Data collection methods

Determinants of data collection methods

 The method of data collection in any research project must be
appropriately described in the data collection section of a
proposal.

 Data collection methods are determined by four important

considerations:
– The general and specific objectives/questions
– The type of data required to achieve the objective/questions
– The type of research design used
– The type of analysis
 Generally, there are two types of methods
– Quantitative methods
– Qualitative methods

 While in behavioural research both quantitative and qualitative

methods may apply, in natural sciences, quantitative methods are
used.
 Tools and Methods of data
collection Quantitative research
 Definition
 A questionnaire is the most common instrument/tool for
collecting data in quantitative research.

 It consists of a set of interrelated questions and other

prompts designed to collect data from respondents for the
purpose of achieving the objectives of a research.

– It is a good tool for collecting data used to describe,

compare, or explain an event or situation, as well as,
knowledge, attitudes, behaviors, and/or socio-
demographic characteristics on a particular target group.

 They are usually designed for statistical analysis of data.

 Two types of questionnaires:

• Structured/closed ended
• semi/unstructured
 Types of questionnaires
 Structured- Closed-ended
– standard answers provided (coded)
– facilitates answering sensitive questions
– generally considered easy for respondent

 Unstructured- Open-ended
– respondent may answer as they choose
– exploratory; allows detailed response
– preferable for complex questions
– may be difficult to code and score
– requires more time and effort from respondent
 Steps to design a questionnaire:
 Designing questionnaires is a complex process that
follows several steps:
– Review out the main goal and specific
objectives/questions of your study;
– Develop concepts/information based on the

6
objectives to be collected;
– Review literature for already validated
questionnaires or aspects of questionnaires that
measure your specific area of interest;
– Compose a draft of your questionnaire;
– Assemble the final questionnaire; and
– Pretest the draft questionnaire for validation.
 To develop an effective questionnaire, 5 principles are
used.
The questionnaire design process
 Principles of questionnaire design
 There are six main principles in questionnaire design that
must be followed:
– Question format (Structured/unstructured)

– Types of questions/responses (measurable for analysis)

– Question content (collectively exhaustive and mutually exclusive

to avoid respondent confusion)

– Question wordings (simple, straight forward, clear and

unambiguous )

– Question placement or sequence (should progress from simple to

complex/from general to specific/personal/private. Questions
must be placed or asked in the order that will minimize loss of
data.

– Question layout (title, attractively formatted, reasonable length,

consistent in wording, correct spelling, legible font etc)
 Pre-testing and correction
 Pretesting is administering a questionnaire to a limited number
of potential respondents and other individuals capable of
pointing out design flaws

 Purpose of pre-test:
– To ensure that the questionnaire meets the expectations in
terms of the information it is intended to collect.

 Some of important issues to be noted during pre-testing are:

– Necessity of questions. Do question serve intended

purposes?

– Are there important issues been included? Match questions

to objectives? and

– Clarity of questions.
Methods of administration of
questionnaire
Methods of questionnaire administration
 If the questionnaire is a method of choice of questionnaire
administration, the researcher has an obligation of explaining in
the proposal how it will be done. This is because in many settings,
certain methods are simply not applicable.
 Three main methods:
– Face to face interview

– Personal /self administration

• Mail/post
• Email
– Telephonic interviews

 The face to face method is the most applicable in most developing

country settings including in Uganda.
 Face to face interview
 The face-to-face interview is a method of data collection that
utilizes a conversation approach between the enumerator and the
respondent.

 Advantages:
– Participation by illiterate people
– Clarification of ambiguity
– Quick answers
– High response rate
– Tolerable longer interview

 Disadvantages:
– Interviewer bias
– Needs more resources
– Difficult for sensitive issues
– Time consuming
 Tools and Methods of data
collection Qualitative research
 Qualitative data collection methods

Four main methods

– Focus group discussions

– In-depth interviews

– Observatory methods

– Content analyses
 The instrumentFocus Group Discussion
of data collection is a Focus Group Discussion (FGD)
guide.

 The guide comprise of a limited set of questions and specific probes.

– The guide collects information aimed at constructing group
meanings and norms, values, perception, attitudes about an issue.

– It should not be intended to collect personal level information and

data that is private and sensitive in nature.

 The FGD guide is administered through an interactive group

discussions with a small number of homogeneous individuals (7-12).

 The method is intended to collect a narrative rather than measurable

data.

 The discussion process must be taped to keep the correct recode of

what is said by participants .

 The discussions should usually not exceed 2 hours if participants are

to remain focused.
 FGD organizing tips
 It is not enough to say that data will be collected by using
FDGs. The researcher should show the step by step process
that will be used, including the following::
– Develop the key themes around which FGD guides will be
developed;
– Identify the target population from whom data will be
collected;
– Identify the relevant participants groups from whom data
will be collected
– Show how the participants will be identified for the FGD
discussion;
– Demonstrate the structure the FGD will take ensuring
homogeneity between groups;
– Demonstrate how the FGD will be conducted by
explaining the roles of parties in the discussion process.
– A statement on the experience of the moderator and note
taker should also be provided.
 In-depth interviews
This is a detailed one to one interview.

Uses a conversational approach aimed at

obtaining in-depth information based on the
individuals construction of reality in terms
of knowledgeable and practice.

It uses in-depth interview guide as a tool of

data collection.

 The interview guide can be unstructured or

semi-structured with probes.
 In-depth interviews-tips
 In assessing proposals, reviewers should ensure
researchers present the following if in-depth
interviews are proposed as methods of data
collection:
– Present the main themes around which in-depth
interviews will be conducted;
– Identify key informants who will constitute the
sample for in-depth interviews;
– Present how the informants will be sampled and
identified for interview;
– Present the settings under which in-depth
interviews will be conducted; and
– Ensure that in-depth interviews are recorded to
ensure correct and verbatim record of the
process.
Discussion
Data analysis methods
 Methods
 A sound research proposal must clearly explain the methods
of analysis.

 The choice of analysis is determined by among other things

the:
 Research design
 Data collection methods (qualitative/quantitative)
 Type of data (qualitative/quantitative)
 Data formatting/data structure (continuous/categorical)
It is important to note that software such as
SPSS/STATA/Atlas-ti are not methods of
analysis.
 Planning for Analysis

Type of Type of
Data Formatting

Type of
Analysis
 Quantitative analysis
 Quantification
 Before we can do any kind of analysis, we need to quantify our data
 “Quantification” is the process of converting data to a numeric
machine readable format.

– Reduce a wide variety of information to a more limited set of

variable attributes by codes.
 Example of Coding
– “What is your occupation?”
• Use pre-established scheme: Professional, Managerial,
Clerical, Semi-skilled, etc.

• Assign value to each category in the scheme: Professional =

1, Managerial = 2, etc.
 Quantitative analysis
 Data entry
– The software for data entry should be mentioned. This is
usually Excel, EPI-INFO or SPSS.

– Computer entered data should be upload to an

appropriate analysis software: SPSS, STATA, SAS.
 Quantitative Analysis
 A Choice a level of analysis that is appropriate for the
research objectives/question should be described.

 The type of statistical analysis appropriate for the variables to

be measured in each objective/question should be described.

 This is determined by the variable structure which can take

the following formats:
 Nominal/Categorical
 Ordinal or
 Continuous

 Data processing is a critical step before data analysis.

 Quantitative Analysis
 In quantitative analysis, there are usually three levels of analysis.

 Each level of analysis is intended to address/solve a different issue

in the analysis process. These are:
 Univariate - simplest form, describe cases by single variable.

 Bivariate - subgroup comparisons, describe cases in terms of

two variables simultaneously.

 Multivariate - analysis of two or more variables

simultaneously.

 More Details required in practical sessions

 Univariate Analysis
 Describing cases in terms of the distribution of attributes
that comprise it.
 A.K. A Simple statistics summarize values of one variable
 Usually presented as:
– Frequency distributions
– Mean, Median, Mode

 This is important for checking data were entered

properly, planning further data transformation and for
planning the type of further analysis.
 No conclusions about a research can be made at this
level.
 Bivariate Analysis
 Describe cases in terms of two variables simultaneously.

 Often used as a measure of associations between the

independent variables and the dependent variable.

 Proof of association or no association is a correlation

matrix (continuous variables) or a chi-square statistic
(categorical variables) which ever is appropriate.

At this point of analysis some conclusions can be

made.
 Multivariate Analysis
 Analysis of more than two variables simultaneously.
 Can be used to understand the relationship between
multiple variables more fully.
 This level of analysis is also called regression analysis.
 Type of regression analysis will depend on the type of
variables
– OLS (continuous)
– Logistic (other types)
– Several other regression methods ….
 It is only at this level of analysis that conclusions about
research results can be meaningfully made.
 It is only at this level that results from a sample can be
meaningfully generalized to the target population.
 Analysis of FGD and in-depth interview data
 Data analysis is usually done by the use of the master sheet
approach. In the analysis section, a researcher should clearly
state the steps bellow.
 This involve the following steps:
– Transcribe data verbatim from tape recorder;
– Organize transcribed data into themes;
– Code data within themes using unique codes;
– Summarize coded data by themes for each focus group
manually or using a qualitative data analysis software eg
Atlas-ti.
– Interpret results and write up;
– Use topical statements from the transcripts verbatim to
support findings.
 An example of a manual master sheet analysis is indicate below :
 Master sheet approach
Participant 1 Participant 2 Participant 3 Summary

Themes

1
2
3
4
5
 Mixed Methods
A combination of quantitative and qualitative
techniques.

 Mixed methods research

– A research that uses both quantitative and qualitative
procedures of data collection.

– The purpose of the method is to build upon the synergy

and strength of both quantitative and qualitative
methods in order to more fully understand a given
phenomenon.
– To use this approach a good justification must be provided
in the introduction/background of the proposal.
 The research design
 The descriptive non- experimental design is most appropriate.
 It can however be ordered in such a way as to draw greater emphasis
either on both or on one of the approaches:
 These can take the following approaches:
– QUAL-Quan Model
• Qualitative data are collected first and are more heavily
weighted

– QUAN-Qual Model
• Quantitative data are collected first and are more heavily
weighted

– QUAN-QUAL Model
• The triangulation mixed methods design
• Quantitative and qualitative data are collected concurrently
and both are weighted equally
 Ten Characteristics of Mixed Methods
 The title of the research includes terms that suggest more than one
method is being used
– Mixed methods
– Integrated
– Triangular
– Quantitative – qualitative
 The researcher describes the kinds of mixed methods being used
 The data collection section indicates narrative, numerical, or both
types of data are being collected
 The purpose statement or the research questions indicate the types
of methods being used
 Questions are stated and described for both quantitative and
qualitative approaches
 Ten Characteristics of Mixed Methods
 The researcher indicates the sequencing of collecting
qualitative and/or quantitative data (i.e., QUAN-
Qual, QUAL-Quan, or QUAN-QUAL)
 The researcher describes both quantitative and
qualitative data analysis strategies
 The writing is balanced in terms of quantitative and
qualitative approaches
Discussion
Research Ethics
 A statement of ethics
 A statement of research ethics is required in all research proposals on
human and animal populations or that have direct and indirect impact on
these research subjects.

 These include:
 Social science research
 Medical research
 Environmental research
 All population based physical science research

 The purpose of the statement is to demonstrate how adverse research

impacts on humans and animal will be prevented or mitigated.

 In many cases specific statements of adverse impact prevention and

mitigation are required.

 However in the more general form, evidence that the research protocol
has been cleared by responsible IRB is adequate.
 A statement of ethics
 Usually, the following aspects of research ethics is statement:
 Approval will be obtained by University IRB;
 Approval will be obtained from UNCST in the case of Uganda;
 Informed consent will be obtained from the respondent or
guardian if the respondent is a minor or incapacitated;
 All rights of participants will be observed including:
• Privacy
• Confidentiality of information collected
• Anonymity
 Other assurances against abuse and exploitation may be
required on a case by case basis.
 Researchers should as much as possible refer to the
international research ethics guideline.
Citation and referencing formats
 Citation
 This is an anti plagiarism requirement.

 Used to acknowledge the use of other authors work within

the text.

 There are many different citation formats.

 Citations and referencing management software: Eg Zotero,

Mendely, EndNote

 Example of APA:
– In-text citation: One author (Omuri, 2016).
– In-text citation 2 authors (Omuri and Nalwanga, 2016).
– In-text citation three or more authors (Omuri et., 2016).
– Quotation (Omuri, 2016, p 20).
 Reference
 Book
– Lumby, J. (2001). Who cares? The changing health care
system. Sydney, Australia: Allen & Unwin.

 Book chapter
– McKenzie, H., Boughton, M., Hayes, L., & Forsyth, S. (2008). Explaining
the complexities and value of nursing practice and knowledge. In I.
Morley & M. Crouch (Eds.), Knowledge as value: Illumination through
critical prisms (pp. 209-224). Amsterdam, Netherlands: Rodopi.

 Journal article
– Boughton, M., & Halliday, L. (2008). A challenge to the menopause
stereotype: Young Australian women's reflections of 'being diagnosed'
as menopausal. Health & Social Care in the Community, 16(6), 565-
572. doi: 10.1111/j.1365-2524.2008.00777.x

 Webpage with an author

– Welch, N. (2000, February 21). Toward an understanding of the
determinants of rural health. Retrieved
from http://www.ruralhealth.org.au/welch.htm
 Reference
 Webpage with no author
– ANCI national competency standards for the registered nurse and the
enrolled nurse. (2000).  Retrieved
from http://www.anci.org.au/competencystandards.htm

 Newspaper article
– Bagnall, D. (1998, January 27). Private schools: Why they are out in
front. The Bulletin, pp. 12-15.

 Government publication
– The Health Targets and Implementation (Health for All) Committee.
(1988). Health for all Australians.  Canberra, Australia: Australian
Government Publishing Service.

 Company and Industry Reports

– Magner, L. (2016). IBISWorld Industry Report OD5381. Coffee Shops in
Australia. Retrieved from IBISWorld database.
Discussion