Chapter 6 - Sampling

Chapter 7 - Measurement and scaling

Chapter 8 - Designing the questionnaire
Chapter 11 - Basic data analysis for quantitative research
Chapter 12 - Examining relationship in quantitative research

5 questions. Question 1* (compulsory) and ANY three other questions.

Note: * question on SPSS result

Read textbook and focus chapter 8,11,12

CHAPTER 6 – SAMPLING

Sampling involves selecting a relatively small number of elements from a larger defined
group of elements and expecting that the information gathered from the small group
will enable accurate judgments about the larger group.

Sampling is often used when it is impossible or unreasonable to conduct a census. With

a census, primary data is collected from every member of the target population.

Sampling decisions influence the type of research design, the survey instrument, and the
actual questionnaire.

General idea of the target population and the key characteristics that will be used to draw the
sample of respondents, researchers can customize the questionnaire to ensure that it is of
interest to respondents and provides high-quality data.

Population is the identifiable set of elements of interest to the researcher and pertinent
to the information problem.

Defined target population consists of the complete group of elements (people or

objects) that are identified for investigation based on the objectives of the research
project. A precise definition of the target population is essential and is usually done in terms
of elements, sampling units, and time frames.

Sampling units are target population elements actually available to be used during the
sampling process. Sampling frames are eligible sampling units.

One of the major goals of researching small, yet representative, samples of members of a
defined target population is that the results of the research will help to predict or estimate
what the true population parameters are within a certain degree of confidence.
Factors Underlying Sampling Theory

Central Limit Theorem (CLT) describes the theoretical characteristics of a sample

population. The CLT is the theoretical backbone of survey research and is important in
understanding the concepts of sampling error, statistical significance, and sample sizes.

With an understanding of the basics of the CLT, the researcher can do the following:

1. Draw representative samples from any target population.

2. Obtain sample statistics from a random sample that serve as accurate estimates of the
target population’s parameters.
3. Draw one random sample, instead of many, reducing the costs of data collection.
4. More accurately assess the reliability and validity of constructs and scale measurements.
5. Statistically analyze data and transform it into meaningful information about the target
population.

Tools Used to Assess the Quality of Samples

There are numerous opportunities to make mistakes that result in some type of bias in any
research study. This bias can be classified as either sampling error or non sampling error.

Two difficulties associated with detecting sampling error are

(1) a census is very seldom conducted in survey research and

(2) sampling error can be determined only after the sample is drawn and data collection is
completed.

Sampling error is any bias that results from mistakes in either the selection process for
prospective sampling units or in determining the sample size. Moreover, random
sampling error tends to occur because of chance variations in the selection of sampling
units.

Nonsampling error occurs regardless of whether a sample or a census is used. These

errors can occur at any stage of the research process. For example, the target population
may be inaccurately defined causing population frame error; inappropriate question/scale
measurements can result in measurement error; a questionnaire may be poorly designed. In
general, the more extensive a study, the greater the potential for non sampling errors.

There are two basic sampling designs:

(1) probability and

(2) nonprobability.
Probability sampling, each sampling unit in the defined target population has a known
probability of being selected for the sample. The actual probability of selection for
each sampling unit may or may not be equal depending on the type of probability
sampling design used. Specific rules for selecting members from the population for
inclusion in the sample are determined at the beginning of a study to ensure

(1) unbiased selection of the sampling units and

(2) proper sample representation of the defined target population.

Probability sampling enables the researcher to judge the reliability and validity of data
collected by calculating the probability that the sample findings are different from the
defined target population.

Nonprobability sampling, the probability of selecting each sampling unit is not known.
Therefore, sampling error is not known. Selection of sampling units is based on intuitive
judgment or researcher knowledge. The degree to which the sample is representative of the
defined target population depends on the sampling approach and how well the researcher
executes the selection activities.
Types of Probability and Nonprobability Sampling Methods

Probability Sampling

Simple random sampling – every sampling unit has a known and equal chance of being
selected.
Systematic random sampling requires that the defined target population be ordered in
some way; it creates a sample of objects or prospective respondents that is very similar in
quality to a sample drawn using simple random sampling.

Researchers must be able to secure a complete listing of the potential sampling units that
make up the defined target population.

Stratified random sampling involves the separation of the target population into different
groups, called strata, and the selection of samples from each stratum. Stratified random
sampling is similar to segmentation of the defined target population into smaller, more
homogeneous sets of elements.

To ensure that the sample maintains the required precision, representative samples must be
drawn from each of the smaller population groups (stratum). Drawing a stratified random
sample involves three basic steps:

1. Dividing the target population into homogeneous subgroups or strata.

2. Drawing random samples from each stratum.
3. Combining the samples from each stratum into a single sample of the target population.

Two common methods are used to derive samples from the strata

Proportionately stratified sampling, the sample size from each stra- tum is dependent on
that stratum’s size relative to the defined target population. Therefore, the larger strata are
sampled more heavily because they make up a larger percentage of the target population.
Disproportionately stratified sampling, the sample size selected from each stratum is
independent of that stratum’s proportion of the total defined target population. This approach
is used when stratification of the target population pro- duces sample sizes for subgroups
that differ from their relative importance to the study.

Optimal allocation sampling. In this method, consideration is given to the relative size of
the stratum as well as the variability within the stratum to determine the necessary sample
size of each stratum.

Multisource sampling is being used when no single source can generate a large or low
incidence sample. While researchers have shied away from using multiple sources, mainly
because sampling theory dictates the use of a defined single population, changing
respondent behaviors (e.g., less frequent use of e-mail and more frequent use of social
media) are supporting multi source sampling.

Cluster sampling is similar to stratified random sampling, but is different in that the
sampling units are divided into mutually exclusive and collectively exhaustive subpopulations
called clusters. Each cluster is assumed to be representative of the heterogeneity of the
target population.

\
A popular form of cluster sampling is area sampling. In area sampling, the clusters are
formed by geographic designations.
Nonprobability Sampling

Convenience sampling is a method in which samples are drawn based on convenience. In

reality, it is difficult to accurately assess the representativeness of the sample. Given self-
selection and the voluntary nature of participating in the data collection, researchers should
consider the impact of nonresponse error when using sampling based on convenience only.
Judgment sampling, sometimes referred to as purposive sampling, respondents are
selected because the researcher believes they meet the requirements of the study. The
assumption is that the opinions of a group of experts are representative of the target
population.
Quota sampling involves the selection of prospective participants according to prespecified
quotas for either demographic characteristics (e.g., age, race, gender, in- come), specific
attitudes (e.g., satisfied/dissatisfied, liking/disliking, great/marginal/no quality), or specific
behaviors (e.g., regular/occasional/rare customer, product user/nonuser). The purpose of
quota sampling is to assure that prespecified subgroups of the population are represented.
Snowball sampling involves identifying a set of respondents who can help the researcher
identify additional people to include in the study. This method of sampling is also called
referral sampling because one respondent refers to other potential respondents. Snowball
sampling typically is used in situations where (1) the defined target population is small
and unique and (2) compiling a complete list of sampling units is very difficult. While
membership in these types of social circles might not be publicly known, intra circle
knowledge is very accurate. The underlying logic of this method is that rare groups of people
tend to form their own unique social circles.
Determining Sample Size

Consider how precise the estimates must be and how much time and money are available to
collect the required data, since data collection is generally one of the most expensive
components of a study. Sample size determination differs between probability and
nonprobability designs.

Probability Sample Size

1.The population variance, which is a measure of the dispersion of the population, and its
square root, referred to as the population standard deviation. The greater the vari- ability in
the data being estimated, the larger the sample size needed.

2. The level of confidence desired in the estimate. Confidence is the certainty that the true
value of what we are estimating falls within the precision range we have selected. For
example, marketing researchers typically select a 90 or 95 percent confidence level for their
projects. The higher the level of confidence desired is the larger the sample size needed.

3. The degree of precision desired in estimating the population characteristic. Precision is

the acceptable amount of error in the sample estimate. The more precise the required
sample results, that is, the smaller the desired error, the larger the sample size.
Sampling from a Small Population

In the previously described formulas, the size of the population has no impact on the deter-
mination of the sample size. This is always true for “large” populations. When working with
small populations, however, use of the earlier formulas may lead to an unnecessarily large
sample size. If, for example, the sample size is larger than 5 percent of the population then
the calculated sample size should be multiplied by the following correction factor:

N/(N + n − 1)

where
N = Population size
n = Calculated sample size determined by the original formula

Thus, the adjusted sample size is

Sample size = (Specified degree of confidence × Variability/Desired precision)2 × N/(N + n −
1)

Nonprobability Sample Sizes

Sample size formulas cannot be used for nonprobability samples. Determining the sample
size for nonprobability samples is usually a subjective, intuitive judgment made by the
researcher based on either past studies, industry standards, or the amount of resources
available.

Steps in Developing a Sampling Plan

A sampling plan is the blueprint to ensure the data collected are representative of the
population. A good sampling plan includes the following steps:

(1) define the target population;

(2) select the data collection method;

(3) identify the sampling frames needed;

(4) select the appropriate sampling method;

(5) determine necessary sample sizes and overall contact rates;

(6) create an operating plan for selecting sampling units; and

(7) execute the operational plan.

Step 1: Define the Target Population in any sampling plan, the first task of the researcher
is to determine the group of people or objects that should be investigated. With the problem
and research objectives as guidelines, the characteristics of the target population should be
identified. An understanding of the target population helps the researcher to successfully
draw a representative sample.

Step 2: Select the Data Collection Method Using the problem definition, the data
requirements, and the research objectives, the researcher chooses a method for collecting
the data from the population. Choices include some type of interviewing approach (e.g.,
personal or telephone), a self-administered survey, or perhaps observation. The method of
data collection guides the researcher in selecting the sampling frame(s).

Step 3: Identify the Sampling Frame(s) Needed A list of eligible sampling units must be
obtained. The list includes information about prospective sampling units (individuals or
objects) so the researcher can contact them. An incomplete sampling frame decreases the
likelihood of drawing a representative sample. Sampling lists can be created from a number
of different sources (e.g., customer lists from a company’s internal database, random-digit
dialing, an organization’s membership roster, or purchased from a sampling vendor).

Step 4: Select the Appropriate Sampling Method The researcher chooses between
probability and nonprobability methods. If the findings will be generalized, a probability
sampling method will provide more accurate information than non probability sampling
methods. As noted previously, in determining the sampling method, the researcher must
consider seven factors: (1) research objectives; (2) desired accuracy; (3) availability of
resources; (4) time frame; (5) knowledge of the target population; (6) scope of the re-
search; and (7) statistical analysis needs.

Step 5: Determine Necessary Sample Sizes and Overall Contact Rates In this step of a
sampling plan, the researcher decides how precise the sample estimates must be and how
much time and money are available to collect the data. To determine the appropriate sample
size, decisions have to be made concerning (1) the variability of the population
characteristic under investigation, (2) the level of confidence desired in the estimates,
and (3) the precision required. The researcher also must decide how many completed
surveys are needed for data analysis.

At this point the researcher must consider what impact having fewer surveys than initially
desired would have on the accuracy of the sample statistics. An important question is “How
many prospective sampling units will have to be contacted to ensure the estimated sample
size is obtained, and at what additional costs?”

Step 6: Create an Operating Plan for Selecting Sampling Units The researcher must
decide how to contact the prospective respondents in the sample. Instructions should be
written so that interviewers know what to do and how to handle problems contacting
prospective respondents. For example, if the study data will be collected using mall-intercept
interviews, then interviewers must be given instructions on how to select respondents and
conduct the interviews.

Step 7: Execute the Operational Plan This step is similar to collecting the data from
respondents. The important consideration in this step is to maintain consistency and control.
CHAPTER 7 – Measurement and scaling

Structured Data Unstructured Data

Organized and stored in a specific, defined do not have values and cannot be
format, and categorized, organized, and processed
managed in fixed fields

The predefined data structure is in a analyzed using conventional statistical

spreadsheet format, referred to as a methods.
relational database Examples: video or audio files, mobile
activity, text, social media posts, satellite
images, and so forth.

Structured data are often referred to as Often identified as qualitative data.

quantitative data.

immediately useful in decision making. Unstructured data are converted into

structured data and analyzed based on
relevant characteristics

Measurement Process
Measurement is the process of determining the intensity (or amount) of information about
constructs, concepts, or objects.
The measurement process consists of two tasks.
● Construct selection/development with the goal to precisely identify and define what is
to be measured.
● Scale measurement determines how to precisely measure each construct.

What Is a Construct?
● A construct is an abstract idea/concept formed in a person’s mind through a
combination of construct characteristics and the characteristics are the variables that
collectively define the concept and make measurement of the concept possible.
● For example, to measure “customer interaction from the seller’s perspective,”
researchers may use Agree–Disagree Scales.
● The resultant score is called a scale, an index, or a summated rating.
● So to identify restaurant satisfaction, researchers may conduct a literature review,
perform interviews, and use personal experience to form the framework for
measuring the construct.
Construct Development
● A construct is an unobservable concept measured indirectly by a group of related
variables.
● The construct is measured with scale measurements of each individual indicator
variable.
● Construct development begins with an accurate definition of the purpose of the study
and the research problem.
● Researchers identify characteristics that define the concept and then develop
methods of indirectly measuring the concept.
● At the heart of construct development is the need to determine exactly what is to be
measured.
● If an object’s features can be directly measured, then the feature is a concrete
variable and not an abstract construct.

Construct Dimensionality and Validity

● Researchers use qualitative data for insights into the unobservable characteristics of
the construct of interest.
● In contrast, domain observables are the set of identifiable and measurable
components associated with the construct.
● Unobserved constructs are either unidimensional or multidimensional.
● Construct validity assessment begins when researchers prepare the operational
definition of the construct.
● Assessment of validity is ongoing and happens after-the-fact as well.
Researchers use content validity, convergent validity, discriminant validity, and
● nomological validity.

Categorical Types of Information

● State-of-being is physical, demographic, or socioeconomic characteristics of
individuals, objects, or organizations
● State-of-mind represent a person’s mental attributes or emotional feelings about an
object; not observable, so specific questions are asked but limited to respondents’
honesty.
● State-of-behavior represents the current observable actions or recorded past
actions of individuals or organizations, example researchers use surveys, trained
observers, or existing records.
● State-of-intention represents the expressed plans to undertake specified future
behavioral actions.
Properties of Measurement Scales
The amount of information obtained depends on which scaling properties are used in the
scale design.
Assignment property: the use of unique descriptors to identify each object in a set.
Order property: establishes “relative magnitudes” between the descriptors creating
hierarchical rank-order relationships among objects.
Distance property: enables the researcher and respondent to identify, understand, and
express absolute (or assumed) differences between objects.
Origin property: a unique scale descriptor designated as being a “true natural zero” or “true
state of nothing.”

Scale Measurement
● Involves assigning a set of scale descriptors to represent the range of possible
responses.
● The scale descriptors are a combination of labels and numbers which are assigned
using a set of rules.
● Scale measurement assigns degrees of intensity to the responses.
● The degrees of intensity are commonly referred to as scale points.

There are four scale levels.

● Nominal, scale:
● Uses questions requiring respondents to provide some type of descriptor as the raw
response.Ordinal, scale:
○ Let respondents express relative magnitude between the answers.
○ Mode, median, frequency distribution, and range can be applied to ordinal scales.
● Interval and Scale
○ Example 1:
How likely are you to recommend the Santa Fe Grill to a friend?
Definitely Will Not Recommend/Definitely Will Recommend
○ Example 2:
Using a scale of 0–10, with “10” being Highly Satisfied and “0” being Not Satisfied At
All, how satisfied are you with the banking services you currently receive from
(read name of primary bank)?
Answer: Blank
○ Example 3:
Please indicate how frequently you use different banking methods. For each of the
banking methods listed below, click the number that best describes the frequency
you typically use each method.
An interval scale demonstrates absolute differences between each scale point.
● Ratio Scale
○ A ratio scale allows the researcher not only to identify the absolute differences
between each scale point but also to make comparisons between the responses.
○ This enables a “true natural zero” or “true state of nothing” response.
Example 1:

Please circle the number of children under 18 years of age currently living in your household.
 0 1 2 3 4 5 6 7 If more than 7, please specify: Blank

Example 2:

In the past seven days, how many times did you go online to shop at Amazon.com?

Blank # of times

Example 3:

In years, what is your current age?

Blank # of years old

Evaluating Measurement Scales—Scale Reliability

● Scale reliability is the extent a scale can reproduce the same or similar
measurements in repeated trials which is consistency.
● The test–retest technique involves repeating the scale measurement with either the
same sample at two times or two different samples.
● The equivalent form technique creates two similar yet different scale measurements
for the given construct.
● Internal consistency is the degree to which the individual questions of a construct
are correlated.
● The split-half test divides questions into two groups and correlates scores against
each other.
● The coefficient α calculates the average of all possible split-half measures resulting
from differences in dividing the scale questions.

Evaluating Measurement Scales—Validity

● Scale validity assesses whether a scale measures what it is supposed to measure
accuracy.
● Face validity, researchers use expert judgment.
● Content validity is a measure of the extent to which a construct represents all
relevant dimensions.
● Convergent validity occurs when multiple measures of the same construct have a
variance of more than 50%.
● Discriminant validity is the extent a single construct differs from others and is
unique.
● Two ways to obtain data to assess validity via pilot study or a panel of experts.

Developing Measurement Scales Require;

● Understanding the research problem and the specific research information
objectives.
● Identifying and developing constructs.
● Establishing detailed data requirements such as state of being, mind, behavior, and
intention.
● Understanding the scaling properties.
● Selecting the appropriate measurement scale.
Criteria for Scale Development
● Respondents may not understand the questions—use a pretest.
● The discriminatory power of scale descriptors is the scale’s ability to differentiate
between the scale responses.
● A balanced scale has an equal number of positive and negative response
alternatives.
● A forced-choice scale does not have a neutral descriptor.
● Include negatively worded statements to verify attention.
● Measures of central tendency locate the center of a distribution.
● Measures of dispersion include: frequency distribution, range, and standard
deviation.

Relationships Between Scale Levels and Measures of Central Tendency and

Dispersion

General steps in construct development/scale measurement.

Step 1: Identify and define the construct.
Step 2: Use qualitative research to identify a theory.
Step 3: Refine problem using qualitative judgment and analysis.
Step 4: Design scales, pretest, and pilot test.
Step 5: Evaluate reliability and validity.
Step 6: Purify scales by eliminating poorly designed statements.
Step 7: Complete the final scale evaluation.
Scale formats:
● Likert-type scales
○ Ordinal scale asking respondents to indicate the extent they agree or
disagree with a series of mental or behavioral belief statements about a given
subject.
○ For each listed statement below, please select the one response that best
expresses the extent to which you agree or disagree with that statement.

Definitely Somewhat Slightly Slightly Somewhat Definitely

Statements Disagree Disagree Disagree Agree Agree Agree

I buy many things with a

credit card.

I wish we had a lot more

money.

My friends often come

to me for advice.

I am never influenced
by advertisements.

● Semantic differential scales

○ We would like to know your opinions about the expertise, trustworthiness, and
attractiveness you believe Jennifer Garner brings to Capital One bank card
service advertisements. Each dimension below has five factors that may or
may not represent your opinions. For each listed item, please click the
number that best expresses your opinion about that item.
Expertise:

Knowledgeable 7 6 5 4 3 2 1 Unknowledgeable

Expert 7 6 5 4 3 2 1 Not an expert

Skilled 7 6 5 4 3 2 1 Unskilled

Qualified 7 6 5 4 3 2 1 Unqualified

Experienced 7 6 5 4 3 2 1 Inexperienced

Trustworthine
ss:

Reliable 7 6 5 4 3 2 1 Unreliable

Sincere 7 6 5 4 3 2 1 Insincere

Trustworthy 7 6 5 4 3 2 1 Untrustworthy

Dependable 7 6 5 4 3 2 1 Undependable

Honest 7 6 5 4 3 2 1 Dishonest

○ A Semantic Differential Scale is a unique bipolar ordinal scale that captures

attitudes or feelings on a give object.
○ A problem in designing semantic differential scales is the inappropriate
narrative expressions of the scale descriptors.
○ A well-designed scale has truly bipolar anchors.
○ A negative pole descriptor is not truly an opposite, creating confusion.
○ Another problem is the use of an odd number of scale points, creating a
so-called neutral response dividing the poles.
○ The neutral response has no diagnostic value to researchers.
○ To avoid, use an even-point (or forced-choice) scale point and incorporate a
“nonapplicable” response out to the side.
 ● Behavioral intention scales
○ A behavioral intention scale captures the likelihood a person demonstrates
predictable behavior with future purchases.

Noncomparative Rating Scales

● Noncomparative rating scales get respondents’ attitudes, behavior, or intentions
about an object without referencing another object.
● Graphic rating scales use a scaling descriptor format that presents a respondent with
a continuous line as the set of possible responses.

Comparative Rating Scales

● Respondents’ attitudes, feelings, or behaviors about an object on the basis of some
other object.
● Rank-order scales enables comparing objects by indicating their order of preference
or choice from first to last.
● Constant sum scales require the respondent to allocate a given number of points,
usually 100, among each separate attribute or feature relative to all the other listed
one.
Other Measurement Scale Issues
● A Single-Item Scale collects data about only one object attribute.
● A Multiple-Item Scale collects data on several attributes of an object.
● Two factors on whether to use a single or multi item scale is number of construct
dimensions, reliability, and validity.
● A Reflective Scale is a sample of all possible indicators which should be highly
correlated.
● A Formative Scale has scale items that are not correlated so all indicators must be
included for accurate measure.

Misleading Scaling Formats

● A double-barreled question includes two or more attributes, but responses allow
comment on a single issue.
● A leading question influences the respondent’s answers.
● A loaded question suggests a socially desirable answer or involves an emotionally
charged issue.
● Ambiguous questions involve possible responses that can be interpreted a number
of ways.
● Complex questions are worded in a way making the respondent unsure how to
respond.
● A double negative question contains two negative thoughts in the same question.
● Scale responses should be mutually exclusive.
CHAPTER 8 – Designing the Questionnaire

● Descriptive questionnaires collect data that can be turned into knowledge about a
person, object, or issue.
● Predictive survey questionnaires predict changes in attitudes and behaviors, and
test hypotheses.
● A questionnaire designed to gather primary data.

Pilot Studies and Pretests

● A pilot study is a small-scale version of the intended main research study, including
all subcomponents, such as the data collection and analysis, from 50 to 200
respondents who are representative of the main study’s defined target population.
● A pretest is a descriptive research activity representing a small-scale investigation of
10 to 30 subjects representative of the main study’s defined target population but
focus on a specific sub­component of the main study.

Questionnaire Design
Step 1: Confirm research information objectives and data requirements.
Step 2: Select appropriate data collection method.
Step 3: Develop questions and scaling.
Step 4: Determine layout and evaluate questionnaire.
Step 5: Obtain initial client approval.
Step 6: Pretest, revise, and finalize questionnaire.
Step 7: Implement the survey.

Step 1: Confirm Research Information Objectives and Data Requirements

The American Bank in Baton Rouge, LA has these objectives.
● To identify preferred banking services, as well as attitudes and feelings toward those
services.
● To capture from customers and potential customers their satisfaction with and
commitment to current banking relationships.
● To collect data on selected lifestyle dimensions and social media usage patterns to
better understand American Bank customers and their banking habits and those of
potential customers.
● To collect data on selected demographic characteristics that can be used to create a
profile of current American Bank customers as well as potential future customers.
Step 2: Select Appropriate Data Collection Method
Data requirements for the American Bank study include:
Section 1: Banking Services.
Section 2: Banking Relationships.
Section 3: Lifestyle Dimensions.
Section 4: Demographic Characteristics.

Bank management and the researcher decide on:

● An initial contact via email or text message.
● Followed by an online survey for current customers.
● A research panel to obtain data from potential customers.

Step 3: Develop Questions and Scaling

● Question format: unstructured or structured.
● Wording of questions, scales, and instructions for responding to questions and
scales.
● Type of data required: nominal, ordinal, interval, or ratio.
● Researchers then develop screening questions and responses.
● They must consider how the data will be collected.

Unstructured questions Structured questions

open-ended closed-ended

Responses are unaided and unlimited but often They require a choice from a set of responses or
skipped. scale points.

These questions are more difficult to code for Reduces respondent’s effort.
analysis.

They also require more thought from Easy to answer, easy to code.
respondents.

Work best in self-administered and in-person Used in quantitative surveys.

surveys.

● Avoid ambiguous words and phrases or difficult vocabulary.

● When unsure, use a pretest for questionable words.
● Words and phrases influence a respondent’s answer.

Bad questions prevent or distort communications between the researcher and the
respondent.
● Unanswerable: either because the respondent does not have access to the needed
information or none of the answer choices apply.
● Leading (or loaded): when the respondent is directed to a response that would not
ordinarily be given.
● Double-barreled: when the respondent is asked to address more than one issue at a
time.
Step 4: Determine Layout and Evaluate Questionnaire
● An introductory section gives respondent’s an overview and screening questions
identify qualified respondents.
● The second section is the research questions section based on the objectives and
arranged from general to specific.
● The last section includes lifestyle, social media usage, and demographics, then
ends with a thank-you statement.

Design questions to minimize response order bias.

● Common methods variance may be present when responses to independent and
dependent variable questions are highly correlated.
● Reduce CMV by collecting data at different points in time or from different sources.

Review questionnaire
● Focus on whether each question is necessary and if length is acceptable.
● Check that the survey meets the research information objectives.
● Make sure the scale format and instructions work well.
● Check that questions move from general to specific.
● With online surveys, view it on the screen as a respondent would.
● Physically inspect mail, drop-off, and self-reporting surveys.
● Self-administered questionnaires should look professional and visually appealing.

Online Survey Considerations

● The number of questions appearing on the screen must be considered.
● Some measurement scale formats may not work on all devices.
● Online survey researchers must plan how respondents are solicited.
● Consider the length of time needed to complete the survey.
● “Straight liners” may be removed from the usable sample.
● Address the effect of response box size on length of answers.
● Address the use of radio buttons versus pull-down menus.
● Address the appropriate use of visuals.

Step 5: Obtain Initial Client Approval

When conducting research using human respondents, academic researchers must pass the
Institutional Review Board (IRB).
Many IRBs require a consent form, which includes:
● A statement the study involves research, the purpose of the research, the
procedures, and amount of time required.
● Any risks or discomforts, including a possible confidentiality leak.
● Any possible benefits; if none, say so.
● How confidentiality is maintained.
● Contact information.
● A statement that participation is voluntary and the person may stop.

Step 6: Pretest, Revise, and Finalize the Questionnaire

In the pretest, the questionnaire is given to a small group (10 to 30) to fill out and provide
feedback to researchers about:
● How much time respondents need to complete the survey.
● Whether to add or revise instructions and questions.
● What to say in the cover letter.
● Problems with question formatting or understanding.
A pilot study is a small-scale version of the main study, with data collection and analysis,
given to 50 to 200 respondents, sometimes used to examine specific subcomponents of the
overall research plan and to see if refinements are needed.

Step 7: Implement the Survey

Focus on the process followed to collect the data using the agreed-upon questionnaire.
● Process varies depending on whether the survey is self-administered or
interviewer-completed.
● Methods are used in increase response rates of self-administered surveys.
● With online surveys the format, sequence, skip patterns, and instructions must be
thoroughly checked.

The Role of a Cover Letter and Introductory Section

● The cover letter explains the nature of the survey and obtains the respondent’s
cooperation in participating in the survey.
● With personal or telephone interviews, interviewers use a verbal statement.
● Online surveys have an introduction designed to motivate respondents, explain the
survey, and communicate contact information.
● Self-administered surveys often have low response rates but good cover letters
increase those rates.

Other Considerations in Collecting Data—Supervisor Instructions

A supervisor instruction form includes the following:
● Detailed information on the nature of the study.
● Start and completions dates.
● Sampling instructions.
● Number of interviewers required.
● Equipment and facility requirements.
● Reporting forms.
● Quotas.
● Validation procedures.

Other Considerations in Collecting Data—Interviewer Instructions

Interviewer instructions include the following:
● Detailed information about the nature of the study.
● Start and completion dates.
● Sampling instructions.
● Screening procedures.
● Quotas and number of interviews required.
● Guidelines for asking questions.
● Use of rating cards.
 ● Recording responses and reporting forms.
● Verification form procedures.

Other Considerations in Collecting Data—Screening Questions and Quotas

● Screening questions ensure the respondents are qualified and representative of the
defined target population.
● Occurs frequently when a person’s occupation or a family member’s occupation
eliminates the person from inclusion in the study.
● Quotas are used to ensure the data are collected from the correct respondents.
● When a particular subgroup reaches the quota, questionnaires for that subgroup are
no longer completed.

Other Considerations in Collecting Data—Call or Contact Records

Call records are used to estimate the efficiency of interviewing—they gather the following
information.
● Number of calls or contacts made per hour.
● Number of contacts per completed interview.
● Length of time of the interview.
● Completions by quota categories.
● Number of terminated interviews.
● Reasons for termination.
● Number of contact attempts.
CHAPTER 11 – Basic Data Analysis for Quantitative Research
Measures of Central Tendency

● Frequency distributions can be useful for examining the different values for a
variable.
● Researchers use descriptive statistics to summarise information
● The mean, median, and mode are measures of central tendency.
● These measures locate the center of the distribution.
● The mean, median, and mode are sometimes also called measures of location.
● MEAN is the average value within the distribution and is the most commonly
used measure of central tendency, used for interval and ratio tapi bila ada outliers,
guna median/mode
● MEDIAN is the middle value of the distribution when ordered in either ascending
or descending sequence, usually used in ordinal data
● MODE is the value that appears in the distribution most often, usually used in
nominal data

Measures of Dispersion

● The range defines the spread of the data, the endpoints of the distribution of values.
● Standard deviation is the average distance of the distribution values from the mean.

● Standard deviation (s) =

● The average squared deviation, called the variance, is used in a number of statistical
processes.

How to Develop Hypotheses

● A hypothesis is an unproven supposition or proposition that explains facts

● Statistical techniques are used to confirm or deny a hypothesis.
● Hypotheses are developed prior to data collection.
● Different tests are appropriate when testing hypotheses comparing two or more
groups, depending on how the samples are related.
● The null hypothesis is no difference in the group means.
● Marketing researchers typically test the null hypothesis.
● Another hypothesis, called the alternative hypothesis, states the opposite of the null
hypothesis.

Analyzing Relationships of Sample Data:

1) Sample Statistics and Population Parameters

● The purpose of inferential statistics is to make a determination about a population
on the basis of a sample of that population.
● Used to make inferences about the population’s parameters.
 ● A population parameter is a variable or some sort of measured characteristic of the
entire population but the actual population parameters are unknown.
● A frequency distribution is commonly used to summarise the results of the data
collection process.
● When it displays a variable in terms of percentages, then the distribution represents
proportions within a population.

2) Choosing the Appropriate Statistical Technique

● Univariate statistics used when the population using only one variable.
● Multivariate statistica when many lah
● Scale of measurement.
● Different types of scales of measurement used in different situations.
● When the data are measured using an interval or ratio scale and the sample size is
large, use parametric statistics
● When a normal distribution cannot be assumed, the researcher must use
nonparametric statistics.

Type of Scale and Appropriate Statistic

Type of Scale Measure of Measure of Dispersion Statistic

Central Tendency

Nominal Mode None Chi-square

(nonparametric)

Ordinal Median Percentile Chi-square

(nonparametric)

Interval or Mean Standard deviation t-test, ANOVA (parametric)

ratio

Univariate and Bivariate Relationship— Selecting the Appropriate Test

Relationship
Independent Dependent Variable
Examined Statistical Test
Variable (I V) (DV)
(hypothesis)

Nominal Nominal Differences

Ordinal Ordinal Association Spearman correlation (NP)

Metric Ordinal Association Spearman correlation (NP)

Ordinal Metric Association Spearman correlation (NP)

Metric Metric Association Pearson correlation (P)

Two groups Ordinal Differences Median test (NP)

Three or more Kruskal-Wallis one way

Ordinal Differences
groups analysis of variance (NP)
 t-test (independent samples)
Two groups Metric Differences
(P)

Three or more One way analysis of variance

Metric Differences
groups (P)

Single Variable
Blank Blank
Nominal

3) Univariate Statistical Tests

● Univariate tests of significance test hypotheses about a sample characteristic against
a known or given standard.
● The average monthly electric bill in Miami, FL, exceeds $250.00.
● Translate that into a null hypothesis and test it

4) Bivariate Statistical Tests

● Usually used by marketing researchers
● Three types of bivariate hypothesis tests are discussed, Chi-square, which is used
with nominal data, the t test (to compare two means), analysis of variance
(ANOVA) (compares three or more means).
● Both the t test and analysis of variance are used with either interval or ratio data.

5) Cross-Tabulation
● Cross-tabulation is useful for examining relationships and reporting the findings for
two variables.
● Purpose to determine whether differences exist between subgroups.
● Simplest methods for describing sets of relationships.
● It is a frequency distribution of responses on two or more variables.
● To conduct cross tabulation, the responses for each of the groups are tabulated and
compared.
● Chi-square (X²) analysis enables us to test whether there are any statistical
differences between the responses for the groups that nominally scaled variables
● Chi-square (X²) analysis statistic answers questions that cannot be analyzed with
other types of analysis, such as ANOVA or t tests.
6) Comparing Means—Independent Versus Related Samples
● A frequently examined question is whether the means of two groups of respondents
are significantly different.
● An independent samples comparison would be the results of interviews with male
and female coffee drinkers.
● A related samples compares the average cups of coffee per day by male students
with average soft drinks per day by the same male students.
● Although questions are independent, the respondents are the same.

7) Using the t Test to Compare Two Means

● The bivariate t test requires interval or ratio data.
● Useful when the sample size is small (n < 30) and when the population standard
deviation is unknown.
● When comparing means, the t test is the difference between the means divided by
the variability of the means.
● The t value is a ratio of the difference between the two sample means and the
standard error.

●
● Sometimes marketing researchers want to test for differences in two means for
variables in the same sample.
● To examine this, researchers use the paired sample test for the difference in two
means.
● Examines whether two means from two different questions using the same scaling
and answered by the same respondents are different.

8) Analysis of Variance (ANOVA)

● Researchers use analysis of variance (ANOVA) to find the statistical difference
between three or more means.
 ● One-way ANOVA is used to examine group means when only one independent
variable is involved.
● Researchers also can use ANOVA to examine the effects of several independent
variables simultaneously.
● ANOVA requires the dependent variable be interval or ratio and the independent
variable be categorical (nonmetric).
● ANOVA examines the variance within a set of data.
● Researchers use the F test with ANOVA to evaluate the differences between group
means for statistical significance.
● The F test in ANOVA will tell us if these observed differences are statistically
significant.
● The total variance in a set of responses is made up of: Between-group variance
measuring how much the sample means of the groups differ from one another and
within-group variance measuring how much responses within each group differ
from one another.
● A weakness of ANOVA is the test does not identify which pairs of means are
different.
● n-way ANOVA is a type of ANOVA that can analyze several independent variables at
the same time.
● Using multiple independent factors creates the possibility of an interaction effect.
● That is, the multiple independent variables can act together to affect dependent
variable group means.
● The use of experimental design may require n-way ANOVA.
● Such as marketers measuring responses to different prices or ads.
● Conceptually the same as one-way ANOVA, but the mathematics is more complex.

Perceptual Mapping
● Perceptual mapping is a process that is used to develop maps that show the
perceptions of respondents.
● Tahu dah kan cane buat ni
● Ada rankings, medians, and mean ratings.

●
● Identifies gaps in perceptions, helping position new products development.
● Identifies a company’s image, helping position them against competitors
● Can assess advertising effectiveness in positioning the brand for advertising
● Could be used to assess similarities of brands and channel outlets in distribution
CHAPTER 12 – Examining Relationship in Quantitative Research

Statistical Significance: Tests whether observed relationships in data are likely to be

genuine or due to random chance.

Direction of Relationship: Once a relationship's presence is confirmed, the next aspect to

consider is its direction. This can either be:

● Positive: Where an increase in one variable leads to an increase in another. For

instance, higher ratings of food quality at Santa Fe Grill are associated with greater
customer satisfaction.
● Negative: Where an increase in one variable results in a decrease in another. For
example, an increase in service problems at Santa Fe Grill might lead to a decrease
in customer satisfaction.

Strength of Association: The strength of the relationship indicates how closely the
variables are related and is typically quantified by the correlation coefficient:

● No Relationship: No consistent pattern is observed between the variables.

● Weak Relationship: The variables share very little common variance.
● Moderate Relationship: There is a noticeable but not perfect alignment in the
variations of the variables.
● Strong Relationship: A very tight relationship where variations in one variable are
closely mirrored by variations in the other.

For example, a strong positive correlation coefficient between time spent in the store and
amount spent indicates that customers who spend more time also tend to spend more
money.

Type of Relationship: The nature of the relationship can significantly affect how it is
analyzed:

● Linear Relationship: The relationship between variables remains consistent across

the spectrum of the data. In a linear relationship, a straight-line formula (Y = a + bX)
can accurately predict the outcome variable (Y) based on the predictor (X).
● Curvilinear Relationship: The relationship changes in strength or direction across
the range of variables. For instance, moderate levels of fear appeal in advertising
might be highly effective, whereas too little or too much fear can be ineffective. Such
relationships require more complex models to describe and predict outcomes
accurately.

Covariation The amount of change in one variable that is consistently related to the change
in another variable of interest.
Scatter diagram A graphic plot of the relative position of two variables using a horizontal
and a vertical axis to represent the values of the respective variables.

The scatter plot might show a random distribution of dots, resembling a circle with no
discernible pattern.

Interpretation: Such a diagram indicates no systematic relationship or covariation between

the variables. For instance, random placements of Y and X values suggest that knowing one
variable does not help predict the other, and the covariation measure would be close to zero.

The relationship shown is more complex, where the pattern of dots changes direction;
initially, small increases in Y correspond with increases in X, but as Y increases further, X
begins to decrease.

Interpretation: This curvilinear relationship indicates that the association between the
variables varies depending on their levels. Such patterns are challenging because the
direction of the relationship changes, making linear statistical methods inadequate for a full
understanding.
Similar to linearity but with an opposite pattern; increases in Y correspond with decreases in
X.

Interpretation: This scenario reflects a negative linear relationship, still indicating high
covariation but in opposite directions. This relationship is crucial for scenarios where one
variable inversely affects another.

The dots align in a pattern that could be depicted as a straight line or an elongated ellipse,
indicating that as values of Y increase, values of X also increase.

Interpretation: This type of diagram illustrates a positive linear relationship where the
variables change in the same direction. The covariation is considered high, suggesting a
strong and direct association between the two variables.
Correlation Analysis

Scatter Diagrams: Scatter diagrams are graphical tools used to visualize the relationship
between two quantitative variables. They help identify patterns in data such as trends,
clusters, and potential outliers, which indicate how one variable changes in relation to
another.

Pearson Correlation Coefficient (PCC): The Pearson Correlation Coefficient is a statistical

measure that quantifies the degree of linear relationship between two interval or ratio-scale
variables. The coefficient's value ranges from -1.00 to +1.00:

● -1.00: Indicates a perfect negative linear correlation.

● 0: Implies no correlation.
● +1.00: Indicates a perfect positive linear correlation.

The sign of the PCC indicates the direction of the relationship (positive or negative), while its
magnitude shows how strong the association is. Higher absolute values of the coefficient
(closer to 1) signify a stronger relationship.

For example, a calculated PCC of 0.61 between Starbucks coffee consumption and income
levels suggests a moderately strong positive relationship, implying that higher income is
associated with greater coffee consumption.

Statistical Significance of PCC: To determine whether a computed Pearson correlation

coefficient represents a true correlation in the population, researchers test the statistical
significance of the coefficient:

● Null Hypothesis (H0): There is no association between the variables (PCC is zero in
the population).
● Alternative Hypothesis (H1): There is an association between the variables (PCC is
not zero in the population).

Statistical significance is typically assessed using a p-value. If the p-value is less than the
chosen level of significance (commonly 0.05), the null hypothesis is rejected, affirming that
the observed correlation is likely not due to random chance.

For instance, if the p-value for the correlation between coffee consumption and income is
0.05, it suggests that there's only a 5% probability that such a correlation would occur if there
was actually no association in the population, thus providing grounds to claim a real
relationship exists.
Rule of Thumb for Interpreting the Pearson Correlation Coefficient:

When evaluating the strength of association between two variables using the Pearson
correlation coefficient, it's helpful to apply these general guidelines:

For example, a Pearson correlation coefficient of 0.61 between Starbucks coffee

consumption and income suggests a strong relationship, indicating that as income
increases, so does the expenditure on Starbucks coffee. Conversely, a correlation coefficient
close to 0 would indicate a lack of any meaningful association between the two variables.

These thresholds provide a quick way to assess the significance of the correlation
measured, aiding in swift and effective decision-making or further statistical analysis.

Substantive Significance of the Correlation Coefficient

Coefficient of determina- tion (r2) A number measuring the proportion of variation in one
variable ac- counted for by another. The r2 measure can be thought of as a percentage and
var- ies from 0.0 to 1.00.

Pearson correlation coefficient is a primary tool used to quantify the strength and direction of
a linear relationship between two variables. This coefficient can range from -1 (perfect
negative correlation) to +1 (perfect positive correlation), with 0 indicating no correlation.

Strong and Significant Correlation:

A strong correlation coefficient (e.g., 0.776 in the case of Santa Fe Grill relating customer
satisfaction to the likelihood of recommending the restaurant) suggests a robust linear
association between the variables.

When this correlation is squared, it yields the coefficient of determination (r²), which in this
example is 0.602. This indicates that approximately 60.2% of the variability in the likelihood
of recommending the restaurant can be explained by variations in customer satisfaction.
Statistical vs. Substantive Significance:

Statistical Significance: Refers to the likelihood that the correlation observed is not due to
random chance. This is often determined through hypothesis testing where a p-value less
than a threshold (commonly 0.05) suggests significant results.

Substantive (Practical) Significance: Considers whether the strength of the correlation,

even if statistically significant, is large enough to be of practical use. For instance, a
correlation coefficient of 0.20 might be statistically significant but only explains 4% of the
variance (r² = 0.04), which might not be practically useful for making business decisions.

Spearman rank order correlation coefficient A statistical measure of the linear

association between two variables where both have been measured using ordinal (rank
order) scales.

In cases where correlations are weak or the data scales are not interval or ratio (e.g., ordinal
or nominal scales), alternative methods like the Spearman rank order correlation coefficient
are recommended. This measure is more appropriate for data that do not meet the
assumptions required for the Pearson correlation coefficient.

What Is Regression Analysis?

Correlation and regression are fundamental statistical tools used in analytics to describe
relationships between variables. The correlation coefficient quantifies the direction and
strength of a linear relationship between two variables, while regression analysis builds upon
this to predict values and test causal theories.

Correlation Coefficient:

● Purpose: Determines if a relationship exists between two variables and describes

the strength and direction of this relationship.
● Application: Used by managers to assess relationships that affect business
outcomes, such as the impact of a price increase on sales or market share.

Regression Analysis:

Bivariate Regression: A statistical technique that analyzes the linear relationship between
two variables by estimating coefficients for an equa- tion for a straight line. One variable is
designated as a dependent variable and the other is called an indepen- dent or predictor
variable.

Involves a simple relationship between two variables (one independent and one dependent).

For example, predicting sales volume based on price per unit can be modeled by the linear
equation: Sales Volume (Y)=a+b×Price per Unit (X)Sales Volume (Y)=a+b×Price per Unit (X)
Here, aa represents the intercept (often zero if no sales occur without pricing), and bb is the
slope, indicating how sales volume changes with price adjustments.
Evaluating Regression Models:

Model Assessment: The effectiveness of a regression model is evaluated by comparing

predicted values with actual values from the data. This comparison helps gauge how well the
model captures the underlying relationship and its predictive accuracy.

Assumptions: Key assumptions in regression analysis include:

1. Linearity: The relationship between variables should be linear.

2. Interval/Ratio Measurement: Variables should be on interval or ratio scales.
3. Normal Distribution: Variables should ideally follow a normal distribution.
4. Independence: Error terms should be independently and normally distributed.

Least Squares Method: This common estimation technique minimizes the sum of the
squared differences (errors) between observed and predicted values, ensuring the best fit
line through the data points.

Multiple Regression:

Concept: Extends bivariate regression to include multiple independent variables, allowing a

more comprehensive analysis of factors affecting the dependent variable.

Example: In a business setting, a manager might use multiple regression to understand how
various factors like age, income, employee satisfaction, and service speed collectively
impact customer satisfaction.

Developing Regression Coefficients:

OLS Estimation: Ordinary Least Squares (OLS) is employed to estimate the regression
coefficients that will predict the dependent variable with the least error. These coefficients
(denoted as bb) represent the individual impact of each independent variable on the
dependent variable.

Error in Regression:

Error Term (ei): Represents the discrepancy between actual and predicted values of the
dependent variable for each observation. The sum of these squared errors across all
observations gives a measure of the overall model error, reflecting the accuracy of the
regression model.

Practical Implications:

Decision Making: By understanding and applying both correlation and regression analysis,
managers can make more informed decisions, predict future trends, and allocate resources
more effectively.

Model Limitations: It's crucial to recognize that statistical significance does not imply
causality. Regression models provide predictive insights and should be interpreted within the
context of sound theoretical reasoning and business logic.
Significance

Statistical Significance:

Coefficient Significance: Initially, determining the statistical significance of regression

coefficients answers whether a real relationship exists between the dependent and
independent variables. A statistically significant coefficient suggests a meaningful
association beyond random chance.

T-tests and F-statistics: These tests assess the validity of the coefficients and the overall
regression model, respectively. The t-test evaluates if individual coefficients significantly
differ from zero, while the F-statistic tests the overall model fit, comparing the variance
explained by the model to the unexplained variance.

Substantive Significance:

Coefficient of Determination (r²): This measure reflects the proportion of variance in the
dependent variable explained by the independent variables. For example, an r² of .230
indicates that 23% of the variability in customer satisfaction can be explained by variations in
perceptions of price reasonableness.

Strength of Relationship: Examining r² and the size of the regression coefficients helps
gauge how changes in independent variables affect the dependent variable. Substantive
significance considers whether the size and impact of these coefficients are meaningful in
practical, real-world terms.
Multiple Regression Analysis

A statistical technique which analyzes the linear relationship between a dependent variable
and multiple independent vari- ables by estimating coefficients for the equation for a straight
line.

Extension of Bivariate Regression: Multiple regression allows for the inclusion of various
independent variables to better understand their collective impact on a dependent variable.
This is crucial in complex scenarios where multiple factors influence outcomes such as
business, economics, and social sciences.

Analysis of Coefficients: Each independent variable's effect is quantified by a unique

regression coefficient, which provides insights into how much a unit change in the
independent variable changes the dependent variable, assuming other variables are held
constant.

Beta Coefficients: Standardized regression coefficients, or beta coefficients, facilitate the

comparison of the impact of variables measured on different scales. They reflect the change
in the dependent variable for each standard deviation change in the independent variable.

Challenges and Assumptions

Homogeneity of Variance (Homoskedasticity): Assumes that the variance around the

regression line is constant across all levels of the independent variables. Heteroskedasticity,
where variance differs across values, can lead to inefficient estimates and affect the
trustworthiness of significance tests.

Normality: The model assumes that the distribution of residuals (errors) is normal. This
underpins the validity of many statistical tests, including the computation of confidence
intervals and hypothesis tests.

Linearity: The relationship between the independent and dependent variables is assumed
to be linear. Non-linear relationships require different analytical approaches, such as
transformations or non-linear modeling.

Evaluating Model Effectiveness

Interpreting r² and Adjusted r²: These statistics are crucial for understanding the
effectiveness of the model. While r² indicates the percentage of variance explained by the
model, adjusted r² accounts for the number of predictors and the sample size, providing a
more accurate assessment in the context of multiple regression.

Comparative Influence: Beta coefficients and their statistical significance help in comparing
the relative influence of different independent variables on the dependent variable.
Structural Equation Modeling (SEM):

Complex Relationships:

SEM allows for modeling of complex relationships where variables can be both dependent
and independent. This is useful in layered structures where, for example, an employee's
commitment might influence their performance but is itself influenced by factors like pay,
teamwork, and work environment.

Path Models:

SEM utilizes path models to illustrate the relationships among variables. These models are
comprised of multiple stages and can include intervening or mediating variables that help
elucidate indirect effects.

Partial Least Squares SEM (PLS-SEM):

For cases involving complex models with multiple stages, PLS-SEM is used. It is similar to
ordinary least squares but is designed to handle path models with multiple constructs
measured by multiple variables. This method focuses on maximizing the variance explained
in the dependent variables.

Software Tools:

Tools like SmartPLS facilitate the use of PLS-SEM by providing a user-friendly interface
where researchers can draw path models, import data, and execute the analysis with ease.
These tools also provide various outputs such as Cronbach's Alpha for reliability, validity
testing, and explained variance (R²).

Statistical Significance and Model Evaluation:

Coefficient of Determination (R²):

Reflects the amount of variance in the dependent variables explained by the independent
variables. For instance, in the Santa Fe Grill example, the R² values indicated strong
predictive power of the model regarding employee commitment and performance.

Statistical vs. Substantive Significance:

SEM analyses both the statistical significance of paths (using t-tests for individual paths and
F-statistics for overall model fit) and the substantive significance, which considers the
practical importance of the relationships identified.
Advantages of PLS-SEM in Research:

1. Flexibility in Measurement: Handles all types of data scales—nominal, ordinal,

interval, and ratio.
2. Nonparametric Nature: Suitable for data that do not follow a normal distribution.
3. Sample Size Versatility: Effective with both small and large samples, adaptable based
on model complexity.
4. Complex Model Capability: Can analyze intricate models with numerous constructs
and indicators.
5. Comprehensive Analysis Options: Supports various analyses like mediation,
moderation, and importance-performance mapping.

Practical Considerations and Recommendations:

Corrective Measures for Weak Constructs: If certain variables or constructs perform

poorly (low reliability or weak loadings), they should be reevaluated, adjusted, or potentially
removed to enhance model accuracy.

Predictive Capability: The ultimate test of a structural model's utility is its ability to predict
outcomes of interest effectively, as demonstrated by the predictive R² values in the models.