70 degrees. For the first hour, the
temperature will be -10. For the second hour, it will be 140
degrees. In this case, it looks as though you will end up with no
grandmother, a funeral, and possibly an inheritance.

(Note: for our purposes, degrees Celsius were translated into degrees
Fahrenheit to make the point from the original example clear.) The
website goes on to state:

While this example is rather extreme, it does show one

important principle that is highly relevant…It is dispersion
around the mean that matters, not necessarily the mean itself.

382 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=429

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.4-2.5 Understanding Skew and Kurtosis

Understanding Graphical Representations: Categorical Variables

There exist at least 3 different ways of presenting the very same

information about categorical variables: as text (words) only, as a
table of information, or in graphs/figures. We can compare these

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 383

approaches using a real-world data example: suicide rates by age
group, as reported by the Centers for Disease Control and Prevention
(CDC, 2017). A “rate” is similar to a frequency: rather than being
an actual count (frequency), the frequencies are converted into
population proportions. So, we often see a rate described as some
number per 1,000 (or per 10,000, or per 100,000) of the population,
instead of seeing the actual number of individuals.

Text: During 2016, middle-aged adults between the ages of 45 and 54

represented the group with the highest suicide rate (19.7 per 100,000
persons in the population). This was closely followed by individuals
aged 85 or older (19.0 per 100,000), adults aged 55 to 64 (18.7), and
older adults in the 75 to 84 age range (18.2 per 100,000). Next were
adults in the 35 to 44 age group (17.4 per 100,000), adults in the 65
to 74 age range (16.9 per 100,000), and adults aged 25 to 34 (16.5 per
100,000). The suicide rate among adolescents and young/emerging
adults aged 15 to 24 was somewhat lower (13.2 per 100,000), and the
very lowest rate occurred among persons under the age of 15 (1.1
per 100,000). The rate for all ages combined was 13.42 per 100,000
persons in the population (age-adjusted).

Table: Table 2-6 presents the same information described in the text
above, but much more succinctly.

Table 2-6. Suicide rate per 100,000 by age group during 2016.

384 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

Age Group Rate per 100,000
Under 15 1.1
15 – 24 13.2
25 – 34 16.5
35 – 44 17.4
45 – 54 19.7

55 – 64 18.7
65 – 74 16.9
75 – 84 18.2
85 or older 19.0
overall age-adjusted rate 13.42

Bar Charts: The same information is presented once again in Figure

2-6, but graphically this time, where you can visually compare the
height of the bars representing each age group.

Figure 2-6. Suicide rate bar chart.

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=429

Module 4 Chapter 2: Working With Quantitative Descriptive Data | 385

Chapter Summary

The emphasis in this chapter surrounded univariate descriptive

statistics. You learned about describing frequency and percentage
results. Then, you learned about the “Central Tendency” statistics
most commonly used in social work, social, and behavioral research:
mean, median, and mode. You also learned about the statistics
concerned with how data are distributed: variance and standard
deviation. Related to the issue of data distribution, you learned about
normal distribution (and histograms), as well as the nature of skew
and kurtosis in relation to non-normal distribution. Finally, you
witnessed the power of graphs and table to present univariate
statistics information compared to written text. In the next chapter,
we begin to explore bivariate analyses.

386 | Module 4 Chapter 2: Working With Quantitative Descriptive Data

Module 4 Chapter 3:
Univariate Inferential
Statistical Analysis
In this chapter we introduceinferential statistics strategies,
beginning with univariate analyses and expanding to bivariate
analyses. Why these statistical approaches are called “inferential”
may seem, on the surface, to be a simple question. Simple or not, it
is an important question. The word “inferential” is used because
investigators and statisticians use data about samples to make
inferences about the population-as-a-whole. Working with sample
data is necessary when it is not feasible to measure or observe every
single individual in the population. The idea with inferential
statistics is that investigators can estimate the degree of confidence
they can reasonably place in their conclusions about a population,
conclusions based on data from a sample.
In this chapter you will learn about:

• the importance of distinguishing between populations and

samples;
• confidence intervals for univariate population parameters based
on sample statistics;
• degrees of freedom;
• distribution of the t-statistic.

Populations and Samples

Our prior discussions about study designs, variables, measurement,

and study participants (Module 3) addressed how investigators

Module 4 Chapter 3: Univariate

Inferential Statistical Analysis | 387
gather data from observations about a sample drawn from a
population. Depending on the nature of the design and sampling
strategies applied, data about the sample will be either more or less
representative of what is going on for the whole population. To
understand the logic of hypothesis testing with statistics, it is
important to understand the distinction between what we know
about the sample and what we are guessing about the population
based on the sample. This brings us to understanding the difference
between the statistics we explored in Chapter 2 (mean, median,
mode, variance, and standard deviation) and their corresponding
“guessed” values in the population: the difference between sample
statistics and population parameters.

Consider the example from Chapter 2 about student absenteeism. We

calculated the mean number of days absent for the entire classroom
to be 6.93 days. This is the population parameter for the entire
classroom—it was feasible to measure the absent days for everyone
in the class. Now consider the possibility that an investigator wished
to use that class as a sample to represent absenteeism for the entire
school or school district without having to measure absenteeism for
everyone in the entire school or district population. In this case, the
class is no longer a population, it is a sample. Data about the sample

388 | Module 4 Chapter 3: Univariate Inferential Statistical Analysis

are used to make inferences about the population from which the
sample is drawn.

Statisticians like to use different symbols for population parameters

and sample statistics, just to keep straight which one they are
working with at any time; the symbols used in statistical computation
differ somewhat from those used in APA-style reporting of statistics,
as well. For example, sample mean is M in APA-style reporting and X̅
in statistical formulas and computations. Here are some commonly
used symbols compared:

sample population
description pronunciation pronunciation
statistic parameter
mean X̅ (or M) x-bar μ mu

standard
s (or sd) s σ sigma
deviation

variance s² s-squared σ² sigma squared

It would be useful to have a way to decide how well that sample’s

statistics estimate the larger population’s parameters. Sampling from
the population always introduces the possibility that the sample data
are not a perfect representation of the population. In other words,
investigators need to know the degree of error encountered in using
the sample statistics to estimate the population parameters, and
what degree of confidence can be placed on the population estimates
based on the sample statistics. This leads to an exploration of the
confidence interval (CI) for univariate statistics.

Confidence Interval (CI) for Univariate Statistics

Although no one can ever know the true population parameter for the
mean on a specific variable without measuring every member of the
population, a range of values can be specified within which one can
be reasonably confident the actual population parameter lies (though

Module 4 Chapter 3: Univariate Inferential Statistical Analysis | 389

not 100% certain). This range of “confidence” values can be calculated
if certain information about the sample is known, and with awareness
about certain assumptions being made. These include:

• knowing the sample mean (X̅, or M), since this will be used as an
estimate for the population mean;
• knowing the sample standard deviation (designated as s, or sd);
• deciding the level of confidence desired (commonly, 95%
confidence is used)
• an assumption that the sample was randomly drawn from a
population normally distributed on the variable of interest;
• a formula for calculating the range of values for the confidence
interval (more about computing the t-statistic below);
• information about the distribution of t-values to assess the
calculated values (more about this below).

t-Statistics. Based on probability theory (beyond the scope of this

course), statisticians developed an understanding of what they called
the t-distribution. The t-distribution is very similar to a normal curve,
but the curve differs slightly depending on the number of data points
(the sample size, or N). The t-distribution is relevant when sample
sizes are relatively small; with an infinite number of observations, or
data points (N=∞, or N=infinity), the t-distribution becomes identical
to the normal distribution. This is another way of saying that the
entire possible population is normally distributed for that variable
(one of our assumptions in the list above). Statisticians use a set
of values from a standard table to evaluate the t-statistic computed
from the formula that we will examine soon. However, to use the
standard table of t-values, the number of degrees of freedom must be
known.

Degrees of Freedom (df). As previously noted, t-values are dependent

on the number of observations in a sample (n). The goal in the present
situation is to estimate the mean for a population based on a sample’s

390 | Module 4 Chapter 3: Univariate Inferential Statistical Analysis

mean. The number of degrees of freedom involved in calculating that
estimate concerns the number of values that are free to vary within
the calculation. Starting with the sample mean as our estimate of
the population mean, once we know the sample mean and all of the
values except one, we know what that last value would be—it is the
only possible value left once the other values have been identified or
locked in. Therefore, the only degrees of freedom we have would be
represented as (N-1) degrees of freedom (abbreviated as df=N-1).

A simplified example comes from working Suduku-type puzzles. The

goal in this simplified Suduku example is to fill in the blank cells with
numbers 1, 2, or 3 so that no number appears twice in the same row
or column and each number is used once. There is only one possible
answer for the remaining empty cell that meets the requirements of
the puzzle: the number 2. When the puzzle was entirely empty, many
degrees of freedom for that cell existed—any of the allowed numbers
could be placed in that empty cell. Once the values were known for
the other cells, however, only one possibility remained, leaving no
degrees of freedom. Thus, that cell started out with the number of
degrees of freedom for the entire range of possible values (3), but
once the other 8 values were filled in, there were zero degrees of
freedom remaining—only one answer would meet all criteria.

1 3
2 3 1

3 1 2

In a different example, imagine we are trying to achieve a value of

5 by adding together 3 non-negative, non-decimal (whole) numbers;
we are going to use exactly three numbers to add up to that value of
5. Once we know any two values from 0 – 5, we automatically know
the last, third number. In other words, with three possible values to

Module 4 Chapter 3: Univariate Inferential Statistical Analysis | 391

add together, our degrees of freedom are limited to zero once those
other two values are known.

0 + 1 + 4 = 5
0 + 2 + 3 = 5
0 + 3 + 2 = 5
0 + 4 + 1 = 5
0 + 5 + 0 = 5

1 + 1 + 3 = 5
1 + 2 + 2 = 5
1 + 3 + 1 = 5
1 + 4 + 0 = 5
2 + 2 + 1 = 5

2 + 3 + 0 = 5

Starting out, before any values are filled in, for any one cell and if we
were going to add together three non-negative numbers to get 5, our
degrees of freedom would be:

df = (N-1) or df = (3-1), which is df = 2.

Table of t-distribution. Now we are ready to find the comparison

criterion t-value using the standard table mentioned earlier. Table 3-1
present an excerpt of such a table, with only the values for a two-
tailed t-distribution at 95% confidence levels. You would read this
table by going down the 95% confidence column to the row where
the number of degrees of freedom for the sample are presented on
the left. The value in that cell would be the t-value against which
the computed t-statistic would be compared to draw inferential
conclusions.

Table 3-1 presents the values for a two-tailed t-distribution at 95%

confidence, depending on the number of observations minus 1

392 | Module 4 Chapter 3: Univariate Inferential Statistical Analysis

(degrees of freedom). As you can see, the greater the sample size,
the greater the number of degrees of freedom, and the greater the
number of degrees of freedom, the closer to the population normal
distribution value of 1.960 we come. At df=1, we are very far away from
that value (12.71 is much greater than 1.960), while at df=30, we are
much closer (2.042 is close to 1.960). A full t-table would also include
values for other levels of confidence, such as the 90% and 99% levels.

Table 3-1. Values at 95% confidence, for two-tailed t-distribution, by

df.

Module 4 Chapter 3: Univariate Inferential Statistical Analysis | 393

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=430

Calculating a Confidence Interval for the Mean.

This work has led up to being able to calculate the confidence interval
for the population parameter mean from the sample statistic mean.
At this point, we simply need to plug the proper values into the
appropriate formula and compute the answer. In application, we
make two computations using the formula: once to identify the top/
highest value in the confidence interval’s range and again to identify
the bottom/lowest value in the confidence interval’s range. As a
result, we will have identified the range of values based on our sample
statistics where, with 95% confidence, the true (unknowable)
population mean lies. We will have inferred the population mean from
the sample statistics.

Consider what the formula asks us to do. To compute the 95%

confidence interval (CI95%) highest value, we start with the sample
mean (X̅), then add a computed value that takes into consideration

394 | Module 4 Chapter 3: Univariate Inferential Statistical Analysis

our sample size and the degree of variation observed in the sample
values: that computed value is calculated from our t-value (located in
the table using our degrees of freedom) multiplied with our standard
deviation (sd) which has been divided by the square root of our
sample size (√n)–see the formula written below. What that division
does is change the overall standard deviation figure into an average
degree of variation around the mean—this refers to the formula we
used in Chapter 2 to calculate variation and standard deviation in the
first place.

Finally, we compute our lowest value in the confidence interval range

by doing the same thing but subtracting the computed value instead
of adding it. Since we already know the values for everything to the
right of X̅ in the formula, we do not have to recalculate it—this time
we simply subtract it from X̅ instead of adding.

In mathematical terms, this set of steps looks like this:

Working an Example. This looks a lot more complicated than it really

is, so let’s work a simple example. Imagine that the 28 students in our
school absenteeism study were drawn randomly from the population
of students in a particular school district, and that we safely can
assume normal distribution of the days absent across the population
of students in an entire school district. From our earlier work, we
know the following information:

• X̅=6.93 days absent

• sd=10.452
• n=28

Module 4 Chapter 3: Univariate Inferential Statistical Analysis | 395

 • df=(28-1)=27
• comparison criterion t-value=2.052 (using 27 df at 95%
confidence fromTable 3-1)
• square root of 28 = 5.2915 (using a square root computation
program)

The highest value in our 95% confidence interval for the population
mean would be:

CI95% = 6.93 + [(2.052) * (10.452/√28)]

= 6.93 + [(2.052) * (10.452/5.2915)

= 6.93 + [(2.052) * (1.9752)]

= 6.93 + [4.053]

= 10.98 with rounding.

The lowest value in our 95% confidence interval for the population
mean would be:

CI95% = 6.93 – [4.053]

= 2.88 with rounding.

We would report this as: CI95%= (2.88, 10.98). This is interpreted as

the school district’s population mean days absent falling between
2.88 and 10.98 days, based on 95% confidence and the sample
statistics computed for these 28 students.

Interactive Excel Workbook Activities

396 | Module 4 Chapter 3: Univariate Inferential Statistical Analysis

 Complete the following Workbook Activity:

• SWK 3401.4-3.1 Computing Confidence Interval for the

Mean (CI 95%)

Chapter Summary

In this chapter, you learned important distinctions between

univariate sample statistics and population parameters—that we
estimate population parameters using sample statistics. You also
learned about strategies for assessing the adequacy of our estimates
and our confidence in the how well the sample statistics represent
true values for the population. The essential information for
developing confidence intervals was presented, including how
statisticians select the comparison criterion t-value and degrees of
freedom for computing and evaluation results from the statistical
formula. A more advanced statistics course could help you better
understand the nature of the t-distribution (and how this relates to
normal distribution), the rationale behind degrees of freedom, and
the assumptions we mentioned in relation to computing confidence
intervals. The next chapter examines how some of these concepts
apply to bivariate statistical analyses.

Module 4 Chapter 3: Univariate Inferential Statistical Analysis | 397

Module 4 Chapter 4:
Principles Underlying
Inferential Statistical Analysis
In addition to using inferential statistics to estimate population
parameters from sample data, inferential statistics allow investigators
to answer questions about relationships that might exist between
two or more variables of interest (bivariate analyses). Inferential
statistics help test initial hypotheses about “guessed at” relationships
between study variables. In other words, investigators infer answers
to quantitative research questions about populations from what was
observed in sample data. In prior modules, you learned about study
design options that maximize the probability of drawing accurate
conclusions about the population based on sample data. Now we
examine how to use inferential statistics to test bivariate hypotheses.
To understand our level of confidence in the conclusions drawn from
these statistical analyses, we first need to explore the role played by
probability in inferential statistics.

In this chapter, you will learn about:

• probability principles and the role of probability in inferential

statistics
• principles of hypothesis testing and the “null” hypothesis
• Type I and Type II error.

Understanding Probability Principles

You think about and calculate probability often during the course
of daily living—what are the chances of being late to class if I stop

398 | Module 4 Chapter 4: Principles

Underlying Inferential Statistical
Analysis
for coffee on the way, of getting a parking ticket if I wait another
half hour to fill the meter, of getting sick from eating my lunch
without washing my hands first? While we generally guess at these
kinds of probability in daily living, statistics offers tools for estimating
probabilities of quantitative events. This is often taught in terms of
the probability of a coin toss being heads or tails, or the probability of
randomly selecting a green M&Ms® candy from a full bag.

Let’s see what happens when we consider the probability of drawing

a blue card from a deck of 100 Uno® cards—Uno® game card decks
have equal numbers of Blue, Yellow, Red, and Green cards: 25 each
(ignoring the un-numbered action and wild cards). In drawing one
card from the deck, we have 100 possible outcomes—our draw could
be any of the 100 cards. You may intuitively see that we have a 1 in 4
probability of blue being drawn (25%).

The following formula captures the

probability of drawing a blue card
Back to Basics: Adding
(Pblue) in this scenario:
Fractions

Pevent = # of times event is possible /

# of total possible

In other words: To add fractions

together, you
Pblue = # of blue cards/# of total
simply add the
possible, or Pblue = 25/100 = ¼ = 0.25 =
numerators (top
25%
numbers) and then
use the
denominator
(bottom numbers) if
the denominators

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 399
 It is only
slightly more
complicated to
are the same:
determine the
1/3 + 1/3 = 2/3.
probability of a
But, if the blue card being
denominators are drawn later in
not the same, you the game,
will need to convert when the
them to a common shuffled deck
denominator that has been drawn down in play: for
can be kept intact. example, only 8 cards remain in the
Example: 1/3 + 2/5 draw pile: 3 blue, 2 red, 2 green, and 1
Step 1. Multiply yellow card. Now what is the
each fraction by 1 probability of drawing a blue card?
(any number Again, we apply the probability
multiplied by 1 is formula, plugging in the different
still that number). values.
The value you
select to equal 1 will Pevent = # of times event is possible /
be a fraction where # of total possible
the numerator and
Pblue = # of blue cards / # of total
denominator are
possible
the same: 5/5 and
Pblue = 3/8 = 0.375 = 37.5%
3/3 both equal 1.
This is how you use It is up to you to decide if you want to
that knowledge: risk the outcome of the game on this
1/3 x 5/5 = 5/15 chance of drawing a blue card!
and 2/5 x 3/3
=6/15 Probabilities have at least three
interesting characteristics. It is helpful
Now that the
to consider each of these three in a bit
more detail.

400 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
Probabilities exist on a
continuum. The continuum of
probabilities is from 0 to 1.0—from denominators are
absolutely no chance to a 100% the same, you can
guarantee. add the results:
5/15 + 6/15 = 11/15
(which divided is
0.7333)

You can check this

by converting your
initial fractions to
decimals and
adding them:
Probability differs from chance. This 1/3 = 0.3333 and 1/
rule is at the heart of what is called the 5 =0.4, 0.333 + 0.4
gamblers fallacy. Assuming that two =0.733
possible outcomes have equal chances
of happening (heads/tails on a coin, or
having a girl versus a boy baby), it is
possible to calculate the probability of any particular sequence
occurring. Imagine that a couple has 3 daughters and are expecting
their 4th child. They want to know whether they can expect the next
child to be a boy. The answer is: the chance of having a boy remains
the same—whether this is their first baby or their 10th baby. The
chance remains 1 in 2, or 50%.

The probability question is a bit different than they expressed,

however. They may not really be asking the probability of having
a boy, they are actually asking the probability of having a family
composition that is girl, girl, girl, boy. In that case, we can multiply
the chance of each individual event happening to get the probability
of the total combination happening:

1/2 girl x 1/2 girl x 1/2 girl x 1/2 boy

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 401
 The result is 1 x 1 x 1 x 1 =1 for the
numerator, divided by 2 x 2 x 2 x 2 = 16
Back to Basics: for the denominator. In other words,
Multiplying Fractions the probability of this combination
happening in this order of births is 1/
16, or 0.625 or 6.25%. Remember there
are many other possible combinations

To multiply that the family could have had if the

fractions, first order does not matter, including no

multiply the boys, one boy, two boys, three boys, or

numerators (top four boys among the girls, and each

numbers), then possible combination having the same

multiply the probability of occurring in specified

denominators order (1 in 16):

(bottom numbers).
The gambler’s fallacy confuses the
You may need to
chances of the next child being a boy (1
resolve the
in 2, or 50%) with the probability of a
resulting fraction to
whole sequence/pattern occurring
a simpler number
(girl, girl, girl, boy being 1 in 16, or
or do the fraction
6.25%).
division.

Example. Multiply
1/3 x 2/5
1 x 2 = 2
(numerators) and
3 x 5 = 15
(denominators), so
1/3 x 2/5 = 2/15
(which divided =
0.1333)

402 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
Remember: this only applies if the chance of the event happening
is random—if there is a biological reason why boys are more or less
likely than girls, the entire formula for probability will need to be
revised. It is called the gambler’s fallacy because people mistakenly
bet on certain outcomes, confusing chance with probability. A real-
life example presented in an online article about the gambler’s fallacy
(www.thecalculatorite.com/articles/finance/the-gamblers-
fallacy.php) described a spectacular event that occurred in August
of 1913 at a Monte Carlo casino: the roulette wheel turned up the
color black 29 times in a row, the combined probability of which
was 1 in 136,823,184! The article’s authors stated, “After the wheel
came up black the tenth time, patrons began placing ever larger bets

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 403
on red, on the false logic that black could not possibly come up
again.” Obviously, it did, 19 more times. The gamblers’ logic was false
because the chances of black coming up was always 18 out of 37
times on this type of roulette wheel (48.6%). Therefore, each of those
29 times had an 18/37 chance of being black, regardless of what
happened in prior spins. Do you see how this chance of an event (18/
37) differed from the probability of the whole sequence occurring (1 in
136,823,184)?

Inverse probability. When there exist only mutually exclusive events,

the probability of a specific event happening and the probability of
that specific event NOT happening add up to 1. Thus, the probability
that it will happen is the mathematical inverse of the probability that
it will not happen. In other words, subtracting the probability that it
will happen from 1.0 gives you the probability that it will not happen;
and, vice versa, subtracting the probability that it will not happen
from 1.0 gives you the probability that it will happen.

As an example, let’s go back to our probability of drawing a blue

card from the whittled down play deck with only 8 cards left. We
computed the probability of a blue card (Pblue) as being 0.375.
Therefore, we can apply knowledge of inverse probability if we want
to know the probability of drawing something other than a blue card
(Pnotblue is the combined probability of drawing red, green, or yellow):

404 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
Pnotblue = 1.0 – Pblue

Pnotblue = 1.0 – 0.375 = 0.625

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=431

Probability Principles Applied in Statistics

Probability is about the likelihood of an outcome. Remember that

with inferential statistics we can never know with 100% certainty
that our sample data led us to the correct conclusion about the
population. This is where probability comes into play in the process
of engaging in inferential statistical analyses. The following scenario
helps make the case:

• Imagine you draw a random sample from the population for

your study, accurately measure the variable of interest, and
compute your sample statistics (for example, the mean and
standard deviation for an interval variable).
• Now, imagine that you “put back” those sampled participants
and draw a new sample. With random selection, any of the
previous participants and all the other possible participants in

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 405
 the population still have an equal chance of being drawn (we
learned about this in Module 3, chapter 6 about random
sampling).
• You repeat your measurement process and compute your new
sample statistics. Because there are different participants this
time, and there exists variability between individual
participants, your sampling statistics for the different sampling
groups will most likely be somewhat different—maybe a little
different, maybe a lot different.

Which computed sample value is more

“right” in describing the full
Back to Basics: Normal population? The answer is: neither or
Distribution both, and there is no way of knowing
the population’s true value without
measuring the entire population (we
learned about this in Module 4, Chapter

Remember that one 3). So, imagine your next step is to keep

defining drawing samples, measuring, and

characteristic of a computing the sample statistics for

normal curve is that each sample drawn. Eventually, for a

95% of the values variable that is distributed normally in

fall within 2 the population-as-a-whole, the sample

standard deviations statistic values that you computed over

of the mean (above hundreds of trials (samples) will be

or below). This is approximately normally

the probability of a distributed—you learned about normal

value falling in that distribution in Chapter 4 of this

range (95%). This module.

leaves a 5%
For example, imagine that you were
probability of the
computing the mean on a variable
“number of hours each week spent
working on issues related to someone’s

406 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
alcohol misuse” for each sample of
social workers drawn 1,000
times—some social workers spent
values falling
virtually none of their time in this type
outside of that
of activity while other spent almost all
range—4% at 2-3
their time this way. Your first sample’s
standard deviations
mean number of hours was 15, your
away from the
second sample’s mean was 12, your
mean (99% –
third sample’s mean was 22, and so on
95%=4%), and 1% at
until you have 1,000 mean values for
more than 3
your drawn samples. If this variable is
standard deviations
normally distributed across the
away (100% – 99%
population of all social workers and you
= 1%). Furthermore,
drew each of your samples randomly
5% is the inverse of
from the population of social workers,
95%, since
those mean values would come close to
1-.05=.95 and
looking like a normal “bell” curve. At
1-.95=.05.
this point:

• You could compute the mean-of-

means for the whole set of 1,000 samples. This sample mean-of-
means would be a reasonable estimate for the population mean
(population parameter).
• Knowing the mean-of-means value, you could compute the
standard deviation for the mean number of hours across your
1,000 samples. The sum of squared distances for each sample’s
mean from the mean-of-means divided by 999 would be the
variance, and the square root of the variance is the standard
deviation (you learned about variance and standard deviation in
Chapter 2 of this module).

Figure 4-1 is the same normal curve introduced earlier, this time with
the notation that the x-axis (horizontal) represents the computed

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 407
values for the 1,000 sample statistics, and the y-axis (vertical)
represents the frequency each computed value for the sample
statistic was the result. Note that our unit of analysis is no longer
each individual social worker’s hours—that was the unit of analysis in
each of our 1,000 samples. Now the unit of analysis is the mean for
each of those 1,000 samples.

Figure 4-1. Normal distribution around the mean

In reality, no investigator would set out to repeat a research study

this many times. However, the investigator who completed the study
the first time wants to know: is my sample statistic falling within
the 95% range in relation to the true population mean, or does it lie
in one of the two tails: the 2.5% below or 2.5% above two standard
deviations away, adding up to the remaining 5%? This should sound
familiar: it represents the 95% confidence interval (CI) for the sample
statistic!

Using what we learned about the inverse probability relationship,

selecting a 95% confidence level means accepting a 5% error rate
(100% – 95% = 5%): or, in decimal terms, 1.0 – .95 = .05. In statistics,

408 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
this is often referred to as applying a decision rule of “alpha” equals
.05 (5%)—symbolized as α=.05. By definition, there exists an inverse
relationship between the probability value (p-value) and the value of
α. In other words, both of the following statements are true:

α = (1– p)

and

p = (1- α).

The social work, social science, and behavioral research convention

is to adopt a 95% probability figure as its decision rule concerning
conclusions drawn from sample data. In other words, we generally
accept a 5% probability of being wrong in our conclusions. This is
written as: α=.05.

Research Questions and the Null Hypothesis

Statistical decision rules apply to hypotheses being tested

statistically, hypotheses drawn from the research questions (you
learned about this question/hypothesis relationship in Module 2,
Chapter 1). For example, an investigator wanted to know if readiness
to change scores differ for men entering batter treatment programs
voluntarily versus through court order. The investigator’s research
question might be:

Is there a relationship between “program entry status” and

“readiness to change” scores?

Research questions often need to be rephrased in terms of

statistically testable hypotheses. For example, this investigator’s
research question needs to be rephrased as a statistical hypothesis:

A statistically significant relationship exists between the

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 409
 variables “program entry status” and “readiness to change”
scores.

The investigator would know if the hypothesis is true is by

determining whether the difference in readiness to change scores
for the two sample groups is meaningfully different from 0—zero
difference indicating that entry status is not meaningfully related to
readiness to change scores.

The purpose of statistical analysis is to draw conclusions about

populations based on data from a sample, making it very important to
carefully phrase the hypothesis to be tested statistically. Two ways of
phrasing research hypotheses are relevant here: they are the inverse
of each other. One is called the null hypothesis, the other is called the
alternate hypothesis. Let’s start with the alternative hypothesis first
because it is more intuitively easy to understand.

Alternate Hypothesis. An investigator develops hypotheses about

possible relationships between variables. In the Safe at Home
example, investigators hypothesized that a difference in readiness
to change scores exists between men entering batterer treatment
programs voluntarily compared to men entering the programs under
court order. This research hypothesis needs to be re-worded into
statistical terms before applying statistical analyses to evaluate its
accuracy. If there exists a meaningful difference between the two
groups, the mean readiness score for the voluntary entry group
would be different from the mean for the court-ordered group—in
other words, subtracting one mean from the other would give a result
that is not zero. The alternative hypothesis, symbolized as Ha,would
be phrased as:

Ha: The difference in mean readiness scores for the population

of men entering batterer treatment voluntarily compared to

410 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
 men entering under court order is significantly different from
zero.

You might be wondering why this re-phrasing of the research

hypothesis is called the alternate hypothesis. It is the alternate (or
inverse) to the null hypothesis, which is the one actually tested
statistically. At first, this may seem like confused logic—why test the
inverse of what we really want to know? The next sections explain
this logic.

Null Hypothesis. A statistically tested hypothesis is called the null

hypothesis. The null hypothesis (typically symbolized as H0) is stated
in terms of there being no statistically significant relationship
between two variables among the population—in the Safe At Home
example, program entry status and readiness to change scores. While
the alternate hypothesis (Ha) was expressed in terms of the
difference in mean readiness scores for the two groups in the
population being different from zero, the null hypothesis specifies
that the difference between readiness scores for the two groups is
zero. The null hypothesis is the inverse of the alternate hypothesis
and would be presented like this:

H0: The difference in mean readiness scores for the population

of men entering batterer treatment voluntarily compared to
men entering under court order is 0.

Then, if a statistically significant difference in the sample means

for the two groups is observed, the investigator rejects the null
hypothesis. In other words, based on sample data, the investigator
concludes support for the alternative hypothesis as it relates to the
population.

When an observed difference is found to be non-statistically

significant, the investigator fails to reject the null hypothesis. In other
words, the true difference in the population could include zero.

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 411
Failing to reject the null hypothesis, however, is not the same as
determining that the alternative hypothesis was wrong! Why not?
Because the investigator can only conclude failure to detect a
significant difference—it is possible that no difference truly exists, but
it also possible that a difference exists despite not being detected for
some reason. This point is so important that it warrants repeating:

Failure to reject the null hypothesis does not mean that the
alternate hypothesis was wrong!

Remember that inferential statistics are about what (unknown)

differences might exist in the population based on what is known
about the sample. Just because a difference was not observed this
time, with this sample, does not mean that no difference really exists
in the population. This conversation between a 6-year old and her
father demonstrates the point:

Dad: Why are you afraid to go to bed in your own room?

Daughter: I am afraid the zombies will come and get me.

Dad: You don’t need to be afraid. There is no such thing as

zombies.

Daughter: How do you know?

Dad: Because no one has ever seen a real zombie.

Daughter: That doesn’t mean there aren’t any—maybe no one

looked in the right places!

The 6-year old is logically more correct than her father—just because
something was not observed does not mean that it does not exist.
Failing to reject the null hypothesis means that the investigator
recognizes that chance could be responsible for the outcome
observed in the sample. Rejecting the null hypothesis means that the

412 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
investigator is confident that a difference observed in the sample
accurately reflects a true difference in the population, a difference
not due to chance.

No Double-Barreled Hypotheses. Only one statistical hypothesis can

be tested at a time. You may recall from our discussion about
measurement and survey/interview questions (Module 3, chapter 4),
the concept of a “double-barreled” question. The same applies in
terms of testing the null hypothesis. If there were two parts to the
hypothesis, it would be impossible to interpret the outcome: either or
both could be rejected, and either or both could fail to be rejected. An
investigator would not know whether to draw conclusions in terms
of one or both parts of the hypothesis. Therefore, each needs to be
phrased as a separate, unique null hypothesis.

For example, the following is not an analyzable hypothesis:

H0: There is no statistically significant difference in either the

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 413
 readiness scores or number of weeks since the intimate partner
violence incident for clients who enter the program voluntarily or
by court order.

Instead, this should be presented as:

First H0: There is no statistically significant difference in

readiness scores for clients who enter the program voluntarily or
by court order.

Second H0: There is no statistically significant difference in

number of weeks since the intimate partner violence incident for
clients who enter the program voluntarily or by court order.

Take a moment to complete the

following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=431

Statistical Significance in Hypothesis Testing

Notice that these hypotheses mention a “statistically significant”

difference. What if the difference is not EXACTLY equal to 0, if there

414 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
exists a tiny difference? Does that mean there is a meaningful
relationship between the variables? Investigators want to know if an
observed difference in the sample is MEANINGFUL before
concluding that such a relationship really exists in the population.
This is where we consider the meaning of statistically significant
results—the observed difference in the sample is sufficiently
different from zero to have statistical significant meaning (leading to
the conclusion about rejecting or failing to reject the null hypothesis
about the population). Statisticians work with probability logic in
their statistical analyses of sample data.
Thinking again about the Safe At Home data, the investigator knows
the categorical status of each participant (n=75 voluntary, n=328
non-voluntary) and each participant’s overall readiness to change
score (higher scores indicate greater readiness to change one’s
behavior). Before we get to the actual statistics used to answer this
question, we need to think about the question in terms of probability
for drawing the right conclusion about the population of men
entering treatment based on the sample statistics. The investigators
in our Safe At Home example need a decision rule to help decide
whether an observed difference for the two groups of clients
(voluntary and court-ordered) is meaningful. This leads to examining
the principles of hypothesis testing—remember we are statistically
testing the null hypothesis: if a statistically significant result for the
difference in scores for the two groups is observed, the null
hypothesis of no difference (difference = 0) will be rejected.
Testing the Null Hypothesis. As previously noted, the decision rule
convention in social work, social science, and behavioral research is
to use α=.05 as the criterion for what is a significant result. This
means that:

• if the computed significance level for a test statistic (the

probability or p-value) is .05 or less, the decision is to reject the
null hypothesis;

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 415
 • if the computed significance level for a test statistic (probability
or p-value) is greater than .05, the decision is failure to reject
the null hypothesis.

Here is why. Inferential statistics draws on probability principles for

help in interpreting data from samples. In any one single study,
chance plays a role in the adequacy of the data for representing the
population. For example, if investigators observe a difference in
mean readiness to change scores between the sampled clients who
entered batterer treatment voluntarily and sampled clients who
were court-ordered into batterer treatment, they want to know if
that observed difference happened just by chance or if it is
reflecting a true difference in the population. By chance, with a
normal distribution of readiness scores, 95% of the time an observed
difference would fall within 2 standard deviations of the true mean.
Therefore, any observed differences within that 2-standard
deviation range could be the result of chance—so, investigators
cannot safely conclude that a meaningful difference exists (i.e., they
fail to reject the null hypothesis). However, if the observed
difference is far enough out in the tails of the distribution (2.5%
above or 2.5% below two standard deviations), investigators are
comfortable with rejecting the null hypothesis of no difference and
concluding that the observed difference reflects the population’s
reality.

416 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
This normal curve demonstrates our use of the α=.05 decision
rule—we are willing to be wrong 5% (.05) of the time by concluding
that a meaningful difference exists when none really does. In other
words, an observed difference far enough out in one of these tails is
considered statistically significant. So, we state that, if the
probability that the observed difference happened by chance is less
than 5% (α=.05), we are willing to reject the null hypothesis of no
difference. The statistical approaches presented in Chapter 5 show
how different test statistics are computed and how the probability is
determined for these decisions. There is nothing magical about .05,
it is simply a common convention in social work, social, and
behavioral science research: sometimes investigators apply more
stringent criteria (α=.01) or less stringent criteria (α=.10). When more
stringent α=.01 decision criteria are applied, being 99% certain, it is
harder to reject the null hypothesis: the observed values would have
to be even further out in the tails of our distribution, out past three
standard deviations. When less stringent α=.10 decision criteria are
applied, being 90% certain, it is easier to reject the null hypothesis:
the observed values do not have to be quite as far out in the tails of
our distribution.
Turning back to our Safe At Home example, let’s see what the
investigators learned. The null hypothesis being tested was:

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 417
 H0: the difference in mean readiness scores between men
entering treatment voluntarily compared to men entering under
court order is equal to 0.
The investigator’s sample statistics (using SPSS statistical analysis
software) showed the following:

This table shows the mean value on readiness to change scores for
the 328 men court-ordered to treatment was 4.56 and the mean for
the 75 who entered the program voluntarily was 4.90 (with rounding).
Clearly the difference between these two groups is not exactly zero:
there exists a 0.34-point difference between the two group means
(4.90 – 4.56 = 0.34). But is this 0.34-point difference observed in
the sample sufficiently far from zero for investigators to reject the
null hypothesis of there being no difference and to conclude that
a population difference exists? This is another way of asking if the
difference from zero is statistically significant.

The investigators in this case could use a t-test analysis approach

for comparing the two group means (more about this later). The
observed 0.34 difference in the two group means was significant at
the p=0.029 level (probability=.03 with rounding) in this analysis. How
does this translate into a conclusion about the results?

• The investigators adopted the tradition of using α=.05 criterion

as a decision rule—accepting a 5% chance of drawing the wrong
conclusion, being 95% confident of their conclusion.
• The computed statistical significance for the difference between
the groups was 0.03 (rounding the p=0.029 level).
• The computed significance level (α=.03) is less than the accepted
α=.05 criterion value, so the decision is to reject the null

418 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
 hypothesis of no difference (concluding that a difference exists).
• Additionally, the confidence interval for the sample t-statistic
did not include the value 0, since both the lower and upper
range values were negative numbers/less than zero: CI95%= (-
.641, -.034). Thus, the investigators are 95% certain that the real
population difference does not include zero as a possibility
and rejected the null hypothesis of no difference: the conclusion
is that a statistically significant association between program
entry status (voluntary versus court order) and readiness to
change scores exists in the population of men entering batterer
treatment programs.
• Looking at the observed sample means shows that individuals
entering the programs voluntarily had higher readiness to
change scores than did individuals court-ordered to enter the
programs (M=4.90 and 4.56 respectively). This could be reported
as the voluntary entry group has significantly higher mean
readiness to change scores than the group entering treatment
through court order.

One-tailed and Two-tailed Hypotheses. Up until this point, we have

worked with testing hypotheses about a “difference.” The difference
could be in either direction, either group being more or less than the
other. For example, an alternative hypothesis (Ha) could be:

Ha: A significant difference exists in the readiness to change

scores between men voluntarily entering batterer treatment and
men court-ordered to treatment.

In this case, the null hypothesis (H0) being statistically tested would
be:

H0: No significant difference exists in the readiness to change

scores between men voluntarily entering batterer treatment and
men court-ordered to treatment (the difference is zero.)

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 419
This null hypothesis could be rejected in either of two scenarios:
either the voluntary entry group has higher readiness to change
scores OR the voluntary entry group has lower readiness to change
scores compared to the court-ordered group. In statistical terms,
this means that we would be working with the probability of the
result being either in the 2.5% tail above the mean or in the 2.5% tail
below the mean (if we apply the α=.05 decision rule). This is what is
meant by a two-tailed test—the null hypothesis would be rejected by
a result landing in either of the two tails (values falling in either of
the red areas of the curve below); values landing anywhere else would
result in failure to reject the null hypothesis.

However, there exists another possible alternative hypothesis (Ha),

depending on what the investigators think about the research
question:

Ha: the voluntary entry group’s readiness scores will be

significantly greater than the court-ordered group’s scores.

This changes the null hypothesis (H0) to the following:

H0: the voluntary entry group’s readiness scores will not be

significantly greater than the court-ordered group’s scores.

420 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
Now there is only one tail that can lead to rejecting the null
hypothesis: the resulting statistic would have to be in the 2.5% at the
top tail of the distribution to reject. If it is in the 95% middle or in the
bottom 2.5%, the investigator would fail to reject this null hypothesis.
Since there is only one way to reject this time, not two, this is called a
one-tail test (rejecting the null hypothesis for values falling in the red
area of the curve, failing to reject the null hypothesis for values falling
anywhere else).

In summary, if the hypothesis is directional, the decision rule needs

to be based on one-tailed test criteria; if the hypothesis is bi-
directional (difference could be either way), the decision rule should
be based on two-tailed test criteria.

Type I and Type II Error

While the investigators in our Safe At Home example would be happy

to share their results and conclusions with the world, it is important
to understand the possibility that their conclusions could have been
drawn in error. Remember: being 95% confident means accepting a

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 421
5% probability of being wrong (α=.05 criterion). In any situation, 4
possible combinations of decisions and realities exist: two represent
accurate decisions and two represent mistaken decisions. The
mistake possibilities include a Type I error and a Type II error (see
Figure 4-2 adapted from Begun, Berger, & Otto-Salaj, 2018, p. 13)
depending on whether the investigator rejects or fails to reject the
null hypothesis.

Figure 4-2. Type I and Type II errors for the null hypothesis (H0=no
difference)

Type I Error. The risk of making a Type I error concerns the risk of
rejecting the null hypothesis of no difference when in reality there
exists no significant difference in the population. For example, our
Safe At Home investigator rejected the null hypothesis of no
difference in readiness to change scores between the two groups. A
Type I error would occur if, in fact, there is no difference in readiness
to change within the population of men entering batterer treatment
programs.

Sometimes, investigators decide to apply a looser criterion: α=.10

means a 10% chance of being wrong, or 90% confidence. This would

422 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
increase the probability of making a Type I error since it is “easier”
to reject at this level. There are few instances in social work, social
science, and behavioral research where 99% certainty is preferable
to 95% certainty (α=.01 rather than α=.05). The reason being that we
also want to reduce the probability of a Type II error.

Type II Error. The risk of making a Type II error concerns the risk
of failing to reject the null hypothesis when, in reality, a significant
difference in the population does exist. For example, if an investigator
uses the α=.05 decision criterion and computes a significance value
p=.06 from the sample data, the null hypothesis would not be
rejected. In reality, however, a difference may truly exist in the
population. This is problematic with intervention research conducted
with too few study participants; with small samples it is harder to
reject the null hypothesis. A very promising intervention may get
“scrapped” as a result of a Type II error. Thus, study design is an
important strategy for reducing the probability of making a Type II
error.

Trochim (2005, p. 204) explains Type I and Type II error this way:

Type I Error: Finding a Relationship When There Is Not One (or

Seeing Things That Aren’t There)

Type II Error: Finding No Relationship When There Is One (or

Missing the Needle in the Haystack).

Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis | 423
 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=431

Chapter Summary

This chapter laid the groundwork for understanding inferential

statistical approaches presented in Chapter 5. In this chapter you
learned basic principles of probability and how these probability
principles play a role in the logic and assumptions underlying
inferential statistics. As a result, you should now have a basic
understanding of the decision rules that are commonly applied in
social work, social science, and behavioral statistics—the role of p and
α, for example. You learned about the logic behind the null hypothesis
(H0) and how it relates to a research question or “alternative”
hypothesis (Ha),as well as what it means to reject or fail to reject the
null hypothesis. You also learned about the distinction and meaning
of Type I and Type II errors. Now you are prepared with a base for
understanding specific statistical analyses.

424 | Module 4 Chapter 4: Principles Underlying Inferential Statistical

Analysis
Module 4 Chapter 5:
Introduction to 5 Statistical
Analysis Approaches
Prior modules and chapters laid the groundwork for understanding
the application of inferential statistics approaches commonly used to
evaluate research hypotheses based on quantitative data. Inferential
statistics may answer either univariate or bivariate hypotheses.

In this chapter, you will learn about:

• Specific statistical approaches for testing hypotheses

(parametric tests of significance)

▪ one-samplet-test
▪ independent samples t-test

◦ one-way analysis of variance (Anova) test

◦ chi-square test of independence
◦ correlation test
• Approaches to consider when data are not normally distributed
(non-parametric options)

What These Tests Have in Common

The five statistical analysis approaches presented here have several

things in common. First, they each are based on a set of assumptions
and/or have specific requirements related to the type of data
analyzed. Second, with the exception of the chi-square test of
independence, they rely on assumptions related to normal
distribution in the population on at least one variable. Third, they
all assume that the sample was randomly (properly) drawn from the

Module 4 Chapter 5: Introduction to 5

Statistical Analysis Approaches | 425
population, and therefore are representative of the population. These
last two factors characterize these are parametric tests (non-
parametric tests are briefly discussed at the end of this chapter). For
each of the statistical approaches, the task involves computing a test
statistic based on the sample data, then comparing that computed
statistic to a criterion value based on the appropriate distribution
(t-distribution, F-distribution, χ2 distribution, or correlations). This
comparison leads the investigator to either reject or fail to reject the
null hypothesis.

One Sample t-Test

One of the simplest statistical tests to begin with allows investigators

to draw conclusions based on a single group’s data about whether
the population mean for the group is different from some known or
expected standard mean value. For example, the standard IQ test has
a mean value of 100—the test is designed so that the population is
approximately normally distributed for IQ around this mean. Imagine
that an investigator has concerns that a specific group in the
population—perhaps children living in rental households—has a mean
IQ significantly different from that standard mean of 100. In theory,
this could happen from children being exposed to environmental
toxins, such as lead in drinking/cooking water, and ingesting
particles contaminated by lead-based paint or playground soil
contaminated from heavy traffic where lead was once in the gasoline
of cars driving in the area.

426 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

Test Purpose: Assess the probability that the population subgroup’s
mean IQ differs from the standard population mean of 100.

Test H0: The group’s mean IQ is not significantly different from 100.

Type of Data: IQ scores are interval data.

Test Assumptions and Other Test Requirements: The univariate

one-sample t-test is based on the following assumptions and
requirements.

• Type of Variable: The scale of measurement is either ordinal or

continuous (interval).
• Normal Distribution: The dependent variable is normally
distributed in the population, or a sufficiently large sample size
was drawn to allow approximation of the normal distribution.

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 427

 • Independent Observations: Individuals in the sample are
independent of each other—random selection indicates that the
chances of one being sampled are independent of the chances
for any other being sampled.

The test statistic: The t-test analysis is based on the t-distribution.

The t-statistic is computed by first subtracting from the sample mean
(X̅) the comparison value called “delta” (delta is indicated by the
symbol Δ, which in our example has Δ = 100), then dividing this
resulting value by the value computed as the sample’s standard
deviation (s) multiplied by a fraction with 1 divided by the square root
of the sample size. The formula looks like this:

t(df)= (X̅ – Δ)/[(s) * 1/√n)]

Back to Basics: Dividng It should remind you somewhat of the

by a Fraction formula we saw back in Chapter 3 for
the 95% confidence interval using the
t-value—we are using the same
components of sample mean (X̅), t-
Dividing by a value, standard deviation (s), and the
fraction is the same square root of the sample size (√n):
as multiplying by its
inverse. For CI95% = X̅ + [(tfor 95% confidence) *

example, dividing (s/√n)] (computes highest value)

by ½ is the same as
CI95% = X̅ – [(tfor 95% confidence) *
multiplying by 2.
(s/√n)] (computes lowest value
In our formula,
Completing the example:In our
(sd)/√n is
hypothetical example, imagine the
following: our sample mean (X̅)for IQ
was 92, the standard deviation (s) was

428 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

21.00, and the sample size was 49
children. This means that our degrees
of freedom (df) is 49 -1, or 48 (because equivalent to (sd) *
df=n-1). We can consult the 1/√n.
t-distribution table (from Chapter 3) to
determine our decision value for
t-values based on 48 degrees of
freedom and α=.05 (95% confidence). Since our table does not show a
value for 48 degrees of freedom, we estimate the criterion value as
2.01 for purposes of our example.

Plugging these values into the formula, we obtain our t-value:

t(48)= (92 -100)/ (21.00/√49)

t(48)= (-8)/ (21.00/7)

t(48)= (-8)/3

t(48)= (-2.67)

Now, we take that t(48)= -2.67 value and compare it to the criterion
value from the t-distribution table (2.01). Since 2.67 (the “absolute
value” of -2.67) is more extreme than our criterion value of 2.01,
we reject the null hypothesis of no difference. In other words, we
conclude that a significant difference exists in the mean IQ score for
this subpopulation compared to the expected standard mean of 100.

On the other hand, our statistical Back to Basics: Absolute

computations using the computer Value

might have generated a p-value for this

test. Instead of using the t-distribution
table at 95% confidence to determine
a criterion value for comparing the

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 429

 t-statistic we calculated, the p-value
computation will lead to the same
conclusion. In this example, imagine
In many instances,
that the computer informed us that the
we are concerned
probability of a chance difference
with how far from
(p-value) was .024: p=.024 in the
zero a value lies.
output. Using the α=.05 decision
Thus, we work with
criterion (95% confidence), we see that
figures
.024 is less than .05. Thus, we reject the
representing the
null hypothesis of no difference, just as
absolute value. For
we did using the t-statistic compared
example, -4 and +4
to the criterion value.
are each 4 points
away from 0. Thus, Note: we used a two-tailed test in this
the absolute value example. If the investigator
of -4 is 4, as the hypothesized that the subgroup has a
absolute value of +4 lower IQ than the mean, one-tailed
also 4. test criteria would need to be applied.
With comparing a
test statistic to its
criterion value, we Interactive Excel Workbook Activities
generally work with
the statistic’s
absolute value.
Complete the following Workbook
Activity:

• SWK 3401.4-5.1 One-Sample t-Test Exercise

Comparing Means for Two Groups: Independent Samples t-Test

The next level of complexity involves comparing the mean for two

430 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

independent groups, rather than what we just saw in comparing a
single group to a standard or expected value. This moves us into
a bivariate statistical analysis approach: one independent and one
dependent variable. The independent variable is the “grouping”
variable, defining the two groups to be compared. The dependent
variable is the “test” variable, the one on which the two groups are
being compared.

For example, we have been working throughout the module with the
question of whether there is a meaningful difference in readiness
to change scores for clients entering batterer treatment voluntarily
versus through court order. In this case, we have one interval variable
(readiness to change score) and one categorical variable with 2
possible categories—a dichotomous variable (entry status).

Test Purpose: Assess the probability that the readiness to change

score in the population of clients entering batterer treatment
voluntarily differs from the change scores for the population of
clients entering batterer treatment through a court order.

Test H0: The mean readiness to change scores for the two groups are
not significantly different (the difference equals zero).

Type of Data: Independent “grouping” variable (entry status) is both

categorical and dichotomous; dependent variable (readiness to
change scores) is interval data.

Test Assumptions and Other Test Requirements: The independent

samples t-test is based on the following assumptions and
requirements.

• Type of Variable: The scale of measurement for the dependent

variable is continuous (interval).
• Normal Distribution: The dependent variable is normally
distributed in the population, or a sufficiently large sample size

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 431

 was drawn to allow approximation of the normal distribution.
Note: the “rule of thumb” is that neither group should be smaller
than 6, and ideally has more.
• Independent Observations: Individuals within each sample and
between the two groups are independent of each other—random
selection indicates that the chances of any one “unit” being
sampled are independent of the chances for any other being
sampled.
• Homogeneity of Variance: Variance is the same for the two
groups, as indicated by equal standard deviations in the two
samples.

The test statistic: The independent samples t-test analysis is based

on the t-distribution. The t-statistic is computed by first computing
the difference between the two groups’ means: subtracting one mean
(X̅courtorder)from the other (X̅voluntary). This computed difference is
then divided by a sample standard deviation value divided by the
square root of the sample size, just as we did in the one mean test.
However, this is more complicated when we have two groups: we
need a pooled estimate of the variance (a pooled standard deviation)
since neither standard deviation alone will suffice. Furthermore, we
need a pooled sample size since both groups 1 and 2 contributed to
the variance estimate. Our earlier formula looked like this:

t(df)= (X̅ – Δ)/[(s) * 1/√n)]

Our new formula looks like this:

t(df)= (X̅1 – X̅2)/[(spooled) * 1/√n1+ 1/√n2).

Before we can start plugging in values, we need to know how to

find the pooled variance estimate—the pooled standard deviation.
The pooled standard deviation is computed as a sort of weighted
estimate, with each group contributing the proportion of variance
that is appropriate for its sample size—the smaller group contributing

432 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

less to the estimate than the larger group. Thus, s1 is the standard
deviation for group 1, and s2 is the standard deviation for group
2. Making a series of mathematical adjustments for the equivalent
aspects of multiplying and dividing fractions, we end up with a
formula for the pooled standard deviation (sp) that works like this:

• take the square root of the entire computed value where

• the first group’s degrees of freedom (sample size minus one, or
n1-1) is multiplied by its standard deviation (s12) and
• this is added to the second group’s degrees of freedom
(n2-1)multiplied by its standard deviation (s22)
• and that addition becomes the numerator of a fraction where
the two sample sizes are added, and the 2 extra degrees of
freedom are subtracted (n1+ n2-2)

In terms of a mathematical formula, it looks like this:

sp= √[(n1-1)*s12+ (n2-1)*s22/ (n1 + n2 – 2)]

Then, we take this pooled standard deviation and plug it into the
t-test formula above. The degrees of freedom for this t-statistic will
be the (n1+ n2– 2) that was in the denominator of our pooled variance.

Completing the example: The first step is to compute the pooled

standard deviation value, then plug it into the formula for computing
the t-statistic. In lieu of calculating this by hand, we can let Excel or
another statistical program perform the calculation. The resulting t-
statistic for comparing the mean readiness to change scores for our
two groups, where the degrees of freedom were 401 (df=401):

t(401)= 2.189

The criterion value for 95% confidence was 1.960 in the t-distribution
table. Therefore, the decision would be to reject the null hypothesis
of equal means. The mean for the voluntary group was greater than

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 433

the mean for the court order group: X̅=4.90 and X̅=4.56 respectively,
a difference in the means of .34 (X̅1– X̅2= 4.90 – 4.56 = .34).

Two other ways of drawing the same conclusion (reject the null
hypothesis) are offered by the computer output. The analysis
indicated a 95% confidence interval of the difference in the two
group means (.34) does not include zero as a possible value since the
range includes only positive values: CI95% = (.64, .03). In addition, the
two-tailed significance level computed for the test statistic (t-value)
was p=.029. Using the criterion of α=.05, seeing that .029 is less than
.05, the same conclusion is drawn: reject the null hypothesis of no
difference between the groups. All three of these approaches came
to the same conclusion because they are simply different ways of
looking at the very same thing.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.4-5.2 Independent Samples t-Test Exercise

Comparing Means: One-Way Analysis of Variance (Anova) Test

The one-way analysis of variance (Anova) test is another bivariate

statistical analysis approach. Like the independent samples t-test, it
is used to compare the means on an interval (continuous) variable
for independent groups. Unlike the independent samples t-test, it is
not restricted to comparing only 2 groups. For example, the one-
way analysis of variance would allow an investigator to compare

434 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

the means between multiple racial/ethnic groups, or between 3
treatment conditions (e.g., low, medium, high dose intervention
groups, or treatment innovation, treatment as usual, no treatment
control groups). In this case, we still have a bivariate statistical
analysis approach: one independent and one dependent variable. The
independent variable is still the “grouping” variable, defining two or
more groups to be compared. The dependent variable is the “test”
variable, the one on which the groups are being compared.

Two points are worthy of noting here. First, because they are
mathematically related, the one-way Anova test, when conducted
on a dichotomous (two-group) categorical variable yields a result
identical to the independent samples t-test. The F-statistic is the
square of the t-statistic in this 2-group case, and the F-distribution
(as well as the criterion value) are squared, as well. Thus, the
conclusion drawn and the p-value are identical.

Second, in a one-way Anova involving three or more groups, the

result indicates whether a difference in mean values exists—it does
not indicate where the difference lies. For example, it could be that
one group differs from all the others, that all of the groups differ from
each other, or some combination of differences and no differences.
A significant result leads to further “post hoc” analyses to determine
which pairs differ significantly.

Extending our example with readiness to change scores for men

entering batterer treatment, consider the possibility that
investigators wanted to explore the relationship between readiness
to change scores and substance use at the time of the referring
incident of intimate partner violence. The “grouping” independent
variable was coded for who used substances at that time: (0)
substances were not involved, (1) the client only, (2) the partner only,
and (3) both client and partner. Thus, the analysis involves 4 groups

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 435

being compared on the dependent variable (readiness to change
battering behavior).

Data were available on both the dependent and independent

variables for 391 clients. Here is what the mean and standard
deviations looked like, as well as the 95% confidence interval for each
group mean (see Table 5-1).

Table 5-1. Results of descriptive analyses for group means.

You can see the number of individuals in each group under the
column labeled “N” and the mean readiness to change score for
each group in the next column (labeled “Mean”). The group with the
highest mean score was the group where both partners were using
alcohol or other substances at the time of the incident, followed by
the group where the client was the only one using at the time. The
lowest readiness to change scores were among men whose partner
was the only one using alcohol of other substances at the time of the
incident, lower than the group where no substances were used by
either partner. The 95% confidence intervals for these mean values
all excluded zero as a possibility, since lower and upper bounds of the
confidence interval (range) were positive values.

Test Purpose: Assess the probability that the readiness to change

scores differ for the population of clients entering batterer treatment
following an incident of intimate partner violence where no

436 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

substances, client only substance use, partner only substance use, or
both using substances was involved.

Test H0: The mean readiness to change scores for the four groups
are equal (the differences equal zero). As previously noted, this multi-
group test will not tell us which groups differ from the others, only
whether there exists a difference to further analyze.

Type of Data: Independent “grouping” variable (substance use status)

is categorical; dependent variable (readiness to change scores) is
interval data.

Test Assumptions and Other Test Requirements: The one-way

analysis of variance (Anova) test is based on the following
assumptions and requirements.

• Type of Variable: The scale of measurement for the dependent

variable is continuous (interval).
• Normal Distribution: The dependent variable is normally
distributed in the population, or a sufficiently large sample size
was drawn to allow approximation of the normal distribution.
Note: the “rule of thumb” is that no group should be smaller
than 7, and ideally each group has more.
• Independent Observations: Individuals within each sample and
between the different groups are independent of each
other—random selection indicates that the chances of any one
“unit” being sampled are independent of the chances for any
other being sampled.
• Homogeneity of Variance: Variance is the same for each group, as
indicated by equal standard deviations in the groups’ samples.

The test statistic: The Anova test analysis is based on the

F-distribution. The F-statistic is computed as a complex difference
value divided by a complex value computed with standard deviations

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 437

and group sample sizes. This is the same logic used in our t-statistic
computation, just more complexly computed because there are more
groups involved. The F-statistic is then compared to the criterion
values for the F-distribution, the same way we did this comparison
using the t-distribution to find the criterion t-statistic values. It is
beyond the scope of this course to compute the F-statistic manually;
it is sufficient to understand what it means.

One important facet of the F-statistic evaluation is understanding the

degrees of freedom involved. In fact, with the F-statistic there are
two different types of degrees of freedom involved: between groups
and within group degrees of freedom. Between groups degrees of
freedom is related to the number of groups, regardless of group size.
In our example involving 4 groups, the between groups degrees of
freedom is 4-1, or df=3. Within groups degrees of freedom is related
to the number of individuals in the entire sample, but instead of
being n-1, it is n minus the number computed for between groups
degrees of freedom. In our example with data provided by 390 men
from 4 groups, the within groups degrees of freedom is 390 – (4-1) =
387. This is written as df=(3,387)—this is not read as three thousand
three hundred eighty seven degrees of freedom, it is read as 3 and
387 where the first number is always the between groups degrees of
freedom and the second number is always the within groups degrees
of freedom.

Completing the example: The next task is to apply our computed

degrees of freedom to an F-distribution table to determine the
criterion value for df=(3, 387) at 95% confidence (α=.05): this turns
out to be a criterion value of 2.60. Then, we allow the computer to
generate the comparison F-statistic for our data, as presented in this
table (see Table 5-2).

Table 5-2. Results of one-way analysis of variance, 4 group mean

readiness to change scores.

438 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

The values in the Sum of Squares and Mean Square cells relate to
what would be plugged into the formula that we are not delving into
in this course. The F-statistic computation answer would be written
as F(3, 387)=2.92, p=.034 for this analysis. Therefore, the decision
would be to reject the null hypothesis of equal means since the
F-statistic value of 2.92 is greater than the criterion value of 2.60 from
the F-distribution table.

Another way of drawing the same conclusion (reject the null

hypothesis) is offered by the computer output. The analysis
indicated a significance level computed for the test statistic
(F-statistic) was p=.034. Using the criterion of α=.05, seeing that .034
is less than .05, the same conclusion is drawn: reject the null
hypothesis of equal means. Both of these approaches came to the
same conclusion because they are simply different ways of looking
at the very same thing. Post-hoc comparisons would be required to
determine which of the differences in means listed in Table 5-1 are
significantly different; remember that the test only told us that a
statistically significant difference exists .

Interactive Excel Workbook Activities

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 439

 Complete the following Workbook Activity:

• SWK 3401.4-5.3 One-Way Analysis of Variance (Anova)

Exercise

Chi-square Test of Independence

The chi-square test (often written as the χ2test, pronounced “chi-

square”) is a bivariate statistical analysis approach. It is used to
evaluate whether a significant relationship exists between two
categorical (nominal) variables. Data are conceptualized in the format
of a contingency table (see Table 5-3), where the categories for one
of the two variables are represented as rows and categories for the
other variable are represented as columns. The chi-square test is
concerned with analyzing the relative proportions with which cases
from the sample are sorted into the different cells of the contingency
table.

For example, we can look at the Safe At Home data concerning

whether there exists a significant relationship between treatment
entry status (voluntary or court order) and treatment completion
(completed or dropped out). The data would be entered into a 2 x
2 contingency table, where Oi is the observed number of cases in
each cell, with i being 1, 2, 3, or 4 depending on which cell is being
identified (see Table 5-3).

Table 5-3. Contingency table for chi-square analysis of treatment

entry status and treatment completion.

440 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

 Program Completion
yes (completed) no (dropped) row total
court-ordered O1 O2
Entry Status
voluntary O3 O4

column total

Test Purpose: Assess the independence of treatment entry status and

treatment completion variables.

Test H0: No association between the variables exists (i.e., the

variables are independent of each other if differences in the cell
proportions equal zero).

Type of Data: Two variables are categorical, with two or more

possible categories each.

Test Assumptions and Other Test Requirements: The chi-square

test is based on the following assumptions and requirements.

• Type of Variable: The scale of measurement for both variables is

categorical (nominal).
• Distribution: Normal distribution is not assumed with
categorical variables. The χ2-distribution is used to evaluate the
test statistic.
• Independent Observations :Individuals within each sample and
between the different groups are independent of each
other—random selection indicates that the chances of any one
“unit” being sampled are independent of the chances for any
other being sampled.
• Estimating Expected Cell Counts: In order to compute the χ2-test
statistic, the number of cases expected in each cell of the
contingency table if the two variables are independent of each
other must be calculated; this requires sufficient sample size

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 441

 and a minimum of 5 expected cases per cell.

The test statistic: The chi-square test analysis is based on the

χ2-distribution. The χ2-test statistic is computed in terms of the
observed counts in each cell compared to what would be the
expected counts if no relationship exists between the two variables.
In essence, the formula for computing the χ2-test statistic is
relatively simple: for each cell, compute the difference between what
was observed and what was expected, square the difference, and
divide the result by the expected value for that cell; then, add up the
results for the individual cells into a single value. In mathematical
terms, the formula looks like this:

χ2= ∑(Oi– Ei)2 / Ei

• χ2 is the relevant test-statistic

• the ∑ symbol means add the values up (compute the sum)
• Oi is the observed number in the ith cell (i designates each cell in
the contingency table; 1 through 4 in a 2×2 table; 1 through 6 in a
2 x 3 table; 1 through 9 in a 3×3 table; and so forth)
• Ei is the expected number in the ith cell (i designates each cell in
the contingency table; 1 through 4 in a 2×2 table; 1 through 6 in a
2 x 3 table; 1 through 9 in a 3×3 table; and so forth)

The only thing we still need to know is how to calculate the expected
(Ei) value for each of the contingency table cells. Again, this is a simple
calculation based on the row, column, and sample size total values
related to each cell.

Ei= (row total * column total)/sample size

Completing the Example: In the Safe At Home study example

investigators wanted to know if there exists an association between
program entry status and program completion. (The null hypothesis

442 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

being tested is that no association exists; that the two variables are
independent of each other). Table 5-4 presents the completed 2 x 2
contingency table with the observed (Oi) values are presented, as are
the calculated expected (Ei)values for each cell.

Table 5-4. Data distribution for chi-square analysis of treatment

entry status and treatment completion.

Program Completion
yes (completed) no (dropped) row total

O1 = 174 O2 = 234
court-ordered 408
E1 = 173.7 E2 = 234.3
Entry Status
O3 = 38 O4 = 52
voluntary 90
E3 = 38.3 E1 = 51.7

column total 212 286 498

• O1 is court-ordered clients completing treatment: E1=(212 *

408)/498=173.7
• O2 is court-order clients dropping out of treatment: E2=(286 *
408)/498=234.3
• O3 is voluntary clients completing treatment: E3=(212 *
90)/498=38.3
• O4 is voluntary clients dropping out of treatment: E4=(286 *
90)/498=51.7

We now have all the information necessary for computing the χ2-test
statistic.

χ2= ∑(Oi– Ei)2/Ei

χ2=[(174-173.7)2/173.7]+[(234-234.3)2/234.3]+[(38-38.3)2/38.3]+[52-51.7)
2
/51.7]

χ2=[.32/173.7]+[-.32/234.3]+[-.32/38.3]+[.32/51.7]

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 443

 χ2=[.09/173.7]+[.09/234.3]+[.09/38.3]+[.09/51.7]

χ2=.00052 + .00038 + .0023 + .0017

χ2=.0049

What remains is to evaluate the χ2-test statistic which we computed

as .005 (with rounding up from .0049). This value can be compared
to the criterion value for χ2-distribution based on the degrees of
freedom at α=.05. In order to do so, we need to know the degrees of
freedom for this analysis. In general, degrees of freedom for a chi-
square test are computed as the number of rows minus 1 and the
number of columns minus 1. In a mathematical formula, this would
be:

df=(#rows-1, #columns-1).

Thus, for our 2 x 2 contingency table, the degrees of freedom would

be 2-1 = 1 for each, rows and columns: we would write this as df=(1,1).
The criterion value is then looked up in a chi-square distribution
table. For df=(1,1) at 95% confidence (α=.05), the criterion value is
3.841. In our example, the computed .005 is less than the criterion
value, therefore we fail to reject the null hypothesis. In other words,
there could very well be no association between the two
variables—they could be operating independently of each other.

Another way of drawing the same conclusion is to have computed the

probability value (p-value) for our analysis. In this case, the statistical
program computed a p-value of .941 for our χ2-test statistic. Because
.941 is greater than our decision rule of α=.05, we fail to reject the null
hypothesis.

Interactive Excel Workbook Activities

444 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

 Complete the following Workbook Activity:

• SWK 3401.4-5.4 Chi-Square (χ2) Analysis Exercise

Correlation Test

Correlation is used to analyze the strength and direction of a linear

relationship between two interval (continuous) variables. Similar to
the chi-square test, it assesses the probability of a linear association
between the two variables, but in this case the variables are not
categorical, they are interval numeric variables. The correlation
analysis involves determining the single line that best fits the
observed sample data points. In other words, the line which has the
shortest possible total of distances each point lies from that line.
Figure 5-1 shows just such a line drawn through hypothetical data and
the first few distances from data points to that line are shown in red.
The goal is to find the line where the total of those distances is the
least possible.

Figure 5-1. Best line through data points with correlation coefficient
0.7, with one outlier removed. (Note: image adapted from
https://statistics.laerd.com/spss-tutorials/pearsons-product-
moment-correlation-using-spss-statistics.php)

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 445

The correlation coefficient is designated in statistics as r. The value
for rcommunicates two important dimensions about correlation:
strength of the association or relationship between the two variables
and direction of the association.

Strength of Relationship. The possible range of the correlation

coefficient is from -1 to +1. If two variables have no linear relationship
to each other, the correlation coefficient is 0. A perfect correlation
between two variables would be either -1 or +1—in other words, the
absolute value ofris 1. Neither of these extremes usually happens, so
the strength of the association is determined by how far from zero
the correlation coefficient (r) actually lies. This diagram might help
you visualize what this means:

446 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

The histograms we saw in Chapter 2 depicted univariate data—the
frequency with which specific values on a single variable were
represented in the sample data. Bivariate analyses are depicted using
a scatterplot where each data point (dot) on the scatterplot indicates
an individual participant’s value on each of two variables, the
x-variable and they-variable (height and weight, for example). Figures
5-2 and 5-3 graphically depict hypothetical examples of data points
with zero or no correlation (no evident line) and a perfect +1
correlation.

Figure 5-2. Hypothetical scatterplot with no correlation.

Figure 5-3. Hypothetical scatterplot with perfect +1 correlation

(45o angle line).

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 447

Direction of Association. There is a meaningful difference between
a negative and a positive correlation coefficient. A perfect positive
correlation of +1, as seen in Figure 5-2 indicates that as the values of
one variable increase (x-axis), so do the values of the other variable
(y-axis). On the other hand, a perfect negative correlation of -1 would
indicate that as the values of one variable increase (x-axis), the values
of the other variable decrease (y-axis). The perfect negative
correlation is presented in Figure 5-4.

Figure 5-4. Hypothetical scatterplot with perfect -1 correlation

(45o angle line).

448 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=436

Test Purpose: Assess the strength and direction of association

between two variables.

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 449

Test H0: The two variables are unassociated in the population (i.e.,
their correlation is 0).

Type of Data: Both variables are interval for a Pearson correlation; if

either/both is ordinal, a Spearman correlation is preferable.

Test Assumptions and Other Test Requirements: Correlation tests

require the following conditions and the following considerations.

• Each case analyzed must have complete data for both variables;
if more than one correlation is computed (multiple pairs of
variables), cases can be deleted from the analysis wherever this
is not true (pairwise deletion) or can be deleted from all analyses
(listwise deletion).
• Data on each variable should be approximately normally
distributed.
• Data should be assessed for “outliers” (extreme values) which
can affect the correlation analysis. Outlier cases, typically
having values more than 3.25 standard deviations from the mean
(above or below), should be excluded from correlation analysis.
• Distribution of the data should be in a relatively tubular shape
around a straight line drawn through the data points on a
scatterplot; for example, a cone shape violates this assumption
of homoscedascity. Homoscedascity refers to the degree of
variance around the line being similar regardless of where along
the x-axis the points are mapped. A cone shape would mean
that low values on the xvariable have data points closer to the
line than high values on the xvariable (a case of
heteroscedascity).
• Distribution of data should be relatively linear. For example, a
curved line reflects a non-linear relationship between the
variables. Non-linearity violates an assumption necessary for
correlation analysis.
• Statistically significant correlation coefficients are increasingly

450 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

 likely as sample sizes increase. Therefore, caution must be used
in interpreting relatively low correlation coefficients (below
about 0.3) which are flagged as statistically significant if the
sample size is large.
• Correlation simply indicates an association between two
variables, it does not in any way indicate or confirm a causal
relationship between those variables. Always remember that
correlation does not imply causality (you have read this before
in other modules). Causal relationships require different kinds of
study designs and statistical analyses to confirm.

The test statistic: As previously noted, correlation analysis is based on

identifying from among an infinite array of possible lines, the one line
that best defines the observed data—the line where the cumulative
distance from data points to that line is minimized. The formula looks
more complicated than it actually is. The correlations coefficient (r)
is computed from a numerator divided by a denominator, just as our
previous coefficients have been. The numerator is made up of the
sample size N (the number of pairs for the two variables) multiplied
by the sum of each pair of values (x times y for each case, summed)
minus the result of multiplying the sum of the x values times the sum
of the y values. The denominator is the square root of multiplying two
things together: first, the sample size N times the sum of the x2 values
minus the square of the sum of the x values, then doing the same for
the y values—sample size N times the sum of the y2 values minus the
square of the sum of the y values. In a mathematical formula, it looks
like this, where the final step will be dividing the numerator by the
denominator:

r numerator = N * ∑x*y – (∑x)*(∑y)

r denominator = square root of [N*∑x2 – (∑x)2] * [N*∑y2 – (∑y)2]

With large numbers of cases, hand calculating this correlation

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 451

coefficient is cumbersome—computer software programs can do the
computation very quickly.

Completing the example:Consider for example that the Safe At Home

project investigators wanted to assess the association between
readiness to change scores and the length of time between client
entry into the treatment program and the intimate partner violence
incident leading to treatment entry (measured in weeks). In our
example, the computed correlation coefficient was r=-.060 for N=387
clients entering batterer treatment for whom completed data were
available and following removal of three outliers with extremely long
time lag between the incident and program entry.

The negative sign on the correlation coefficient suggests that greater

number of weeks between incident and program entry might be
associated with lower readiness to change scores. This could be
explained in several different causal models, so it is not acceptable to
infer that one caused the other:

• More time passing leads to less impetus to change.

• Low impetus to change leads to longer lag time before entering
a program.
• Some other factor or factors influence both time to program
entry and readiness to change scores.

The strength of the correlation coefficient needs to be assessed. The

coefficient is not far from zero (.06 is not even 1/10th of the way
to 1.0), suggesting that the association is weak. This is particularly
notable since the sample size was so large (N=387): even a small
correlation should be more easily detected with a large sample size.
Allowing us to reject the null hypothesis is the computed two-tailed
significance value where p=.241. Since .241 is greater than our
decision rule of α=.05, we fail to reject the null hypothesis. In other
words, we conclude that there may be no association between the

452 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

two variables. In reality, this finding could be explained by some
other factor or factors influencing the length of time before entering
a program—events such as incarceration for the offenses could
interfere with entering treatment, for example.

Interactive Excel Workbook Activities

Complete the following Workbook Activity:

• SWK 3401.4-5.5 Correlation (r) Exercise

Choosing the Right Statistical Test for the Job

You have now been introduced to 5 powerful statistical tools for

analyzing quantitative data. The guidelines for choosing the right tool
for the job have to do with (1) the nature of the hypothesis being
tested, and (2) the nature of the variables being analyzed. This flow
chart might help (see Figure 5-5).

Figure 5-5. Decision flow chart for 5 statistical analytic approaches.

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 453

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can
view it online here:
https://ohiostate.pressbooks.pub/
swk3401/?p=436

454 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

The Role of Non-Parametric Statistical Approaches

With the exception of the chi-square test, all of the approaches we

explored are based on assumptions of normal distribution across the
population on at least one variable; an assumption assessed based on
the distribution of data from the sample. In many instances, true
normal distribution is not required, as long as the sample size is
sufficiently large to approximate normal and extreme outliers have
been addressed. However, it is important to know that there exist
alternative approaches to statistical analysis when the assumption of
normal distribution does not hold up. The alternatives are called
non-parametric tests.
For example, when a t-test is the right option but the data are non-
normal in distribution, a non-parametric alternative is the rank sum
test—also called the Wilcoxon-Mann-Whitney test. The test statistic
in this case is called U instead of t and is compared to a different
distribution table. This test has four main steps:

• Rank order all of the cases from lowest to highest on the

dependent variable, regardless of which group they
represent—all are in one single pool.
• Sum the ranks for one of the two groups (it does not matter
which is selected)—this becomes R1 in the formula below.
• Compute the result of [number of cases in group1 * (number of
cases in group1+1)] divided by 2.
• The test statistic is the result of subtracting this last
computation from the R1 value.

U = R1 – [n1 * (n1 + 1)]/2

The test statistic is then compared to the distribution for U and/or a
p-value can be computed.
Similarly, the Kruskal-Wallis H test is a one-way Anova computed on
ranks as a non-parametric alternative to the one-way Anova we
previously learned. As in the case of the Wilcoxon-Mann-Whitney U

Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches | 455

test, the Kruskal-Wallis H test operates using the rank orders of
cases rather than the actual case and mean values for the dependent
variable. In this way, variance is removed as a potential problem in
the computation.

Chapter Summary

In this chapter you learned about 5 different statistical analysis

approaches: one-sample t-test, independent sample t-test, one-way
analysis of variance (Anova), chi-square, and correlation. You learned
about the principles and assumptions underlying these statistical
approaches, and how to determine which might be the best choice
for different scenarios. Throughout the chapter you learned how
statistical probability and the application of decision criteria are
applied in drawing conclusions about populations based on
samples—rejecting or failing to reject the null hypothesis. You were
introduced, as well, to basic principles concerning non-parametric
tests that can be used when the underlying assumptions necessary
for parametric inferential statistical analyses are violated.

456 | Module 4 Chapter 5: Introduction to 5 Statistical Analysis Approaches

Module 4 Summary
This information-loaded module began with an introduction to how
investigators manage and analyze qualitative data. You learned a bit
about transcription and coding practices and issues, and are
encouraged to pursue more advanced training in these skill sets. The
remainder of the module introduced a sizeable number of topics,
issues, and skills related to analyzing quantitative data. One
important distinction you learned about concerns univariate and
bivariate analyses. In the univariate domain you learned univariate
techniques for describing data representing different types of
variables (categorical, dichotomous categorical, ordinal, and
interval/continuous variables). These included frequency and
proportion descriptive statistics, as well as central tendency (mean,
median, mode) and variability (range, variance, standard deviation)
statistics. You also learned how this type of information might be
presented in tables and figures, and learned the basic elements of
normal distribution. This module explained the probability logic
underlying inferential statistics and null hypothesis testing, as well
as Type I and Type II errors. You learned to craft and interpret null
hypotheses, and how information about statistical criterion values
are managed (distribution tables, confidence intervals, p-values, and
α levels). Five basic parametric approaches for inferential statistical
analysis were introduced along with their underlying assumptions
and how they are interpreted: one-sample t-test, independent
samples t-test, one-way analysis of variance (Anova), chi-square, and
correlation analyses. Finally, you were introduced to non-parametric
approaches which might be used in situations where the assumptions
underlying the parametric approaches might be violated. Throughout
the module, you were provided with opportunities to practice new
skills using Excel. Our next (and final) course module addresses how

Module 4 Summary | 457

results from social work, social science, and behavioral research
might be presented to varied audiences.

458 | Module 4 Summary

Module 4 Key Terms and
Definitions
α=.05: convention in social work, social, and behavioral sciences for
decision to reject/fail to reject the null hypothesis when analyzing
quantitative data.

alternate hypothesis: a research hypothesis phrased in terms of the

statistical analysis to be performed on observed differences with
quantitative data; often designated as Haor H1 (see null hypothesis for
more information).

a priori coding: qualitative data analysis utilizing a pre-determined

set of coding categories (in contrast to open coding).

bivariate analysis: quantitative data analysis approaches evaluating

the association or relationship between two variables.

central tendency: quantitative univariate statistical approaches (e.g.,

mean, median, mode) that serve as indicators of how data cluster
around a “common” value.

chi-square (χ2) test: bivariate quantitative statistical analysis for

evaluating an association between two categorical variables based
on the observed and expected proportions in a variable x variable
contingency table; evaluated using criterion values based on the
χ2-distribution and degrees of freedom.

coding: means of creating categories out of themes present in

qualitative data for analysis.

confidence interval (CI): range of statistical test values that would

Module 4 Key Terms and

Definitions | 459
arise by chance at the determined probability level (usually 95%
confidence interval in social work, social, and behavioral science).

context: descriptive information in qualitative research that aides the

interpretation of coded data.

correlation: bivariate quantitative statistical analysis for evaluating an

association between two interval (continuous) variables based on the
“best” fitting line for the observed data points.

correlation coefficient (r): the statistical value computed in

correlation analysis which helps determine statistical significance,
strength, and direction of the relationship between two variables.

degrees of freedom: a component of many inferential statistics

approaches based on a computation related to the number of cases in
a data set and, depending on the statistical test, the number of groups
being compared.

descriptive statistics: a set of univariate analysis approaches used

to describe how a sample is distributed along various dimensions,
particularly frequency, central tendency, and variability.

F-statistic: sample statistic computed using group means, standard

of deviation, and both within and between group degrees of freedom;
compared to criterion values based on the F-distribution.

field notes: supplemental information in qualitative data and analysis

that provide insight into the context of data collected from
participants and the thinking process of the investigator analyzing
the data.

frequency: univariate quantitative statistical information concerning

how a sample distributes on a (categorical) variable or some
combination of categories for multiple categorical variables
(subgroups) based on the count of numbers in each group.

460 | Module 4 Key Terms and Definitions

histogram: a univariate graphing approach showing the frequency
with which values appear in a sample.

homoscedascity: principle in quantitative data analysis involving data

in relation to a “best” fitting line, reflecting how the data points
cluster around that line; refers to the degree that these distances
between data points and the line are similar.

independent samples t-test: bivariate quantitative statistical analysis

for comparing mean scores (interval variable) for exactly two groups
(categorical variable); evaluated using criterion values based on the
t-distribution and degrees of freedom.

inferential statistics: set of quantitative data analysis approaches

based on drawing inferences about a population based on data from a
sample; these approaches have specific assumptions and limitations
that need to be evaluated and addressed for accurate interpretation
of the results.

Kruskal-Wallis H test: a non-parametric quantitative statistical

analysis used when assumptions related to normal distribution and
sample size are violated; analog to the parametric one-way analysis of
variance (Anova) compares rank order of data points in groups rather
than actual data values and group means.

kurtosis: one dimension for evaluating a sample’s distribution on an

interval (continuous) variable, specifically how it meets criteria for
normal distributions regarding concentration of data in the center
versus tails of the distribution curve.

mean: univariate quantitative statistic indicating one facet of central

tendency (often called “average”); particularly susceptible to
distortion from outliers and small sample sizes.

median: univariate quantitative statistic indicating one facet of

Module 4 Key Terms and Definitions | 461

central tendency (often called “50thpercentile”); half of cases fall
below and half fall above this value.

mode: univariate quantitative statistic indicating one facet of central

tendency, represents the most common value in the data (note there
may be more than one mode present).

non-parametric tests: inferential quantitative statistical analyses that

are not dependent on assumptions of normal distribution or large
sample size (in contrast to parametric tests).

normal distribution: distribution of values on an interval (continuous)

variable that falls on a curve that is (a) symmetrical around the mean,
and (b) specific proportions of cases fall within 1 standard deviation,
2 standard deviations, 3 standard deviations and more from the mean
(symmetrically above and below the mean); often depicted as a “bell-
shaped” graph.

null hypothesis: a research hypothesis phrased in terms of no

statistically significant difference based on statistical analysis to be
performed with quantitative data; often designated as H0 (see
alternative hypothesis for more information).

one-sample t-test: univariate quantitative statistical analysis for

comparing a group’s mean score (interval variable) to a specific
standard value; evaluated using criterion values based on the
t-distribution and degrees of freedom.

one-tailed test: comparison statistic is based on probability related

to one tail of the normal distribution, rather than both tails, because
the hypothesis being tested is directional (more than or less than
hypothesis) instead of a bidirectional “difference” hypothesis (in
contrast to a two-tailed test).

one-way analysis of variance (Anova) test: bivariate quantitative

462 | Module 4 Key Terms and Definitions

statistical analysis for comparing mean scores (interval variable) for
two or more groups (categorical variable); evaluated using the
F-distribution and degrees of freedom.

open coding: qualitative data analysis utilizing coding categories that

emerge out of the data being analyzed (in contrast to a prioricoding)

outliers: data points that are exceptionally different from the majority
of cases in a sample; having potential to greatly distort central
tendency measures like the mean.

p-value: a probability value computed in inferential quantitative

statistical analyses, used to interpret statistical test values in terms of
rejecting or failing to reject the null hypothesis.

parametric tests: inferential quantitative statistical analyses that are

dependent on assumptions of normal distribution and sufficiently
large sample size (in contrast to non-parametric tests).

percentage: univariate quantitative statistical information concerning

how a sample distributes on a (categorical) variable or some
combination of categories for multiple categorical variables
(subgroups) calculated as the proportion of the total in each group
(out of 100%).

population parameters: values for central tendency and variability

(dispersion) for an entire population; when data are not available
for the entire population, sample statistics are utilized to estimate
population parameters while recognizing the probability of error in
these estimates.

probability: a mathematical/statistical concept concerned with

computing the likelihood of an event occurring; the inverse (the
likelihood of an event not occurring) may also be computed;
probability forms the logic basis underlying inferential statistics.

Module 4 Key Terms and Definitions | 463

projective test: an observational measurement technique used to
reveal internal thought processes or personality traits as individuals
respond to ambiguous stimuli (words, images, objects, scenarios/
situations, music, or other forms of presentation).

proportion: the segment or portion something represents of the

whole (usually presented in fraction or percent format).

range: the mathematical difference between the lowest and highest

values in a sample data set, or the lowest to highest possible values
for a measurement tool.

rank sum test/Wilcoxon-Mann-Whitney U test: a non-parametric

quantitative statistical analysis used when assumptions related to
normal distribution and sample size are violated; analog to the
parametric independent samples t-test, compares rank order of data
points in groups rather than actual data values and group means.

sample statistics: values for central tendency and variability

(dispersion) for a sample of cases drawn from a whole population;
when data are not available for the entire population, sample
statistics are utilized to estimate population parameters while
recognizing the probability of error in these estimates

scatterplot: a bivariate graphing approach showing data points for the

paired values of each case in a sample (x-axis and y-axis reflect the
values for the two variables).

skew: one dimension for evaluating a sample’s distribution on an

interval (continuous) variable, specifically how it meets criteria for
normal distributions regarding symmetry of data around the mean.

standard deviation: a sample statistic used to estimate population

variance, computed as the square root of the sample variance.

t-statistic: sample statistic computed using group means and

464 | Module 4 Key Terms and Definitions

standard of deviation; compared to criterion values based on the
t-distribution.

two-tailed test: comparison statistic is based on probability related

to both tails of the normal distribution because the hypothesis being
tested is bidirectional (unequal) instead of a unidirectional “greater”
or “less” than hypothesis (in contrast to one-tailed test).

transcription: a common step in preparing qualitative data for

analysis, constructing a verbatim written record of the information
provided by each study participant.

univariate analyses: quantitative data analysis approaches evaluating

data for one variable at a time (e.g., frequency, proportion, central
tendency, variability/dispersion).

variance: quantitative univariate computed statistical value that

serves as an indicator of how values are dispersed (vary) for either a
population or a sample (see standard deviation).

verbatim: word-for-word reporting, in exactly the same words, to

preserve the original meaning of what was originally said.

Wilcoxon-Mann-Whitney test (U test)/rank sum test: a non-

parametric quantitative statistical analysis used when assumptions
related to normal distribution and sample size are violated; analog to
the parametric independent samples t-test, compares rank order of
data points in groups rather than actual data values and group means.

Module 4 Key Terms and Definitions | 465

Module 4 References
Begun, A.L., Barnhart, S.M., Gregoire, T.K., & Shepherd, E.G. (2014). If
mothers had their say: Research-informed intervention design for
empowering mothers to establish smoke-free homes. Social Work
in Health Care, 53(5), 446-459.

Corcoran, J., & Secret, M. (2013). Social work research skills workbook:
A step-by-step guide to conducting agency-based research. NY:
Oxford University Press.

Creswell, J.W., & Poth, C.N. (2018).Qualitative inquiry & research

design: Choosing among five approaches, 4thed.Thousand Oaks, CA:
Sage.

Engel, R.J., & Schutt, R.K. (2018). The practice of research in social
work, 3rded. Thousand Oaks, CA: Sage.

Glaser, B.G., & Strauss, A.L. (1967).The discovery of grounded theory:

Strategies for qualitative research. New Brunswick, NJ: Aldine.

Graham, H. (no date). APA format for statistical notation and other
things. Retrieved from http://users.sussex.ac.uk/~grahamh/
RM1web/
APA%20format%20for%20statistical%20notation%20and%20othe
r%20things.pdf

Hinduja, S., & Patchin, J.W. (2008). Personal information of

adolescents on the Internet: A quantitative content analysis of
MySpace. Journal of Adolescence, 31(1), 125-146.

McInroy, L.B., & Craig, S.L. (2015). Transgender representation in

offline and online media: LGBTQ youth perspectives. Journal of
Human Behavior in the Social Environment, 25, 606-617.

466 | Module 4 References

Nadelman, L. & Begun, A.L. (1982). Effects of the newborn on the
older sibling. In M. E. Lamb & B. Sutton-Smith (Eds.), Sibling
relationships: Their nature and significance across the lifespan, (pp.
13-37). New Brunswick, NJ: Lawrence Erlbaum.

Rubin, A., & Babbie, E. (2013). Essential research methods for social
work, 3rded.Belmont, CA: Brooks/Cole, Cengage Learning.

Salkind, N.J. (2017). Statistics for people who (think they) hate
statistics, 6thedition. Thousand Oaks, CA: Sage.

Taleb, N. (2012). Antifragile: Things that gain from disorder.NY:

Random House.

Tracy, S.J. (2013). Qualitative research methods: Collecting evidence,

crafting analysis, communicating impact. Malden, MA: John Wiley &
Sons, Inc.

Trochim, W.M.K. (2005). Research methods: The concise knowledge

base. Mason, OH: Atomic Doc/CENGAGE Learning.

Module 4 References | 467

MODULE 5

Module 5 | 469
Module 5 Introduction
In our final module we learn how data and evidence about diverse
populations, social phenomena, and social work problems are
prepared for presentation to various audiences. As social work
professionals you will be called on to help different types of
audiences understand complex social work problems, social
phenomena, and diverse populations. In this module, some powerful
tools are explored for their communication potential. We not only
emphasize the ways investigators present evidence, we also address
how professional reports and presentations can help our audiences
identify or recognize social justice issues and can help advocate for
change. In this module we learn to construct tables, figures, graphs,
and graphic materials using commonly accessed software like
Word®, PowerPoint®, and Excel®. We continue to practice the
American Psychological Association (APA) style guidelines for
professional writing.

READING OBJECTIVES

After engaging with these reading materials and learning resources,

you should be able to:

• Describe parts of an empirical manuscript or report;

• Use the American Psychological Association (APA) style
guidelines for professional writing;
• Create a strong presentation for various audiences;
• Develop graphs and charts of different types for different
communication purposes;
• Design an infographic communication tool;
• Summarize what was learned in this first course in the 2-course
research and evidence sequence;

Module 5 Introduction | 471

 • Define key terms related to communicating social work
evidence to varied audiences.

472 | Module 5 Introduction

Module 5 Chapter 1: Parts of
an Empirical Manuscript or
Report
In Module 2 you learned about the different parts of an empirical
article from the standpoint of critically reviewing them. The
emphasis in this module is from the other side of the coin: how they
are composed. The same sections are discussed: the title, abstract,
introduction, methods, results, discussion, and references. However,
emphasis in this module is placed on writing issues, and how
professional reports might differ from empirical manuscripts. In this
chapter you will learn:

• What to include in each section of a manuscript or report;

• Important writing issues of language use and avoiding
(accidental) plagiarism;

What to Include and Where

In this section, writers are coached about what pieces of information

belong in each section of an empirical manuscript or professional
report. A great deal of this information is available in the Publication
Manual of the American Psychological Association 6thedition (APA,
2009).

Title. A good title conveys to readers what the manuscript or report

is about. It is also no more than about 12 words in length. A good
title is fairly specific and concrete rather than vague and general.
Consider these “good” examples—note that each describes the main
topic, methodology, and/or population addressed.

Module 5 Chapter 1: Parts of an

Empirical Manuscript or Report | 473
 • Understanding Latino grandparents raising grandchildren
through a bioecological lens (Mendoza, Fruhauf, Bundy-Fazioli,
& Weil, 2018)
• Parent, teacher, and school stakeholder perspectives on
adolescent pregnancy prevention programming for Latino
youth. (Johnson-Motoyama et al., 2016).
• Innovative ethics: Using animated videos when soliciting
informed consent of young people for online surveys (McInroy,
2017).
• Self-Identified strengths among youth offenders charged with
assault against a non-intimate family member (Mengo et al.,
2017).
• Gender differences in pathways from child physical and sexual
abuse to adolescent risky sexual behavior (Yoon, Voith, &
Kobulsky, 2018).
• Pathways to age-friendly communities in diverse urban
neighborhoods: Do social capital and social cohesion matter?
(Parekh, Maleku, Fields, Adomo, Schuman & Felderhoff, 2018).
• Using social determinants of health to assess psychological
distress and suicidal thoughts and behavior among youth on
probation (Quinn, Liu, Kothari, Cerulli, & Thurston, in press).
• The role of family financial socialization and financial
management skills on youth saving behavior (Kagotho, Nabunya,
Ssewamala, Mwangi, & Njenga, 2017).

Notice that these report titles include details about when or where
the information relates:

• 2015-2016 Ohio State’s LiFE Sports annual review. Columbus,

OH: College of Social Work. Ohio State University (Anderson-
Butcher, Wade-Mdivanian, Davis, Newman, & Lower, 2016).
• Connecting Youth in Child Welfare to Behavioral Health Services
in Franklin County: Provider Perspectives. Evaluation report

474 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

 written for Franklin County Children Services (OH), (Bunger,
Robertson, Hoffman, & Cao, 2015).
• Evaluation of Marathon County Community Response Program:
Report to Marathon County Department of Social Services
(Maguire-Jack & Campbell, 2015).

For more information, consult the Publication Manual (APA, 2009) on

page 2.

Abstract or Executive Summary. Although the abstract or executive

summary appears first to readers, it is one of the very last things
written by the author(s). This is because it summarizes all the
remaining information. For a manuscript (or professional conference
presentation), an abstract allows readers to quickly review what the
manuscript is about, to see if it is sufficiently relevant for them
to continue reading the entire manuscript. A manuscript abstract
is usually 150-200 words. Abstracts for professional conference
presentation proposals vary in length, depending on the preferences
of the conference organizers. Typically, they range from 200 to 500
words in length. The abstract typically follows the format of the
manuscript: introduction, methods, results, and discussion. Hitting
just the right balance between these sections of the abstract is
important: the introduction and discussion each are usually only one
or two sentences long, with the bulk of the space devoted to methods
and results. In addition to the APA manual (APA, 2009), a useful
resource is available on line: https://www.verywellmind.com/how-
to-write-an-abstract-2794845.

On the other hand, if the written product is a professional report

(to a funding agency or board, for example), then the abstract is
replaced with an initial summary statement, often called an executive
summary. Like the abstract, it highlights key points, especially a
project’s process, results, and recommendations/implications.
However, it is written in a more “bullet point” style, and less like

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 475

a journal abstract. The report’s executive summary ideally fits on a
single page. If you want to see some examples, search for “executive
summary” or “technical report” samples.

For more information, consult the Publication Manual (APA, 2009)

pages 25-27.

Introduction. The introduction to a manuscript is usually organized

in the following manner:

• An introduction of the topic or problem addressed—its

background and significance (i.e., the problem dimensions,
costs, impact on other problems, impact on which populations).
This is the “so what” statement—why was it worth the effort to
address this problem and why should readers care.
• Summary of literature and prior research, with a critical review
of what has been done, relevant theory, and important
knowledge gaps.
• Research questions and the rationale for them; if a quantitative
study, hypotheses are also identified here.

A strong literature review and introduction to an empirical

manuscript (adapted from John Elliott, an instructor at Ohio State
University):

• Is clearly organized
• Flows logically
• Uses primary literature sources
• Specifies types of studies reviewed (qualitative, quantitative,
mixed methods)
• Includes recent studies, although classic studies can be included
if they help set the stage
• Provides a balanced appraisal of the material presented, not
leaving out sources that might disagree with an author’s point of

476 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

 view and discussion of strengths/weaknesses of reviewed
studies is balanced
• Presents what is relevant for providing context related to the
current issue/project, does not include the entire universe of
possible materials
• Provides a framework for what is known and what is not known
• Leaves the reader aware of what has come before, what is
needed, and why it is needed, and leads logically to the specific
research questions and hypotheses.

A professional report, on the other hand, might rely less on the

literature in the introduction (although the project should be
informed by literature), emphasizing instead the background and
context of the project’s initiation and development.

For more information, consult the Publication Manual (APA, 2009)

pages 27-29.

Methodology. The methodology or methods section is written with

the intent of describing for readers what was done to conduct the
study so that (1) it could be replicated by another investigator and
(2) evaluation of the methods can be performed by a reviewer or
potential consumer of the evidence. This section typically begins
with a brief introductory paragraph describing the study design and/
or sources of data (e.g., if it is secondary or administrative data set,
where the original data came from). It should be clear to a reader
what “model” is being tested in a quantitative study (e.g., univariate
descriptives or bivariate relationships between specified
independent and dependent variables).

Next, the study participants are described along with details about
their recruitment, including participation rates and sampling
strategy where appropriate. The description includes information
about numbers, characteristics, how they reflect or differ from the

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 477

population they represent, and how many fit into each group or
category if there were quantitative analyses conducted.

The next subsection of methods describes how data were

collected—the measurement tools and procedures. These are
described in terms of the variables that they represent (in
quantitative projects) or in terms of qualitative methods employed.
This section also describes any manipulation of the data that needed
to be made (e.g., creating new variables out of measured variables or
transcription of qualitative data, for example).

For more information, consult the Publication Manual (APA, 2009)

pages 29-32.

Results or Findings. The art of writing the results section is to clearly

and succinctly present the findings in an objective manner—this is
not the place where the results are analyzed, interpreted, or critiqued
(that is in the discussion section). This is the place to tell readers
what was observed in the data collected. It is important that all
relevant results be reported—those that fail to confirm hypotheses
are just as important as those that confirm. The tables, graphs, and
charts that you learned about in earlier sections of this course are
valuable tools for reporting details; they replace the need to describe
the details in the test of the manuscript or report. This section may
need to explain HOW findings were generated (what analyses were
conducted), in addition to WHAT the finding might have been. In
a research manuscript, the label “results” is usually applied; in a
professional report, the label might either be “results” or “findings.”

For more information, consult the Publication Manual (APA, 2009)

pages 32-35. In addition, the Publication Manual provides a great
amount of detailed advice concerning how to report numbers and
statistics, as well as how to display results in tables and figures (pages
111-161).

478 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

Discussion or Recommendations. The final segment of an empirical
manuscript is the author’s discussion of the study’s interpretation or
analysis of the study and its results. It usually begins with a brief
summary of key results and how they relate to the study aims initially
presented in the introduction. Then, the results are discussed in
terms of their implications for theory, future research, and/or
practice. These conclusions need to be justified by the evidence
presented in the study—it is not acceptable to reach far afield of what
the data actually support. Included in the Discussion section is an
analysis of any potential limitations of the study, such as the sample,
the measurement tools, or the data quality. The Discussion section
typically concludes with a statement of what the study contributes to
knowledge and what more might need to be explored in the future.
In a professional report, the focus of the concluding discussion is a
bit different: it emphasizes specific recommendations based on the
findings.

For more information, consult the Publication Manual (APA, 2009)

pages 35-36.

Writing Tips

The APA (2009) Publication Manual discusses writing style for a

manuscript that apply to most professional writing: write clearly,
write concisely, and organize the content (see pages 61-70). One
of the best ways to improve your professional writing is to read
others’ manuscripts with an eye to style as well as to content. Several
additional tips are worth considering, as well.

Verb Tenses. For the most part, an empirical manuscript or

professional report is written in the past tense. This is because the
work has been completed. Even in summarizing the literature, past
tense is preferable. For example, in their review of literature
concerning integrative body, mind, and spirit practices in social work,

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 479

the authors (Lee, Wang, Liu, Raheim, & Tebb, 2018) generated past-
tense statements such as:

• “Studies were identified through electronic databases…” (p. 252)

• “The review process consisted of …” (p. 252), and
• “The review process involved two independent reviewers rating
each study” (p. 253).

Here is an example of confused verb tense (see underlined verbs),

along with a reasonable repair:

Begun, Early and Hodge (2016) reported that 82.5% of women

preparing for release from jail engagedin pre-incarceration
substance misuse that placed them at risk of a substance use
disorder; 64% areat very high risk.

Begun, Early and Hodge (2016) reported that 82.5% of women

preparing for release from jail engagedin pre-incarceration
substance misuse that placed them at risk of a substance use
disorder; 64% wereat very high risk.

Regardless of your choice of verb tense, it is critically important that

the tense be applied consistently.

Parallel construction. The concept of parallel construction applies at

two different levels. First, it applies to the different sections of a
manuscript or report. For example, if you describe three research
questions or hypotheses in the introduction (1, 2, and 3), then it is
important to maintain that structure in that order in the methods,
results, and discussion sections, as well (i.e., 1, 2, 3 throughout). That
structure becomes the skeleton on which the rest of the manuscript
or report is built.

Second, parallel construction also applies to sentences. For example,

if a sentence begins with a singular subject (e.g., a person) then the

480 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

verb and other parts of the sentence need to be in the singular,
as well (e.g., was, not were). Likewise, if a sentence begins with a
plural subject (e.g., adolescents), the verb and other parts of the
sentence need to be plural, too (are, not is). One more example of
parallel construction: notice that the last two sentences are similarly
constructed—the “bridge” word came first (i.e., For example and
Likewise), followed by the “if” clause which was followed by a “then”
clause.

Language use. In its most recent editions, the Publication Manual

emphasized the importance of using language that is respectful, free
from bias, and non-stigmatizing (APA, 2009, pp. 70-77). Reviewers,
editors, and audiences are increasingly sensitive to the need for using
language that reflects an appreciation for diversity, particularly as
related to being inclusive around gender, age, race, ethnic group,
national origin, religion/spirituality, sexual orientation, and
residence, as well as other social, health, functional, and personal
circumstances, conditions, or diversity (Begun, 2016; Broyles et al.,
2014). For example, strong efforts are made to move away from
stigmatizing terms like “mental retardation” and labels that identify
people by diagnosis: mentally retarded, schizophrenic, addict, and
epileptic. Instead, a predominating theme is to “put people first” by
using terms such as “persons experiencing addiction” or “individuals
diagnosed with schizophrenia.” In this way, we avoid making the
diagnosis “be the person,” and instead make it a part of the person’s
experience. In addition to avoiding the use of labels, the APA (2009)
guidelines recommend avoiding describing groups in stigmatizing
comparisons. The example they offered relates to comparing one
group to a “normal” group, such as comparing lesbians to “normal
women” or “the general public” is inappropriate.

This awareness of language use in social work professional writing

extends beyond diagnosis. Reducing bias, stigma, and
microaggression that stems from careless use of language is an

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 481

important goal. One strategy recommended in the APA (2009)
Publication Manual is to ask persons from the groups represented
in a manuscript or report what terms are preferred and to have
them review what is written for the presence of inadvertent bias.
For example, staff in an Africentric program may wish to avoid the
term Afrocentric, believing the latter relates to a hairstyle and the
former describes a program grounded in African cultural traditions
(i.e., the seven core principles of Kwanzaa being Umoja, Kujichagulia
Ujima, Ujamaa, Nia, Kuumba, and Imani). Others may prefer the term
Afrocentric, believing that there is no such word as Africentric. Both
groups may prefer these terms to being described as an African
American program. As another example, in reference to designated
parking, many individuals prefer the term “accessible parking” rather
than “handicapped parking” because the emphasis is on the
accommodation rather than the disability.

The APA (2009) guidelines also suggest describing compared groups

using parallel terms. They gave the example of unparalleled nouns:
man and wife. This is considered problematic because the woman is

482 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

described only in terms of her relationship to the man; preferable is
the use of parallel terms, such as man and woman or husband and
wife. Or, a non-gendered term like “spouse” might be appropriate.

Paraphrasing and citing. It is always important to avoid plagiarism.

Period. Always. Knowing how to paraphrase and cite your sources are
two skills to help avoid accidental plagiarism. Here is an example
of paraphrasing with the use of citation. In the original article, the
authors (Mengo & Black, 2015, p. 244) wrote:

Sexual victimization appears to have a stronger negative

impact on students’ academic performance than physical/
verbal victimization.

A first choice for applying this information in your own writing is to

use it as a direct quote with an appropriate citation. For example:

Mengo and Black (2015) concluded that, “Sexual victimization

appears to have a stronger negative impact on students’
academic performance than physical/verbal victimization” (p.
244).

Because it is a direct quotation, the page number is also cited.

However, the APA (2009) publication manual discourages over-use of
direct quotes. Used too often, they create a choppy, poorly integrated
result. Sometimes a previous author has stated something so
perfectly that a direct quote is desirable. Most of the time, you will
need to convert your references to others’ work by paraphrasing.
Here is the example paraphrased and accompanied by the proper
citation:

Students’ academic performance is more negatively affected

by sexual victimization than by victimization through acts of
physical or verbal violence (Mengo & Black, 2015).

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 483

Formatting citations. In the APA style, citations are either “in text”
or “parenthetical” in nature. The first example shows how an “in text”
citation is formatted. The authors’ names are connected using the
word “and” since you are reading it as part of sentence, and the
publication date appears in parentheses (2015). The second example
shows how a “parenthetical” citation is formatted. Notice here that
the authors’ names are combined with the “&” sign, not the work
“and”’ also, their names appear in the set of parentheses along with
the publication date of their article. The paraphrased statement ends
with a period afterthe reference citation in parentheses. Many
students believe that formatting citations is the main contribution
of the APA (2009) style guidelines. While this is covered (p. 169-192),
there are many other important style tips and formatting guidelines
presented in the manual.

References. The full reference for any citation appears in the

reference list at the end of your manuscript or report. References
are placed in alphabetical order by the first author; initials are used
rather than full first names. Here is the reference listing for our
example’s citation:

Mengo, C., & Black, B.M. (2015). Violence victimization on a college

campus: Impact on GPA and school dropout. Journal of College
Student Retention: Research, Theory & Practice, 18(2), 234-248.

If you check the reference list at the end of this module, you will see
this reference is placed ahead of the next entry authored by the same
first author, Mengo, Lee, Bolton, Lehmann, & Jordan (2017). Similarly,
you will see that the reference for Begun, Early, and Hodge (2016)
comes after the reference for Begun (2016) but before the reference
for Begun and Rose (2007). This is because the second author’s names
are the next step in alphabetizing the references.

The format requirements for articles, books, book chapters, websites,

news or magazine articles, and other sources are presented in the

484 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

APA (2009) publication manual (see pages 193-216). On line resources
are also available to help with formatting references, one of the most
commonly used is called Purdue Owl, distributed by Purdue
University through the website https://owl.purdue.edu/owl/
research_and_citation/apa_style/
apa_formatting_and_style_guilde/general _format.html.

For each item identify if the

paraphrasing is appropriate for writing
your own manuscript or report (True)
or not (False). The paraphrasing is
based on the following material
published in an article by Doogan and
Warren (2017, p. 1430). The article
reports findings from a social network
analysis into the dynamics of
communication exchanges between
group members participating in
therapeutic community (TC) sessions as part of residential
treatment while incarcerated.

These findings support the TC use of the community of peers

as the method of treatment (De Leon, 2000). They suggest that
the best way for TC clinical staff to foster an active unit is
to encourage residents to intervene with each other. The
increased likelihood of residents to correct peers after
receiving a peer affirmation supports De Leon’s (2000) claim
that affirmations serve to energize residents and increase
program participation. Finally, this pattern fits the definition
of generalized reciprocity; residents who have received help
from peers show a tendency to pass it along (Molm, Collett
& Schafer, 2007, Stanca, 2009), perhaps because they feel an

Module 5 Chapter 1: Parts of an Empirical Manuscript or Report | 485

 increase in their own positive feelings (Isen & Levin, 1972).
There is evidence that generalized reciprocity increases
feelings of self-competence (Alvarez & van Leeuwen, 2015).
Generalized reciprocity in response to peer affirmations may
therefore promote positive clinical outcomes.

An interactive or media element has been excluded from

this version of the text. You can view it online here:
https://ohiostate.pressbooks.pub/swk3401/?p=599

Chapter Summary

This chapter identified what content belongs in which sections of an

empirical manuscript or professional report. Additionally, important
tips about writing professionally were presented, with specific
attention to using language free of bias and stigma. Writers were
repeatedly referred to the information presented in the APA (2009)
Publication Manual since the guidelines for writing in APA style are
applied prevalently in social work.

486 | Module 5 Chapter 1: Parts of an Empirical Manuscript or Report

Module 5 Chapter 2: Creating
Strong Presentations
Presenting evidence to audiences depends a great deal on the
audience. A room full of scientists will be looking for a different
emphasis in the information shared than an agency board, group
of practitioner colleagues, policy decision makers, or a community
group like the League of Women Voters, Parent Teachers Association,
or Kiwanis service club. However, all audiences want to be respected
with interesting, accurate, concise, and clearly presented
information. The major difference between preparing a written
report and a live presentation has more to do with style than content.
In this chapter, you will learn:

• Basic elements of strong presentations;

• PowerPoint® tips;
• Using images in presentations, including copyright issues and
locating open source materials;
• Presentation style tips.

Elements of Strong Presentations

One of the most important aspects of a strong presentation is its

organization—how content flows from start to finish. However, a
strong presentation requires appropriate and adequate preparation
up front. Here are six important tips (based on experience and tips
from http://thevisualcommunicationguy.com/2015/10/30/how-to-
give-a-powerful-presentation-eight-steps-to-an-awesome-
speech/ and https://www.entrepreneurship.org/learning-paths/
powerful-presentations).

Module 5 Chapter 2: Creating Strong

Presentations | 487
 1. Preparation. Effective presenters prepare for their specific
audiences. At the very least, they develop an understanding of
who the audience is, what the audience expects to learn from
the presentation, and how the presentation fits into the rest of
the day’s experiences. For example, a presenter would want to
know if their presentation is keeping the audience from lunch or
if it is following a meal that makes people sleepy; if it is being
presented at a critical, contentious time in the group’s process;
if are they being forced to attend or attending out of interest; if
it is one of many presentations in a day or stands alone.
Knowing the audience expectations is also important: are they
looking for the science behind the key points (emphasis on the
methodology) or the implications of the evidence? A strong
presentation starts with understanding why the audience wants
to hear what the presenter has to say instead of looking at their
smartphones or laptops as they politely (or impolitely) endure
the presentation.
Timing is another important preparation point. Attention tends
to wane after only a few minutes—is the presentation expected
to be a short 15-minute Ted-style talk or an hour-long lecture?
In either case, preparing to engage the audience at multiple
points is critical for maintaining attention, interest, and
learning. Having appropriate and interest-grabbing visual,
auditory, and/or action-oriented aids planned into the
presentation is important. Long gone are the days when
audiences tolerate presentations from lecturers who “wing
it”—they expect the respect of strong preparation. Realize that it
can take several hours to prepare a 15-minute presentation
when you have a great deal of expertise in the topic—it can take
days or weeks to prepare for one where you are developing the
expertise at the same time.

“It usually takes me more than three weeks to prepare a

488 | Module 5 Chapter 2: Creating Strong Presentations

 good impromptu speech.”
–Mark Twain

2. A strong opening. Audiences are most attentive in the first

minute or two of a presentation. This is the time to capture
attention and interest with a briefstory, anecdote, or fable to
make the topic relevant; a sensitive and appropriate joke; a
thought-provoking question; some “Wow!” statistics.
For example, which lecture would you rather hear in your child
development class:
– Theories of psychomotor development during infancy, or
– Why can’t the baby ride a bike?
The opening should be a form of “Wow!” statemen, one or two
sentences—never more than three (according to Kauffman
Founders School, www.entrepreneurship.org). You have learned
in this course how to locate powerful, informative, “oh, wow”
statistics and data—evidence-supported facts that grab
attention. This is also an option for a strong opening. Or,
perhaps a powerful quote—maybe from a client or study
participant, maybe from history or pop culture (but be very
certain about citing the person properly—there are many
misquoted and miscited quotations circulating on the internet).
Showing the audience an object or image is the way to capture
their attention. Pro tip: if you start with a story, be sure that you
end with the story’s resolution.

Module 5 Chapter 2: Creating Strong Presentations | 489

3. Integrate Visual and/or Auditory Elements. According to
thevisualcommunicationguy
(http://thevisualcommunicationguy.com/2015/10/30/how-to-
give-a-powerful-presentation-eight-steps-to-an-awesome-
speech/), people recall information better when an image is
attached to the message—almost twice as well as when the
message is delivered in text alone. The trick with visual tools is
to make them clear and simple. If the message is complex, build
from a simple visual toward a complex one. Here is an example
adapted from a lecture “Three Strikes: Female, In Jail, and
Substance Abusing” (Begun & Rose, 2007)—the images build
towards the final complex visual aid and the presenter is able to
discuss what is relevant to each sphere as it is added.

Images (pictures, charts, diagrams) may come from open access

resources, or you may have to create them specifically for your
message (see Chapter 3 of this module). Sources to consider

490 | Module 5 Chapter 2: Creating Strong Presentations

 include pexels.com, pixabay.com, and 123rf.com where all or a
limited number of images may be downloaded for free. These
are royalty-free sources where it is relatively easy to identify the
copyright conditions for use of the images; capturing images
from the wider internet or blogs or advertisements can be risky
with regard to potential copyright infringement. Note that
cartoons are usually copyright protected, as well. You want to
focus your search on objects with “Creative Commons” status
that has no restrictions or where your presentation meets the
restrictions in place. In this screenshot example, see what it
says under the CC0 License information on the far right: free for
personal and commercial use, and no attribution required.

If you decide to incorporate audio (or video) components to

your presentation, it is important to make sure the platform and
facilities where you will be presenting can support the
presentation. For example, ensuring sufficient connectivity if
the material is being pulled from the internet and not locally
embedded and that speakers and volume control are sufficient
for the facility. Some agencies where you might be asked to
present lack projection equipment and screens, as well.
4. Progress Logically Start to Finish. Follow an outline of key

Module 5 Chapter 2: Creating Strong Presentations | 491

 points and follow a consistent format for making each point.
Within the first minutes, it is important to tell the audience
what you are going to tell them: a statement of the presentation
goal and objectives is helpful. This provides the audience with a
roadmap for where you are taking them with this presentation.
In the opening you tell the audience what you are going to tell
them about. Next, in the middle, you tell them (the “meat” of the
sandwich). Here you provide “signposts” to cue them as to
where on the roadmap they are. Make clear when the topic has
shifted to a new point with transition cues (new slide headings,
numbered points, and transition statements). Next, see how to
finish the presentation “sandwich.”

5. Wrap up/Conclude. In conclusion, you summarize or remind

the audience of what you told them. This is not the place for
new content to be introduced. If you followed a story line
throughout your presentation, this is the place to deliver the
“punch line” (for a joke) or the story’s resolution. Or, it may be
the place for a new (brief) story with a different outcome. Or,
perhaps a (different) quote. It is the place to summarize the key
points if your purpose was to share information, or to identify
key action points if your purpose was to stimulate an action

492 | Module 5 Chapter 2: Creating Strong Presentations

 response. You might even end with a group “evaluation”
activity—a quick survey or quiz that everyone can now get
correct after hearing the presentation. This is effective if you
chose to open with a couple of “quiz” questions that they might
not know how to best answer or where their opinions could
change in the course of the presentation.

6. Prepare for Q&A. After a presentation concludes there may be a

formal period for question and answer by the audience. Prepare
for these questions—try to predict what they might ask and have
a response prepared. And, have a response ready for questions
that you did not anticipate and are not prepared to answer, too!

Tips for Presentation Slides

PowerPoint® is a commonly used tool for creating visual

presentations. Presentations created using this software can be
relatively simple or quite complexly structured. Suggestions for a

Module 5 Chapter 2: Creating Strong Presentations | 493

strong PowerPoint® presentation are presented in this example
“slide.”

Create the first 3 slides of a simple

PowerPoint presentation. The first
should be a title slide, the second a list
of either what the presentation covers
(if it were a full presentation) or
learning objectives for the audience to
achieve from the presentation, and the
third should include an image (it is the
first slide of your information content).
Assess your presentation slides for
clarity, design, and succinct writing.
Suggested topic: the most useful knowledge you have learned from
this course.

494 | Module 5 Chapter 2: Creating Strong Presentations

Chapter Summary

In this chapter we examined the elements of strong presentations

and some tools to help create and deliver strong presentations to
various audiences. You were introduced to six steps for developing
strong presentations, tips for using visual aids. Copyright concerns
and resources for avoiding them were identified. Next, you will learn
tips for creating your own visual aids (figures graphs) when just the
right thing is not readily available.

Module 5 Chapter 2: Creating Strong Presentations | 495

Module 5 Chapter 3: Creating
Figures and Graphs
Just as you learned in Module 4, information is often better
understood by audience when presented in a table, figure, or graph
compared to lengthy verbal descriptions. Previously, you learned how
to interpret different types of graphs; in this chapter you are
introduced to building them using Word®, PowerPoint®, and Excel®
software. In this chapter you will learn:

• Simple tools for creating figures;

• Tools for creating graphs.

Creating Figures

Word® and PowerPoint® have tools available for creating basic

figures or diagrams. For example, the “Three Strikes” Venn
diagram presented in Chapter 2 was created using the SmartArt
design feature in Word®.

496 | Module 5 Chapter 3: Creating

Figures and Graphs
In the Word® top menu bar is a tab called “Insert.” If you select this
tab, then select “SmartArt” from the next menu, an array of options
becomes available which are preformatted for making simple figures.
It is also possible to use a combination of “Shapes” from the menu to
create your own figure, but this can be more challenging, time and
effort consuming, and difficult to format.

Another “Insert” option for creating your own figures is the “Icons”
menu. Here is an example of the icons available under the heading of
“people.”

Module 5 Chapter 3: Creating Figures and Graphs | 497

 Follow the steps outlined for creating a
simple figure using SmartArt. Some
suggestions: a hierarchy diagram of
who lives in your household (including
pets), a process diagram explaining
your current mood, a relationship
diagram depicting your various
interests, or other ideas. See what you
can create on your own.

Creating Graphs

In the Word® “Insert” menu is an option called “Charts.” This option

provides tips and tools for creating the kinds of charts and graphs we
studied in Module 4: bar charts, pie charts, scatterplots, histograms,
and line charts for example. If you need charts or graphs in a
PowerPoint® presentation, you can also create them the same way
or you can create them in Word® and then perform a “copy & paste”
operation to place them in a PowerPoint® slide. Furthermore, you
can create a graph in Excel® and copy it into a Word® or
PowerPoint® document, as well. The three programs are compatible
in this way. In addition to these types of quantitative charts and
graphs, you are introduced to one additional quantitative type
(bubble charts) and one for demonstrating more qualitative types of
information (word clouds).

Basic Steps. The steps for creating graph or chart are, basically,
selecting “Insert” and “Charts” from the toolbar menus, select the

498 | Module 5 Chapter 3: Creating Figures and Graphs

type of graph or chart you wish to create, enter the data for the chart
or graph in an Excel® spreadsheet that opens for this purpose, then
fine tune the “look and feel” features of the chart or graph (labels,
fonts, color scheme, proportions, and more). Here is a screen shot
of the options for types of graphs or charts available in Word® and
PowerPoint®.

Pie Chart

A pie chart is used to represent univariate distribution (proportion)

on a categorical variable: the proportion out of 100% (the full pie)
of individuals in each group on that variable. In our hypothetical

Module 5 Chapter 3: Creating Figures and Graphs | 499

example of children living in rental versus family owned homes (n=50
rental, n=128 family owned, from the “count” Excel workbook exercise
in Module 4), the pie chart would be created in the following manner.

Step 1. Open the “Insert” and “Chart” menu, select “Pie” chart option
and click “OK.”

Step 2. A small Excel spreadsheet box opens for you to enter the
relevant information. The “Sales” column is waiting to be overwritten
with your data; the 1stQtr, 2ndQtr, 3rdQtr, and 4thQtr are also place
holders waiting to be overwritten with your data. Replace the number
8.2 in cell B2 with the number 50, and the cell B3 number 3.2 with
128. Then delete the rows for 3rdQtr and 4thQtr since we only had 2
categories in our variable.

Step 3. Change the place holder label “Sales” to the variable name you
want to have showing in your chart (try Family Housing) and the place
holder labels 1stQtr and 2ndQtr with the category names you want to
have showing (try rent and own).

Step 4. Close the Excel box by clicking on the X in the upper right
corner—this puts the data entry routine in the background but you
will not lose your work.

Step 5. Now is the time to adjust the appearance of the pie chart.
Right now, it is unclear to an audience what they are seeing. You can
start by clicking on the “legend” which is the area of the chart where
the color for rent and the color for own are being defined. This should
open a menu to the right of the chart with four symbols.

• The second symbol, a + sign in a box allows you to manipulate

the labels. Select each option to see what it changes in your
chart—”data labels” inserts the values of the data you entered.
• The third symbol, a paintbrush in a box allows you to manipulate
the design and color scheme of your pie chart.

500 | Module 5 Chapter 3: Creating Figures and Graphs

 • The fourth symbol, that looks like a funnel in a box allows you to
manipulate more about the appearance.

Additionally, if you right click on one of the pie segments, you should
see a drop-down menu with one of the options being “format data
labels.” This allows you to select the “percentage” as what shows in
the chart instead of the actual raw numbers. Selecting this should
change the 50 and 128 to 28% and 72%. You could also delete the
legend and check “category name” to make the chart easier to
interpret. Then, you can double click the name and number in a slice
of the pie chart to edit the font (different font size, bold, different
color, different font) using the regular Word tools. You can also drag
the box around to place it where you like.

Another option is to work with the toolbar menu running across the
top of the screen once you have clicked on the chart (it will appear
within a box). The Chart Tools” menu is what you want to work with
in design or format. Note that there is also an option to edit the data
(this will reopen that Excel box) or even to change the chart type if
you decide a different type is preferable (a bar chart, for example).

Here is an example of what it might look like when you are

finished—many options exist. You want to make the chart as easy to
interpret as possible.

Module 5 Chapter 3: Creating Figures and Graphs | 501

 Follow the steps outlined for creating a
pie chart for the children from rental or
family owned homes. See what you can
create on your own.

Bar Chart

A bar chart, like the pie chart, is used to represent univariate

distribution (frequency) on a categorical variable: the number
(frequency) of individuals in each group on that variable. When the
frequencies do not represent a total 100% in proportions, a bar chart
is more easily interpreted than a pie chart; it is also easier to interpret
a bar chart if there are more than a few categories involved–wedges
of the pie might become very small and too numerous to understand
easily. For example, according to the Henry J. Kaiser Family
Foundation (www.kff.rg/other/state-indicator/distribution-by-
age/) the U.S. population census can be reported in terms of age
groups: children aged 0-18 years (24%), adults aged 19-25 years (9%),
adults aged 26-34 years (12%), adults aged 35-54 years (26%), adults
aged 55-64 years (14%), and those aged 65 and older (15%) as of 2017.
The percentages total 100%, so it could be presented as either a
pie or bar chart. If, however, the data were in the form of numbers

502 | Module 5 Chapter 3: Creating Figures and Graphs

(not percentages/proportions), a bar chart would be preferable
regardless of there being more than a few categories represented,

A bar chart of this age group information would be created in the

following manner.

Step 1. Open the “Insert” and “Chart” menu, select “Column” chart
option and click “OK.” Note: the “bar chart” option is a sideways
presentation.

Step 2. A small Excel spreadsheet box opens for you to enter the
relevant information. The “Series” columns are waiting to be
overwritten with your data. The series would be the 6 age groups.
Categories would be relevant if you had the same series in multiple
years or multiple locations, for example (1997, 2007, and 2017
perhaps). In our example, we have only the one category, so you could
begin by deleting the rows for category 2 through 4 from the Excel
spreadsheet.

Step 3. Enter the data into the Excel spreadsheet. Cell B2 would
be the percent for children aged 0-18 (24%), C2 is 9%, D2 is 12%,
E2 is 26%, F2 is 14%, and G2 is 15%. Change the place holder label
“Category 1” with the variable name (try Percent U.S. Population, 2017)
and the “series” place holder labels with category labels (try 0-18 yrs,
19-25 yrs, and so forth).

Step 4. Close the Excel box by clicking on the X in the upper right
corner—this puts the data entry routine in the background but you
will not lose your work.

Step 5. Now is the time to adjust the appearance of the bar

chart—right now, it is a little confusing to an audience. You can start
by clicking on the chart itself. This should open the “Chart Tools”
menu on the upper tool bar, as well as the four symbols to the right
of the chart.

Module 5 Chapter 3: Creating Figures and Graphs | 503

 • The second symbol, a + sign in a box allows you to manipulate
the labels. Select each option to see what it changes in your
chart—”data labels” inserts the values of the data you entered
and moving the legend to the right might help. The “format data
series” option allows you to alter the spacing between the bars
(columns) which can make the chart a bit more interesting.

Another option is to work with the toolbar menu running across the
top of the screen once you have clicked on the chart (it will appear
within a box). The Chart Tools” menu is what you want to work with
in design or format. Note that there is also an option to edit the data
(this will reopen that Excel box) or even to change the chart type if
you decide a different type is preferable (a bar chart, for example).

Here is an example of what it might look like when you are

finished—many options exist. The goal is to make the chart as easy to
interpret as possible.

A word of caution about bar charts: their interpretation is influenced

by the scale of the y-axis. In other words, if you increase the spread

504 | Module 5 Chapter 3: Creating Figures and Graphs

of the values on the y-axis, it makes even small differences between
columns seem impressive; if you compress the spread of values, it
makes the columns seem more similar. This point is made in images
from a book called How to Lie with Statistics (Huff & Gels, 1954); here
is the authors’ example demonstrating that the very same numbers
(a difference of 2 billion dollars) can be interpreted as “great” or
“alarming” news. Imagine the y-axis is about billions of dollars spent
on [insert any category of spending you like here] during different
months of a single year; the x-axis reflects which month the data
represent. The chart on the left uses a scale where each increment
represents 2 billion dollars. The movement from 20 billion to 22
billion does not look like very much; the line is only slightly angled
(Huff & Gels, 1954, p. 61). The chart on the right uses a scale indicating
0.2 billion dollars for each increment (Huff & Gels, 1954, p. 63). The
movement from 20 billion to 22 billion looks like a whole lot; the line
is very dramatic. Yet, both charts are representing the same 2 billion
dollar difference! So, the moral of this story is that scale matters in
graphing.

Module 5 Chapter 3: Creating Figures and Graphs | 505

 Follow the steps outlined for creating a
bar chart for the population by age
group. See what you can create on your
own.

Histogram

In Word® and Powerpoint®, histograms essentially are column/bar

charts that include many columns—they are univariate description
tools just like the bar chart. As previously explained, the histogram
depicts the frequency with which each value on an interval
(continuous) variable appears in the data. With enough data points,
it is possible to assess how close to normal the distribution appears.
Turning to the student absenteeism data with which we worked in
Module 4, these are the steps involved in creating a histogram using
Word.

Step 1. Open the original data file and sort the data by the variable
of interest (days absent). Then, copy the column of data for that
variable. This step would not be engaged if you are entering the data
from “scratch” into the Excel® spreadsheet in the histogram creation
routine.

Step 2. Open the “Insert” and “Chart” tabs from the Word® menu bar.

506 | Module 5 Chapter 3: Creating Figures and Graphs

Replace the “series” placeholder data with the data copied from your
Excel® data file. Or, replace the “series” placeholder data with the
data you are entering from “scratch.”

Step 3. Now you will specify the increments that you want included
in each column of the histogram. The “placeholder” increments are
“chunking” your data in increments of 12. However, you want a more
refined picture. To change the increments, double click on the area of
the graph where [0,12] (12,24] and so on appear (horizontal axis). This
should open a “format axis” menu to the right of your work. In Excel®
the increments are called “bins” and you want to adjust the bin width.
If we change the 12.0 in the Bin width box to 1, the chart shows the
frequency with which each value for number of days appears among
our 28 students—since the lowest value was 0 and the highest value
was 39, the result is a total of 49 bins (increments) in our chart. This
seemed to be confusing because of that one outlier at 49 days absent.
Checking “overflow bin” in the format axis menu and changing the
default value (38) to 27 retains that one student in the picture but
doesn’t spread the axis out as much: the student is in the >27 bin (all
values greater than 27). Then, the format axis menu can be closed.

Step 4. The final steps relate to the chart’s appearance. Double

clicking on the Chart Title placeholder allows overwriting of those
words with a meaningful title. It also opens a Format Chart Title
menu where color, position, and other aspects can be modified.
Clicking on the chart opens the three symbols menu where axis titles
may be added.

Here is an example of how a histogram of these data created in

Word® might appear.

Module 5 Chapter 3: Creating Figures and Graphs | 507

Another option is to create the histogram in Excel® and perform a
copy & paste operation into the Word® document or PowerPoint®
presentation slide.

Follow the steps outlined for creating a

histogram for the student absenteeism
data (located in the Excel file called
student absenteeism start.xlsx). See
what you can create on your own.

508 | Module 5 Chapter 3: Creating Figures and Graphs

Scatterplot

As previously discussed, a scatterplot is used to visually depict the

relationship between two variables (bivariate). Using a hypothetical
set of data from 50 undergraduate social work students in a research
and statistics course, a scatterplot is created depicting the
relationship between their predictions about their course grades
generated at the start of the course and their actual final course
grades.

Step 1. Open original data file (unless entering data from “scratch”
into the chart development routine); copy each column and replace
the “placeholder” values in the “Scatterplot” Excel® spreadsheet that
opens with “Insert” and “Chart” selected from the toolbar.

Step 2. Double click on the chart to access the formatting tools—for

example, the format axis tool allows you to adjust the “y” axis values
(“Number” option) to remove the irrelevant decimal place; it allows
you to adjust the scale of the “y” axis, as well. Not only does it
allow adjusting of the scale (e.g., 10-point increments versus 20-point
increments), it allows deletion of the bottom or top range where
there are no values in the data—instead of starting at 0.0 (no one
predicted a score this low), the scale can start at 40 instead. Instead
of extending to 120 (the highest possible value was 100%), it can end
at 100.

Step 3. Formatting the “x” axis allows similar functions. These

formatting tools also allow insertion of labels for the “x” and “y” axis.
Changing the chart title is likely to be necessary. If the axes are
labelled, the “legend” can be deleted.

Here is how a scatterplot of these data might appear.

Module 5 Chapter 3: Creating Figures and Graphs | 509

Line Graph

In many instances the point that a presenter wishes to make relates

to a trend in data, for example how some feature has changed over
time. Here is a modified line graph with two lines showing the trend
in income shares between 1980 and 2016 (see the World Inequality
Report, Alvaredo et al., 2018, p. 8). In this chart, you can quickly see
that households in the bottom 50% of income level have declined in
their share of income and households in the top 1% have climbed. In
other words, the income inequality gap increased considerably.

510 | Module 5 Chapter 3: Creating Figures and Graphs

These charts can be created using the same “Insert” and “Charts”
menu we have been working with so far. The choice is “Line” in the
menu of options. The data are inserted into the drop-down Excel®
worksheet, just as in the previous chart examples. Editing the
features of the chart are the same as in previous examples, as well.
The income inequality line graph shows an interaction between the
two lines, one for each group. Line charts can be as simple as tracking
one variable at two time-points or can be very complex, showing
a trend for multiple groups. Here is an example from the National
Survey on Drug Use and Health (SAMHSA, 2017). This figure depicts
the percent of adults receiving different forms of mental health
services each year from 2008 to 2017. The legend at the bottom shows
the different types of services—a separate line is presented for each.
Not only can you see the trend for each type, you can see the types in
relation to each other (e.g., inpatient services are less common than
prescription medicine).

Module 5 Chapter 3: Creating Figures and Graphs | 511

Bubble Chart

A type of chart with a different emphasis is called a bubble chart.

It is basically a type of scatterplot, but in a bubble chart the “dots”
(bubbles) vary in size. The size of the bubbles represents an added
dimension of information. As in the scatterplot, we still have an x-axis
and y-axis. In a bubble chart, there is a 3rd dimension (variable) in
the diagram. The 3rd dimension is represented by the size (area)
of the bubble, or sometimes by colors. Bubble charts do not work
if there are too many bubbles (cases) represented (see
https://datavizcatalogue.com/methods/bubble_chart.html).
Microsoft hosts a site where you can learn more about how to use
Excel to create a bubble chart: https://support.office.com/en-us/

512 | Module 5 Chapter 3: Creating Figures and Graphs

article/present-your-data-in-a-bubble-
chart-424d7bda-93e8-4983-9b51-c766f3e330d9

Here is an example of a bubble chart based on National Geographic

data, posted on the website www.aploris.com/blog/charts/
category/bubble-chart/. The x-axis represents the percent of
population in each country who lack access to water from pipes or
well protected from contamination. The y-axis represents the GDP
(gross domestic product, indicator of national wealth) in billions of
U.S. dollars for each nation in the sample. The size of each bubble
reflects the number of persons affected in each nation (this is
different from percent and reflects a 3rd variable).

You can see that relatively wealthy nations, like the U.S. have very
small bubbles far on the left of the graph (very small percent of
population affected). Four African nations have the largest percent of
population without access to water from protected sources (Congo,
Tanzania, Ethiopia, and Nigeria) and these are also nations in the
lower quadrant which indicates they have lower GDP. Nigeria has
a larger sized bubble than Tanzania, for example, indicating that
the problem affects a larger number of persons in the nation, even
though a larger percent of the Tanzanian population is affected. The
“legend” on this chart shows the size of a bubble representing 60
million persons, allowing for comparisons on this dimension. The size
of the bubble for the U.S. is very small compared to this 60 million
persons bubble, whereas China and India have larger bubbles, even
though their percent of affected population is near the U.S. value.

Module 5 Chapter 3: Creating Figures and Graphs | 513

Word Cloud

You have seen them: Word Clouds. A word cloud is a graphic image
composed of a cluster of words that fit together under a general
theme or topic. A resource for learning to create them using Word®
software with an add-in from Microsoft Office Store (created by
Orpheus Technology, called Pro Word Cloud) is available at
https://www.youtube.com/watch?v=my1JRX84tyc, and there exist
several free software programs for creating them, as well. They also
can be created in PowerPoint® with an add-in, or you could create
one in Word® that you then copy and paste into a presentation slide.

A word cloud begins with a set of words. These can be entered as a

list or a body of text can be copied into the program. They can be
created with frequency of each word’s appearance determining its
positioning and prominence. Common words like “a,” “the,” “and,” can
be ignored in the options, and other features can be modified (color,
orientation, and others).

Here is an example of a word cloud used as an online textbook cover

for a different course.

514 | Module 5 Chapter 3: Creating Figures and Graphs

 Take a moment to complete the
following activity.

An interactive or media
element has been excluded
from this version of the text. You can

Module 5 Chapter 3: Creating Figures and Graphs | 515

 view it online here:
https://ohiostate.pressbooks.pub/swk3401/?p=601

Chapter Summary

In this chapter you were introduced or re-introduced to 7 types

of visual graphs and charts that are useful in communicating
information to an audience. Not only did you refresh your skills at
interpreting them yourself, you also learned how to set about
creating them. These tools can be easily embedded in a written
report or presentation. The next chapter describes another
communication tool gaining in popularity for communicating
powerful messages: the infographic.

516 | Module 5 Chapter 3: Creating Figures and Graphs

Module 5 Chapter 4: Creating
Infographics
Unlike the research manuscript, professional report, or presentation,
an infographic presents data and implications in a brief but impactful
format. The challenge involves making complex information easily
understood by diverse audiences. Infographics can be incorporated
into presentations, but are often free-standing tools. As the
producers of the venngage blog suggest, “they also make dry
information more engaging” (https://venngage.com/blog/create-
infographics). As social work professionals, you may elect to present
empirical evidence in this format, depending on the audience and
purpose of the communication effort. In this chapter you will learn:

• What infographics are and when they might be useful;

• Steps in creating an effective infographic;
• Tool for developing infographics;

Infographics Described

An infographic is a brief informational graphic tool (usually 1 page or

computer screen shot). According to the venngage blog introduction
to infographics:

They are often used to visualize data, step-by-step processes,

timelines, research findings, organizational hierarchies, and
many other types of information (https://venngage.com/
blog/create-infographics).

Infographics tend to be more engaging to audiences than lengthy

text—like reading a book with pictures often is more interesting than

Module 5 Chapter 4: Creating

Infographics | 517
reading a book with only words. Imagine reading Hawthorne’s The
Scarlet Letteras a high school student with pictures to make the
characters and their situations come alive for you, compared to
reading the text alone. This helps explain the popularity of graphic
novels.

Infographics tap into a fact long recognized in psychology: images are

more easily interpreted and recalled than the written word (Paivio,
1971). For example, thenextweb blog presented the example below
where you can quickly see how costs (of being a superhero) exploded
between 1939 and 2013. Imagine if you had to read through this
information in text form?

https://thenextweb.com/wp-content/blogs.dir/1/files/2013/10/
batman-then-and-now.jpg

518 | Module 5 Chapter 4: Creating Infographics

Module 5 Chapter 4: Creating Infographics | 519
Steps in Creating an Infographic

The following presents 4 steps in creating an infographic for

communicating with your audiences.

Step 1. Translate your topic into a burning question that interests

your audience. As an example, venngage blog
(https://venngage.com/blog/create-infographics) highlights the use
of an infographic by the Bill and Melinda Gates Foundation to
generate interest in a topic of little interest to many people in
Northern climates: malaria. They framed it in terms of the world’s
deadliest animals—if you are a fan of “shark week,” have a fear of
crocodiles, snakes, spiders, or creatures, or are concerned about
interpersonal violence rates, this infographic might peak your
interest.

520 | Module 5 Chapter 4: Creating Infographics

Module 5 Chapter 4: Creating Infographics | 521
Step 2. Gather the information to be presented in the infographic.
Most likely, you will collect more information than is needed. You will
need to be selective and/or creative in how you combine different
pieces of information. However, it is also critically important that
you maintain truthfulness in your selection and presentation of
information. You also need to cite your sources!

Step 3. Design the infographic. One feature to keep in mind as you

design an infographic is the medium through which it will be
distributed. If it will be presented in print, you may have the luxury
of using both sides of a page; if a poster or electronic “screen” is
the medium, you are limited to one side/one page. Various software
packages are available, many for free, to help in the design process
(e.g., Canva Infographic Maker; Easel.ly; Piktochart; Snappa;
Venngage; Visme; Visualize). Most of these are free to use in their
“base” form; advanced options are available with a fee. Another option
is to design the tool “from scratch” using either Word® or
PowerPoint® with or without design templates. Remember to be
concise, precise, and accurate in what you write. Consider image and
graph/chart tools previously discussed, as well.

Step 4. Produce the infographic in your selected medium (print,

electronically, or both) and disseminate your information to your
audience.

522 | Module 5 Chapter 4: Creating Infographics

 Spend some time in a “scavenger hunt”
viewing and critiquing infographics
that others have developed. Consider
what seems to work and what does not
for the different examples you locate.
Use your observations to generate a set
of guidelines for developing your own
infographics (e.g., layout, color, design,
content, writing style).

Chapter Summary

Infographics as a powerful communication tool were introduced. You

learned a bit about why they are effective means of communicating
with diverse audiences, and you learned 4 steps in their production.
In addition, several free tools were identified for assisting in their
design.

Module 5 Chapter 4: Creating Infographics | 523

Module 5 Chapter 5: Course
Conclusions
This concludes the reading activities for the first course in our two-
course sequence, SWK 3401. In these five modules you covered a
great deal of information related to the course title: Research &
Statistics for Understanding Social Work Problems and Diverse
Populations. Here is a brief synopsis of what you learned in each
module.

Module 1.The Importance of Research to Social Work Practice; Research Ethics

The first course module provided an orientation to the important

relationship research and evidence play in social work practice and
promoting social justice, including the profession’s mandates
regarding the place for research in social work. This orientation
analyzed different types and sources of knowledge (“ways of
knowing”) used by social workers in developing an understanding of
social work problems, diverse populations, and social phenomena.
It also provided a context for understanding the place of research
and empirical evidence in the wider realm of knowledge we use. The
module presented principles of critical thinking engaged by social
workers, and presented key distinctions between science,
pseudoscience, and opinion. Within this orientation, the
philosophical roots of different approaches to evidence were
identified (positivism, empiricism, and ethnoepistemology in
quantitative and qualitative research), as were relationships between
theory, research, and practice. Strategies for locating sources of
empirical evidence were explored in this first module. Finally, a great

524 | Module 5 Chapter 5: Course

Conclusions
deal of attention was directed to the ethical conduct of research and
important considerations for research involving human participants.

Module 2. Identifying and Applying Evidence

Our second course module extended what was learned in the first
by developing a framework for understanding different types of
research questions and study specific aims. The distinction between
research questions and research hypotheses was explained, as were
differences between background/foreground, exploratory,
descriptive, and explanatory research questions. The relationship
between types of questions and scientific approach was addressed. A
translational science framework explained differences between basic,
intervention, and implementation science, as well as “research for
and about research.” These categories were developed as strong
influences on all subsequent decisions made by investigators. This
module analyzed the role of theory in research and identified ways
of identifying and assessing empirical evidence regarding social work
problems, diverse populations, and social phenomena. Furthermore,
a great deal of emphasis was placed on evaluating empirical literature
and distinguishing between empirical and non-empirical literature.
In the course of completing Module 2, you were introduced to
working with Excel spreadsheets.

Module 3. Research Approaches, Designs, and Methods

Our third course module extended from the background presented

in Modules 1 and 2 to explore specific approaches, designs, and
methodologies used in qualitative, quantitative, and mixed methods
research. The relationship between research questions and research
approaches was elaborated in this module with an overview and
details for each approach (qualitative, quantitative, mixed methods).
Six major traditions in qualitative research were explored: narrative,

Module 5 Chapter 5: Course Conclusions | 525

phenomenological, grounded theory, ethnographic, participatory
action, and content analysis research. Design issues of
generalizability, internal and external validity, cross-sectional and
longitudinal study were examined with regard to quantitative
research traditions. Distinctions between descriptive, correlational,
and experimental research were identified, as was the distinction
between correlation and causality. Delving deeper into quantitative
methods, a great deal of attention was directed to recognizing
different types of variables and their implications for study design,
measurement, and analysis (demographic, independent, dependent,
categorical/nominal, ordinal, interval/continuous). Measurement
issues in quantitative science were addressed, including
measurement validity and reliability, as well as cultural competence
in measurement approaches. Different strategies for collecting
qualitative and quantitative data were explored: naturalistic
observation, Photovoice, artifact/content analysis, administrative
and secondary data, geographic information systems (GIS), key
informant information, surveys, clinical screening or assessment
tools, interviews, focus groups, journaling, ecological momentary
assessment (EMA), concept maps, and social network analysis. The
final methodology chapter was devoted to study participants.
Distinctions were made between samples and populations, and key
issues related to the recruitment and retention of study participants
were explored. Not only were matters of sample size and diversity
addressed, strategies for participant selection were analyzed
(random selection as probability sampling, convenience and snow-
ball selection as non-probability sampling). In the course of
completing Module 3, your introduction to working with Excel
spreadsheets and data files was extended.

Module 4. Understanding Descriptive and Inferential Statistics

The fourth module in this course was devoted to approaches used

526 | Module 5 Chapter 5: Course Conclusions

in analyzing data collected from the methods described in Module
3—qualitative, quantitative, and mixed methods analyses. In the
chapter concerning qualitative analysis, you were introduced to data
preparation, field notes, coding, coding confirmation, and cross-
checking coding decisions. The subsequent chapters introduced
quantitative analysis strategies and issues. These included
distinguishing between univariate and bivariate analyses, use of
inferential statistics, the role and interpretation of a null hypothesis,
Type I and Type II error, and different types of analyses related to
the nature of the variables and the research questions. You learned
to conduct and interpret univariate descriptive analyses involving
the frequency, proportion, mean, median, mode, variance, standard
deviation, probability, and normal distribution of data. In addition,
you learned to conduct and interpret five types of bivariate analyses:
one-sample t-test, independent samplest-test, one-way analysis of
variance (Anova), chi-square, and correlation analysis. The idea
behind non-parametric analysis and two examples were also
developed in this module. In the course of studying this content,
you learned to conduct different types of analyses using Excel
spreadsheets and data files.

Module 5. Presenting Evidence

By now you may recognize that the course has been structured,
loosely, along the structure of an empirical report: an introduction
(Module 1 and Module 2), Methods (Module 3), Results (Module 4),
and Discussion (Module 5). Module 5 itself concerned ways that social
workers communicate with various types of audiences about
empirical evidence and scientific data concerning social work
problems, diverse populations, and social phenomena. The module
discussed important aspects of empirical manuscripts and
professional reports, features of strong presentations, creating
figures (graphs and charts), and creating infographics. With this

Module 5 Chapter 5: Course Conclusions | 527

concluding chapter of Module 5, you are now well-prepared to
succeed in meeting the goals of the second course in our two-course
sequence. In SWK 3402 your skills and knowledge will be applied
to understanding social work interventions. Together the content in
these two courses prepares you to engage effectively with evidence
in social practice, regardless of the nature of practice in which you
engage.

What do you believe was the single best

thing you learned about in this course? Consider sending your
answer to your instructor and/or including it in your course
evaluation comments.

528 | Module 5 Chapter 5: Course Conclusions

Module 5 Key Terms and
Definitions
abstract: a brief, comprehensive survey summarizing contents of an
empirical manuscript or professional report, presented at the front of
a manuscript.

bar chart: a figure used in graphically reporting univariate statistics

for a categorical variable (frequency count for each category
represented in a bar or column).

bubble chart: a form of scatterplot chart (two variables) with a third

variable reflected in the size of the data points (dots or bubbles).

citation: identification of the source of information used in report,

manuscript, or presentation, allows audiences to evaluate the
information and/or locate the original source for themselves.

confused verb tense: a problematic writing practice where two or

more different verb tenses are used in the same sentence, clause, or
paragraph.

Creative Commons: a public copyright license whereby the creator of

a freely distributed work specifies how that work can be shared, used,
or modified by others.

discussion: final section in the body of an empirical manuscript or

professional report where results are briefly summarized,
conclusions and interpretations are presented, strengths and
limitations are analyzed, and implications of the results are
presented.

Module 5 Key Terms and

Definitions | 529
findings: like study results, the findings might be presented in a
professional report.

histogram: a figure used to graphically represent the frequency with

which each value for a variable is represented in the data; reflection
of data distribution for a variable.

infographic: a clear, concise visual tool for demonstrating a key point

of information, data, or knowledge.

initial or executive summary: similar to an abstract, a summary of the

contents of a professional report presented at the front of the report.

introduction: opening section of an empirical manuscript or

professional report that introduces the topic/problem, surveys and
critiques existing literature, and states study aims/research
questions/hypotheses.

line graph: a figure used to graphically represent the relationship

between two variables (the “x” and “y” axes), often used to depict
trends over time (time as the “x” axis) or other demonstrations of
change.

methodology: detailed description in an empirical manuscript or

professional report that describes how a study or project was
conducted and sources/methods of data collection; includes
descriptions of study approach/design, study participants, and
measurement tools/procedures.

parallel construction: a writing technique where two or more

sentences, clauses, phrases, paragraphs, or sections have the same
grammatical structure.

paraphrase: using different words to reflect the meaning of someone

else’s ideas, writing, or verbalized speech, usually to either clarify or
summarize what was originally stated.

530 | Module 5 Key Terms and Definitions

pie chart: a figure used to graphically represent the univariate
statistic of proportions for a categorical variable (proportion for each
category represented in a “slice” or “wedge” of the pie).

plagiarism: using someone else’s words or ideas without properly

identifying them as belonging to that person rather than being one’s
own.

put people first: a “use of language” practice using descriptors that

begin with the person rather than a condition, circumstances, or
label.

recommendations: part of the final discussion in a professional report

where conclusions are offered in terms of advisable actions or next
steps based on the project findings.

reference: full information about a resource, usually one cited in a

manuscript or report, providing sufficient detail that readers can
evaluate the source and/or locate the source for themselves.

results: section of an empirical manuscript or professional report

where data analysis procedures and their outcomes are described.

scale: the units used in a graph, chart, or other figure where the
distance between points has systematic meaning in numerical terms.

scatterplot: a figure used to graphically represent the relationship

between two variables (the “x” and “y” axes), often used to depict the
strength and direction of their correlation or association.

SmartArt: a tool available in Word® and PowerPoint® with ready-

made templates for creating figures.

Venn diagram: a type of figure depicting the existence or extent/

degree of overlap between different groups, constructs, themes,
ideas, or events where what is shared in common is represented

Module 5 Key Terms and Definitions | 531

in the area of overlap; usually constructed from circles, ovals, or
spheres, but may be formed from other geometric shapes.

532 | Module 5 Key Terms and Definitions

Module 5 References
Alvaredo, F., Chancel, L., Piketty, T., Saez, E., & Zucman, G. (2018).
World inequality report: Executive summary.
https://wir2018.wid.world/files/download/wir2018-summary-
english.pdf

Anderson-Butcher, D., Wade-Mdivanian, R., Davis, J., Newman, T., &

Lower, L. (2016). 2015-2016 Ohio State’s LiFE Sports annual review.
Columbus, OH: Ohio State University College of Social Work.

Begun, A.L. (2016). Considering the language we use: Well worth the
effort. Journal of Social Work Practice in the Addictions, 16(3),
332-336.

Begun, A.L., Early, T.J., & Hodge, A. (2016). Mental health and
substance abuse service engagement by men and women during
community reentry following incarceration. Administration &
Policy in Mental Health & Mental Health Services Research, 43(2),
207-218.

Begun, A. L., & Rose, S.J. (2007).Three strikes: Substance abusing,

female, and in jail.University of Wisconsin-Milwaukee Morris
Fromkin Memorial Lecture, Milwaukee, WI.

Broyles, L.M., Binswanger, I.A., Jenkins, J.A., Finnell, D.S., Faseru, B.,
Cavaiola, A., … Gordon, A.J. (2014). Confronting inadvertent stigma
and pejorative language in addiction scholarship: A recognition and
response. Substance Abuse, 35, 217-221.

Bunger, AC, Robertson, H, Hoffman, J, Cao, Y. (June 2015). Connecting

youth in child welfare to behavioral health services in Franklin
County: Provider perspectives. Evaluation report written for
Franklin County Children Services (OH). Columbus, OH.

Module 5 References | 533

Doogan, N.J., & Warren, K.L. (2017). Saving my life: Dynamics of peer
and staff corrections among therapeutic community residents.
Substance Use & Misuse, 52(11), 1429-1438.

Huff, D., & Gels, I. (1954). How to lie with statistics.NY: W.W. Norton &
Co.

Johnson-Motoyama, M., Moses, M., Kann, T. K., Mariscal, E. S., Levy,

M., Navarro, C., & Fite, P. (2016). Parent, teacher, and school
stakeholder perspectives on adolescent pregnancy prevention
programming for Latino youth. Journal of Primary Prevention, 37(6),
513-525.

Kagotho, N., Nabunya, P., Ssewamala, F., Mwangi, E. N., & Njenga,
G. (2017). The role of family financial socialization and financial
management skills on youth saving behavior.Journal of Adolescence,
59, 134- 138.

Lee, M.Y., Wang, X., Liu, C., Raheim, S., & Tebb, S. (2018). Outcome
literature review of integrative body-mind-spirit practices for
mental health conditions. Social Work Research, 42(3), 251-266.

Maguire-Jack, K., & Campbell, T. (2015). Evaluation of Marathon

County Community Response Program: Report to Marathon
County Department of Social Services. Columbus, OH: Ohio State
University College of Social Work.

McInroy, L.B. (2017). Innovative ethics: Using animated videos when

soliciting informed consent of young people for online surveys.
Social Work Research, 41(2), 121-128.

Mendoza, A.N., Fruhauf, C.A., Bundy-Fazioli, K., & Weil, J. (2018).

Understanding Latino grandparents raising grandchildren through
a bioecological lens. International Journal of Aging and Human
Development, 86(3), 281-305.

534 | Module 5 References

Mengo, C., & Black, B.M. (2015). Violence victimization on a college
campus: Impact on GPA and school dropout. Journal of College
Student Retention: Research, Theory & Practice, 18(2), 234-248.

Mengo, C.W., Lee, J., Bolton, K., Lehmann, P., & Jordan, C. (2017). Self-
identified strengths among youth offenders charged with assault
against a non-intimate family member. Journal of Human Behavior
in the Social Environment, 27(8), 881-893.

Paivio, A. (1971). Imagery and verbal processes. NY: Holt, Rinehart, &
Winston.

Parekh, R., Maleku, A., Fields, N., Adomo, G., Schuman, D., &
Felderhoff, B. (2018). Pathways to age-friendly communities in
diverse urban neighborhoods: Do social capital and social cohesion
matter?Journal of Gerontological Social Work, 61(5), 492-512.

Quinn, C.R., Liu, C., Kothari, C., Cerulli, C., & Thurston, S. (in press).
Using social determinants of health to assess psychological distress
and suicidal thoughts and behavior among youth on probation.
Adolescent Psychiatry, 7.

Substance Abuse and Mental Health Services Administration

(SAMHSA). (2018). Key substance use and mental health indicators
in the United States: Results from the 2017 national Survey on
Drug Use and Health (HHS Publication No. SMA 18-5068, NSDUH
Series H-53). Rockville, MD: Center for Behavioral Health Statistics
and Quality, Substance Abuse and Mental Health Services
Administration. Retrieved from www.samhsa.gov/data/sites/
default/files/cbhsq-reports/NSDUHFFR2017/
NSDUHFFR2017.pdf).

Yoon, S., Voith, L.A., & Kobulsky, J.M. (2018). Gender differences in
pathways from child physical and sexual abuse to adolescent risky
sexual behavior. Journal of Adolescence, 64, 89-97.

Module 5 References | 535