HMPYC80_2024_TL_104_0_B

HMPYC80/104/0/2024
Tutorial Letter 104/0/2024
Research Methodology
HMPYC80
Year module
Department of Psychology
Tutorial letter 104 contains various important
research resources.
2
Contents Page
A. RESEARCH AND LEARNING TOOLS

About this resource 5
Learning activity: What is it? 6
Project planning and management 8
Research and course journal 11
Portfolio 15
B. RESEARCH OVERVIEW
Research: What is it? 21
C. ETHICS AND RESEARCH

Ethics 24
Academic integrity and avoiding plagiarism 29
Declaration of academic honesty 31
D. REFERENCING AND FORMATTING

Academic writing tips and suggestions 32
American Psychological Association formatting and style (APA7) 36
E. PROPOSALS, ARTICLES AND PRESENTATIONS

Research proposals 41
Scientific journal articles 46
Conference presentations 55
F. SOCIAL SCIENCE RESEARCH PARADIGMS

Paradigms 62
Positivism 66
Interpretive research 68
Constructionist research 70
Triangulation 72
G. QUANTITATIVE RESEARCH DESIGNS AND STATISTICS

Quantitative research problems and questions 73
Experimental research designs (Relational‑causal designs) 76
Variables and measurement levels 79
Online statistical resources and software 82
Relationship between variables 83
Hypotheses 95
Hypothesis testing 99
Data entry form for quantitative data analysis 103
<Selecting the appropriate statistic> . 105
<Correlation coefficient> . 109
<T-test> . 110
<One-way analysis of variance> . 111
<Factorial analysis of variance> . 112
<Simple regression analysis> . 113
<Multiple regression analysis> . 114
<Manifest and latent variables> (Intercorrelation matrix and Factor Analysis) . 115
T-test 117
Correlation coefficient 126
One-way analysis of variance 131
HMPYC80/104 3
HMPYC80/104/0/2024
Factorial analysis of variance 135

Regression analysis 147
Multiple regression analysis 151
Latent and manifest variables (Intercorrelation matrix and Factor Analysis) 161
H. QUALITATIVE RESEARCH DESIGNS AND METHODS

Qualitative analysis 178
Text: What is it? 183
Observation 184
Interviews 185
Content analysis: An example 186
Thematic Analysis 190
Discourse analysis 194
Ensuring Trustworthiness of the Data 199
Bracketing 201
I. APPLICATION OF THE RESEARCH REVIEW INVENTORY (R.R.I. revised)

Application of the Research Review Inventory (R.R.I. revised): Example 1 202
J. Example research proposal with Turnitin report

Research proposal with the Turnitin mark-up and report 249
4
About this resource

This tutorial letter is a resource that complements the readings in the prescribed book. Each section
consists of various resources that have a title that briefly indicates what that resource is
about. The section covers a particular theme, that is sorted into the following categories:
A. Research and learning tools
B. Research: Overview
C. Ethics and research
D. Referencing and formatting
E. Research proposals, articles and presentations
F. Social science research paradigms
G. Quantitative research design and statistics
H. Qualitative research designs and methods
I. Application of the Research Review Inventory (R.R.I. revised).
Resources can refer to other resources, and such references are indicated by enclosing the title of the
resource between angular brackets (< ... >). For example, if the Correlation Coefficient resource has a
reference to Hypothesis Testing the reference would be embedded in the text as follows: “In <hypothesis
testing> there are a number of different methods ...”. This means you can learn more about hypothesis
testing in a resource called: Hypothesis testing.
The learning activities in Tutorial Letters 102 and 103 have references to these resources. Each learning
activity has a resource section. The resource sections contain references to the resources in question. Use
the 5 sections and their various resources when you need the information for a particular learning
activity.
HMPYC80/104 5
HMPYC80/104/0/2024
Learning activity: What is it?

A Learning Activity is a key element of the learning model used in this module. In generic terms, a
learning activity can be defined as the range of activities promoted to achieve learning. However, the
learning activities are shaped by a learning model. The learning model is the way in which the subject
material is presented and the way in which you have to study the material. You do not have to study the
learning model itself.
Let us begin by clarifying the terminology that is associated with the learning model. These terms are:
 Learning outcome: A learning outcome describes the purpose of learning. The entire learning
process is geared towards achieving learning outcomes. Learning outcomes
are achieved by producing specific outcome products.
 Outcome product: An outcome product is the result of the learning activities a learner engages in
during the learning process. In other words, the learner produces particular
outcome products during the process of learning. The way in which a learner
produces an outcome product is a production method
 Resource reference: One requires resources to produce outcome products. For example, one may
require knowledge about certain issues before one can discuss these issues. A
resource reference indicates where a required resource can be found.
 Learning activity: The learning process is realised through learning activities. A learning
activity is an opportunity to work towards achieving the required learning
outcomes by producing particular outcome products. A learning activity has
three elements, namely (a) an outcome product (b) the method for producing
the outcome product, and (c) and a reference to the resource required for
producing the outcome product. Note that a learning activity often contains
other learning activities in its method. In other words, a number of learning
activities can be grouped into a single learning activity.
The outcome product is the most important element of a learning activity. It regulates the activities that
occur during the course of learning. In other words, each action of a learning activity is geared towards
the production of the outcome product. An outcome product has to fulfil certain standards. If a learner
produces an outcome product that fulfils certain minimum criteria the learner can be declared competent.
However, if the outcome product does not realise the minimum standards the learner has to be declared
not yet competent.
An outcome product can be produced in different ways. If one is experienced and competent one can be
flexible about the exact method one wishes to use to produce a particular outcome product. However, if
one is in a process of learning it is better to follow a prescribed method. In this module, we prescribe
specific methods of production by outlining particular activities and actions. By executing these actions
you learn the most efficient and valid methods to produce the outcome products in question.
Resource material is material that is necessary for the successful execution of activities and actions. The
resources required by the production method are referenced in the resource reference section of a
learning opportunity. You have to study the required resource material in order to complete the actions
6
successfully.
The following is a generic description of a learning activity:
Learning activity
Outcome product
A description of the outcome product and the standards that the product must fulfil. The standards
are often indicated in terms of the range of the product and criteria for assessing the quality of the
product.
Production method
Activity 1 (consisting of one or more actions)
Activity 2 (consisting of one or more actions)
etc.
Resource reference
References to resources required for executing the various activities and actions
The next table depicts the HMPYC80 course as a single learning activity. Tutorial Letters 102 and 103
describe the method of this learning activity and each learning activity.
Learning activity
Study research methodology
Outcome product
An initial research proposal and a follow-up research proposal
Production method
Activity 1. Formulate the initial (1st) research proposal
Evaluate a research article
Evaluate research proposals
1.1 Plan a research proposal
1.2 Find a research problem
1.3 Formulate a research statement
1.4 Design a feasible research project
1.5 Plan data gathering and analysis
1.6 Write a research proposal
1.7 Evaluate research proposals and an article
Activity 2. Study the research data collection and interpretation processes
Evaluate a research article
Formulate a follow-up (2nd) research proposal
Evaluate the research proposals
2.1 How to plan the research project
2.2 How to gather and analyse research data
2.3 How to structure research information for presentation
2.4 How to report research in the form of a written article
2.5 Evaluate the follow-up (2nd) research proposal and a research article
Resource reference
Fouché, C. B., Strydom, H., & Roestenburg, W. J. H. (2021). Research at grass roots: For the
social sciences and human service professions (5th ed.). Van Schaik.
Tutorial Letters: See HMPYC80 Tutorial Letters 102 and 103.
HMPYC80/104 7
HMPYC80/104/0/2024
Project planning and management
Project management has as its aims to keep a project on course, on time and budget. It may best be
defined as the planning and controlling of resources (time, energy, people, funds, etc.) for a relatively
short period of time to achieve a specific goal or objective. Furthermore, it often involves managing a
team who are working toward a final complex outcome or product. You are probably only managing
yourself within the context of your research project and learning experiences. However, these principles
are an invaluable aid to successfully completing any project. The project approach is an attempt to
effectively manage the scarce resources (your time, money, motivation and effort) in order to achieve a
successful project outcome.
Any project planning starts by determining what needs to be done by whom and by when in order to
fulfil the assigned responsibility. To do this one has to define your objectives and goals, defining the
exact outcomes that one expect and how they can be assessed. You have to set up a programme of
actions, schedule your time planner, set up a budget (if necessary), and then continuously monitor,
evaluate and correct your progress.
PROJECT PLAN STRUCTURE
A good research project management plan includes the following:

 A definition of the goals/objectives of the project (for both your research proposal and for your
successful completion of the course as a learning experience). A list of guiding principles may
also be helpful.
 A definition of the exact outcomes that you expect and how they can be assessed.
 Listing all the important activities needed to complete the project.
 Project plan: develop a time schedule, budget and process plan.
 Implement the plan.
 Monitor and control the project.
 Evaluate.
JOE'S PROJECT PLAN (example):
GOALS/OBJECTIVES
My primary goal is to finish a high-quality research proposal on time, which will be complemented
by a Research and Course Journal and a Portfolio. I aim to finish all the activities in this course on
time to pass the module. My ultimate objective is to learn how to do good and enjoyable research.
I am currently exploring the topic of recreational kite flying in a materialistic world.
The principles which I find important are: to have fun, to learn a lot, to do my work on time, to open
myself to new learning experiences, to always reflect on what I have done and want to do next, to
always respect the people involved in my research, and to spend quality time with my family and
friends.
8
EXACT OUTCOMES AND HOW THEY CAN BE ASSESSED
The exact outcomes are specified in each Learning Experience and the Assignments. They are
always linked to my Research and Course Journal entries and the Portfolio that I will have to
submit.
ACTIVITIES NEEDED TO COMPLETE THE PROJECT

The outcomes indicated in each Learning Experience will be my guiding structure. The activities
needed are:
· Plan and manage my research proposal: Create and continuously update a Research and
Course Journal, plan my Portfolio, and design a project plan and schedule.
· Formulate a research interest or problem: .... (list)
· Formulate a research statement: .... (list)
· Design a feasible study: .... (list)
· Identify appropriate information: .... (list)
· Select data analysis technique: .... (list)
· Write the proposal: .... (list)
· Evaluate and improve the proposal: .... (list)
In addition, the activities in the assignments will be important, and I need to remember the
scheduled assignment dates.
TIME SCHEDULE
BUDGET
Photostat and printing costs for portfolio assignments: +- 300 pages @ R1.00 per page = R300
Photostat and printing costs during literature search: +- 500 pages @ R1.00 per page = R500
Internet use for 10 months: contract/pay-as-you-go bundle x 10 months @ R250.00 = R2500
TOTAL = R3300
PROCESS PLAN
The activities needed to complete the project can be linked as a mind map or diagram, showing
which actions are dependant or prior actions. Link the activities and show how they join to form
final research products.
IMPLEMENT THE PLAN

Use the ADORE principle <Research and course journal> to trouble shoot problems during
implementation.
MONITOR, CONTROL AND EVALUATE THE PROJECT

Continuous monitoring of my learning and of the research will be achieved through reflection
using the ADORE questions in my journal. The final evaluation will be a combination of all the
activities and the assignments of Module A (Research proposal) and Module B (Research article)
in my Portfolio. The portfolio will rely heavily on my entries in my Research and Course Journal.
TIME SCHEDULING AND PLANNING
Time planning and scheduling are essential parts of your project plan. Here are some principles that you
will have to use to create your very own time planner:
 First plan your schedule in rough.
HMPYC80/104 9
HMPYC80/104/0/2024
 Indicate the dates for non-negotiable deadlines where a specific product will have to be
delivered. Refer to the [Assignments] due dates and the [Study Schedule] in your first tutorial
letter.
 Determine which activities are dependent on the successful completion of other activities.
Indicate these and mark them as potential bottlenecks in your research plan.
 Allow for postage and delivery, moving your schedule a few days earlier to accommodate these.
 Now move your entire time schedule another 2 to 3 weeks earlier, in order to allow for
unforeseen circumstances which always seem to occur (such as a personal, family or work
crises). Remember that it will only take 2 weeks of slacking down to pull your entire project plan
apart with possible dire consequences for your course work.
 Finalise the dates when certain products need to be delivered and write the final proposed dates
clearly in black ink. Use a black pen in order to facilitate copying for assignment submission
purposes.
 Display the time schedule prominently whilst you are busy with activities on this course (e.g. on
a piece of cardboard on the wall) and adjust the planning whenever necessary.
Project planning and management is a well-developed field and this introduction is a very sketchy
beginning to the field. If you are interested in more information the following key terms will be useful:
need assessment/analysis, situation analysis, project management, programme management, network
analysis, work breakdown, problem structure analysis, network analysis, bar charts, financial and
economic sustainability, community participation, management by objectives, programme evaluation,
monitoring and evaluation, information dissemination. There is also a myriad of resources on the World
Wide Web. Any reputable search engine will be able to point you to these resources and references.
SNAFU'S and Murphy's law:
You are probably already familiar with Murphy's law that goes something like this: IF
SOMETHING CAN GO WRONG IT PROBABLY WILL. This describes the normal state of
affairs as described in that great book by Joseph Heller "Catch 22": Situation Normal, All Fouled
Up (SNAFU).
Assuming that things will or can easily go wrong is a great help to good planning. The reverse of
Murphy's law also holds equally true: if you plan for something to go wrong it probably won't
(because you are prepared and have a problem-solving strategy). This is why you should be such an
excellent planner when doing research.
How to make Murphy's reverse law work for you:

* If you do your work on a notebook or personal computer always make backups of your
work every time you finish a work session. According to Murphy your computer, if not
well-managed, is a disaster waiting to happen.
* Don't leave your work or your notebook in a parked motor vehicle. It might get stolen.
* Don't eat or drink over your computer.
(You should be getting the picture by now!)
* Don't leave everything until the last minute – there usually isn't enough time to finish it.
* Make contingency plans, so that if Plan A does not work you can try Plan B or Plan C.
* Don't be a victim of circumstances - empower yourself with good planning, timely action
and a wise backup strategy!
10
Research and course journal
Getting into the habit of journaling is indispensable for any portfolio-based learning. The act of writing
is central to this approach. For example, one has to make notes when observing phenomena, jot down
ideas when reflecting on issues, formulate plans of action and define important concepts. Eventually, all
of this has to be documented properly. Experience has shown that keeping a journal is one of the most
effective ways of organising and managing the writing process. It is therefore essential that you form this
habit as soon as possible!
The professional journal is to be distinguished from the typical logbook as well as the well-known diary.
It not only includes factual descriptions (logbook) and free-flowing impressions (diary), but also
analytical and interpretative notes.
It could be useful to start each new entry on a separate page and to date it (upper right-hand corner of the
page). This allows for easy access, cross-referencing and proper sequencing of events when writing up
reports. (A MS Word document or a two-ring binder with loose sheets of paper is ideally suited to this
purpose.) Always write in black ink in order to facilitate copying and photostating of your work for
assignment submission purposes. Or, type your work on an electronic device such as a laptop, desktop
computer or other mobile device.
In addition to a date on each page, indicate with headers the topic/index of the description. Using an
action research approach to journaling, the index categories could be: Plan, Act, Observe, (Document),
Reflect and Envisage. The two main categories involved are (a) your planning and management of the
research project and (b) your experience of the course as a learning process. Write too much rather than
too little.
Log your experiences diligently. Unless you get into the habit of journaling, all your attempts will be in
vain. Failing to record your observations, impressions and ideas on a continual basis will surely result in
the loss of indispensable “raw” data! Needless to say, without the latter neither plotting your position nor
changing course will be possible.
Plot your position regularly. Keeping your nose to the proverbial grindstone can be ensured by taking
stock at set intervals. This is greatly facilitated by completing a Research and Course Journal entry using
the ADORE questions on a weekly/ bi-weekly basis. In doing so you are actually writing regular
progress reports.
AIM (What did you want to achieve?)

DO (How did you go about it?)
OBSERVE (What happened as a result?)
REFLECT (What did you make of it?)
ENVISAGE (What do you consider to be your next step?)
Also consider the KIS(-S) principle: Keep It Simple, but not Stupid!
In other words, make the argument and academic debate as simple as possible, but not over-simple or
dumbed-down to the point where it becomes ridiculous and irrelevant. (Hint: look up “Ocam’s Razor”
on the internet.)
HMPYC80/104 11
HMPYC80/104/0/2024
Some other questions that you could address include the following reflective statements:
· What did I do?
· What does this mean?
· What have I learnt?
· What have I not learnt?
· What must I unlearn?
· What is the value or purpose of that learning?
· What insights have I had from this learning?
· How has my ability to accomplish new activities or tasks grown from these insights?
· What aspects have I not explored in depth?
· What aspects have I explored in depth?
· What are the problems that will prevent me from achieving my objectives?
· What can I do to deal with these problems?
· How can I do things differently?
· How can I apply this knowledge to my individual situation - in other words: What is in it for me?
· What aspects would I like to pursue in future?
·
(Also refer to <portfolio>)
DISCERNING PATTERNS
It should not take you long to record a sufficient number of useful observations. Once that is done you
will be in a position to examine the available “raw data” more carefully. The object of this exercise is to
detect recurring patterns of behaviour. Since every action on your part is bound to be followed by some
reaction from the environment, you need to identify the typical sequence of behaviours that ensue in
dealing with others. Whatever you say or do affects the way others respond to you - and vice versa. In all
probability, your focal concern relates to a situation where you are not achieving the desired reaction.
Hence the need to explore its essential dynamics. This may sound like a tall order but it really is not. All
you need to do is to leisurely scan your recorded observations a number of times. After a while, the
essential pattern will show itself. If not you can consider discussing it with a friend to help interpret it.
Consider the example of Ms Academia who is concerned about her progress in her studies. She always
seems to be running behind schedule. She, therefore, decides to document her experiences during the
process of her studies and learning. Here are some of her journal entries:
(1) Reflection - Research problem

21 April, 19:00
Today I started formulating my research issue. I found it quite difficult to select a specific
problem, perhaps because I see so many issues that interest me. I think I would like to focus
on the problem of violence and why older people are less likely to engage in violence, as
opposed to younger people (18-35 years). But I also fancy looking at domestic violence and
why some spouses show resilience to violence and others just go into a downward spiral. I
know this sounds a bit gory, and it makes me feel a bit scared to want to investigate these
topics that have always been a taboo in our house. It reminds me of the time that I lost my
temper and took it out on my youngest child - not that he was seriously hurt - but it's just that
it seems so easy to get seriously violent. I wonder how I could have handled it better? The
literature I've been reading suggests a "time out" period for both parent and child. I think I'll
give it a try.
12
(2) Observation - Course
22 April, 22:00
I have just revised my time schedule and must admit that I am quite irritated that the study
materials arrived so late. This has upset my plans and I must now work harder to complete
all the steps for my first few assignments. But I suppose this is a normal Snafu that occurs
regularly in the real world. I have high hopes for this course as it seems to be based on very
real concrete outcomes that will really build my research skills. The course is structured in a
very different way to all the other courses I have done. I feel quite anxious, working in this
strange learning experience way. Well, the only way to overcome my fear is to "plan my work
and work my plan", and continue steadily each week. We are doing a lot of fascinating
things, aren't we!
Reflection - I'm Stuck

(3) 25 May, 23:30
Oh dear, I don't know what to do. My research proposal needs to be submitted in two weeks
time and I requested the library materials too late. Where will I find enough literature? OK,
breathe calmly and deeply - and relax ...
I seem to have two options - either submit the proposal late and lose a lot of assignment
credits, or work amazingly hard over the weekend, perhaps taking Monday off to finish the
work. I didn't want to get into this spot, and really tried so hard to stay on schedule! Think I
need to go and jog or gym to get all this frustration out of my system. Afterwards, I'll visit my
friend Marvin and see if he can do a quick Internet search for me to find some good
references. Maybe I will check out some articles in our local library as well. Well, I'm off for
a long jog with the dog and then I am sure the answer and way forward will present itself to
me.
Completing the task of “pattern detection” is bound to arouse your curiosity. (Nothing stimulates the
mind like becoming more aware of self-in-action!) Your natural response will be to ask questions like:
What is the meaning of this?, Why is this happening?, Am I my own worst enemy?, Am I at fault, or does
it take two to tango?, or What should I do? And this is how it should be! Just think about it - without
questions, serious thinking is impossible and without serious thinking, a layered and context-specific
nuanced understanding of the phenomenon is not possible.
Establishing cause: Linking cause and effect is one of the hallmarks of scientific investigation. It is also
part and parcel of daily living. Unless we know what causes a particular problem, it is difficult to solve
it. Yet, establishing a cause is seldom easy - especially when dealing with human behaviour. Admittedly,
considering all possible aspects (variables) poses a formidable task. Nevertheless, we have to try our
best. Returning to Ms A’s journal should prove to be instructive.
(4) Reflection - aims and doing

2 June, 19:50
In reviewing my work up to date (using the reflective questions), the following pattern seems
to emerge. I seem to start off with great enthusiasm but little knowledge of what a task or
learning experience entails. This means that my planning is always too optimistic. The time
schedule and my list of principles have been a great help, as they usually show me by the end
of the week that I need to put in more effort. Then I work myself into a tizz on Friday evenings
and find that I have to rest the whole of Saturday. I think I will add to my list of principles that
I will be more realistic in my planning and work more steadily each evening during the week
so that I have more time for myself over weekends.
HMPYC80/104 13
HMPYC80/104/0/2024
There is an important lesson to be learnt here. When things do not go our way a bit of soul-searching is in
order. But that in itself is not sufficient in carrying the day. We need to go much further. We need to
generate as many different explanations as possible for the situation in which we find ourselves. Enough
said! The ball is now in your court. Follow the above-mentioned procedures in analysing your own
particular situation and then summarise the results with the help of the ADORE questions.
14
Portfolio
Written by MH (Daleen) van Niekerk (Teaching advisor at the Unisa Bureau for University Teaching).
Adapted for this course by Johan Kruger.
WHY PORTFOLIO ASSESSMENT?
A portfolio will help you as a learner to integrate your experience into the learning process and to move
from dependent to more independent/self-regulated learning. It provides an assessment tool in an
authentic learning environment. The portfolio as an authentic assessment tool provides the opportunity
to assess your performance in contexts more like that encountered in real life, that would include basic
skills such as your ability to: frame problems, find information, evaluate alternatives, create ideas and
products, invent new answers, plan cooperatively, analyse, and solve problems.
Therefore, the focus in the portfolio is on assessing your performance and development as a dynamic
learning process rather than merely your ability to recall and recognise facts or other static learning
products. You are given the opportunity to provide evidence of reflection and self-assessment. Portfolio
work can enable you to use your own experience, while exploring ideas in depth, and to evaluate the
process you go through and the product you create. Portfolio work will provide an opportunity for
authentic assessment - that is, a sample of your performance in a context more like that encountered in
real life.
1. Portfolio can be used to demonstrate professional development and learning (you can use it
when applying for a job to demonstrate your competence!).
2. Portfolio work provides a representation of your personal growth;
3. Portfolio work offers a foundation for introspection, reflection, future goal setting and
action.
PUTTING A PORTFOLIO TOGETHER
Although you will work towards final products, the process of learning is of equal importance. You will
be able to investigate how your skills have changed over time, rather than just evaluating the final
product.
Very often in a traditional teaching/learning environment learning takes place where the task or goal is
determined by the lecturer, the plan of action for reaching the goal is often prescribed and so are the
assessment strategies. The learner is only involved in carrying out the plan.
With the portfolio the ideal is to involve students in all of the following activities:
· deciding on tasks or goals
· planning the stages of various tasks
· carrying out plans and adjusting them as necessary
· assessing the success of tasks.
Portfolio work unfolds, starting with ideas and plans, changing into drafts, undergoing revision, settling
into its near-final form and zigzagging between these different moments as well. The process of
refinement and improvement will, therefore, be a crucial aspect of the development of a portfolio.
HMPYC80/104 15
HMPYC80/104/0/2024
You are asked to reflect critically on specific aspects of learning to do research as you work through the
course. By asking you to critically reflect on what you have learnt, the role of reflection in learning is
promoted when you are asked to answer questions about why and what you are learning. This is a
process-oriented approach that is part of any portfolio development. (Refer to your <Research and
course journal>).
Learning is a rather "messy" process because of refinements and revisions in the processes of
comprehension and understanding. Similarly, a portfolio is not a neatly typed document from beginning
to end. The final product should tell the "story" of what and how you have learnt during the course
(including the "messy" parts).
Selecting a framework
You need to spend time thinking about how to organise and sequence the inclusions in your portfolio.
However, before selecting a framework you need to have a clear understanding of the focus of your
portfolio based on your goals and objectives, as the structure is sensitive to the context of the goals and
objectives. The following organising frameworks could serve as a starting point for developing a
framework:
1. Chronological
The portfolio is organised by date, keeping track of what you include in the portfolio by dating
the items. This will give an indication of how your thinking changed over the course of the year.
In many cases, this is the easiest way to organise a portfolio.
2. Thematic
This means that the portfolio is arranged around themes or functions. For example, you may
want to have a section called "revised assignments" and another section called "personal
reflections on my learning". Alternatively, you may want to divide the portfolio into
"compulsory portfolio activities" and "optional portfolio inclusions" or according to specific
themes in the curriculum, such as "open learning".
This method often makes more sense to learners although it may not seem as easy as the
chronological framework.
3. Problem-oriented
The portfolio is organised around problems or issues. As students work through the study
material and activities, you would have a list of problems always in mind. Each problem would
have a section in the portfolio and your reflections, answers, guesses, subsidiary problems,
related questions, etc, will be filed in the same section.
4. Your own organising principle

This means that you organise your portfolio in a combination of any or all of the above ways - or
in a new way of your own.
Everyone's portfolio is a unique picture of that person's learning, reflection and development in a course,
so no two portfolios will be exactly alike.
16
Documentation anhd archiving - Keeping track of the portfolio
One of the first decisions that students will have to make is to decide on a collection device for the
portfolio materials (eg file folders, spiral binding, hanging files, boxes). A ring binder is recommended
as it will allow you freedom to alter or add to portfolio contents.
The framework you have chosen for the content could then serve as a guideline for organising and
keeping track of your portfolio. Students should be advised to date every item you include and to label it
clearly, for example, "Portfolio activity 1.2" or "Personal contribution". Remember to write clearly in
black ink, as you will have to photostat this work later in order to submit it for assignment purposes.
One way to start is to collect everything that you do and think about the course in a file, box or a drawer
of a desk. Label this file or box as your Documentation Centre or Archive. You can consider creating
separate files for different issues, as they occur to you.
Also, you will have to start keeping a "research and course journal" (See <Research and Course
Journal>) where you write down your self-reflections on activities, assignments and so on for possible
inclusion in the portfolio.
The portfolio then evolves from this into a final product that contains only the selected documents that
display appropriate evidence of learning and growth. The careful selection of the final items to be
included is part of the learning process.
What items to include
One of the first questions may be: "What items can I include in the portfolio?". Here is a list of possible
items to include:
observations, rough drafts, summaries, mind maps, results from interviews, assignments,
feedback from lecturer, final copies of work, action plan, self-edited drafts, self-awareness
entries, criteria for assessment, work-in-progress, self-assessment feedback/reports,
self-reflective entries, checklists, revised drafts, a survey of your study skills and habits, goals
and objectives
Although you should use your own discretion when deciding what items to include in your portfolio, the
lecturer should state the minimum requirements, such as items that are indicative of your active
participation and learning throughout the course and that reflect the attainment of your personal
objectives and the course’s learning outcomes.
Such evidence could

· indicate your goals and objectives for the development of the portfolio as well as the course
· indicate your proposed plan of action for achieving your objectives for the portfolio as well as
the course
· include self-selected and lecturer-selected portfolio content
· include reactions to feedback
· include elements of self-reflection
· include self-generated, lecturer-generated and collaboratively generated criteria and
performance standards for evaluating the portfolio, as well as certain assignments you will be
asked to evaluate
HMPYC80/104 17
HMPYC80/104/0/2024
Goals and objectives

Students should be made aware of the overall goal with portfolio work, that is, to move towards a more
authentic assessment model, away from the traditional assessment model which may prevent you from
becoming thoughtful respondents to, and judges of your own work.
Some important aims with portfolio work should be to:

· keep track of your progress in the course
· determine the extent to which the learning outcomes of the course and your personal objectives
have been achieved
· provide you with an opportunity to value and assess your own accomplishments
· provide you with the opportunity to show what and how you have learnt
It is also important that you include your own goal and objectives for the portfolio work.
Plan of action
There are different strategies students can follow when planning a portfolio, for example to:
· use a mind map, or diagram to draw up a plan
· consult with peers and/or other specialists in the field and then draw up a plan
· make use of any strategy you feel comfortable with
There are a few basic issues you may need to consider before drawing up a plan, such as:
· identification of the "tasks" and requirements for the course
· identification of your goals and objectives
· exploration and listing of options that will help you meet goal and objectives
Lecturer-selected content
This could include a number of compulsory portfolio activities, while assignments could also be
included in the portfolio.
Reactions to feedback on assignments and activities

By including feedback in the portfolio a more complete picture is ensured, as the purpose of the feedback
should be to help and support students in your learning. It should be
· realistic and not make suggestions that will be impossible to adhere to
· specific so that students will know what it means
· sensitive to your goals
· timely
· accurate in the interpretation of your inputs
It is unrealistic to expect students to agree with the feedback all the time. Students should be encouraged
to voice their opinion or reactions to the feedback. As a student, you should feel free to argue with the
lecturer. The only requirement for disagreeing with any contentions or feedback from the lecturer should
be that you justify your opinion.
Elements of self-reflection
It is important to include evidence of self-reflection. This reflection refers to those moments when
students return to their work, taking the stance of an informed critic, noticing any or all of the following
things:
· What do you believe is important and why?
· What is characteristic of your learning?
· What has changed with time?
18
· What remains to be done?
· How have different kinds of experiences contributed to your growth and learning?
· What personal insights you have gained from the experience?
· What arguments are held up in the light of your own evolving standards?
· How your opinions and ideas have changed, refined or not changed as you progress through the
course?
When constructing reflective statements students should be encouraged to ask themselves questions,
such as the following:
· What did I do?
· What does this mean?
· Why have I chosen this particular item to be included?
· What have I learnt?
· What have I not learnt?
· What must I unlearn?
· What is the value or purpose of that learning?
· What insights have I had from this learning?
· How has my ability to accomplish new activities or tasks grown from these insights?
· What aspects have I not explored in depth?
· What aspects have I explored in depth?
· How can I do things differently?
· What are the problems that will prevent me from achieving my objectives?
· What can I do to deal with these problems?
· How can I apply this knowledge to my individual situation - in other words: What is in it for me?
· What aspects would I like to pursue for future learning?
Reflective statements are essential components of any portfolio as you allow students to "examine" the
thinking behind your decisions and accomplishments. Critical academic reflection should become an
automatic activity in the learning process. Portfolio work allows an excellent opportunity to exercise this
skill.
Other items to include

Students should be encouraged to choose any other related/applicable work. Apart from the above
suggested items you may select items to show something you did that:
· shows evidence of learning,
· you found interesting,
· shows work-in-progress with written plans for revision,
· you found especially difficult,
· shows your ability to apply particular principles,
· shows your problem-solving ability,
· you enjoyed doing.
It is also important that students develop a rationale for selecting any of the above materials. Answering
the following questions may help you develop this rationale (Shackelford 1996:32):
· Why have I decided to include this particular item?
· Does it show growth? How and what?
· What skills or abilities do I wish to document?
· Do these abilities match identified needs?
· What evidence can I include to substantiate these abilities?
· Do materials adequately and accurately demonstrate my abilities?
· Does the item match or reflect my goal(s) and objectives?
· Does the portfolio reflect what I have learnt about my ability to solve problems?
HMPYC80/104 19
HMPYC80/104/0/2024
CONCLUSION
Portfolio work could and should be an enjoyable experience. It provides a new way of demonstrating
how learning experiences were experienced and helps one to reflect on one’s past, present and future
learning.
References
Barnett, B.G., & Lee, P. (1994). Experiential learning: a new approach. In Jackson L. & Caffarella R.S.
(Eds.) New directions for adult and continuing education. Jossey-Bass.
Boud, D. (ed.). (1981). Developing student autonomy in learning. Kogan Page.
Dewey, J. (1938). Experience and education. Collier.
Evans, N. (1992). Experiential learning as learning to effect. In Barnett R. (Eds.) Learning to effect.
Open University Press.
Hamilton, R. J. (1994). Adult learning, currency and subjectivity in prior learning assessment: semantic
and conceptual ambiguities in prior learning assessment. Journal of the National Institute on the
Assessment of Experiential Learning, 11, 4-11.
Knowles, M.S. (1970(. The modern practice of adult education: andragogy versus pedagogy.
Association Press.
Paris, S.G., & Ayres, L.R. (1994). Becoming reflective students and teachers with portfolios and
authentic assessment. American Psychological Association.
Paulson, F.L., Paulson, P.R., & Meyer, C.A. (1991). What makes a portfolio a portfolio? Educational
Leadership, 48, 60-63.
Shackelford, R.L. (1996). Student portfolios: a process/product learning and assessment strategy. The
Technology Teacher, May/June, 31-36.
Thorpe, M. (1995). Bringing learner experience into distance education. In Sewart D. One world many
voices: quality in open and distance learning. OUUK (17th World Conference for Distance
Education)
Wolf, D.P. (1989). Portfolio assessment: sampling students work. Educational Leadership, 46, 35-39.
20
B. RESEARCH OVERVIEW
Research: What is it?

Research is (not) what most people think it is!
Most students approach the study of research methodology with a sense of apprehension. Terms like
research and methodology often evoke images with rather intimidating connotations. You can quickly
determine where you stand on this issue by playing the following game: What do you see in your mind’s
eye when you hear the word research? A laboratory, or perhaps a bespectacled individual with a white
coat? And what would this imply? Probably the perception that research is an extremely complicated
matter that pertains to a very special domain of human activity! If so, you are bound to believe that only
a few clever and perhaps eccentric individuals can do research. And secondly, what does the term
methodology conjure up? The face of your high-school maths teacher, or perhaps the idea of a pocket
calculator? Indeed, generations of students have had to battle with statistical methods and have felt they
have lost - thereby adding all kinds of negative bits and pieces to the collective academic unconscious
about research!
Perceptions like these are rather unfortunate - especially because learning about research is part and
parcel of tertiary education. In fact, if the academic enterprise is to thrive at all one can hardly afford an
attitude of "war with statistical (and qualitative) techniques" on the part of the novice researcher. The
question that therefore confronts us is how to entice students into a love affair with the subject of
research.
The key to the answer lies in the way we conceive of the research enterprise. In as much as the
impression is created that it is a highly complex endeavour, all is lost. Arguably, under certain conditions
research may involve highly sophisticated procedures - but this should not lead one to believe that
complexity is a prerequisite for producing useful knowledge. Dispelling this myth is our first and
foremost task.
So, what is “Research”?
In the < Scientific journal articles > resource we describe 5 types of research:
(a) Relational causal research where research questions and hypotheses are investigated; with the aid
of specified variables and the postulated relationships between these variables, as well as the results that
are expected and why they are expected. In experimental and quasi-experimental studies the causal
relations between variables are studied. These studies aim to predict and control the variables for the use
of scientific and objective research.
(b) Descriptive interpretative research where the aim is to better understand and gain insight into a
socio-psychological phenomenon or dynamic. The study aims to illuminate a psycho-social
phenomenon or psychological dynamic. (In the case of a descriptive study avoid words that imply that
you are looking for causal relations between variables.)
(c) Action research is where an individual or a group acts on a problem or a group concern. The
action-reflection research cycle is used to create a research learning context.
(d) There are other research types, such as: mixed methods, participatory action research, program
evaluation research, and intervention research. Some of these are discussed in Section E of the
prescribed textbook by Fouché, et al. (2021).
HMPYC80/104 21
HMPYC80/104/0/2024
(e) An exciting and dynamic research area is in the domain of indigenous research methodology, and
decolonising research methodologies. The reason for doing this kind of research is stated by Chilisa
(2008, quoted in Denzin, et al., 2008, p. 60):
“Postcolonial indigenous research methodologies provide an important framework through
which Western-educated researchers can explore the possible biases in the literature we read,
identify the knowledge gaps that have been created because of the unidirectional borrowing of
Euro-Western literature, and bring to a halt the continuing marginalization of other knowledge
systems that occurs because of the dominant Euro-Western research paradigms and their
discourses on what can be researched and how it can be researched. Applying indigenous
research methodologies to research with and about the colonized Other should involve going
back and forth to retrieve marginalized and suppressed literatures to review, analyze, and
challenge colonizing and deficit theorizing and interpretation, to create counternarratives that see
the past differently, and to envision a transformative agenda with the researched. It also involves
defining what literature and theorizing in the context of former colonized societies is.
Postcolonial indigenous research methodologies perceive literature as language, cultural
artifacts, legends, stories, practices, songs, rituals poems, dances, tattoos, lived experiences such
as the people’s fight against HIV/AIDS, personal stories, and community stories told in
weddings, funerals, celebrations and wars.” (Denzin, et al., 2008, p. 60).
This approach challenges the deficit thinking and pathological descriptions of the formerly colonized
and reconstructs a body of knowledge that carries hope and promotes transformation and social change
among the historically oppressed. Chilisa (in Denzin, Lincoln, & Smith, 2008) further states that:
“Western-educated scholars need to investigate the psychological harm, humiliation, embarrassment,
and other losses that these theories and body of knowledge caused to the researched colonized Other.
They also need to use the body of indigenous knowledge about the researched to counter theories and
other misinformation that may cause communities humiliation and embarrassment (p. 60).
A good introduction to indigenous approaches to research, inquiry and relationship in relation to
historical trauma in Canada is presented by Yellow Knife Dene Nation psychologist Dr Suzanne L.
Stewart of the Department of Applied Psychology, OSEI University of Toronto: “Intergenerational
trauma and education” at https://www.youtube.com/watch?v=yHQSJaP3liM.
For more readings on indigenous research methodologies and decolonising research methodologies see
the following resources:
Chilisa, B. (2012). Indigenous research methodologies. Sage.
(Prof Bagele Chilisa hails from the University of Botswana and has published extensively.)
Denzin, N.K.; Lincoln, Y.S.; & Smith, L.T. (2008). Handbook of critical and indigenous
methodologies. Sage.
(The first chapter can be downloaded for free from the Sage website.)
Smith, L. T. (1999). Decolonizing methodologies: Research and indigenous peoples.

University of Otago Press.
There are also numerous keynote presentations by Prof Linda Tuhiwai Smith, and other indigenous
scholars on the Internet available for viewing. We encourage you to look into this resource!
22
Research is about curiosity, interests, solving behavioural puzzles, understanding (and sometimes
solving) complex social issues, and ultimately attaining better, deeper and more valid way(s) of
understanding and contextualising human behaviour.
All of us are researchers and inquirers, whether we know it or not! Admittedly, this process may be
carried out with different degrees of sophistication - hence the distinction between naive research and
formal research. The term formal research applies to the academic context where problem-solving is
characterised by a more precise and serious attitude. In this respect, one could use the description
“deeper understanding” or serious problem solving”, since a deliberate attempt is made at being
thorough, systematic and accurate or situation-valid.
The basic elements of “Research”
Human inquiry involves a series of basic steps, namely observing a situation, reflecting on the situation,
planning how to engage with (or handle) the situation, acting to address or attempting to solve a problem
or challenge (if the research is problem-based), observing the result or effects of one’s actions and
reflecting on the implications of these results or effects, planning the next action, observing the effects of
this action .... Note how the four words, observe, reflect, plan and act repeat. These problem-solving
steps are interdependent and form a second-order (reflective) learning spiral in action learning and
action research.
In coping with the challenges of daily living, one is not always conscious of following this routine but it
nevertheless applies. However, when engaging in deep and reflexive human inquiry a deliberate attempt
is made at executing the various steps as thoughtfully as possible. But this does not mean that the process
has to be complex and difficult. By characterising research as a highly sophisticated endeavour the
matter is complicated unnecessarily. Of course, complex situations, complex dynamics and complex
situations (“wicked problems” or “messy problems”) do exist and these studies call for the use of
sophisticated research methods. However, not all problems can be solved or fixed – and many solutions
create unexpected and new problems or challenges!
Recommended self-study
Dr Cecile Badenhorst has developed a useful collection of short research presentations that is available
on this Youtube channel: https://www.youtube.com/channel/UCDXlZhpn7iJcw9BdgzUXWbA/videos
Scroll through the titles and view any relevant presentations. Useful topics that are covered include
research basics, what is the proposal, and writing skills.
Additional sources
Chilisa, B. (2012). Indigenous research methodologies. Sage.
Denzin, N.K., Lincoln, Y.S., & Smith, L.T. (2008). Handbook of critical and indigenous
Fouché, C. B., Strydom, H., & Roestenburg, W. J. H. (2021). Research at grass roots: For the social
sciences and human service professions (5th ed.). Van Schaik.

HMPYC80/104 23
HMPYC80/104/0/2024
Ethics
For the first couple of years of one’s psychological studies, one is mainly concerned with learning about
psychological knowledge created by others. However, at postgraduate level one is required to show that
one is able to actively participate in the creation of new scientifically acceptable knowledge. This means
one has to design a research project based on a topic that you deem to be personally important and
meaningful; and that aligns to the academic literature as a valid and important psychological research
project.
Students typically respond with considerable enthusiasm to this opportunity and often produce work of
high quality. Even if students do not end up creating the best possible research proposal and report, they
will most likely find themselves learning more about the practicalities, the joys and the disappointments
of research than they did during undergraduate studies.
Experiencing the research process firsthand one quickly discovers that research is a process of continual
compromise so that what one eventually ends up doing is never the same as what one at first envisaged.
The challenge is to design a project that one finds meaningful, but also one that is feasible and ethical.
IT IS UNACCEPTABLE TO RE-USE ANY RESEARCH!
It is not acceptable to reuse or recycle any research that was developed by yourself and/or by others.
You may therefore not use research again that was previously used in any academic course/module or
any formal research project in any setting. This is a policy that is agreed on by all universities, all
accredited editorial committees of journals, and reputable publishing houses across the world.
[Secondary data analysis of existing research is a different matter, and is acceptable.]
Note that it is possible to plagiarise yourself (i.e. your own work). To avoid this all your previous work
must be appropriately cited and referenced, including research projects or proposals submitted to
universities or other institutions. You may not copy and paste your previous work indiscriminately into
a new proposal. Every new proposal must be developed from scratch and from the ground up.
For this module HMPYC80 you are expected to develop a new proposal from scratch. You must develop
a psychological research proposal that is unique and is the fruit of your own and independent work.
Research expectations in the HMPYC80 module.
You are doing this course via distance learning, that means that you are essentially on your own – you do
not have a research supervisor or an institutional ethics panel to back you up should things go wrong. For
this reason, you have to be cautious, and not become involved in just any type of project. You will only
be producing a research proposal in this module. You may not execute the empirical aspect of the
research, or for that matter any other research without proper institutional ethical oversight, approval and
an official research and ethics clearance certificate.
24
In general, if you use questionnaires or interviews you will need to complete the following checklist for
each research participant.
Table 01
Research ethics checklist
Name of research participant:
Signature of research participant:
Signature of researcher:
Date on which participant partakes in research:
Action Yes No
I explained the research study to the research participant

I explained the purpose of the research study to the research participant
I indicated what kinds of questions the research participant can expect to get
I informed the research participants that they may leave questions unanswered
should the question make them feel uncomfortable
I informed the research participants that they may withdraw from the study at
any stage
The research participant indicated that they understood the information
provided
The research participant indicated that they were willing to partake in the
research project.
If the research is done at an organisation: I obtained written permission from
the research participant’s organisation for the research participant to partake in
the research study
If minors are involved in the research project: I obtained written permission
from the research participant’s parents/primary caregiver for the research
participant to partake in the research study
The sponsoring research organisation or institution (for example, the relevant
university):
(a) I obtained a certificate of permission and approval to execute the research
from the relevant Higher Degrees Committees (or Research Review
Committees).
The sponsoring research organisation or institution (for example, the relevant
university):
(b) I obtained a certificate of permission and approval to execute the research
from the official Research Ethics Committees.
HMPYC80/104 25
HMPYC80/104/0/2024
Only psychological practitioners who are registered with the Health Professions Council of South
Africa are allowed to perform any psychological therapeutic interventions. (This excludes
psychometricians and researchers as they are also not allowed to perform psychological therapeutic
interventions.)
Although there should not be any problems with using for your research project the sources of
information listed below, it is a good idea to keep in mind the cautions indicated above. Always treat
information about people as confidential, and handle the information in a manner that would not cause
them any harm.
Publicly available information and data sources are good and ethical sources of information. These
can include:
- textual material that is publicly available – for example, classified advertisements in newspapers,
advice columns in magazines, or anything openly published on the internet in such a way that it can
be viewed by anybody.
- audio and video materials that are publicly available – for example, radio talk shows, television news
bulletins, internet media posts (for example on Twitter, Facebook, Instagram, websites, blogs,
interviews and videos), and broadcasts on television or radio.
- quantitative data that are publicly available – for example, statistics provided by Statistics SA, the
South African Police Service, other government or municipal departments, and parastatal state
owned entities.
- textual, audio or video material or quantitative data drawn from records kept by an organisation or
company – but only if you obtain written permission from the person(s) in charge of the material and
keep to any conditions they have
- simple observation in public spaces – for example, what you observe when visiting a shopping mall,
walking down a street, or at a local park (but not private spaces such as organisations, schools,
churches, hostels or private residences).
You may compile a research proposal on the following or any other topics for this
module (HMPYC80). Note that you will not be performing the empirical data
gathering, analysis and interpretation for this module (HMPYC80).
In general, you must obtain various levels of ethical oversight to perform empirical research on the
following topics (i.e. obtain institutional oversight and permission). The following are deemed sensitive
topics, that either involve expert intervention modalities, and/or relate to vulnerable groups, and/or relate
to risk to researcher and participants.
Section 1: SENSITIVE RESEARCH TOPICS

1.01 research about parenting
1.02 research investigating sensitive personal issues
1.03 research investigating sensitive cultural issues
1.04 explorations of grief, death or serious/traumatic loss
1.05 depression, mood states, anxiety
1.06 gambling
1.07 eating disorders
1.08 illicit drug taking
1.09 substance abuse
1.10 self-report of criminal behaviour
26
1.11 any psychological disorder
1.12 suicide
1.13 gender identity
1.14 sexuality
1.15 race or ethnic identity
1.16 any disease or health problem
1.17 fertility
1.18 termination of pregnancy
Section 2: SENSITIVE RESEARCH PROCEDURES

2.01 use of personal data obtained from external agency without participant’s knowledge
2.02 deception of participants
2.03 concealing the purposes of the research
2.04 covert observation
2.05 audio or visual recording without consent
2.06 recruitment via a third party or agency
2.07 withholding from one group specific treatments or methods of learning,
from which they may "benefit" (e.g., in medicine or teaching)
2.08 any psychological interventions or treatments
2.09 administration of physical stimulation
2.10 invasive physical procedures
2.11 infliction of pain
2.12 administration of drugs
2.13 administration of other substances
2.14 administration of ionising radiation
2.14 administration of ionising radiation
2.15 tissue sampling or blood taking
2.16 collecting body fluid
2.17 genetic testing
2.18 use of records where participants can be identified or linked
2.19 drug trials and other clinical trials
2.20 administration of drugs or placebos
Section 3: PARTICIPANTS VULNERABILITY (SENSITIVE RESEARCH)

Participants who fall within the following targeted categories:
3.01 suffering a psychological disorder
3.02 suffering a physical vulnerability
3.03 people highly dependent on medical care
3.04 minors without parental or guardian consent
3.05 people whose ability to give consent is impaired
3.06 resident of a custodial institution
3.07 persons whose ability give free informed consent is restricted because of
difficulties in understanding information statements (e.g. language difficulties)
3.08 members of a socially identifiable group with special cultural or religious
needs or political vulnerabilities
3.09 those in dependent relationship with the researchers
(e.g. lecturer/student, doctor/patient, teacher/pupil, professional/client)
3.10 will it be possible to identify any participant in any final report when specific
consent for this has not been given?
Section 4 - RESEARCH SETTINGS THAT ARE SENSITIVE

4.01 research being undertaken in a politically unstable area
4.02 research involving sensitive cultural issues
HMPYC80/104 27
HMPYC80/104/0/2024
4.03 research in countries where criticism of government and institutions might

put participants and/or researchers at risk
In general, by sticking to these guidelines, you will greatly reduce your chances of doing an unethical
research project. However, the guidelines do not guarantee that your research will be ethical - you still
have to pay close attention to any ethical issue that my crop up in your research project. Always make
sure that you are knowledgeable about ethical guidelines.
You may compile a research proposal that uses any of the above sensitive topics or
aspects for this module (HMPYC80) as you will not perform the actual empirical data
gathering, analysis and interpretation.
Ethical guidelines
You can read about ethics and ethical conduct in the prescribed book. The following document provides
additional information. The document is based on the state of research ethics in the social sciences in
India. The arguments, state of the field and suggestions mirror much of what is happening in South
Africa and Africa. As such it is a useful document that should elicit further discussion and debate. Ethics
is never a completed state of affairs - this is particularly obvious in qualitative-interpretive,
hermeneutical, feminist, (participatory) action and other research approaches. The following document
omits important issues on minority and/or vulnerable groups (such as racial or ethnic minority groups;
institutionalised children and adults; orphans; gay, lesbian, bisexual, intersex and transgender people;
disabled people, and people living in other contexts of oppression or exclusion.)
Study the South African Department of Health’s document on Ethics in Health Research Principles,
Processes and Structures (2015) at
https://knowledgehub.health.gov.za/elibrary/ethics-health-research-principles-processes-and-structures
Study the Health Professions Council of South Africa’s guidelines for good practice in the health care
professions: General ethical guidelines for health researchers at
https://www.hpcsa.co.za/Uploads/professional_practice/ethics/Booklet_13_Gen_Ethical_Guidelines_f
or_Health_Researchers.pdf
Also carefully study the policies and guidelines for research at Unisa at
https://www.unisa.ac.za/sites/corporate/default/Apply-for-admission/Master%27s-&-doctoral-degrees/
Policies,-procedures-&-forms
• Policy for copyright infringement and plagiarism (PDF)
• Intellectual property policy (PDF)
• Language policy (PDF)
• Policy on Research Ethics (PDF)
• Academic Integrity (PDF)
Resources:
Psychological Society of South Africa (PsySSA). (2007). South African professonal conduct
guidelines in psychology. http://www.psyssa.com/ethics/
28
Academic integrity and avoiding

plagiarism
Academic dishonesty involves plagiarism or cheating -- as stated in Tutorial Letter 101.
These involve the following:
Plagiarism
Plagiarism is the act of taking the words, ideas and thoughts of others and presenting them as your own.
It is a form of theft which involves several dishonest academic activities, such as the following:
• Cutting and pasting from any source without acknowledging the source.
• Not including or using incorrect references.
• Paraphrasing without acknowledging the original source of the information.
Cheating
Cheating includes, but is not limited to, the following:
• Completing assessments on behalf of another student, copying from another student during an
assessment or allowing a student to copy from you.
• Using social media (e.g. WhatsApp, Telegram) or other platforms to disseminate assessment
information.
• Submitting corrupt or irrelevant files (intentionally).
• Buying completed answers from “tutors” or internet sites (contract cheating).
Note that all your assignments and research proposals must be your own work. If you take words, ideas
and thoughts of others and pass them off as your own you are committing an act of plagiarism.
Plagiarism is a form of theft that involves a number of dishonest academic activities. Plagiarism is
forbidden and there are serious consequences for submitting any plagiarised work.
The following caveat that was mentioned in the previous discussion on ETHICS bears repeating (see
<C. ETHICS AND RESEARCH: Ethics>):
IT IS UNACCEPTABLE TO RE-USE ANY RESEARCH!
It is not acceptable to reuse or recycle any research that was developed by yourself and/or by
others. You may therefore not use research again that was previously used in any academic
course/module or any formal research project in any setting. This is a policy that is agreed on by all
universities, all accredited editorial committees of journals, and reputable publishing houses across
the world.
Note that it is possible to plagiarise yourself (i.e. your own work). To avoid this all your previous
HMPYC80/104 29
HMPYC80/104/0/2024
work must be appropriately cited and referenced, including research projects or proposals submitted
to universities or other institutions. You may not copy and paste your previous work indiscriminately
into a new proposal. Every new proposal must be developed from scratch and from the ground up.
For this module HMPYC80 you are expected to develop a new proposal from scratch.
We also expect you to submit a “declaration of academic honesty” with your research proposal. The
template for the declaration is provided in the next section (See C. ETHICS AND RESEARCH:
DECLARATION OF ACADEMIC HONESTY).
For further information see: Disciplinary Code for Students. The Disciplinary Code for Students is given
to all students as part of your study materials. Students are advised to study the Code, especially Sections
2.1.13 and 2.1.14 (pp. 3-4). Kindly read the University’s Policy on Copyright Infringement and
Plagiarism as well.
The Unisa library has a guide on research and plagiarism at http://libguides.unisa.ac.za/Research_skills
RECOMMENDED STUDY
Please view or listen to the following useful and important presentations online:
Badenhorst, C. (2020, June 9). What is Plagiarism? Thesis/Research Writing and Plagiarism. [Video].
Youtube. https://www.youtube.com/watch?v=bPmDiRf4wrA
Badenhorst, C. (2020, June 9). Understanding citation practices. Research writing and plagiarism.
[Video]. Youtube. https://youtu.be/QXwfNu7KHG8
Badenhorst, C. (2020, June 9). Using sources effectively: Thesis/Research Writing and Plagiarism
[Video]. Youtube. https://youtu.be/HNfrOzTa-aY
Hastings, C. (2015, Feb 28). Avoiding Plagiarism, writing with integrity [Video]. Youtube.
https://www.youtube.com/watch?v=F1S1FZ-bn5E
Hastings, C. (2018, June 15). Their Cheating Hearts (Plagiarism) [Video]. Youtube.
https://www.youtube.com/watch?v=cD9_FG7QUlY
30
DECLARATION OF ACADEMIC
HONESTY
DECLARATION OF ACADEMIC HONESTY
Name: …
Student number: …
Module Code: … (eg. HMPYC80)
Module Name: … (eg. Honours in Research Metholdology)
Title of your proposal: …
DECLARATION
1. I know that plagiarism means taking and using the ideas, writings, works or inventions of another
as if they were one’s own. I know that plagiarism not only includes verbatim copying, but also
the extensive use of another person’s ideas without proper acknowledgement (which includes the
proper use of quotation marks). I know that plagiarism covers the use of material found in textual
sources and from the Internet.
2. I acknowledge and understand that plagiarism is wrong.
3. I understand that my research proposal must be accurately referenced.
4. This research proposal is my own work. I acknowledge that copying someone else’s research, or
part of it, is wrong, and that submitting identical work to others constitutes a form of plagiarism.
5. I have not allowed, nor will I in the future allow, anyone to copy my work with the intention of
passing it off as their own work.
6. I confirm that I have read and understood UNISA’s Policy for Copyright and Plagiarism -
https://www.unisa.ac.za/sites/corporate/default/Apply-for-admission/Master%27s-&-doctoral-d
egrees/Policies,-procedures-&-forms
7. I understand that the proposal will be submitted to Turnitin (or similar software) and a plagiarism
check will be performed. In cases where the threshold of 40% similarity (or 50% similarity in the
case of an exception made for HMPYC80) with other existing publications is surpassed this will
likely mean that the proposal mark will be nullified and this may lead to a disciplinary hearing.
Depending on the findings of the disciplinay hearing this may lead to my expulsion from the
university for a lengthy period of time.
________________________ _____________________
SIGNATURE BY STUDENT/ DATE
SURNAME AND INITIALS
HMPYC80/104 31
HMPYC80/104/0/2024
Academic writing tips and suggestions
The following is provided as a guide and is provided courtesy of a former editor of the New Voices in
Psychology journal at the University of South Africa (New Voices Journal, 2014). These guidelines
are provided as a guideline to improve your writing.
• All paragraphs are left-aligned

• Text in 12-point Times New Roman font, or 11-point Arial font.
APA7 (update): A variety of fonts are acceptable, with focus on accessibility for readers. Acceptable
fonts include sans serif fonts such as Calibri 11, Arial 11, and Lucida Sans Unicode 10, as well as
serif fonts such as Times New Roman 12, Georgia 11, and Computer Modern 10.
• All margins 2.5 cm.
• Line spacing 1.5 cm.
• Start each paragraph on the left. Do not start each paragraph with a tab or spaces.
• Use a single space and not a double space after each sentence. Use single spaces throughout the text.
• Leave a line open below each paragraph.
• Do not indent anything.
• Include a page number on every page. The title page starts as page 1. This also applies to dissertations
and thesis manuscripts.
• Use double quotes in all cases ("there were no exceptions to this rule", Maya, 2017, p.15), except for
quotes within quotes in which case the inner quotes are single ("the participants referred to this as
'anomalies' and were unable to explain this phenomenon", Voster, 2015, p.127).
• APA7 (update): Commonly used technology terms are listed and should be spelled as follows:
"email," "internet," and "webpage."
Statistics, tables and figures

• Statistical symbols: Put all symbols in italics, including n, N, p, t, F, M, SD.
• APA7: Table number is bold; table title is title case and set in italics. The table number and caption
are on separate lines and are placed above the figure. Table titles should be in sentence case, left
aligned, and appear at the top of the table. They look like this --
Table 1
Comparison of standard RCS procedure and the adjusted RCS procedure
• APA 7 (update): Figure number and caption are on separate lines and are placed above the figure, and
the style matches that for tables: Figure number is bold, figure caption is title case and set in italics.
Figure titles are in sentence case, left aligned. They look like this --
Figure 1
Scatter plot graph of standard RCS procedure and the adjusted RCS procedure
• In-text citation of tables and figure:
"as seen in Table 6, the value of …" and "as Figure 7 shows, the value of …"
32
Punctuation
• Where commas are used in the body text to separate items listed in a sentence, the last item is
preceded by an and, and there is a comma before the final and. For example: Apples, pears,
oranges, bananas, and peaches are healthy or Ratele, Duncan, and Roos (1998) claim…
• There are full stops after initials, and each full stop is followed by a space.
• Expand the MS Word ellipsis character (…) to three full stops with a space after each one ( . . . ).
• Only use e.g. or i.e. in brackets and always put a comma after them, for example: (e.g., i.e., ).
In the body text expand them to for example (or such as) and that is and do not put a comma after
them. Bear in mind that “such as” often reads better than “for example”.
• Use a hyphen (-) to link compound words and a dash (--) to link parts of a sentence. When linking
compound words do not leave spaces on either side of the hyphen. When linking parts of a sentence
leave spaces on either side of the dash.
• In the following exceptional case use a dash rather than a hyphen to link words: When the two
elements being linked are of equal status, for example, non-toxic but mother--daughter pairs.
• Do not use dashes in ordinary body text to replace the words to or and. So it should be from the ages
of 15 to 20 (not from the ages of 15-20) and between 12 and 14 (not between 12-14).
• Use double quotes in all cases, except for quotes within quotes in which case the inner quotes are
single.
• Where longer quotations of more than 40 words are presented, please use an indented block with no
quote marks. Example:
If it is not possible to identify an individual author for a source, you may be able to use the name
of an organisation or company as the author. For example, if you want to reference an article on
the website of the British Psychological Society, and there is no obvious author, you could use the
British Psychological Society as the author in your reference. (Davies, 2020, p.14).
Capitalisation
• Capitalise terms such as West and Western, and East and Eastern when used in a cultural sense, but
not when used in a geographic sense (direction east, west, south and north).
• The term apartheid is not capitalised.
• Indian and African are capitalised. The racial terms Black, White and Coloured are or can be
capitalised in this module (see next comment) when referring to race.
“Race refers to physical differences that groups and cultures consider socially significant. For
example, people might identify their race as Aboriginal, African American or Black, Asian,
European American or White, Native American, Native Hawaiian or Pacific Islander, Māori, or
some other race.” (American Psychological Association APAStyle, 2019, 2022, para 3.)
“Language that essentializes or reifies race is strongly discouraged and is generally considered
inappropriate. For example, phrases such as ‘the Black race’ and ‘the White race’ are essentialist in
nature, portray human groups monolithically, and often perpetuate stereotypes.” (American
Psychological Association APAStyle, 2019, 2022, Avoiding essentialism, para 1.)
In the light of the above we suggest that the terms Coloured and White are also capitalised when
referring to racial group constructs that formed part of specific legislation during colonial times,
during apartheid and post-apartheid.
HMPYC80/104 33
HMPYC80/104/0/2024
With regard to contested term “Coloured” it is important to note that this is a contested notion and there
is an active debate on the use of this historical terms in South and Southern Africa. Scholars often
mention very clearly that the use of these labels (for example: “Coloured”) does not imply that they or
their institution endorse the legitimacy of these labels (See for example, Dannhauser, 2006, p.1).
Abbreviations
• Always write out an abbreviation or acronym the first time it is used in the body text and give the
abbreviation in brackets, for example, Stress Symptom Checklist (SSCL). After that use the
abbreviation. Even though an abbreviation is written out in the abstract, it should be written out in
full again in the body text.
• In the body text, spell out i.e. (that is) and e.g. (for example) and put commas before and after that is
and for example.
Numbers
• Use a comma to show thousands, e.g. 18,321 students and 1,100 lecturers, or ( n = 15,123).
• Use decimal points, not commas (SD = 20.35), 1.47 meters, 89.5 kilograms, or (M = 7.8) .
• If a p-value is indicated in the first zero is omitted (p < .05) and p < 0.001.
• Use words for numbers below ten (e.g., there were five participants, and one of them was disabled);
numerals for ten and above (e.g., there were 20 participants and 11 were male). In cases where
there is a considerable number of numbers, they can be kept all the same to avoid distracting variety.
• APA7 update: Numbers representing time are written as numerals, not written out, regardless of
whether the time is exact or approximate (e.g. “about 7 weeks,” “3 decades,” or “approximately 5
years ago”).
• An exception: Use the word (rather than number) form for centuries, example, “nineteenth century”.
• Use “per cent” in the text but “%” in brackets and tables. It is also OK to use “%” where percentages
are thickly clustered in the text.
Words
• Use South African English spelling except when quoting American or other sources and in the titles
of articles, books and so on.
• In South African English -ise is preferred to –ize (e.g., generalise, analyse, utilise, and so on) but
there are exceptions such as size. If the spell checker will accept a spelling with s rather than z, then
make it s, for example analyse rather than analyze.
• It is research method, not methodology (except in some rare cases - - methodology refers to the
science and philosophy of research methods).
• It is Cronbach’s alpha not Cronbach alpha.
• Consider using use rather than utilise.
• Both cannot and can not are acceptable, but the first is much more usual.
• Beware of comprised -- it is often used incorrectly. Consisted of is safer.
• In the light of is British and South African English; in light of is American. The same goes for on the
one hand (South African) versus on one hand (American).
• Use first, second and so on rather than firstly, secondly.
• Either a historical or an historical is OK, but the latter is more usual in South African English.
34
• It is modelled not modeled in South African English.
• The past tense of focus is focused not focussed.
• Once upon a time it was the 1980s and 1960s, not the 1980’s.
• Avoid and/or -- try to use either and or or.
• Centre not center.
• Use its not it’s (unless it is a contraction, in which case it should usually be expanded to it is).
• Avoid etc. -- rather use and so on.
• Expand contractions such as we’ve (we have) and aren’t (are not), except in a quotation. Expand
he/she to he or she.
• Use gender neutral (they) or gender balanced (he or she) language, unless referring to somebody
with a specific gender.
• Anything that can be counted, except money, is fewer rather than less. So -- There were fewer
participants in the first than in the second study (not less participants).
• Use hyphens liberally when tempted to write words separately, for example, use well-being rather
than well being. However, do not unnecessarily split words using hyphens, so it is coefficients not
co-efficients.
• Preferably use terms such as isiZulu rather than Zulu when referring to the language.
• Avoid furthermore.
• Data are plural, datum is singular. Media are plural, medium is singular.
• Looked at is too informal, rather use considered, investigated, was concerned with or some other
formulation.
• Program has one m if it is a computer program, but two for all other programmes – such as social
intervention programmes and programme evaluation.
Source:
New Voices Journal. (2014). Copy-editing guide. Unpublished document. University of South Africa.
References:
American Psychological Association APAStyle (2019, 2022) Racial and Ethnic Identity.
https://apastyle.apa.org/style-grammar-guidelines/bias-free-language/racial-ethnic-minorities
Updated July 2022.
Dannhauser, P.D. (2006). Representation of Coloured Identity in Selected Visual Texts about
Westbury, Johannesburg. Research report for MA in Arts, University of the Witwatersrand.
HMPYC80/104 35
HMPYC80/104/0/2024

American Psychological Association
formatting and style (APA7)
APA stands for American Psychological Association. The Association outlines the style in the
Publication manual of the American Psychological Association [APA] (7th ed.). The following is
provided as a guide and is provided courtesy of a former editor of the New Voices journal in psychology
at the University of South Africa (New Voices Journal, 2014), with updates for the APA7 style.
The citations and reference list are the most important and should be adhered to precisely by Unisa
students in their research proposals. Note that the APA7 style does not apply to all articles and journals,
especially those outside of psychology. For the sake of clarity the examples provided below have been
printed in a larger font (point size 14). However, you must use the recommended font sizes of 12-point
Times New Roman font, 11-point Arial font, or other fonts as discussed in the previous section.
In-text citations
o In-text citations in brackets will use ampersands and not “and”. For example:
Happiness is difficult to define (Christie et al., 2001), but sadness is easy (Van der
Merwe & Lipkovitch, 2001).
Note that this is different from the reference list where there is a comma before the ampersand even
when there are only two authors.
o APA7 update: For sources with three or more authors, use the first author’s name followed by et al.
in your in-text citations. (The phrase “et al.” is Latin for “and others”.)
Heaney et al. (2015) carried out a systematic review of sport psychology education.
Or, a systematic review of sport psychology education (Heaney et al., 2015) found the
following . . .
o APA7 update: In the case of three or more authors use the first author et al. from the start. For
example: Zama et al. (2006) state that the impact of psychology in South Africa has
been substantial.
o When citing several sources in one set of brackets they should be alphabetically arranged, with
semicolons between sources. If more than one source is listed for the same author(s), the dates are
separated by commas. For example, Existential philosophy is now considered old-hat
(Levinas, 1961, 1985; Marnevik, 1923; Tshido et al., 2006; Zapotek & Orion, 1945,
1957, 1970). Note that the list of authors is sorted alphabetically by first author name.
o Give page numbers when quoting. There is some leeway as to where exactly the page number is
placed. For example, “I am an African” (Mbeki, 2003, p. 23) or In the words of Mbeki
(2003), “I am an African” (p. 23) or In the words of Mbeki (2003, p. 23), “I am an
African” or Mbeki (2003, p. 23) says: “I am an African”. Use pp. rather than p. if the
quotation starts and ends on different pages in the source and put a hyphen or a dash (-) between the
first and last page number, for example, Existential philosophy “is now considered old-hat”
36
(Levinas, 1961, pp. 102-103).
o In cases where there are no page numbers in the resource then include the paragraph and/or the
section heading with the paragraph number, for example:
(Myers, 2000, para. 5) or Myers (2000, para. 5)
(Beutler, 2000, Conclusion section, para.1) or Beutler (2000, Conclusion section,
para.1)
Reference list
Note that you should include a DOI (Digital Object Identifiers) if available. Many publishers, databases
and online journals now use DOIs, however older publications will not have a DOI. Use these websites:
http://dx.doi.org/ and http://www.crossref.org/ to find DOI information.
o The highlights: There are spaces between initials. There are commas before ampersands, even for
only two authors. Titles and volume numbers are in italics. Page numbers are separated by a hyphen
or a dash. Below are some examples.
o We discourage double citations in the form of “Segalo (2003) cited in Dlamini (2006)”, but if
they do occur only the source actually consulted by the author (in this case Dlamini) should be
included in the reference list.
o Except for conference presentations, only sources that are accessible to readers should be listed in
the reference list. Personal communications should not be included in the reference list but just cited
in the text like this: (S. de Villiers, personal communication, 17 November 1999)
Reference list: Articles
Figure 01:
How to referencing a journal article in APA7
Note: From St Mary’s College of California Library. (n.d.) APA style 7th edition. Library guide. St
Mary’s College of California Library. http://stmarys-ca.edu/library. (Page 3.) Permission not sought.
o For all articles provide Digital Object Identifier (DOI) if it is available. This is the unique identifying
number used for academic literature which is published online. Older journal articles do not have a
DOI and for these you should use the online address for a website, picture or other resource called
the Uniform Resource Locator (URL), for example http://aje.sagepub.com.
HMPYC80/104 37
HMPYC80/104/0/2024
Adams, S. A., & Riggs, S. A. (2008). An Exploratory Study of Vicarious Trauma

Among Therapist Trainees. Training and Education in Professional Psychology,
2(1), 26-34. https://doi.org/10.1037/1931-3918.2.1.26
Gabbett, T., Jenkins, D., & Abernethy, B. (2010). Physical collisions and injury
during professional rugby league skills training. Journal of Science and Medicine
in Sport, 13(6), 578-583. https://doi:10.1016/j.jsams.2010.03.007
Morar, T., & Marais, B. (2022). Exploring burnout among psychiatric trainees at a
South African university. South African Journal of Psychiatry, 28(0), a1634.
https://doi.org/10.4102/sajpsychiatry.v28i0.1634
o Note that the URL can be given instead of the DOI when the DOI is unknown, for example:
Miller, R. L. (2014). Community psychology, evaluation, and social critique.

American Journal of Evaluation, 36(1), 89-99. http://aje.sagepub.com/
o Always use a hanging indent from the second line of all references. (In MS Word select the
paragraph menu, highlight your reference list and set the indentation to hanging by 1.25cm.)
Reference list: Books

Printed book
E-book with DOI

Given, L. M. (Ed.). (2008). The Sage encyclopaedia of qualitative research methods.
Sage. https://dx.doi.org/10.4135/9781412963909
E-book with url

Manning-Morton, J. (Ed.). (2014). Exploring wellbeing in the early years. Open
University Press. http://www.vlebooks.com
Reference list: Chapter in an edited book

Printed book
Christie, D. J. (2001). Peacebuilding: Approaches to social justice. In D. J. Christie,
R. V. Wagner, & D. D. Winter (Eds.), Peace, conflict and violence: Peace
psychology for the 21st century (pp. 277-281). Prentice Hall.
https://u.osu.edu/christie/about/peace-conflict-and-violence-peace-psychology-
for-the-21st-century/
38
E-book
Aichberger, M. C. (2019). The Current Role of Intercultural Psychotherapy. In M.
Schouler-Ocak, & M. C. Kastrup (Eds.), Intercultural Psychotherapy for
Immigrants, Refugees, Asylum Seekers and Ethnic Minority Patients (pp. 25-28).
Springer. https://doi.org/10.1007/978-3-030-24082-0
Reference list: Webpages (provide the author or institution details)

Amnesty International. (2019). Living in Limbo: Rights of Asylum Seekers Denied.
www.amnesty.org
Dancey, C. P. (2002). Ten years of IBS.

http://www.mindbodydigestive.com/iperspect.html
Reference list: Online videos (provide the presenter or institution details)

Hastings, C. (2016, Sep 27). Get Lit: The Literature Review [Video]. Youtube.
https://www.youtube.com/watch?v=9la5ytz9MmM
The Texas A&M University Writing Center. (2015, Feb 28). Avoiding Plagiarism:
Writing With Integrity [Video]. Youtube.
https://www.youtube.com/watch?v=F1S1FZ-bn5E
Reference list: Podcasts and radio broadcasts

Taylor, L. (Host). (2019, May 8). Conspiracy theories [Audio podcast]. In Thinking
Allowed. BBC Radio 4. https://www.bbc.co.uk/programmes/m0004sly
Bunn, G. (2011, November 9). The origin of common sense [Radio broadcast]. BBC
Radio 4. https://www.bbc.co.uk/programmes/b016x234
Reference list: Film or movie

Coogler, R. (Director). (2018). Black Panther [Film]. Marvel Studios.
Reference list: If there is no date on the publication

Davies, R. (n.d.). How to navigate Google Scholar. https://nowthis.com/googlescholar/
Reference list: Resoure with no author and no date:

What is psychology? (n.d). BestPsychologyDegrees.com.
https://www.bestpsychologydegrees.com/what-ispsychology/
Reference list: When a publication is still in press and has not yet been released
Davies, R. (in press).
Consult the many resources on the Internet for more details on APA7 referencing style.
Source:
Davies, R. (2020). APA 7th Referencing Guide (2nd ed.). Leeds Trinity University Library.
https://www.barnsley.ac.uk/app/uploads/2021/01/APA-7th-full-PDF-final-version-Aug-2020.pdf
HMPYC80/104 39
HMPYC80/104/0/2024
New Voices Journal. (2014). Copy-editing guide. Unpublished document. University of South Africa.
St Mary’s College of California Library. (n.d.) APA style 7th edition. Library guide. St Mary’s College
of California Library. http://stmarys-ca.edu/library.
40
Research proposals
A research proposal is a written document outlining a proposed research project. The proposal explains
WHAT the project is about, WHY the project is important and HOW the project would be conducted.
The aim of a research proposal is to allow for the project to be evaluated before the researcher embarks
on the study. There are several reasons why the proposed projects should be evaluated. Research
processes are expensive and time-consuming, and one would prefer not to waste time on unnecessary
studies. Studies are unnecessary if they are done for no specific reason or if they merely duplicate what
has already been done. Because research is expensive one would prefer to avoid incorrect procedures
and suspect methodology. The time spent on proper planning and evaluation is worth the effort.
Organisations do not want to waste money on ineffective projects. Studies need to be carefully
coordinated and orchestrated to be useful and productive, which means research projects have to be
properly planned and evaluated before they are put into operation.
A research proposal should be brief and precise. It should aim at very clear communication. The
proposal should be easy to follow and understand, and it should be presented in a well-structured format.
It is a good idea not to include technical explanations and tables in the main text, but rather to place them
in an appendix. The clarity of style, the neatness of the presentation and the fluency and accuracy of
language is of such importance that researchers should never submit the first drafts of their proposals.
Proposals should be rewritten at least once but preferably more than once.
As mentioned in Tutorial Letter 102, the following components are essential for a research proposal:
Table 02
Generic components of a research proposal
Generic descriptions of Specific project descriptions of
main components main components
The outline of the research issue or -
interest -
-
(The context of the project) -
The research statement -

(The research focus guiding the study) -
-
-
The research design -

(The approach to the issue) -
-
-
The method of analysis -

(Technique of dealing with -
-
information) -
HMPYC80/104 41
HMPYC80/104/0/2024
Here is an outline of the main components of a research proposal:
1. Cover letter (optional): A research proposal should be accompanied by a letter stating the
purpose of the proposal. The two main reasons for writing a research proposal are: (a) to
convince other people (eg a research committee, an academic supervisor, your boss or manager,
the executive of a community organisation) that you have a worthwhile and feasible project, and
(b) to apply for money (eg from a bursary committee, a funding organisation, your company’s
research and development budget) to cover the costs of your study.
2. Declaration of academic honesty: A signed declaration of academic integrity and honesty is

required. Use the format provided in the section "Academic integrity and avoiding plagiarism" in
this tutorial letter. See example in this tutorial letter <C. ETHICS AND RESEARCH:
Declaration of academic honesty>
3. Project title: The project title is a brief phrase reflecting the essence of the study. The title
should contain the most important key terms of the research (or problem) statement without
becoming over-inclusive or detailed. Avoid fancy or poetic wording if it does not serve a specific
purpose. The title should be a clearly formulated and easily comprehensible statement conveying
the gist of the project.
4. Abstract (or summary) that provides a brief and concise summary of the proposed study.
5. General introduction: The general introduction is a short paragraph providing a schematic

overview of the project. This paragraph aims to orientate the reader by (a) stating the research
focus, issue or problem, and (b) by indicating very briefly how the researcher plans to study the
problem.
6. The research issue, focus or topic: The research issue is introduced in one or more paragraphs
covering the following topics: (a) a description of the research issue or an outline of the research
focus, (b) justification for conducting the research (why is this of interest; to who or whom and
why and how; what are the reasons for researching this issue) and (c) the current state of affairs
(what, if anything, has research already been done on the issue, for example, other studies or
other attempts to understand, explain or solve the issue).
7. Literature review: A brief but compelling literature review, with a discussion of what key
theories and findings are applicable to the proposed study.
8. Research method: Explain the research method or approach: The research method or approach
is introduced in one or more paragraphs covering the following topics: (a) a description of how
the study is to be conducted, (b) an indication of who or what (e.g. people, animals, documents,
videos) are to be used for the study, (c) an indication of the kind of data required for analysis, (d)
an indication of how this data is to be obtained, (e) an indication of how the information is to be
analysed, and (f) an outline of the project plan indicating dates and costs.
9. Ethics: Evaluate the ethics of the project: An evaluation of the ethics of the project is provided in
one or more paragraphs covering the following topics: (a) issues concerning informed consent
(eg, have sufficient information been provided to allow subjects to decide on their participation),
(b) issues concerning privacy (eg, is the privacy of subject guaranteed) and (c) issues concerning
qualifications (eg, is the researcher qualified to perform the acts and/or use the instruments that
42
may be required for the study). An evaluation of the ethics of the project should be provided even
if no ethical problems are expected or foreseen.
10. State the expected outcomes of the project: The expected outcomes of the project is
introduced in one or more paragraphs covering the following topics: (a) an indication of who
could benefit from the study, and (b) an indication of how these beneficiaries could benefit.
11. A complete reference list that adheres to the APA7 format.
12. Appendices: (a) Project budget with expected costs;

(b) Project timelines.
(c) Participants information brochure or leaflet
(d) Informed consent form for institution and participants.
The "final research proposal checklist" as provided in Tutorial Letter 103 is presented here as well:
Table 03
Final research proposal checklist
FINAL RESEARCH PROPOSAL CHECK LIST x ? √
No +- Yes
1 A "table of contents" with the correct document page numbers is provided.

2 Is there a "declaration of academic honesty" that has been signed?
3 Is the title precise, clear, meaningful and appropriate?
4 Is the abstract clear, comprehensive and succinct?
5 Is it clear for what reason(s) the proposal is submitted?
6 Is it clear what the research issue or interest is?
7 Is the context of the research issue or interest described? Why is the project
important? How is it relevant? To whom is it relevant?
8 Is it clear what research question (or questions) in particular is to be investigated?
9 Is the research question socially and ethically acceptable, or is it possible that
some people may be offended by and/or object to this research question?
10 Is sufficient important literature discussed and cited to provide a comprehensive
theoretical context for the project?
11 Are the key aspects (e.g. constructs, variables or themes) implicated in the
research question, clearly indicated?
12 Is it clear how these key aspects, constructs or variables are to be observed,
measured, accessed or evaluated?
13 Are the data/information sources (subject groups, documents, settings, etc.)
clearly indicated?
14 Are these sources of information sufficient to address and answer the research
question?
HMPYC80/104 43
HMPYC80/104/0/2024
15 Can these data sources (the participants, documents or other sources) be accessed,
or can one expect to experience problems in obtaining the data?
16 If the data sources are people or groups of people, are they going to be fully
informed about the nature of the research project and about their rights to
volunteer and participate in the study?
17 Can the equipment/tools (questionnaires, tests, special apparatus, etc.) that are
required for the project be accessed? Is the researcher qualified to use the
equipment/tools required for the project?
18 Is it clear what kind of data are to be gathered?
19 Will the data be reliable and valid (quantitative studies), or trustworthy,
dependable, credible and believable (qualitative studies)?
For example: can one be sure that the data truly reflect the situation in question
(matter of validity). Would one get the same information if the observations and
interventions were to be repeated (matter of reliability)?
20 Does the researcher know how to analyse the data? Is the method of analysis
indicated in the proposal?
21 Are matters of ethics discussed?
How will potential/committed participants be provided with information about the
research project?
How will the participants be protected, and their welfare ensured during the
research project?
How will the right of subjects/participants to withdraw from the research project be
made clear to them?
How will the privacy of subjects/participants be protected?
How will the confidentiality and anonymity of data be ensured?
How will subjects/participants be debriefed or referred for counseling or therapy
(where relevant)?
How will subjects/participants be informed of the results after the research is
completed?
22 Are possible limitations of the project indicated and discussed?
23 Are the relevance and implications of expected findings/outcomes indicated?
24 Is the proposal written in clear and precise language that is to the point?
25 Is the proposal structured correctly? Do all the paragraphs follow logically on
each other?
26 Are the following Appendices included?
(a) Project budget with expected costs.
(b) Project timelines and time frame (as expected).
27 Are the following Appendices included?
(c) Participant information pamphlet/ handout.
(d) Informed consent form for signature by researcher and participants.
44
28 Has the proposal been self-reviewed by the researcher with the aid of the Research
Review Inventory (items 1 to 45)?
29 Have all omissions, errors or limitations that were identified in the proposal
self-review (Research Review Inventory items 1 to 45) been corrected and
addressed?
30 The final proposal has been submitted via Turnitin or similar software as
prescribed by Unisa before the assignment due-date.
If the threshold of 40% similarity or higher has been reached then the relevant
sections must be corrected to adhere to the requirements of academic honesty and
integrity. The updated proposal must be resubmitted via Turnitin or the prescribed
software and a similarity index of less than 40% must be attained for the assignment
submission. The complete Turnitin Report must also be submitted in Adobe PDF
format with the proposal as part of the proposal Appendices. (In certain cases the
threshold will be lifted to a maximum of 50% similarity in Turnitin.)
Evaluate your proposal using the following checklist, and revise and amend if necessary. Use (√) if the
item is covered in your proposal, (x) if it is not covered, and (?) if you feel unsure about a particular item.
The Research Review Inventory (R.R.I.3)

For a more detailed criteria of what constituted an acceptable research proposal -- carefully study the
items of the Research Review Inventory (R.R.I.3) in Tutorial Letter 102 ! This is very important as it
constitutes the final review method for your own proposal and for peer proposal reviews in this course.
Consult the following resources in this tutorial letter with regard to academic writing and referencing:
<D. REFERENCING AND FORMATTING: Academic writing tips and suggestions>
<D. REFERENCING AND FORMATTING: American Psychological Association formatting and style
(APA7)>
RECOMMENDED STUDY:
Study the following presentations:
Hastings, C. (2016, Sep 27). Get Lit: The Literature Review [Video]. Youtube.
https://www.youtube.com/watch?v=9la5ytz9MmM
Hastings, C. (2016, Sept 27). The Perfect Proposal [Video]. Youtube.

https://www.youtube.com/watch?v=bSCKdmCfzHo
Additional sources
Tutorial Letter 102, HMPYC80, 2020: Research Review Inventory (R.R.I.3). Unisa.
HMPYC80/104 45
HMPYC80/104/0/2024
Scientific journal articles

A certain practice of reporting on research projects has developed over the years. This form of reporting
is known as the scientific article. Although it is not the only way in which one can report on one's project
(delivering papers at scientific congresses is another way to do this: <Reporting research (Oral
presentation)>) it is by far the most popular and respected way in which to inform the scientific
community about one's work. The method is respected because it involves peer review. Scientific
journals select articles for publication on the basis of peer recommendations. In other words, unless your
peer scientists agree that your study was conducted in a proper way, you will not succeed in having your
article published in a respected journal.
Scientific journals do accept theoretical and review articles but such papers are extremely difficult to
write. One really needs to be a specialist in one's field of expertise to succeed with theoretical articles.
Therefore we concentrate on empirical studies.
Empirical articles have a fairly standard format, as we describe below, but this does not mean that there
cannot be minor differences in the way you present your material. One's presentation is influenced by the
nature of one's research. Relational-causal research, descriptive-interpretive research and action
research studies often require different report formats.
In what follows we first pay attention to general aspects of scientific writing and then provide more
detailed information about the structure of a scientific article.
The style and tone of writing
(APA7)>
The style and tone of writing affect the readability of an article. One should never forget that a research
article imparts knowledge and therefore readability is an important factor. Your article will be much
more reader friendly if you make sure that what you write is to the point (relevant) and that it is written in
a lucid and precise style. Also take care that your presentation is structured logically, that information is
presented in an interesting way and that all your sources are quoted.
· Be relevant: Your article should form a coherent whole, which means you have to avoid
unnecessary verbiage. Do not digress from the main issue. Ensure that the information you
present is directly relevant to the topic you deal with. Also, guard against personal and anecdotal
information unless your specific approach, for example action research, requires you to do so. If
your research approach does not require personal and anecdotal information this kind of
information only serves to distract the reader from the main themes and the real methodological
issues that form the core of the article.
· Write lucidly and precisely: The main objective of research articles is to impart knowledge
clearly, accurately and concisely. Furthermore, academic writing requires one to substantiate
what one says by either referring to relevant research findings or by arguing one's case.
46
Therefore, define the terms you use (unless the reader can reasonably be expected to be familiar
with their meanings), chose your words carefully, and formulate your sentences clearly. Avoid
making claims you do not provide evidence for, and in presenting your information aim for
precise writing and clear and direct statements. It is often useful to employ graphical
representations (such as diagrams, tables or graphs), but do not overdo this. Employ such media
only if they serve to clarify your findings and the points you try to make. Finally, it is better to
write in the active voice than to do so in the passive voice because passive sentences tend to have
complex structures and they can easily become more difficult to interpret than sentences written
in the active voice.
· Structure your information logically: A research article can be viewed as an extended argument
in which its author attempts to convince her readers about the relevance of her research topic and
the importance of her findings. The aim of an article is to communicate, and not to confuse or to
mystify. It is, therefore, important to structure the content of your article logically. It is often
useful to employ subheadings as these help to organise the material in terms of different logical
sections so that the overall structure and plan of the article becomes apparent to the reader.
Organised material is always easier to understand than poorly organised or illogically presented
information. Organisation is important not only because it allows the reader to grasp the material
more easily but also because it forces the writer (in other words, you!) to think more clearly about
the topic, and to present the information in a coherent an intelligible manner.
· Present the information in an interesting way: Although academic writing is different from
literary writing, it needs not be terse and dull. You should describe your research findings in a
lucid and concise way but this does not mean that you should not strive to present them in an
interesting and compelling manner. It may be difficult to find a balance between these two styles
but bear in mind that the point of the article is to convince the reader of your perspective on the
research topic and the results of a research study. These results should be presented in such a way
that the reader's curiosity is piqued so that he or she wants to find out more about your research
problem and about how you solved or attempted to solve the problem.
· Acknowledge your sources: Academic writing does not occur in a vacuum. Any research takes
place within a research context. It is important to acknowledge how other researchers have
contributed to your understanding of and approach to the research problem. The way to do so is
to include appropriate references for all sources referred to in your text. Use the APA referencing
technique when you compile your list of references. If you quote verbatim from a source use
quotation marks and give a source reference. If you do not acknowledge your sources you are
guilty of plagiarism! Also note that all sources referred to in the article must be included in your
list of references (that is placed at the end of the article).
The structure of a scientific article
Although one cannot insist on a specific format for scientific articles such articles usually involve most
of the following components:
· a title
· an abstract
· an introduction
· an indication of how the research issue, interest or problem was approached
· a section explaining how data were analysed
· a discussion of the results obtained
· an indication of the conclusions drawn
· a list of source references
· one or more appendices
HMPYC80/104 47
HMPYC80/104/0/2024
· TITLE
One’s first task is to indicate what the article is about. The most economical way to do so is to provide a
title. The title has to be a concise statement about the main topic covered in the article. A title usually
refers to the major variables investigated in a study. It summarises the core of the research issue
addressed in the article. Titles are important because they are used for indexing articles in academic
databases. Take time to think about your article's title because titles may lead to incorrect indexing if
they are misleading or if they only vaguely describe the topic. Not only should a title describe a study
accurately, it should also be meaningful and explanatory when standing alone. From simply reading its
title one should already get a good idea of the content of an article. However, this does not mean that you
should make your title into an essay. If possible a title should not exceed 15 words. Be economical with
words and try to avoid redundancies such as "A study of... " or "An experimental investigation into ... ".
Usually, these phrases do not add explanatory value to your title. Also, refrain from using abbreviations
in the title as these may only serve to confuse the reader.
· ABSTRACT
The title of an article is followed by an abstract. Abstracts are the most difficult part of an article to write.
The abstract is a brief summary of the article. From reading an abstract the reader should gain a good
idea of the research problem and the results and conclusions of the study. The abstract should be self
contained. It should be fully intelligible on its own so that the reader can quickly form an idea of what the
study is about. However, despite being a summary of the entire essay the abstract should not exceed 100
to 120 words in length.
A proper abstract supplies four pieces of information. (a) It states the project's goal; (b) it describes the
research method used; (c) it mentions the study’s principal results and conclusions; and (d) it indicates
the significance and implications of these findings.
Although an abstract is situated between the title and the beginning of the article authors usually
compose their abstracts after they have written their articles. However, it is a good idea to write a
preliminary abstract before one composes the rest of one's article. A preliminary abstract serves to focus
one's ideas. Of course one has to revisit to the preliminary abstract (and more often than not rewrite the
abstract) once the article has been completed. Make sure that the abstract you write is an accurate
reflection of the article.
· INTRODUCTION
Research articles begin with introductory notes in which authors briefly describe the research problem
and develop it with enough breadth and clarity to be understood by as wide an audience as possible. In
doing so authors contextualise the research problem. They discuss what other researchers have done in
the field and they explain why they have decided to investigate the matter further. Open your article with
an introductory section in which you state the research problem and explain why the problem is
important enough to study and how it relates to previous theory and research.
Introductions are usually structured as follows: First explain to the reader what the research issue,
interest or problem is you investigate, and why the issue warrants investigation. Next, contextualise the
research issue, interest or problem by situating it in a theoretical context. In other words, indicate how
previous research contributes to the issue at hand and explain how your particular research research
issue, interest or problem emerges from, or links to, previous research. This is why a literature review is
48
an important part of one's research project. One has to include a survey of relevant literature (i.e. articles
and books published on the research problem in question). However, the literature survey should be kept
short an to the point. Try not to drown your reader in detailed discussions of other studies. Readers can
use the reference information you provide to find and read the original studies. The only thing readers
are interested in when reading your article is a summary of previous findings and how these findings
pertain to your research question or research statement. Show how your work builds on and differs from
studies cited in literature. The last part of the introduction deals with a brief explanation of how you
approached the research issue, interest or problem and accomplished the purpose of your study. In other
words, provide an overview of the method (or methods) you used and what the outcomes of your
approach were with regard to the purpose of your study.
Therefore, having read the introduction your reader should have a general understanding of
- the research question (or research statement) that was investigated

- the academic and/or practical context of the research question
- the contribution of the present study (how it builds on and differs from previous work)
- the approach and methods used to investigate the research issue, interest or problem.
Note that (a) relational-causal, (b) descriptive-interpretative and (c) actions research methodologies
require slightly different angles of approach. The summaries that follow should give you an indication of
the differences in emphasis and how these differences may alter your introduction.
(a) Relational-causal research: It is usually appropriate to include a formal statement of the research
questions or hypotheses being investigated. You should also explain what variables are at stake, what
relationships you postulate between these variables, as well as the results you expect and why.
Furthermore, in the introduction the theoretical motivation for the study should be made clear. The
introduction should clarify the point of your study, mention the rationale or logical link between the
problem and the research design and comment on the significance of the study and its relationship to
previous work in the area. A good introduction answers these questions in a paragraph or two and gives
the reader a firm sense of what the research did, and why.
(b) Descriptive-interpretative research: In the case of a descriptive study it is best to avoid words that
imply that you are looking for causal relations between variables. You do not have to state explicitly
whether you are working from a positivist, a hermeneutic or a semiotic perspective, but it helps if you
adopt a specific stance and keep to it (i.e., be consistent). In the introductory section your main concern
should be to formulate a problem statement and to motivate it. To situate the problem in a context, you
can also use some nontraditional background material such as newspapers, references to films and
television programmes, and web searches.
(c) Action research: In the first part (the reconnaissance section) of the article you should describe the
practical difficulty that prompted the study in the first place. Your introduction should attempt to answer
the following questions. What happened? Who was or were involved? What were the circumstances
reigning at the time? Was there a sense of urgency and why? In reflecting on the issue, what explanations
were initially entertained? What action plans were considered and why? Were any pertinent publications
consulted in this respect? - because of the unique and pressing nature of the problem, this may not
always be feasible. This section concludes with a formal definition of the problem as well as a
concomitant action hypothesis. Also see participatory action research and intervention research as
discussed in Section E of the prescribed textbook by Fouché, et al. (2021).
(d) An exciting and dynamic research area is in the domain of indigenous research methodology, and
decolonising research methodologies. This was discussed earlier in <RESEARCH: What is it?>.
HMPYC80/104 49
HMPYC80/104/0/2024
Note: The position and function of an introduction are fairly obvious, therefore authors usually do not
provide a subheading for their introductions called "Introduction". However, since we are practising
how to structure our research articles it is perhaps not a bad idea to label our introductions with the
subheading, "Introduction". Doing so provides a clear focus on the structure of the article and the logical
presentation of the information.
· APPROACH AND METHOD
The introduction contains an overview of how the research issue, interest or problem was approached
and what methods were used to study the issue. In the approach section the ideas mentioned in the
overview are expanded and discussed fully. The introduction simply mentions the approach and
methods. In the approach section you have to explain why a particular approach was followed and why
certain methods were decided on. In other words, you have to discuss the reasons for deeming
appropriate the approach and methods of your study.
The approach and method section should inform the reader how the study was conducted. What you put
in here depends largely on whether your study is mainly: (a) a relational-causal, (b) a
descriptive-interpretative, or (c) an action research project. Begin this section with a brief explanation as
to why you think the approach you followed (i.e. relational-causal, descriptive-interpretative, or action
research) was the best way to tackle your research question. Then progress to a more detailed description
of the methods used. Read the summaries that follow below to gain an understanding of the differences
in emphasis. Do not read about one approach only. You need to consider all three approaches to see the
similarities and differences among them.
(a) Relational-causal research: In the case of quantitative empirical studies, the method should be
described in sufficient detail to permit an experienced investigator to replicate the study if he or she so
desires. Such a description also enables your reader to evaluate the appropriateness of your methods and
the probable reliability of your results. In quantitative research it is usually convenient to divide the
method section into labelled subsections dealing with a description of the subjects, the materials or
apparatus, and the procedure used.
Subjects: Explain who participated and how many participants there were. Also mention major
demographic characteristics (sex and age), as well as any other relevant info.
Apparatus: Give a short description of the apparatus used and their function in the experiment. Also
provide details about any specific psychometric test or any instruments and technology
used.
Procedure: Here you should briefly explain each step in the execution of the design. Mention
instructions to participants and specific experimental manipulations. Describe
randomisation, counterbalancing and other control features in the design. Remember the
following rule of thumb: Provide only as much information as is essential to comprehend
and replicate the study.
(b) Descriptive-interpretative research: If you conduct a qualitative research study you should describe
the material and research participants involved in your study. Kinds of material often include interview
transcripts, field notes, and documentary material (e.g., letters or newspaper articles). Explain why and
how you included the material you did and left out other potential material. Particularly if your material
was collected through ethnographic fieldwork, participatory observation or a similar method, explain the
process of how you became part of the research setting. Describe the way in which you analysed the
material. Do not use labels such as discourse analysis, but write a paragraph or two explaining how
discourse analysis (or whatever other method) was used to analyse the material in your study. Since the
50
researcher is the most important instrument in qualitative analysis, you could say something about your
own preparation for the analysis. If you used other people to help you with the analysis, explain how that
was done.
(c) Action research: Inform the reader about the procedure you followed to solve the problem. Write a
short paragraph or two in which you explain the concept of action research and inform the reader how
you went about conducting your own action research study. Your main concern here is to provide a brief
exposition of and motivation for your method.
(d) Indigenous research methodology, and decolonising research methodologies. This was discussed
earlier in <RESEARCH: What is it?>. There are many variations of this approach and applications are
method specific.
· ANALYSIS
In experimental studies it is customary to include an analysis section in which one explains the
techniques used to analyse the data and also mentions the results obtained from these analyses. In some
instances it is a good idea to separate the analysis section from a section in which the results are
discussed. Following this custom we provide a second section called "Discussion". However, keep in
mind that it is not always possible to separate the analysis and discussion sections in one's research
article. In discourse analysis it is difficult to separate the analysis of data from discussing the
meaningfulness and implications of data. In action research the analysis of data may by a single step in a
cycle of interventions and analyses.
(a) Relational-causal research: The results should summarise the data collected and your statistical
treatment of them. Describe the kind of data collected and the statistical method used to analyse the data.
Explain why the particular method was chosen. State the main aspects of the results or findings (e.g.,
note which relationships are statistically significant and which are not) and report data in sufficient detail
to justify these statements. Do not in this section discuss the meaning and implications of the results.
Discussions of the implications of findings are usually done under a separate heading indicated as
"Discussion of results" conclusion.
(b) Descriptive-interpretative research: In the case of a qualitative study it is generally unnecessary to

have a separate heading for results and discussion. These two sections are typically combined. See the
discussion section for a more detailed description of how the analysis and discussion of results are
reported when you do descriptive-interpretative research.
(c) Action research: This section starts with an account of the first cycle of problem-solving or research
activities. (i) Explain how the plan of action was implemented; (ii) describe the effects (results) that were
observed; (iii) explain how you reflected on the experience (analysis); and (iv) indicate how you planned
to proceed from there. Then comment on the subsequent cycles. An action research report is unique in
that it describes a series of interconnected problem-solving attempts. It is not always easy to distinguish
the various phases in the process. Action research is often a rather messy process and different cycles of
activities may occur simultaneously. However, it is important to give the reader a sense of how the
project developed over time. You may find that many of the ideas entertained in the course of executing
the project derived from your own store of scientific knowledge and you should therefore give
recognition to the original sources of these ideas when you write up your study.
method specific.
HMPYC80/104 51
HMPYC80/104/0/2024
· DISCUSSION AND CONCLUSION
In the final moments of one's research report one has to indicate to what extent one managed to solve or
answer one's research problem or question. In other words, you have to evaluate and discuss your
findings. Also try to tease out some of the implications your study may have for future research. A
research study often succeeds in answering one question while raising many other questions. These new
questions may lead to further studies, thus generating research. Do not expect your study to answer all
the questions you set out to investigate. Research, and specifically empirical research is seldom a
cut-and-dried affair where everything magically falls into place. The theories we work with are only
approximations of reality, and they inevitably contain a certain degree of error, bias and subjectivity.
A note of caution: One's discussion of the contribution of one's research necessarily contains a
speculative element. Remember that speculation is in order only if it is (i) identified as such, (ii) related
closely to empirical data or follows logically from theory; and (iii) is expressed concisely.
Note that the discussion of results/findings functions as a conclusion in which one draws implications
from one's study and suggests possibilities for further research. However, researchers sometimes feel
they have to include a separate section in which they state their final conclusions. If you feel your project
warrants a separate section for your conclusions you are welcome to separate the discussion section into
two sections, one called “Discussion” and the other “Conclusion”.
After these general remarks on discussion and conclusion we conclude the section with more specific
notes for each of the different approaches.
(a) Relational-causal research: After presenting the results, you are in a position to evaluate and
interpret their implications, especially with respect to your initial hypotheses. In the discussion section
of your article, you have to examine your results, interpret and qualify them, and draw your inferences.
Treat the results and findings with care. Ensure that your discussion and interpretation are consistent
with the results reported in the analysis section. It may be useful to open the discussion with a clear
statement on the support or lack of support for your original hypotheses. Then proceed to interpret your
results and to comment on similarities and differences between your results and the work of others.
Conclude by identifying issues emerging from your study and by indicating possible directions for
future research.
(b) Descriptive-interpretative research: As noted previously, the discussion and results section of a
research article are usually combined in the case of a qualitative research study. Therefore, present your
analysis and provide a systematic discussion of it. However, avoid unnecessary detail as this may only
serve to bore the reader - rather concentrate on the highlights. Qualitative analysis can involve all sorts of
interesting restructuring of the material (such as recasting it as a narrative), and you can try to use some
of these ideas by reporting on the analysis as well (i.e., by conducting a meta-analysis) - but be careful
not to alienate your readers by being too avant-garde! A more conventional way of reporting on
qualitative research is to discuss four or five main themes, categories, or discursive strategies emerging
from your analysis. Also provide illustrative material (e.g., interview extracts) as these help to give life
to the report. Conclude by mentioning some implications of your study, by identifying the insights that
you have generated into the research problem, or a new way of looking at the problem, and indicate
possibilities for further research.
(c) Action research: Action research is basically a problem-solving approach in which the researcher
tries to zoom in on a problem in order to overcome (i.e., solve) it. The discussion section of your article
will allow you to reflect on this research process and to determine whether you have come out of it with
52
new understanding. Hence, in this section of the article you should comment on the insights you have
derived from your research. However, while the purpose of action research often is to overcome a
somewhat subjective problem, you should try to draw out some general principles and guidelines from
your study. In other words, you should indicate not only to what extent you have found the study
illuminating to yourself, but also whether these insights are of value to other people/researchers. You
should therefore try to engage the academic fraternity by discussing the understanding that you have
gained through the experience, even though this may amount to little more than asking challenging
questions. Also note that while the typical research article begins with an overview of the pertinent
literature in order to identify a research question, the action researcher often ends with such a review.
method specific.
· SOURCE REFERENCE LIST
Your article has to conclude with a list of all references that were cited in the text. Do not include
references to works that you consulted but did not actually refer to in your article. The list of references
comes at the end of the article but before any appendices. Make sure that all sources are correctly
referenced (all relevant details such as year of publication and page numbers, in the case of journal
articles, have to be included). Also ensure that the list is alphabetically organised.
(APA7)>
· APPENDICES
Appendices can be helpful and are used for material which could be distracting if included in the main
text of your article. For example, when demonstrating the technique of discourse analysis, one does not
have to include the full analysis in the main text of one's article because one only needs segments of text
to illustrate one's method. In such cases it is customary to include the full analysis as an appendix.
Examples include:
(a) Project budget with expected costs;
(b) Project timelines.
(c) Participants’ information brochure or leaflet
(d) Informed consent form for institution and participants.
IN GENERAL
A research article is really no more (or no less) than a written record of the basic components making up
the research process. Its function is to delineate, as concisely and accurately as possible, the aim and
rationale of the study, the way in which this aim was accomplished, the results obtained how these
results should be interpreted. If you find writing the article difficult then take some consolation from the
fact that for all of us the actual writing of a research article is an arduous, time-consuming and tedious
task. It often turns out to be the least enjoyable aspect of the research process. Still, it is probably the
most important part of doing research. You simply have to communicate your findings to other
researchers, and it is through the medium of this kind of written communication (especially the research
paper) that research results are made available to others.
HMPYC80/104 53
HMPYC80/104/0/2024
THE RESEARCH REVIEW INVENTORY (R.R.I.3)

For the detailed criteria of what constituted an acceptable research article and other research manuscripts
such as a research report, a Master's dissertation or a doctoral thesis. Carefully study the items of the
Research Review Inventory (R.R.I.3) in Tutorial Letter 102 ! This is very important as it constitutes the
review method for article reviews in this course.
Additional sources
HMPYC80 (Tutorial Letter 102) Research Review Inventory. Unisa.
54
Conference presentations
Scientific conferences are often the first forums where researchers present the results of their studies.
This makes conferences very popular amongst scientists - not only are conferences the forums where
scientists meet each other and exchange ideas, they also provide platforms for quick and easy
dissemination of the most current theoretical ideas and research results. The papers scientists present at
conferences are often the first step towards writing more comprehensive papers for publication in
scientific journals (See < Scientific journal articles>).
The purpose of this resource essay is to explain how to structure information for an oral presentation.
The emphasis is on organising information for presentation - the methods and technical aspects, and
different kinds of conference presentations will not be discussed here.
The main points of a typical presentation
Typical presentations often contain the following main points:

- A problem statement: What is the reason for the presentation? What is the research interest, issue (or
problem) under consideration? Why is it an interest, issue (or problem), and for whom?
- Approach to the research interest, issue (or problem) and/or the method of investigating the interest,
issue (or problem): What research has been done on the phenomenon or issue? How has the on the
phenomenon or issue been approached? What methods have been followed to investigate the on the
phenomenon, issue or problem?
- Findings or solutions: What were the insights, understandings and findings? What do the findings
suggest and imply with regard to the phenomenon, issue (or problem)?
- Implications: Who could benefit from this research investigation, and how could they benefit?
Obviously a number of further points would be required to describe each of these main points, and
subpoints might be required to describe some of the further points. Here is an example of a presentation
outline in terms of main points and further points.
Table 04
Conference presentation outline
Main points Further points
Research issue, statement or * The methodology of oral presentation is seldom taught

problem * Professors / senior managers seldom pay attention to the
quality of their own presentations
* Conferences are littered with poorly prepared presentations
* Audiences are often bored by presentations
Unravelling the nature of * Unplanned or unprepared presentation

presentations: * Poorly structured information (i.e. no clear points)
What makes a bad presentation * No visual material (no notes or slides) supporting talk
bad?
(approach to the research
interest, issue or problem)
HMPYC80/104 55
HMPYC80/104/0/2024
Unravelling the nature of * Thorough planning and preparation

presentations: * Not more than three to five main points
What makes a good presentation * Clear formulation of main points
good? * Precise & sufficient information to support each main point
(approach to the research
interest, issue or problem)
How to prepare a good * Structure the presentation information

presentation * Plan the presentation (brief, precise and clear)
(studying the research interest, * Support talk with visual material (notes and slides)
issue or problem)
What is to be gained from good * Personal gains for speaker and members of audience
presentations? * Clear and precise dissemination of information
(implications) * Making contacts and building interpersonal networks
More main points could be added perhaps, and the existing main points could be clarified further by
adding additional further points. It would also be possible to expand some of the further points into
subpoints. The amount of detail of a presentation is a function of the time allowed for presenting the
presentation. In other words the level and amount of detail is governed by the length of the presentation.
It would be extremely bad conference etiquette to exceed the time allowed for one’s presentation.
Structuring information: Spread versus depth
A golden rule of oral presentations is to spread the sequence and to restrict the layers of information. Try
not to exceed three layers of information. In other words structure the presentation in terms of main
points (first layer), further points (second layer) and subpoints (third layer). If you find that your
subpoints need to be divided into sub-subpoints, then you are probably creating a structure with too little
spread (ie too few main points) and too much depth (too many layers). If this is the case you will lose
your audience in your presentation. People find it difficult to keep track of ideas when the speaker
explores too many layers of information. Individuals are much better at keeping up with ideas presented
orally if the ideas are presented in a chronological sequence.
Slides are the building blocks of a typical presentation
A presentation is divided into slides. The number of main points outlined for a presentation gives the
minimum number of slides required for the presentation. Expanding the main points into further points
and subpoints usually requires slotting in more slides.
A slide is a basic unit of structured information. A slide captures an idea and expresses the idea in the
form of one or more points. In other words the points presented on each slide fit together to express a
certain idea. Thus a set of slides represents a set of ideas, and a set of ideas constitutes a presentation.
Slides contain points. These points reflect the way in which the information is structured. A slide could
contain main points, or further points, or subpoints of further points, or a combination of these. For
example, one could have a number of main points, or one could list one main point with a number of
further points and subpoints. However, it is a good idea not to have too many points per slide. In fact, one
should try to avoid presenting more than five points per slide! When listing a number of short items (i.e.
56
the characteristics of a phenomenon) one could have more than five points per slide but when presenting
more substantial points one should refrain from having too many ideas per slide.
The nature and layout of a slide
A slide is the basic unit for organising information for an oral presentation. There are different kinds of
slides, for example, flash cards, overheads, photographic slides and electronic slides. For present
purposes one could see the term, slide, as a generic term encapsulating any of these forms. In its generic
abstract form a slide is a unit of structured information. In its concrete form one can think of a slide as an
ordinary piece of paper, for example an A4 size page, with something written on it, or with a graph or a
picture drawn on it.
Slides are usually rectangular in shape. They can be presented in portrait or in landscape orientation. A
portrait orientation is suitable for listing a number of short items. A landscape orientation is used to
present statements, graphs and pictures.
The points contained in a slide can be textual or graphical or a combination of text and graphical
material. In other words, a point can be expressed as a written phrase, or, it can take the form of a graph
or a picture, or it can incorporate a combination of these.
Outlining the first slide
Use the first slide to outline the presentation. The title of this slide could be the title of the presentation
(or an abbreviated phrase reflecting the title of the presentation) The rest of the slide contains the main
points of the presentation. Remember not to formulate too many main points. Begin with three to five
main points - one can always add more main points if this is necessary. Although it may sometimes be
necessary to have more than five main points this is not the norm. If you have more than five main points
chances are you have formulated the main points at too low a level. In other words, the categories
represented by these main points are not sufficiently general. The have been formulated on a level
containing too much detail. If this is the case, you should pitch your main points at a higher (more
abstract) level of description.
Here is a generic example of a first slide:
The title of the presentation
* The first main point

* The second main point
* The third main point
* The fourth main point
* The fifth main point
Here is an example of a first slide, using the presentation outline referred to earlier:
HMPYC80/104 57
HMPYC80/104/0/2024
Preparing good presentations
* Few people know how to prepare good presentations

* What makes a bad presentation bad?
* What makes a good presentation good?
* How to prepare a good presentation
* What can be gained from good presentations?
A few tips: Note the use of an A4 in landscape format (as per the above slide). Landscape should be used
when one does not have a long list of points. The extra width allows for more space and thus fewer lines
per point. This makes the points stand out more clearly. If you are not using a computer to prepare your
presentation, use A4 sheets to outline the slides. Remember to use large lettering - small lettering is very
difficult to read for somebody sitting at the back of an auditorium. Also remember that lower case letters
are easier to read than capital letters. If you use a word processor use different font sizes to make titles
appear bigger than the rest of the text. If you do not use a word processor use capital letters to distinguish
headings from the rest of the text.
Writing slide notes

The presenter determines the nature of the notes that accompany the presentation slides. Some
presenters write extensive notes, while others hardly write any notes at all. There are, however, at least
two reasons why presenters should write notes to accompany their slides: (a) Slide notes refresh the
presenter’s memory. The slides themselves are summaries. They are not complete texts. When preparing
her slides a presenter may know exactly what she intends to say, but one week later this may not be the
case. By that time she may find her own slides cryptic and obscure. (b) Slide notes constitute the
beginnings of a research paper. If the presentation is suitably structured and if the slides are properly
documented then these slide notes constitute the main paragraphs of a research paper.
Here is an example of the notes accompanying the slide that outlines a presentation. It is not difficult to
see that this kind of slide summarises the introduction to a research paper. It contains the problem
statement, the approach, the method, the results and the implications of the research project. Note how
easy it is to write an introduction to a paper about good presentations by simply expanding on the main
points mentioned in the outline slide.
Notes for slide: “Preparing good presentations”:
We live in the information era - an age in which there is hardly an individual who does not every
day encounter some form of information. Our daily lives are filled with moments in which we
have to absorb information, interpret information and provide information. Yet, in this day and
age we are still confronted with the fact that most people do not know how to handle
information effectively and efficiently. Research studies show that more than half of conference
papers are poorly presented. More than a quarter of all presentations made in large companies
are so badly structured that they hardly convey any information at all. The problem is that few
people actually know how to prepare good presentations. Presentation skills are seldom part of
taught curricula and in practice badly structured presentations become the norm as professors
and senior managers alike fail to reflect on the quality of their own presentations.
58
This paper captures the results of an extensive literature survey on the subject of oral
presentations. It looks at the characteristics of poorly structured presentations and uses these to
formulate the attributes of good presentations. It then explains how to prepare a presentation of
high quality. The paper concludes with a discussion of the advantages and benefits of good
presentations.
Outline the remaining slides
Once one has a first slide listing the main points of the presentation the remaining slides are simply
developed from each of these main points. The problem is to get a clear picture of the substance of each
point. In other words one needs a clear view of the further points and the subpoints sprouting from a
main point. It pays to follow an outline of structure approach. The outline of structure approach requires
one to prepare an outline of main points, further points and subpoints for all the slides, and not to work
single slides at a time.
Here is a generic example of a simplified form of the outline of structure approach. It lists main points
only. In this instance there are no further points or subpoints. The generic example appears in the left
column. The right column shows the main points from the “good presentation” example:
Main point Examples Main points
(1) Few know how to prepare Main point (1)

good presentations
(2) What makes a bad Main point (2)

presentation bad?
(3) What makes a good Main point (3)

presentation good?
(4) How to prepare a good Main point (4)

presentation
(5) What is gained from good Main point (5)

presentations?
To expand the structure one simply includes further points and subpoints. Here is a generic example in
which the basic form of the outline of structure approach is expanded to include further points. Three
main points are listed. Main point (1) has three further points and no subpoints, main point (2) has one
further point and no subpoints, and main point (3) has two further points each with subpoints, the first
further point boasting two subpoints and the second four subpoints
HMPYC80/104 59
HMPYC80/104/0/2024
Main points Further points Subpoints
Further point (1.1) Main point (1)
Further point (1.2)
Further point (1.3)
Further point (2.1) Main point (2)
Further point (3.1) Subpoint (3.1.1) Main point (3)
Subpoint (3.1.2)
Further point (3.2) Subpoint (3.2.1)
Subpoint (3.2.2)
Subpoint (3.2.3)
Subpoint (3.2.4)
If one follows the rule of not putting more than five points per slide it is clear that the fully expanded
main point (3) would not fit on a single slide. It would take three slides to convey main point (3) in a
logically coherent way. The first slide would show main point (3) together with further point (3.1) and
further point (3.2). The next slide would contain further point (3.1) and its subpoints (3.1.1 and 3.1.2),
and the third slide would list further point (3.2) and its subpoints (3.2.1 to 3.2.4). Note how this
arrangement organises information hierarchically, forcing the presenter to move from more inclusive or
abstract main points to more detailed subpoints. The presenter is forced also to provide overviews (eg
first listing both further points (3.1) and (3.2)) before progressing to more detailed information (eg
subpoints 3.1.1 and 3.1.2)).
Here are examples of what the layout would look like on the slides:
This is main point (3)
* This is further point (3.1)

* This is further point (3.2)
60
This is further point (3.1)
* This is subpoint (3.1.1)

This is further point (3.2)

A last tip: Make the effort to write slide notes for each slide. The more effort one puts into constructing
these paragraphs the greater the payoff in the end. It is amazing how the paragraphs of a well structured
presentation simply slot into place to form the skeleton of a research paper.
(APA7)>
HMPYC80/104 61
HMPYC80/104/0/2024
Paradigms
Social science research wants to “draw a picture” of people’s personal and social worlds. However, just
as a particular landscape will be rendered differently by artists working in different artistic styles, even
though each may be reflecting the landscape authentically, the shape of the “pictures” we draw in
research also very much depends on our research paradigm - that is, the style of asking questions and
giving answers which we adopt.
There are many different paradigms, but many authors identify the following three broad classes:
<Positivism>, <interpretivist> research, and <constructionist> research.
[However, there are various other important paradigms, such as indigenous research methodologies…]
Positivism could be compared to a classical, realist style of painting, interpretive research to an

impressionistic or expressionistic style and constructionist research to a cubist or abstract style.
Going purely by what the style of each of the three pictures suggest to you, what kind of paradigm would
you probably like to do research in?
Like different styles of art, each paradigm has a different look and feel, and people often gravitate to one
or the other position because it fits in with the sorts of things they would like to know, and not because
one is better than the others. Thus positivism may suit those who are after objective facts, interpretive
research those who care about the meanings people attach to such facts, and constructionist research
those who wonder how the social world gets constructed in the first place as one that contains “facts”.
62
Some researchers will of course be drawn to all three epistemological positions, and may even apply
them to the same object of analysis for different purposes.
Ontology, epistemology and methodology

Each research paradigm has a particular ontology, epistemology and methodology. Ontology refers to
researchers’ assumptions about the nature of the social world, for example, positivist researchers tend to
assume that the social world operates in a predictable, law-like manner analogous to the physical world.
Epistemology refers to researchers’ assumptions about how to gain knowledge about the social world,
for example positivist researchers assume that one should measure social phenomena objectively.
Methodology refers to the actual procedures used by researchers, for example positivist researchers
frequently collect data by means of objective tests and questionnaires and analyse these using statistical
techniques. The table summarises the ontological, epistemological and methodological characteristics of
the three research paradigms.
Table 05
Positivist, interpretive and constructionist paradigms
Ontology Epistemology Methodology

Positivist The social world can Sociological and Tests, questionnaires,
research be explained in terms psychological facts objective observation,
of causes and effects. can be determined by structured interviews,
means of objective and experiments.
measurement. Quantification,
statistics.
Interpretive The social world is Individuals and groups Participant
research defined by people’s can be understood by observation,
subjective engaging with them in open-ended
experiences. a natural and empathic interviews.
manner. Qualitative
interpretation.
Constructionist The social world is a The ideologies and Feminist research.
research social construction, discourses that make Ideology critique.
that is it is the up the social world can Discourse analysis.
product of how be understood by Cultural critique.
people talk about and analysing what people
act towards what say and do in terms of
they consider to be critical social theory.
real.
Another way of understanding the three paradigms is to think of them as three different ways of making
sense of language (for example an interview transcript):
1) As referring (however inaccurately) to the real world (<positivism>).

2) As reflecting (however indirectly) the intentions, experiences, thoughts and feelings of the
speaker (<interpretive research>).
3) As producing and sustaining certain “truths” about the speaker and the world (<constructionist
research>).
HMPYC80/104 63
HMPYC80/104/0/2024
Suppose you're interested in life in Johannesburg's northern suburbs (traditionally an affluent, mainly
white area). A positivist research question could focus on facts such as what proportion of people in
these suburbs suffer from depression and what the causes are. An interpretivist research question could
involve wanting to describe in rich detail the feelings of isolation that people may experience in their
large, walled properties. A constructionist question could involve wanting to find out how a plausible
image of the northern suburbs is constructed (in the media and in everyday conversations) as a place
where affluent white people live isolated lives on large, walled properties (we would guess, for instance,
that people tend to draw on the "poor little rich boy" stereotype which is quite common in fictional
accounts of wealthy people). Here is another example of three research questions dealing with the same
issue but drawing on different paradigms.
• Is racism among white South Africans related to the degree of interracial contact they have
experienced? (a positivist question);
• How do white South Africans who hold racist views experience interracial contact? (an
interpretivist question).
• What conversational strategies do some white South Africans use to communicate racist
sentiments without appearing to be racist? (a constructionist question).
Finally, here are three sample extracts from articles published in the South African Journal of
Psychology, each operating within a different research paradigm. Can you see why?
Extract 1 (positivist research)

To determine which factors from a range of demographic, perinatal, psychosocial and
hormonal factors were related to postpartum depression, a sample of 81 women between
two weeks and six months postpartum was divided into a depressed group (n=22) and a
non-depressed group (n=59) by means of the Beck Depression Inventory as the main
measure and the Visual Analogue Scale as an additional measure. (Spangenberg &
Pieters, 1991).
Extract 2 (interpretive research)

The aim of this investigation was to gain an understanding of the life-world of people
with diabetes and to acquire insight into the demanding aspects of their lives. (Cleaver &
Pallourios, 1994)
Extract 3 (constructionist research)

Groups of male ex-university students' discourse about drinking was examined to
identify the interpretative repertoires that they used to construct real-seeming versions of
masculinity. (Kaminer & Dixon, 1995)
An indigenous research paradigm

A relatively new research area is in the domain of indigenous research methodology, and decolonising
research methodologies. The reason for doing this kind of research is stated by Chilisa (2008):
“Postcolonial indigenous research methodologies provide an important framework through
which Western-educated researchers can explore the possible biases in the literature we read,
identify the knowledge gaps that have been created because of the unidirectional borrowing of
Euro-Western literature, and bring to a halt the continuing marginalization of other knowledge
systems that occurs because of the dominant Euro-Western research paradigms and their
64
discourses on what can be researched and how it can be researched. Applying indigenous
research methodologies to research with and about the colonized Other should involve going
back and forth to retrieve marginalized and suppressed literature to review, analyze, and
challenge colonizing and deficit theorizing and interpretation, to create counternarratives that see
the past differently, and to envision a transformative agenda with the researched. It also involves
defining literature and theorizing in the context of formerly colonized societies. Postcolonial
indigenous research methodologies perceive literature as language, cultural artefacts, legends,
stories, practices, songs, rituals poems, dances, tattoos, lived experiences such as the people’s
fight against HIV/AIDS, personal stories, and community stories told in weddings, funerals,
celebrations and wars” (Denzin, Lincoln & Smith, 2008, p. 60).
This approach challenges the deficit thinking and pathological descriptions of the formerly colonized
and reconstructs a body of knowledge that carries hope and promotes transformation and social change
among the historically oppressed. Chilisa (in Denzin, Lincoln & Smith, 2008) further states that:
“Western-educated scholars need to investigate the psychological harm, humiliation, embarrassment,
and other losses that these theories and body of knowledge caused to the researched colonized Other.
They also need to use the body of indigenous knowledge about the researched to counter theories and
other misinformation that may cause communities humiliation and embarrassment (p. 60).
For a deeper reading of indigenous research methodologies and decolonising research methodologies
see the following resources:
* Denzin, N.K.; Lincoln, Y.S. & Smith, L.T. (2008). Handbook of critical and indigenous
* Smith, L. T. (1999). Decolonizing methodologies: Research and indigenous peoples.

Additional sources
References
Cleaver, G., & Pallourios, H. (1994). Diabetes mellitus: experiencing a chronic illness. South African
Journal of Psychology, 24(4), 175-183.
Goldberg, R. (1996). I am a new age healer. Quoted in Muller, R.K. Implants from ETs. URL -
http://www.spiritweb/et-implants.html.
Kaminer, D., & Dixon, J. (1995). The reproduction of masculinity: a discourse analysis of men's
drinking talk. South African Journal of Psychology, 25(3), 168-174.
Spangenberg, J.J., & Pieters, H.C. (1991). Factors related to postpartum depression. South African
Journal of Psychology, 21(3), 159-165.
HMPYC80/104 65
HMPYC80/104/0/2024
Positivism
Positivism is one of at least three <paradigms> on which social science research is based. Here is how
Auguste Comte, one of the founding fathers of positivism, describes what it is about:
The law is this: that each of our leading conceptions - each

brand of our knowledge - passes successively through three
different theoretical conditions: the Theological, or
fictitious; the Metaphysical, or abstract; and the
Scientific, or positive.
Positivism has undergone various historical mutations, starting with the work of empiricist philosophers
like Locke, Hume and J.S. Mill, through the “logical positivism” of the Vienna Circle, to its present
position as the reigning orthodoxy in the social sciences. What unifies the different strands of positivism
is the idea that knowledge should be a mirror, map or copy of the real world and that such a copy can be
obtained through empirical observation. Thus positivism is data-driven and can be seen as a reaction
against philosophical systems that start from general assumptions and then try to make the observed
facts fit these assumptions.
The purpose of positivist research design is to obtain distortion-free observations by controlling for all
forms of potential bias, particularly the researcher’s subjective perceptions. The precise manner in which
these observations should be assembled to form constructs, models and theories is controversial (with
some forms of positivism allowing more space for abstraction than others), but empirical observation
remains the corner-stone on which knowledge is built. In the positivist tradition theories should originate
from empirical facts and it should always be possible to test such theories against empirical facts.
Although most research done in the social sciences fits into the positivist mould, it is rare to come across
a research report that labels itself as such. There are two reasons for this. Firstly, positivism is so
pervasive that people hardly feel the need to label it at all - they simply assume that what we have here
called positivism is research. Secondly, there have been so many critiques of positivism that some
people consider it unsophisticated to be seen in that camp. They may therefore prefer to call themselves
post-positivists, eclectic researchers, pragmatists or even realists.
How do you relate to positivism?

(Do you have any lecturers who are positivists?
How would you feel if you were forced to be a positivist in order to obtain
employment?)
☺
Positivists in the human and social sciences tend to model themselves after “hard” sciences such as
biology and, ultimately, physics. (Some would argue that they model themselves after an outdated image
of these 'hard' sciences.) They therefore implicitly believe that personal and social acts can either be
reduced to biological and physical events or are in principle similar to such events, that is that the human
66
life-world is really no more than some kind of mechanism. Of course psychologists and sociologists
have long since learnt that they are unable to make the kinds of definite causal connections that are
sometimes possible in the natural sciences and have therefore developed a sophisticated system of
statistical reasoning based on the probability of social events occurring. That we cannot make definite
predictions (yet) is typically attributed either to the immaturity of the social sciences or to the complexity
of the processes that they have to deal with. Positivism also tends to be mechanistic in another sense,
namely that it hopes to obtain accurate readings of the social world by applying set methodological
procedures.
One tenet is positivistic research is “If you can not measure it, it does not exist!”. What are your thoughts
on this assumption and value-proposition?
................................................................................................................................................................
................................................................................................................................................................
Perhaps the main criticism of positivism is that it (especially in its full-blown quantitative variety)
misses out on the richness of human experience. This criticism is often mentioned by researchers
working in the interpretivist paradigm. Another criticism of positivism is that it treats the social world as
consisting of neutral facts that simply need to be discovered, while in fact it is shaped by unequal power
relationships among individuals and groups. This criticism most often comes from researchers working
in the critical paradigm.
Additional sources
HMPYC80 (Tutorial Letter 102) Appendix 02: Research Review Inventory. Unisa.
HMPYC80/104 67
HMPYC80/104/0/2024
Interpretive research
Interpretive research is one of at least three <paradigms> on which social science research can be based.
It is particularly closely associated with qualitative research (but note that not all qualitative research is
interpretive).
Unlike <positivism> that seeks to explain the social world, interpretive research wishes to understand it.
Explanation has to do with seeing things from the outside as objective facts, whereas understanding has
to do with grasping their inner meaning. While the natural sciences should explain nature, the human
and social sciences, it is argued, should attempt to understand culture - and the methods and procedures
appropriate to the former may be of little use to the latter. Positivism is interested in generalisability and
prediction, but hermeneutics is concerned with contextualisation and interpretation.
_____________________________________________________________________
Would you rather be explained or understood?
Think back for a moment to an occasion in the past year when you felt particularly happy
or sad. Would it be sufficient for an objective observer to explain why you felt so
intensely, or would s/he have to understand your feelings empathetically in order to do
justice to the full richness of your experience?
______________________________________________________________________
There are a wide range of highly specific methods available to the positivist researcher. He or she can
draw on standardised instruments that have been shown to be valid and reliable, employ research designs
that guarantee replicability, and analyse the results by means of sophisticated statistical techniques. The
interpretive approach, by contrast, relies on the researcher him- or herself as the primary research
instrument. Anyone can do positivist research (at least in principle) provided they carefully follow
instructions, but one has to become an interpretive researcher, that is, one has to learn the art of
description and interpretation.
Where a proposal to do positivist research may well include a formal hypothesis, interpretive researchers
prefer to start by formulating a research question. It is a good idea to make this as clear cut as possible.
Although this will have the undesirable effect of closing your mind to certain possibilities, it will help
other people (such as a thesis supervisor or research funders) to judge if the proposed project is
worthwhile and feasible.
One technique, suggested by Creswell (1994), that allows you to be both open-ended and quite specific
is to ask a 'grand tour' question followed by more circumscribed sub-questions. Thus you could, for
instance, ask a grand tour question such as How do people come to identify themselves as being gay or
lesbian? followed by a number of more specific questions such as When and how do people discover
their tendency towards homosexuality? How do they respond to this discovery? At what point do they
accept that they are gay or lesbian? How do they “come out” and acknowledge their sexual orientation
to others?
Even the above example is still very wide and could require a greater commitment in time and effort than
the average researcher is able to make. You could perhaps also try another technique for narrowing
68
things down, which is to first do a smaller, more open-ended pilot study and then to select one of the
issues that arise from this for your research project.
Once you have formulated a research proposal, you are ready to start collecting data. One of the first
things to know about data collection in the interpretive tradition is that the term “data”, although
sometimes used, is not very popular. The reason for this is that “data” suggests structured, objective bits
of information of the sort that could be processed by a computer. Interpretive research material, unlike
quantitative data, is usually non-numerical and relatively unstructured and in an “unprocessed” form.
Most commonly interpretive researchers get their material from observation (frequently “participant
observation”), interviewing and documents, and typically the research question will at least to some
extent determine where this material will be collected. For instance, a researcher who is interested in
how white liberals have adapted to the political transformation in South Africa may mix socially and
professionally with people who were at one time involved in liberal causes, interview them, and/or
collect published and unpublished documentary evidence from liberal individuals and organisations.
The central axiom of interpretive data collection is that it is not done from the perspective of the
"outsider looking in", but rather of the "insider looking around". You don't collect little bits of objective,
disconnected information to be stored away for later analysis, but become actively involved with a
phenomenon in order to better understand it.
Sampling
Although the exact extent of the material to be collected in interpretive research often cannot be
specified in advance, one should give some indication in one’s research proposal as to the type and
volume of the data one expects to collect given the time one has available for doing so. In positivist
psychological research it is usually assumed that individuals can be characterised in terms of particular
variables. From this assumption the kinds of sampling strategies needed to ensure generalisability to
larger numbers of individuals follow logically. Interpretive research, by contrast, is often sceptical of the
idea of discrete variables, and for this reason the kinds of sampling strategies used in positivist research
are inappropriate. A sampling strategy that is quite popular in interpretive work is snowball sampling.
This entails interviewing a few people who belong to the group you're interested in and then asking each
interviewee to introduce you to more people and so on until you feel you have a large enough sample.
Interpretive researchers may well be content with including very small numbers of individuals in their
research. It is quite possible to base a Masters or PhD dissertation on five or six, or fewer, “subjects”. A
rich description of how a single person perceives her social world may be worth more than a hundred
superficial accounts. Conversely, however, not all interpretive research need work with small sample
sizes. (As you may have guessed, “subjects” is another word that is not very popular among interpretive
researchers, who often prefer to use less alienating terms such as “research participants”.)
Additional sources
References
Creswell, J.W. (1994). Research Design: Qualitative and quantitative approaches. Sage.
HMPYC80 (Tutorial Letter 102) Resource 02: Research Review Inventory. Unisa.
HMPYC80/104 69
HMPYC80/104/0/2024
Constructionist research
Constructionism is one of at least three social science research <paradigms>. The other two paradigms,
<positivism> and <interpretive research>, both work by privileging what is outside language (either
objective facts or subjective experiences), over language itself - using language merely as a window onto
some other reality. Constructionist research, by contrast, holds that the human life-world is
fundamentally constituted in language and that language itself should therefore be the object of study.
Thus constructionism does not see language as neutral and transparent (as both positivism and
interpretive research may assume), or as a route to underlying realities; rather language helps to
<construct> reality. Another way of putting this is to say that constructionism treats language as “topic”
rather than “resource”.
Imagine that you are doing a research project on stress and that you collect data via interviews with 20
people diagnosed by a psychologist as suffering from stress. As a positivist researcher you may want to
use the interviews as sources of information on variables such as the kinds of events that precede and
follow stressful episodes and the demographic and other characteristics of people who suffer from stress
(you may also want to make your study more objective by using standardised questionnaires and a larger
sample size). As an interpretivist researcher you may want to use the interviews to help you
empathetically understand how people experience stress. As a constructionist researcher you may want
to describe how talk about “stress” functions to construct a social world in which failure to “cope” with
“life events”, and a lack of “social support”, leads to physical or mental “breakdown”. (This, as it
happens, is just what Young, 1980, and Pollock, 1988, attempted to do in their work on stress as a
modern mythology.)
Numerous other constructs used in psychology and the other social sciences have been subjected to
similar treatment. examples include emotion, attitudes, suicide, altruism, childhood and romantic love.
In showing up accepted notions such as these as myths, semiotics does not necessarily claim that they are
not real, but that they exist as social rather than natural objects. Barthes (1972) describes the situation as
follows:
Myth deprives the object of which it speaks of all History. In it, history evaporates. It
is a kind of ideal servant: it prepares all things, brings them, lays them out, the master
arrives, it silently disappears: all that is left for one to do is to enjoy this beautiful object
without wondering where it comes from. (p. 157)
An example of a “beautiful object” of this sort which has increasingly come under the scrutiny of
constructionist researchers is the whole discipline of social psychology. Stainton Rogers et.al (1995), for
instance, analyse social psychology "not in terms of the scientific truths or falsity of its individual
claims, but rather as a series of narratives that rely on certain rules and conventions, and exploit
identifiable rhetorical tactics, in order to produce its meaningfulness" (Collins, 1997). Squire (1990)
similarly suggests that social psychology can be understood as a series of narratives. Conventional social
psychology is like a detective story in that it tries to investigate and resolve problems empirically and
logically. It also doesn't bother much with inner meanings, but tries to understand individuals and groups
through observable behaviours. Another, much smaller, branch of social psychology can be compared to
autobiographical stories in that it tries to work with phenomenological or anthropological
understandings of people's lived experiences. Finally, there is an element of social psychological work
70
which can be likened to science fiction stories, particularly contributions (such as Squire's own work)
which try to step outside the normal scientific procedures to get a wider perspective on what the
discipline is about. Ironically, so Squire claims, mainstream social psychology (the detective story)
needs science fiction to keep going, but tries to deny that this is the case:
Science-fictional speculations generate new hypotheses, models and theories which sustain
social psychology as a discipline. The mainstream discourse's denial of such speculations
leaves processes like hypothesis formulation in the realm of mystery, or reduces them to
deduction. (pp. 44-45)
A word of caution: The idea of exposing social myths suggests that we can construct true accounts of the
social world, a notion which is strongly resisted by many working in the constructionist tradition. We
use the term constructionism quite broadly here to include rather diverse approaches - ranging from
Austin's speech act theory to the structuralist and post-structuralist work of figures such as Saussure,
Levi-Strauss, Barthes, Foucault, Baudrillard, Derrida and Lacan. Each of these theorists has (among
other things) unique ways of dealing with “truth”.
Below is a map of the world devised by the German historian Arno Peters. The Peters' projection is more
accurate than traditional maps in its depiction of the relative sizes of the continents, for example showing
Africa as being considerably larger than North America. However, since a two-dimensional
representation can never do justice to a three-dimensional object, the Peter's projection has to be less
accurate in other respects, for example in reflecting the relative distance between different points on the
earth. Few people would dispute that the earth exists as an objective reality, and that representations of it
can vary in their degree of accuracy. But it is equally important to realise that maps of the world are
signifying practices which construct it as a social object. The same is true, except possibly to a greater
extent, of representations of the world of individual and social psychology. (Those fascinated by
knowledge struggles outside the discipline of psychology may be interested to know that there are
numerous other “revisionist” world maps in addition to Peters' projection. One, which not surprisingly is
gaining popularity in Australia, shows an upside-down globe with Australia centred at the top.)
While some research reports working within the constructionist paradigm advertise themselves as such,
this is not usually the case. By far the most common label you'll see for constructionist work in
psychology is <discourse analysis>, an approach which has been gaining steadily in popularity for more
than a decade. Apart from discourse analysis, other keywords to look out for to help you determine if
research is constructionist in nature are constructivism, deconstruction, narrative and critical.
HMPYC80/104 71
HMPYC80/104/0/2024
Additional sources
References
Collins, A. (1997). Review of Social Psychology: A critical agenda by Stainton Rogers et.al. South
African Journal of Psychology, 27(2), 123-124.
Pollock, K. (1988). On the nature of social stress: Production of a modern mythology. Social Science
and Medicine, 26(3), 381-392.
Squire, C. (1990). Crisis what crisis? Discourses and narratives of the “social” in social psychology. In I.
Parker, & J. Shotter (Eds) Deconstructing Social Psychology. Routledge.
Stainton Rogers, R., Stenner, P. Gleeson, K., & Stainton Rogers, W. (1995). Social Psychology: A
critical agenda. Polity Press.
Young, A. (1980). The discourse of stress and the reproduction of conventional knowledge. Social
Science and Medicine, 14B, 133-146.
Triangulation
Textbooks very commonly advise that researchers should engage in “triangulation”, that is they should
collect their material in as many different ways and from as many diverse sources as possible. The idea is
that this will not only yield a richer and fuller understanding of the topic being researched, but (like a
surveyor finding the exact position of a particular feature of the landscape) help the researcher to “home
in” on a correct understanding by approaching it from several different angles. A researcher wishing to
understand people's views on South Africa's new abortion laws may for instance interview people
opposed to and in favour of abortion on demand as well as women who have had an abortion or decided
not to have one; collect newspaper articles and other documents; and visit clinics where abortions are
performed. However, triangulation is labour-intensive and often beyond the scope of an academic
research project, so be careful not to promise too much in a research proposal.
In addition to data triangulation (as described above), it is also possible to do “method triangulation”
(analyse the material using different methods) and “theory triangulation” (interpret the findings from
different theoretical perspectives).
72
Quantitative research problems and

questions
Researchers who are interested in discovering and confirming causal or correlational patterns and
findings often ask quantitative research questions involving a <relationship between variables> (see
also the discussion on <variables>). Kerlinger (1986) discusses the criteria that a research problem
should meet if one is interested in doing relational research and/or in discovering causes.
It is not always possible for a researcher to formulate his (sic) problem simply, clearly,
and completely. He may often have only a rather general, diffuse, even confused notion
of the problem. This is in the nature of the complexity of scientific research. It may even
take investigator many years of exploration, thought, and research before they can clearly
say what questions he has been seeking answers to. Nevertheless, an adequate statement
of the research problem is one of the most important parts of research. That it may be
difficult or impossible to state a research problem satisfactorily at a given time should not
allow us to lose sight of the ultimate desirability and necessity of doing so.
Bearing this difficulty in mind, a fundamental principle can be stated: If one wants to
solve a problem, one must generally know what the problem is. It can be said that a large
part of the solution lies in knowing what it is one is trying to do. Another part lies in
knowing what a problem is and especially what a scientific problem is.
What is a good problem statement? Although research problems differ greatly and there
is no one "right" way to state a problem, certain characteristics of problems and problem
statements can be learned and used to good advantage. To start, let us take two or three
examples of published research problems and study their characteristics. First, take the
problem of the study by Hurlock mentioned in Chapter 1: What are the effects on pupil
performance of different types of incentives?3 Note that the problem is stated in question
form. The simplest way is here the best way. Also note that the problem states a relation
between variables, in this case between the variables incentives and pupil performance
(achievement). (see "Variable". A variable is a name of a phenomenon or a construct, that
takes a set of different numerical values.)
A problem, then, is an interrogative sentence or statement that asks: What relation exists
between two or more variables? The answer is what is being sought in the research. A
problem in most cases will have two or more variables. In the Hurlock example, the
problem statement relates incentive to pupil performance. Another problem, studied in an
ingenious experiment by Glucksberg and King, is associated with an adage: We
remember what we want to remember, and with Freud's concept of repression: Are
memory items associated with unpleasant events more readily forgotten than neutral
items?4 One variable is items associated with unpleasantness, and the other variable is
remembering (or forgetting). Still another problem, by Jones and Cook, is quite different:
Do attitudes toward blacks influence judgments of the effectiveness of alternative racial
social policies?5 One variable is attitudes toward blacks and the other is judgments of the
effectiveness of social policies.6
There are three criteria for good problems and problem statements. One, the problem
should express a relation between two or more variables. It asks, in effect, questions like:
Is A related to B? How are A and B related to C? How is A related to B under conditions C
HMPYC80/104 73
HMPYC80/104/0/2024
and D? The exceptions to this dictum occur mostly in taxonomic or methodological

research.
Two, the problem should be stated clearly and unambiguously in question form. Instead
of saying, for instance, "The problem is ...," or "The purpose of this study is ...," ask a
question. Questions have the virtue of posing problems directly. The purpose of a study is
not necessarily the same as the problem of a study. The purpose of the Hurlock study, for
instance, was to throw light on the use of incentives in school situations. The problem
was the question about the relation between incentives and performance. Again, the
simplest way is the best way: ask a question.
The third criterion is often difficult to satisfy. It demands that the problem and the
problem statement should be such as to imply possibilities of empirical testing. A
problem that does not contain implications for testing its stated relation or relations is not
a scientific problem. This means not only that an actual relation is stated, but also that the
variables of the relation can somehow be measured. Many interesting and important
questions are not scientific questions simply because they are not amenable to testing.
Certain philosophic and theological questions, while perhaps important to the individuals
who consider them, cannot be tested empirically and are thus of no interest to the scientist
as a scientist. The epistemological question, "How do we know?," is such a question.
Education has many interesting but nonscientific questions, such as, "Does democratic
education improve the learning of youngsters?" "Are group processes good for children?"
These questions can be called metaphysical in the sense that they are, at least as stated,
beyond empirical testing possibilities. The key difficulties are that some of them are not
relations, and most of their constructs are very difficult or impossible to so define that
they can be measured7 (Kerlinger, 1986, pp.15-17).
Suppose that you are interested in how certain things are related or whether something that is happening
is caused by something else. Now see whether you can formulate a question that meets the three criteria
that Kerlinger regards as necessary for a good research problem.
................................................................................................................................................................
................................................................................................................................................................
Let us take these criteria one by one. Start with criterion 2. Have you formulated your problem in
question form? If not, try to do so now. You may even want to break it down into more than one
question.
................................................................................................................................................................
................................................................................................................................................................
Now that was not too difficult, was it? Let’s go on to criterion 1: Underline the variables indicated in
your research question, and mark them A, B, etcetera. Now look at the relationships between these
variables. Are you talking about correlations only, or does your research question refer to causal factors?
For example, are you suggesting that variable A is correlated with B, or are you implying that B is
caused by A? If your question is not clear about the relationships between A, B, etcetera, see whether
you can come up with a simplified formulation of the problem.
74
................................................................................................................................................................
................................................................................................................................................................
Are these relationships conditional? In other words, is A related to B, provided condition C holds? See
whether your problem statement implies specific conditions. If so, write them down.
................................................................................................................................................................
................................................................................................................................................................
Now we get to the most difficult criterion. To satisfy criterion 3 your research question should imply
possibilities for empirical testing. Consider each of the variables A, B, etcetera in your research
question. Can you measure the variable in question? If so, how will you measure the variable?
................................................................................................................................................................
................................................................................................................................................................
If you cannot determine how to measure a variable, can you define the variable so that it implies (points
in the direction of ) some form of measurement? Or can you reformulate your question to come up with
different variables.
................................................................................................................................................................
................................................................................................................................................................
Additional sources
Kerlinger, F.N. (1986). Foundations of behavioural research. Holt, Rinehart & Winston.
HMPYC80/104 75
HMPYC80/104/0/2024
Experimental research designs

(Quantitative: Relational‑causal)
The nature of research activities vary, ranging from passive observations to complex interventions.
Researchers may choose to merely observe what happens, or they may decide to intervene in situations
and then observe the effects of their interventions. The chain of events may be even more complex in the
sense that a researcher may execute a series of interventions and observations in an attempt to actively
change a situation. These are matters of research design. A study needs to be designed carefully to
ensure that the observations made are reliable and valid. Some projects are fairly easy to design, but
others may pose unique difficulties. For example, it is often tricky to spell out a clear design if your
research is aimed at generating a number of research questions (although this often may take the form of
an ex post facto design or the form of the one-shot case study design) or if your project requires an
<action research> approach (although this often takes the form of a series of interventions and
observations). In cases like these you may be able to provide only broad indications of the design of your
project. In action research, for example, you may be able only to plan only the first steps of your study.
A starting point in designing your project is to decide whether the research problem requires active
intervention or whether you are concerned with after the fact (ex post facto) observation. An ex post
facto observation is exactly what it proposes to be, namely an observation of the effects of an event or
situation that has not been planned or purposefully set up. Interviewing the survivors of an air crash
would be an example of an ex post facto approach. Observing woman or smokers are further examples of
ex post facto situations because gender and nicotine dependency are not purposefully introduced before
observations are made. These circumstances already exist before the study in question is planned and
executed. The ex post facto design can be represented as follows:
x   (1)
Here x is the event/occurrence/situation and  represents the observed group. The arrow indicates the
time sequence. If the event or occurrence was purposefully set up (that is, if it is an intervention) we have
a quasi-experimental situation, often called the one-shot case study. Although the name one-shot case
study is very popular one should perhaps stick with a name that ties in logically with the names of the
other designs (as we shall see below). Therefore this design could be called a one-group posttest-only
design. To distinguish this scenario from an ex post facto situation, the event (intervention) can be
indicated with a capital X. The symbolic representation of the quasi-experimental one-group
posttest-only design looks like this:
X   (2)
The one-group posttest-only design is quasi-experimental because one does not know for sure that the
observation is really due to the intervention. An obvious way to try to solve the problem is to observe a
group of individuals, treat them in a particular way and then observe them again to see whether any
changes have occurred. This is known as the quasi-experimental one-group pretest-posttest design.
76
This design can be pictured as follows:
  X   (3)
This design is still quasi-experimental because one does not know for sure that the differences between
the first and second observations are really due to treatment. In other words, the design lacks control. So,
a second way to solve the problem would be to add another group to act as a control group. This would
allow one to compare the observation that follows on the intervention with an observation that does not
follow on the intervention. One could then ascribe any difference between the two observations to the
effects of the intervention. In practice this means one requires two groups of subjects for the research
project. One group is subjected to the intervention, and the other group is not. This is called a
quasi-experimental posttest-only control group design. The quasi-experimental posttest-only control
group design can be represented as follows ( indicates the absence of the intervention, and X/x
indicates that this could be an intervention or an ex post facto event ):
X/x  
   (4)
Although the static-group comparison design does a better job than the previous designs (See (1) and
(2)) to ensure that what is observed is really due to the intervention in question, it still does not do away
with all possible threats to the internal validity of the design. Thus the design is still quasi-experimental.
The problem is that one cannot be sure that the two groups were equivalent in all respects before the
intervention occurred. One way to try to ensure that the two groups are equivalent is to select an initial
group and then assign each subject randomly to one of the two groups. Thus any subject has an equal
change to be in any one of the two groups. This approach ensures that possible differences between the
groups cancel out. Thus the group receiving the intervention is equivalent to the group not receiving the
intervention, and any differences between the groups (after the intervention) are due to the effects of the
intervention. This design is called an experimental posttest-only control group design ( R
represents the random assignment of subjects to groups):
R X  
   (5)
Some researchers feel more confident that groups are equivalent prior to the intervention when they can
observe that this is indeed the case. In doing so they prefer a design called the experimental
pretest-posttest control group design. This is what it looks like:
R   X  
     (6)
In the case of both designs (5) and (6) one compares the final observations between the treatment and
control groups to determine whether the intervention had any effect. Although some researchers may
have more confidence in design (6), in reality the random assignment of subjects to groups is sufficient
to ensure approximate equivalence between groups. Therefore, under most circumstances design (5)
should be sufficient. It is more economical (saving the time and the effort of the first observation) and it
eliminates the possible interaction between the first observation and the intervention. Interaction effects
may occur when the observation prior to the intervention alerts subjects to the nature of the intervention
and causes them to react differently to the intervention than they would have done if they had not been
subjected to observation prior to the intervention. Still, there may be special situations requiring
observations prior to interventions. The Solomon four-group design combines designs (5) and (6) to
counter possible interaction effects. Here is a picture of the Solomon four-group design:
HMPYC80/104 77
HMPYC80/104/0/2024
R   X  
    
X  
   (7)
Most forms of research design are simply variations on these themes, although the variations may
become quite complex. It is a good idea to try to keep the design of your project as simple as possible.
Complex designs often can be broken down into a number of simpler studies. Let your research problem
guide your actions.
Additional sources
78
Variables and measurement levels
A simple definition of variables would be that they are concepts or phenomena that assume different
values or numbers. Let us, however, find out more about variables and their measurement.
Scientists somewhat loosely call the constructs or properties they study "variables."
examples of important variables in sociology, psychology, and education are: sex,
income, education, social class, organizational productivity, occupational mobility, level
of aspiration, verbal aptitude, anxiety, religious affiliation, political preference, political
development (of nations), task orientation, anti-Semitism, conformity, recall memory,
recognition memory, achievement. It can be said that a variable is a property that takes
on different values. Putting it redundantly, a variable is something that varies. While this
way of speaking gives us an intuitive notion of what variables are, we need a more
general and yet more precise definition.
A variable is a symbol to which numerals or values are assigned. For instance, x is a

variable: it is a symbol to which we assign numerical values. The variable x may take on
any justifiable set of values -- for example, scores on an intelligence test or an attitude
scale. In the case of intelligence we assign to x a set of numerical values yielded by the
procedure designated in a specified test of intelligence. This set of values ranges from
low to high, from, say, 50 to 150.
A variable, x, however, may have only two values. If sex is the construct under study, then
x can be assigned 1 and 0, 1 standing for one of the sexes and 0 standing for the other. It
is still a variable. Other examples of two-valued variables are: alive-dead,
citizen-noncitizen, middle class-working class, teacher-nonteacher,
Republican-Democrat, and so on. Such variables are called dichotomies or dichotomous
variables.
Some of the variables used in behavioral research are true dichotomies -- that is, they are
characterized by the presence of absence of a property; male-female, alive-dead,
employed-unemployed. Some variables are polytomies. A good example is religious
preference: Protestant, Catholic, Jew, Other.1 Most variables, however, are theoretically
capable of taking on continuous values. It has been common practice in behavioral
research to convert continuous variables to dichotomies or polytomies. For example,
intelligence, a continuous variable, has been broken down into high and low intelligence,
or into high, medium, and low intelligence. Variables such as anxiety, introversion, and
authoritarianism have been treated similarly. While it is not possible to convert a truly
dichotomous variable such as sex to a continuous variable, it is always possible to
convert a continuous variable to a dichotomy or a polytomy. As we will see later, such
conversion can serve a useful conceptual purpose, but is poor practice in the analysis of
data because it throws information away (Kerlinger, 1986, pp. 27-28).
We see that we can distinguish between continuous and categorical (dichotomous or polytomous)
variables. Depending on the type of measurement, a further distinction is however possible. Firstly we
have categorical or nominal measures. An example is gender - all the research subjects belong to one of
two groups depending on whether they are male or female. The categories are mutually exclusive (each
observation/person cannot fall into more than one category) and exhaustive (there are categories for all
HMPYC80/104 79
HMPYC80/104/0/2024
the observations). We can assign values to males and females, for example, 0 for males and 1 for
females, or perhaps 1 for males and 2 for females, but these numbers do not function as numerals; they
are merely labels: they label the different categories.
The next level of measurement in terms of complexity is the ordinal level of measurement. The rating of
movies in a newspaper is for example measured on this level. It is similar to the categorical level but the
categories are now rank ordered, implying that movies in each category have more or less of some
quality than those in the other categories. The next level is the interval scale. In this case the numbers not
only reflect more or less of a specific quality, but they also indicate the differences in amounts of that
quality. Intelligence test scores are an example of this type of measurement. Ratio scale measures have
an absolute zero value and a person’s mass is for example measured on a ratio scale. Note that most of
the data obtained in empirical psychological research are of a categorical or interval kind.
The following explanation may help you to grasp the intricacies of measurement scale levels:
(1) Here is a row of five dots:
⚫ ⚫ ⚫ ⚫ ⚫
We can name (i.e. label) these dots as follows:
⚫ ⚫ ⚫ ⚫ ⚫
A B C D E
Naming the dots results in a nominal scale. It allows us to categorise objects. Any dot that is similar to dot
C belongs in category C, but this does not mean that category C is more (or less) than, for example, A.
(2) The next level of measurement would be to put these dots in sequence. If A, B, C, D and E can be rank
ordered, one can say that B is more than A, that C is more than B, etcetera.
⚫ ⚫ ⚫ ⚫ ⚫
A B C D E
(3) Next we can look at the differences between dots:
⚫ · ⚫ · · · ⚫
A B C
Now I have an interval scale, because I can count the intervals between dots. I can see that the difference
between B and C (4 intervals) is twice that between A and B (2 intervals). Does this mean C = B x 2, and
B = A x 2?
(4) No, not necessarily, as you can see from the following three examples.
⚫ · ⚫ · · · ⚫
80
1 2 3 4 5 6 7
A B C
⚫ · ⚫ · · · ⚫
2 3 4 5 6 7 8
A B C
⚫ · ⚫ · · · ⚫
10 11 12 13 14 15 16
A B C
How do we manage to “see” that C is not necessarily equal to B times 2? We look at the numbers! In
other words we look at some form of universal scale, and what we see is that our three dots can find
themselves in different locations on this scale. We can locate these locations if we know A’s position
relative to an absolute zero point. If we do not have the zero point reference, the values that A takes on
are arbitrary. Here is an example: Let’s assume that A indicates ground level, which means A is zero
metres above ground level. Then B indicates two metres, and C six metres above ground level. Of course
we can say that C is three times as high as B (6 = 2 x 3), but this is a localised truth, a truth based on a
specific location. A’s absolute value depends on the height above sea level. Unless we know this
absolute value we cannot determine which is the higher of two C’s (which of two six metres above
ground level is really the higher one). In the above examples we have three A’s, worth 1, 2 and 10. If
these values signify the heights above sea level, we can say that a C worth 16 is really twice as high as a
C worth 8, although both are 6 metres above ground level. But we can only do so if we have a universal
point of reference. If we have such a point, the scale we work on is called a ratio scale. can you see why
a person’s mass can be measured on a ratio scale, but not his or her IQ? If Mary scores 160 on an IQ test,
and John manages 80, we can say that Mary did twice as well as John on the particular scale. But this is
a localised truth. It is only true for the particular IQ test. If we want to pronounce Mary twice as
intelligent as John we need some form of “sea level” IQ.
Additional sources
HMPYC80/104 81
HMPYC80/104/0/2024

Online statistical resources and software
STATSOFT : ELECTRONIC STATISTICS TEXTBOOK
The Electronic Statistical Textbook by Statsoft offers training in the understanding and application of
statistics. The material was developed at the StatSoft R&D department based on many years of teaching
undergraduate and graduate statistics courses and covers a wide variety of applications, including
laboratory research (biomedical, agricultural, etc.), business statistics and forecasting, social science
statistics and survey research, data mining, engineering and quality control applications, and many
others. The Electronic Textbook begins with an overview of the relevant elementary (pivotal) concepts
and continues with a more in depth exploration of specific areas of statistics, organized by "modules,"
accessible by buttons, representing classes of analytic techniques. A glossary of statistical terms and a
list of references for further study are included." https://www.uaq.mx/statsoft/aolglfra.html
ONLINESTATS : INTRODUCTION TEXTBOOK

Another good introductory handbook and reference manual for statistics is available on the internet at
http://onlinestatbook.com/2. This is an open source book so you may download it free of charge. There
are different formats available, i.e. pdf, eBook (e-pub) and an interactive version for iPads or Apple
iMac computers.
TEACH YOURSELF STATISTICS (STARTREK)

There is a course called “Teach yourself statistics” at http://stattrek.com/. There are also some
introductory tutorials with Youtube videos that can be viewed from the following link:
https://www.statslectures.com/index.php/general-statistics
JASP - ONLINE FREE STATISTICS PROGRAM

JASP is an open-source project supported by the University of Amsterdam. The “Frequentist analyses”
is appropriate for most research done at psychology honours level. View the getting started page and
video tutorial https://jasp-stats.org/getting-started/ “Before you do anything else, we strongly
recommend that you visit the JASP YouTube channel and the ‘How to Use JASP’ page, where you can
find introductory blog posts and videos. Below is some general information about JASP that may be
helpful as you analyze your first data set. We invite you to post any remaining questions on the JASP
forum or on the JASP GitHub page.” The JASP statistical software is highly recommended.
• See “How to use JASP” https://jasp-stats.org/how-to-use-jasp/
• JASP can compute various transformations (see “Functions & Modules”, and the “How to…” on
https://jasp-stats.org/how-to-use-jasp/ ). Many of the important statistical computations can be
computed with JASP, including parametric and nonparametric statistical tests.
SPSS
Note that Unisa uses the SPSS software and can provide annual licenses to postgraduate students. The
name referes to Statistical Package for the Social Sciences (SPSS) but it was later changed to Statistical
Product and Service Solutions. See https://www.ibm.com/products/spss-statistics
R
A more recent development in the past decade is the R software that is available for free and is very
popular amongst researcher and academics. The R programming language can be used for statistics and
data analysis. See https://www.r-project.org/
82
G. QUANTITATIVE RESEARCH DESIGNS & STATISTICS
Relationship between variables
Graphic portrayal of the relationship between the variables

When doing relational or causal research, we formulate a <hypothesis> about the relationship between
variables. Firstly, we consider the research problem and identify the variables that we are working with.
We then ask ourselves how these variables are related. We formulate a statement of the following kind:
There is a relationship (or a <correlation>) between variable X and variable Y.
Let’s work through an example: A lecturer teaching psychopathology at the University of Cape Town
notices an interesting phenomenon: His students normally enjoy his classes but pay less attention on
those days as they write a test in one of their other subjects. He formulates the following hypothesis:
Anxiety will affect students’ concentration on the day that they write a test. We can identify three
variables in this hypothesis, namely level of anxiety, degree of concentration and writing or not writing a
test. What is the lecturer’s actual interest? He would like his students to pay attention. Therefore “degree
of concentration” is an important variable. We call this the dependent variable and the lecturer wants to
know what is causing the lack of concentration (what the degree of concentration depends on) so that he
can try to do something about it. “Level of anxiety” is therefore also an important variable. We call this
the cause or the independent variable. He could formulate the following hypothesis: There is a
relationship between the level of anxiety (variable X) and the degree of concentration (variable Y).
Formulate your own research question, select the two variables that you think are the most important
ones and formulate a hypothesis that expresses a relationship between these two variables.
................................................................................................................................................................
................................................................................................................................................................
Can you determine the direction of this relationship? In other words, what do you expect will happen
when variable X increases? Will variable Y increase, or will it decrease, or is it impossible to tell? If you
are able to indicate the direction of the relationship you can state a directional hypothesis, if you cannot
do so, your hypothesis is nondirectional. Write down what you think will happen to your variable Y if
variable X increases.
................................................................................................................................................................
................................................................................................................................................................
If Y increases when X increases there is a positive correlation between these two variables (a
directional situation); if Y decreases when X increases, the correlation is negative (still a directional
situation). If you are unable to tell whether Y will increase or decrease when X increases, the correlation
can be either positive or negative (a nondirectional situation). The following graphs depict positive and
negative correlations:
HMPYC80/104 83
HMPYC80/104/0/2024
Let us see how we plot the measurements of two variables on a graph to depict the relationship between
these variables. Below is the outline of a graph. Following convention, we indicate the Y variable on the
vertical axis and the X variable on the horizontal axis. The Y axis takes the dependent variable and the
X axis the independent variable. The terms dependent variable and independent variable are probably
not that unfamiliar to you. Nevertheless, it is not always easy to see which is which, because the
definition depends very much on your experimental setup. A rule of thumb is simply to identify the
independent variable as the primary variable, the one that precedes the other in the sense that it is given.
The dependent variable is the one you are more curious about. Its values are not given, they need to be
determined following the values of the independent variable. However, it is not always easy to
distinguish between the two variables. Take, for example, the variables “attitude” and “motivation”.
Does a positive attitude follow good motivation, or is good motivation a function of a positive attitude?
Surely it depends on one’s point of view, and one’s point of view depends on the particular theoretical
position. We write the name of the X variable beneath the X axis, and the name of the Y variable along or
next to the Y axis. Remember to leave sufficient space between these names and the axes themselves,
because we still need to fill in some numbers along the axes. Here is an example using motivation and
performance. Let’s assume that performance in a research methodology examination is a function of
motivation. In other words, the more motivated the student, the better his/her performance. (It can, of
course, be the other way round as well: Good performance can lead to increased levels of motivation!)
If performance is a function of motivation, then motivation (the independent variable) goes on the X
axis, and performance (the dependent variable) on the Y axis. The short way to indicate performance as
a function of motivation is to write: performance = f(motivation). This corresponds to the equation Y =
f(X). Here is the graph:
84
Now for the numbers that go along the axes. Let’s consider X first. Determine the smallest possible
number as well as the largest possible number that an individual can score on this variable. Write the
smallest number where the Y axis crosses the X axis, and put the largest number way to the right of the X
axis. Now divide the space in between into equal intervals. For example, if the smallest number is 60 and
the largest number 110, put the 60 where the Y axis crosses the X axis, and write the 110 way to the right
on the X axis. Divide this range into equal intervals of say 10 each, which gives you the points 60, 70, 80,
90, 100, and 110 on the X axis.
<Variables> can be measured on different levels. If the X variable is a categorical variable, simply
indicate the categories on the X axis. Here is an example of an X axis reflecting the categories poor
motivation, average motivation and high motivation.
The Y axis is treated in exactly the same way. Here is an example where research methodology
performance scores range from 0 to 100%.
HMPYC80/104 85
HMPYC80/104/0/2024
We have now indicated the possible values (or categories) that an individual can obtain on the X and Y
variables on the two axes. The next step is to obtain measurements on these variables for each individual
in the sample and to plot the values on the graph. Here is an example where 10 individuals obtained the
following scores:
Individual Motivation level (X variable) Performance (Y variable)

1 high 80
2 high 70
3 poor 20
4 average 50
5 average 60
6 poor 30
7 high 90
8 average 40
9 poor 10
10 average 70
This is what the graph looks like
Not everybody with high motivation gets the same score, and likewise for average and low motivation.
The points are scattered, and this is what is known as a scattergram. If we want to draw a single line
indicating the relationship between motivation and performance we have to find a kind of middle road.
There are sophisticated ways to calculate this path, but a simple way is to take the mean scores for each
category. The mean for the poor category is: (20+30+10)/3 = 60/3 = 20; the mean for the average
category is: (50+60+40+70)/4 = 220/4 = 55; and the mean for high motivation is: (80+70+90)/3 = 240/3
= 80. Now the graph looks like this
86
Not quite a straight line, but close enough!
If your X variable is not a categorical variable, you may wish to change it into a categorical variable to
plot a “middle road” graph. Remember, interval or ratio data can always be changed into categorical
(dichotomous or polytomous) variables! The easiest way is to divide the variable into two categories,
namely low and high. For example lets assume we measured motivation on a scale from 6 to 17, and that
the individuals in our example scored as follows:
Individual Motivation (X variable) Performance (Y variable)

1 13 80
2 15 70
3 7 20
4 11 50
5 12 60
6 6 30
7 17 90
8 10 40
9 9 10
10 14 70
HMPYC80/104 87
HMPYC80/104/0/2024
Here is the scattergram
How do we make motivation into a categorical scale? Well, the range is from 6 to 17, which gives 6
points from 6 to 11, and 6 points from 12 to 17. So lets group 6 to 11 as the below average category, and
12 to 17 as the above average category. Now find the means of these two groups. Here is the mean
performance score of the below average motivation group: (20+50+30+40+10)/5 = 150/5 = 30; and here
is the mean performance score of the above average motivation group: (80+70+60+90+70)/5 = 74. Plot
the performance means for the below and above average motivation groups, and connect the points to
get the correlation line.
88
Expressing the relationship between the variables in terms of a linear model
What we did in the previous section was to draw a picture of the relationship between the X and Y
scores. We started off by plotting a point for each X and Y pair of scores. In this way we plotted a
scattergram. For real data (like the data you will get from your observations) such scattergrams can look
pretty scattered. You find points all over the graph, and it is often difficult to make sense of what goes
on. We have already discussed one way of trying to make sense of the situation, by dividing the X values
into two (or more) groups and plotting the points that correspond to the group means. By connecting
these points we got a graph expressing the relationship between the X and Y values. There are much
more complicated ways to plot a line “through” a scattergram. Such lines need not necessarily be straight
lines. They can take a number of different shapes. We will concentrate on a straight-line model. Such a
model is known as a linear model. Models that do not rely on a straight line are known as non-linear
models.
Your data, most probably, will not form a straight line. This does not mean that we cannot apply the
linear model. When we apply the linear model, we assume that the true relationship between the
variables in question takes the form of a straight line. That our data do not reflect a straight line we put
down to error that has crept in at different points in our observations. Thus we assume that should we
look beyond these errors we will see a straight line. So, the question is not whether our data conform to a
straight line. The question is whether the straight line from our data differs from another straight line,
which is the straight line indicating no relationship between the X and Y values. What does the line of no
relationship look like? To find out, we must consider the line graph in more detail. We start with a basic
question: How do we distinguish one straight line from another? Here are three straight lines. How do
they differ?
................................................................................................................................................................
................................................................................................................................................................
................................................................................................................................................................
................................................................................................................................................................
Did you see that lines 1 and 2 have the same slope but they cut the Y axis in different places? Lines 2 and
3 cut the Y axis at the same point but they have different slopes. When we describe a line, we must
therefore indicate at what point it cuts the Y axis as well as what its slope is. For short, let’s use β (beta)
HMPYC80/104 89
HMPYC80/104/0/2024
for the slope and α (alpha) for the point where the line cuts the Y axis (the Y value of the line when X =
0). Now we can say that the two graphs above differ with regard to their α and β values.
Here is the next question: Can we assign values to α and β. Yes we can, and this is how we do it:
Consider the notion of “slope” in everyday terms. The slope of a pathway has to do with how far one
goes forward in relation to how much one climbs. A gentle slope means a lot of forward progress for
little gain in height, whereas a steep slope means a large gain in height for every few forward steps. In
this case X represents our forward steps, while Y measures the gain in height. Here is an example where
2 steps forward give us 1 metre in height (4 - 2 = 2 steps; 1.5 - 0.5 = 1 metre).
So how do we calculate β? β is the ratio of Y over X; 1 metre gained for every 2 steps forward. This gives
1 over 2, which is 0.5. In other words, β = ½ = 0.5.
Of course slopes can also mean going downhill. Here is another example. This time β is -0.5. Can you
work it out?
90
I bet the negative sign causes a problem! You can see the slope is downhill (negative), but you do not
know how to determine the negative sign? Here is how it works: Remember your points on your graph
are actually (X; Y) pairs. For example, the pairs of numbers you deal with here are (10; 6) and (12; 5). To
calculate the differences you subtract the one from the other: (12; 5) - (10; 6) = (2; -1). So the ratio
becomes: -1/2 = -0.5.
If we want to be really fancy we can depict β as y/x. Note that we use x and y to indicate difference
scores: x = x1 - x2, for example 4 - 2 = 2 steps, and y = y1 - y2, for example 1.5 – 0.5 = 1 metre.
Now for α: If X = 0 at the origin of the graph and the line is extended to cut the Y axis, α is simply the
value of Y where the line cuts the Y axis
HMPYC80/104 91
HMPYC80/104/0/2024
To summarise, we can state that our straight-line graphs differ in two respects, namely their α and β
values.
This was our first point. Now for the second point: Remember the expression “Y is a function of X”?
Remember its short cut version: Y = f(X)? Good! Let’s go on to the third point: We want to describe
“function” - in other words, we want to give the “f” in f(X) a bit of body! We know by now that the
function involves two facets, namely a slope (β) and a crossing point (α). So instead of Y = f(X) we can
write Y = α + βX. Don’t worry if you cannot see why the formula appears in this format. As long as you
understand its meaning. It says that one can find a Y score provided one has α and βX. What does it
mean? It means we can find the Y score that corresponds to a particular X score if we know firstly how
the Y scores and the X scores are related, in other words if we know β, and secondly where we are
relative to the Y scale, that is if we know α. Thus the term βX connects X to Y and α anchors this
connection in a specific position on the Y axis.
Here is how the formula works: Let’s say we know that β = 2 and that α = 1. We can now connect any
particular X value with a specific Y value. Let’s choose X to be equal to 10.
Then Y = α + βX
Y = 1 + (2)(10)
Y = 1 + 20
Y = 21
Which means the (X; Y) number pair is (10; 21).
Alternatively, if we have at least two (X; Y) number pairs, we can calculate the α and β values, which
enables us to connect any X value with a corresponding Y value. Here is an example: Given two number
pairs (x1; y1) = (10; 8) and (x2; y2) = (12; 2), we have
β = y/x
β = (y1 - y2)/(x1 - x2)
β = (8 - 2)/(10 - 12)
β = (6)/(-2)
β = -3
92
To find α we begin with Y = α + βX, and juggle the symbols around to give us α = Y - βX. You can now
use any of the two number pair values to determine α. Let’s use both!
α = Y - βX and α = Y - βX
α = 8 - (-3)(10) and α = 2 - (-3)(12)
α = 8 - (-30) and α = 2 - (-36)
α = 8 + 30 and α = 2 + 36
α = 38 and α = 38
Here is what the graph looks like
Remember, we were curious about this graph because we were interested in comparing our own graphs
with a graph that depicts no relationship between X and Y, and we wanted to do so to find out whether or
not our results showed a significant relationship. The key to drawing the line of no relationship lies in the
following statement: β expresses the relationship between X and Y. If there is no relationship between X
and Y, β assumes the value zero. This means there is no slope in our straight-line graph. The slope is
zero. This is what it looks like
Now consider the straight- or linear-line formula: Y = α + βX

If β = 0, we get: Y = α + (0)X
which gives: Y=α
HMPYC80/104 93
HMPYC80/104/0/2024
On the next graph we see what happens when we extend a line with a zero slope.
So what do we see here? The X values are not connected to any Y values. It does not matter what X value
we put into the equation, we always come up with Y = α. It is as if the Y scale does not respond to any X
scale “requests”.
94
Hypotheses
If the aim of the research is to provide proof about a <relationship between variables> (see also
<variables>), a more formal hypothesis is necessary. Kerlinger argues that in this type of research, we do
not test our research question but the hypotheses implied by our question. But first, how do such formal
hypotheses differ from research questions?
There are important differences between problems and hypotheses. Hypotheses, if

properly stated, can be tested. While a given hypothesis may be too broad to be tested
directly, if it is a "good" hypothesis, then other testable hypotheses can be deduced from
it. Facts or variables are not tested as such. The relations stated by the hypotheses are
tested. And a problem cannot be scientifically solved unless it is reduced to hypothesis
form, because a problem is a question, usually of a broad nature, and is not directly
testable. One does not test the questions: Does efficiency enhance organizational
effectiveness and profitability?15 Does differential experience modify the brain?16 One
tests one or more hypotheses implied by these questions. For example, to study the latter
problem, one may hypothesize that animals with different levels of experience will have
different thicknesses of the brain cortex.
Hypotheses direct inquiry. As Darwin pointed out long ago, observations have to be for
or against some view if they are to be of any use. Hypotheses incorporate aspects of the
theory under test in testable or near-testable form. Earlier, an example of reinforcement
theory was given in which testable hypotheses were deduced from the general problem.
The importance of recognizing this function of hypotheses may be shown by going
through the back door and using a theory that is very difficult, or perhaps impossible, to
test. Freud's theory of anxiety includes the construct of repression. Now, by repression
Freud meant the forcing of unacceptable ideas deep into the unconscious. In order to test
the Freudian theory of anxiety it is necessary to deduce relations suggested by the theory.
These deductions will, of course, have to include the repression notion, which includes
the construct of the unconscious. Hypotheses can be formulated using these constructs;
in order to test the theory, they have to be so formulated. But testing them is another,
more difficult matter because of the extreme difficulty of so defining terms such as
"repression" and "unconscious" that they can be measured. To the present, no one has
succeeded in defining these two constructs without seriously departing from the original
Freudian meaning and usage. Hypotheses, then, are important bridges between theory
and empirical inquiry (Kerlinger, 1986, pp. 19-20).
What is the nature of hypotheses? A good hypothesis meets certain criteria, and thus enables the
researcher to make observations, to analyse the results statistically, and then to answer a question about
the relationship between variables.
A hypothesis is a conjectural statement of the relation between two or more variables.

Hypotheses are always in declarative sentence form, and they relate, either generally or
specifically, variables to variables. There are two criteria for "good" hypotheses and
hypothesis statements. They are the same as two of those for problems and problem
statements. One, hypotheses are statements about the relations between variables. Two,
hypotheses carry clear implications for testing the stated relations. These criteria mean,
HMPYC80/104 95
HMPYC80/104/0/2024
then, that hypothesis statements contain two or more variables that are measurable or
potentially measurable and that they specify how the variables are related.
Let us take three hypotheses from the literature and apply the criteria to them. The
first hypothesis seems to defy common sense; Overlearning leads to performance
decrement (or, as the authors say: Practice makes imperfect!). 8 Here a relation is stated
between one variable, overlearning, and another variable, performance decrement.
Since the two variables are readily defined and measured, implications for testing the
hypothesis, too, are readily conceived. The criteria are satisfied. A second hypothesis is
related to the first (though formulated many years earlier). It is also unusual because it
states a relation in the so-called null form: Practice in a mental function has no effect on
the future learning of that function.9 The relation is stated clearly: one variable, practice
in a mental function (like memory), is related to another variable by the words "has no
effect on." On the criterion of potential testability, however, we meet with difficulty. We
are faced with the problem of so defining "mental function" and "future learning" that
they are measurable. If we can solve this problem satisfactorily, then we definitely have a
hypothesis. Indeed, we have a famous one -- but one that has usually not been stated as a
hypothesis but as a fact by many educators of the past and present.
The third hypothesis represents a numerous and important class. Here the relation is
indirect, concealed, as it were. It customarily comes in the form of a statement that
Groups A and B will differ on some characteristic. For example: Middle-class children
more often than lower-class children will avoid finger-painting tasks.10 Note that this
statement is one step removed from the actual hypothesis, which may be stated:
Finger-painting behavior is in part a function of social class. If the latter statement were
the hypothesis stated, then the first might be called a sub-hypothesis, or a specific
prediction based on the original hypothesis.
Let us consider another hypothesis of this type but removed one step further:
Individuals having the same or similar occupational role will hold similar attitudes
toward cognitive objects significantly related to the occupational role.11 ("Cognitive
objects" are any concrete or abstract things perceived and "known" by individuals.
People, groups, the government, and education are examples.) The relation in this case,
of course, is between occupational role and attitudes (toward a cognitive object related
to the role, for example, role of educator and attitudes toward education). In order to test
this hypothesis, it would be necessary to have at least two groups, each representing a
different occupational role, and then to compare the attitudes of the groups. For
instance, we might take a group of teachers and compare their attitudes toward
education to those of, say, a group of businessmen. In any case, the criteria are satisfied
(Kerlinger, 1986, pp. 17-18).
Let us practice the application of these ideas on an example. Here is a research problem: Do universities
discriminate against woman applicants? Try to formulate at least one hypothesis based on this research
question.
................................................................................................................................................................
................................................................................................................................................................
Subject the hypothesis you have formulated to the two criteria a good hypothesis should adhere to. First:
Is your hypothesis a statement about the relationship between variables? Write down the variables
implicated in your hypothesis above as well as the words that express the relationships between these
variables.
96
................................................................................................................................................................
................................................................................................................................................................
Here are two possible research hypotheses: Gender is related to acceptance into universities, or more
males than females will be accepted when applying for admission at universities. Write down the
variables implied in these hypotheses and indicate the relationship(s) between these variables.
................................................................................................................................................................
................................................................................................................................................................
Did you write down gender and application outcome? If not reconsider your answer. Can you see that
gender and application outcome are the two variables in question? Why are these called variables?
Obviously because they vary, they take on different values. Gender has two possible values, namely
male and female, and so does application outcome in the form of acceptance and rejection.
How do the abovementioned two hypotheses differ? The second one says more than the first one; it says
that more males than females are accepted, whereas the first simply suggests that gender is (somehow)
related to university application outcome. The second hypothesis is directional (it gives direction), and
the first hypothesis is nondirectional. Also note that this is a correlational and not a causal description.
The hypothesis does not state that being male causes acceptance; it simply suggests a correlation
between gender and application outcome.
Also note that these hypotheses enable us to test the statement about the relationship between the
variables. The variables are operationalised in such a manner that they can be measured. For example,
the concept “discrimination” mentioned in the research problem is expressed as application outcome,
which renders it measurable in terms of the numbers of applications accepted and rejected. Here is a
picture of the second hypothesis. This picture tells us that 500 males and 300 females were accepted.
A design could also involve more than one independent variable and/or dependent variable. Read the
following in this regard:
Until now the discussion of problems and hypotheses has been pretty much limited to two
variables, x and y. We must hasten to correct any impression that such problems and hypotheses
are the norm in behavioral research. Researchers in psychology, sociology, education, and other
behavioral sciences have become keenly aware of the multivariate nature of behavioral
research. Instead of saying: If p, then q, it is often more appropriate to say: If p1, p2 . . ., pk then q;
or: If p1, then q, under conditions r, s, and t.
HMPYC80/104 97
HMPYC80/104/0/2024
An example may clarify the point. Instead of simply stating the hypothesis: If frustration, then
aggression, it is more realistic to recognize the multivariate nature of the determinants and
influences of aggression by saying, for example: If high intelligence, middle class, male, and
frustrated, then aggression. Or: If frustration, then aggression, under the conditions of high
intelligence, middle class, and male. Instead of one x, we now have four x's. Although one
phenomenon may be the most important in determining or influencing another phenomenon, it is
unlikely that most of the phenomena of interest to behavioral scientists are determined simply. It
is much more likely that they are determined to multiply. It is much more likely that aggression is
the result of several influences working in complex ways. Moreover, aggression itself has
multiple aspects. After all, there are different kinds of aggression.
Problems and hypotheses thus have to reflect the multivariate complexity of psychological,
sociological, and educational reality. Although we will talk of one x and one y, especially in the
early part of the book, it must be understood that contemporary behavioral research, which used
to be almost exclusively univariate in its approach, is rapidly becoming multivariate. (For now,
"univariate" means one x and one y. "Univariate," strictly speaking, applies to y. If there is more
than one x or more than one y, the word "multivariate" is used, at least in this book.)
(Kerlinger, 1986, pp. 22-23.)
Reference
Kerlinger, F.N. (1986). Foundations of behavioural research. London: Holt, Rinehart & Winston.
98
Hypothesis testing
In relational research a <hypothesis> is a statement about the relationship between variables that enables
us to test this statement. In other words, the psychological construct is operationalised in such a manner
that the construct can be measured and the results used to determine the significance of the relationship
between the variables. We obtain measurements for a group of subjects that is a sample of a larger
population. We want to know if the <relationship between the variables> indicated by our data holds for
this population. This is the question that drives hypothesis testing.
Any relationship between X and Y manifests itself in a sloping line, a line that deviates from the
zero-slope line (that indicates no relationship between the X and Y values). This can be a positive or a
negative slope, in other words the line can slant forwards or backwards. There are a number of different
methods we can apply to determine whether or not there is a relationship between variables X and Y, but
in the end, all these methods boil down to testing the significance of the slope of the straight-line graph,
that is the significance of β. What is more, all these tests work in the same way. They look at the slope of
the graph depicting the obtained data, and then ask the question: If for the larger population β is in fact
equal to zero, what is the probability of obtaining the particular (non-zero) β value for a sample of this
population? The larger this probability the more likely it is that the obtained non-zero β value is in fact
equal to zero. The smaller this probability the more unlikely the chances that the obtained (non-zero) β
value is actually equal to zero.
When one wants to know whether the β value obtained via our observations is significant, in other words
whether the obtained β value really deviates from a zero β, one begins with a null hypothesis suggesting
that β is equal to zero.
This is what it looks like in formal language: H0: β = 0
One then considers the probability that one’s obtained (non-zero) β value may in fact come from a
population of β values with a mean of zero. If this probability is fairly large, one must accept the null
hypothesis. But if the probability is small, one can reject the null hypothesis. However, if one rejects the
null hypothesis one must have another hypothesis, otherwise one would not know what to say! The
simplest alternative to the null hypothesis is to suggest that β is not equal to zero.
This is what it looks like in formal language: H1: β  0
This is a nondirectional hypothesis. It suggests β differs from zero, but it does not indicate in which
direction β differs, in other words whether it is more or less than zero. Depending on the nature of the
information one has at one’s disposal, one may be in a position to suggest more specific alternatives.
One may, for example, suggest that β is greater than zero.
This is what it looks like in formal language: H1: β > 0
This is a directional hypothesis. It indicates a direction of difference from zero, namely greater than zero.
Alternatively, one may have sufficient information to suspect that β is less than zero.
This is what it looks like in formal language: H1: β < 0
HMPYC80/104 99
HMPYC80/104/0/2024
This is another example of a directional hypothesis, indicating a direction of difference from zero - in
this case, smaller than zero.
To summarise: In conducting one’s research one obtains data. One then uses the data to calculate a β
value. (Remember there are a number of ways in which β can be determined - the method one uses is a
function of the design of one’s research project.) One then hypothesises that the population β value is in
fact equal to zero, in other words that the non-zero β value obtained for your data is the result of studying
a limited sample of instances of a variable or phenomenon, assessed by a fallible measuring instrument.
One uses a computer to calculate the probability of obtaining the non-zero β value in question should the
population value in fact be equal to zero. If this probability is fairly large we agree that the obtained β
value is an insignificant deviation from zero. But if the probability associated with our obtained β value
is fairly small we suspect that the population β value in reality is not equal to zero. In other words the
question we ask is: What is the probability of obtaining a β value that is equal to or greater than say 0.3
when the particular β value in the population in fact equals zero? Or phrased in a slightly more formal
way: What is the probability of obtaining β  0.3 given that H0 is true? If you are really fancy you write
this question in the following format:
P(obtained β  0.3  population β = 0) = ?
Note that probability values range from 0 (not probable at all) to 1 (absolutely probable).
When do we decide to reject the null hypothesis? In other words, when is the probability associated with
a particular non-zero β value so small that we decide that the population β probably does not equal zero?
There are no hard and fast rules, but we have come to accept the 0.05 cut-off point. Thus when the
probability of a β value drops below 0.05 we agree that the obtained β value is not a chance deviation
from zero and that β, therefore, does not equal zero. Please make sure that you understand when a
number is smaller than 0.05. A value of 0.10 is bigger than 0.05 but 0.01 is smaller than 0.05 and
so is 0.019.
There is another aspect, however, that influences the probability value, namely the alternative
hypothesis. Most computer programmes provide the probability for a nondirectional alternative
hypothesis. If the alternative hypothesis is directional, we must divide the probability value by 2. Thus, if
the computer determines a probability value of 0.34 but your problem requires a directional alternative
hypothesis, you must divide 0.34 by 2, giving a probability value of 0.17. The reasoning behind this is
straightforward as you can see from the following graphs.
100
Picture β = 0 in the middle of the line. The bell-shaped distribution indicates the probability distribution
of obtained or sample β values when the population β value is in fact equal to zero. Remember that we
can obtain β values that are not exactly equal to 0 because of error. The distribution tells us that in most
cases we will find β equal to zero (this is where the graph reaches its apex (highest point)). As the
numbers deviate from β = 0, both to the right and to the left, the distribution curves lower. In other words
as the value of β increases so the chance of obtaining that particular β value decreases. This is quite
logical. If the population β is in fact zero the chances are more likely of obtaining a β value of 3 than a β
value of 8 (3 is closer to zero than 8). And likewise on the negative side, one is more likely to obtain a β
value of -3 than one of -8.
The area under the curve covers the sum of the probabilities of all chance deviations from β = 0. Stating
an alternative hypothesis means one sets bounds to areas under the curve - one cordons off probability
sections. In case of a nondirectional alternative hypothesis one specifies two areas under the curve, one
in the “greater than zero” tail of the distribution and one in its “smaller than zero” tail. This is because
H1: β  0 does not indicate whether the non-zero value of β is larger than or less than zero. When the
computer calculates a probability value, p, for the obtained β value it actually says: The probability of
obtaining this particular β value, given that the null hypothesis is true (in other words, given that the
population β = 0), is p. The nondirectional alternative hypothesis “forces” the probability value, p, into
two sections, each measuring half of p (i.e. p divided by 2, or for short: p/2). This is what it looks like:
In the case of a nondirectional alternative hypothesis one simply compares the computer’s p-value with
0.05. Note that as the obtained β value deviates further from zero the areas representing the probability
of obtaining the β value become smaller. If the probability, p, becomes less then 0.05 one concludes that
the probability is small enough to conclude that the obtained β value deviates significantly from β = 0.
What happens when we have a directional hypothesis? A directional hypothesis places the probability
region in a specific tail of the probability distribution. If H1: β < 0 this region appears in the “below zero”
tail of the distribution. If H1: β > 0 the region is in the “above zero” tail. This is what the H1: β > 0
scenario looks like:
HMPYC80/104 101
HMPYC80/104/0/2024
Note that the directional alternative hypothesis rules out one of the tail regions. In this case H1: β > 0, and
the lower tail region is ruled out. (If H1: β < 0 the upper tail region disappears.) When one of the tail
regions is ruled out we are left with only one probability region, namely p/2. Therefore you have to
divide the computer’s p-value by 2 to obtain the probability value (p/2) for a directional hypothesis.
However, note that although most computer programmes automatically provide a nondirectional p,
some programmes may prompt you, before doing the calculation, to indicate whether the alternative
hypothesis is directional or nondirectional. If you tell the computer to deal with a directional hypothesis,
it will automatically divide p by 2 and present p/2 as the p-value. So be warned: Do not allow a clever
computer to lead you “up the garden path”!
102
Data entry form for quantitative Q

data analysis
Whether you are doing your own data analysis or requesting a statistician to help you, you need to
organise your data in a format that will make it possible to use the various statistical methods of data
analysis. You need to complete a data specification table as well as a data sheet that contains your
numerical data. The following are examples of forms to record this information on:
DATA ENTRY FORM
Name and Surname: ................................................................................................................
Student number: ................................................................................................................
Mail or e-mail address: ................................................................................................................
Fax and telephone numbers: ................................................................................................................
Data analysis technique: ................................................................................................................
Research Question, Statement, or Hypothesis
............................................................................................................................
............................................................................................................................
Operationalisation: ............................................................................................................................
............................................................................................................................
DESCRIPTION OF DATA
<data code sheet>
Name of variable Range of Measurement level Dependent Value and labels (for
(preferably use possible a) (Y) or nominal/categorical) or
only 8 characters) values nominal/categorical Independent the meaning of a higher
b) ordinal (X) variable? score (only for ordinal,
c) interval/ratio interval, ratio)
(Use as many rows as you need for all the variables in your study.)
HMPYC80/104 103
HMPYC80/104/0/2024
DATA SHEET Name: .............................................. Student number: ..............................
Observation/Subject Y X1: X2: X3: X4: X5: X6:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Etc.
104
<Selecting the appropriate statistic>
After completing the <data code sheet> and noting the measurement level, that is nominal/categorical,
ordinal or interval/ratio, of the variables you should be able to decide on a suitable method of data
analysis to test or explore your research hypothesis or questions. Make sure you study the method you
choose before you collect your data. Once you have collected all your data, organise it in a specific
format according to the method of data analysis that you are going to use.
Examples of measurement levels

a) Nominal: Sex, gender, marital status, work status, disability status, city or province, weekday or
month, ethnic group, race, language.
b) Ordinal: Social class, conservatism, alienation, educational level attitudes, ranked 3-level
measures (scores of attitudes, personality traits, and abilities).
c) Interval: Likert-type measures (scores of attitudes, personality traits, and abilities), exam results as
a symbol, Intelligence quotient, intraversion-extraversion score.
d) Ratio: Age, cost, exact income, number of children or persons, exam or test result as a
percentage.
HMPYC80/104 105
HMPYC80/104/0/2024
Variables that use intervals are at the interval level of measurement if the intervals are of equal size
(e.g.. 0-10, 11-20, 21-30, 31-40, etc.) Also note that, Likert type measures (measures of attitudes,
personality traits, and aptitude) are strictly speaking at the ordinal level of measurement, but they may
be treated as interval data as the statistics involved are robust enough to handle this exception. It does
require larger sample sizes (n > 30) and a distribution that is close to normally distributed for the group
or subgroups that are used in the statistical analysis.
The following table: "Making comparisons and exploring for difference (or non-difference) with
statistics typically used in psychology" indicates some of the most commonly used statistics in
psychological research. This list is not exhaustive, but does contain many of the parametrical statistical
tests that you should know. The boxes with the bold borders (#) indicate the parametric statistical test
and are of importance for this course. (Note that the boxes with the dashed borders (@) indicate the
non-parametric statistical test and are presented here for interest’s sake!)
106
Figure 02
Making comparisons and exploring for (non) difference with statistics
HMPYC80/104 107
HMPYC80/104/0/2024
Figure 03
Statistical measures of association and relationship typically used in Psychology
The boxes with the bold borders (#) indicate the parametric statistical test and are of importance for this
course. (Note that the boxes with the dashed borders (@) indicate the non-parametric statistical test and
are presented here for interests sake!).
On the next few pages we present some further examples of statistical methods.
108
<Correlation coefficient>
A researcher is interested in the relationship between two variables, namely age and anxiety about going
to the dentist. He believes that as people get older, they become less anxious and he formulates the
following research hypothesis: People become less anxious about going to the dentist as they become
older. The independent variable (X variable) is age and the dependent variables (Y variable) is level of
anxiety. He decides to calculate a correlation coefficient to determine the relationship between the two
variables. He randomly selects a sample of 39 patients and he obtains the age (in years) and a dental
anxiety score (based on a questionnaire with a minimum score of 4 and a maximum score of 20) for each
patient. This is how he will specify and record his data:
Data specification form
Name of Range of Measurement Dependent Value and labels (for

variable possible level (Y) or nominal/categorical) or
(Do not use more values (e.g. nominal/ Independent the meaning of a
than 8 characters) categorical, (X) higher score (for ordinal,
ordinal, interval, or variable? interval, ratio)
ratio)
age 17 to 71 ratio X higher value = older
anxiety 4 to 20 interval Y higher score = more
anxiety
Data sheet
Observation/subj Y: anxiety X1: age

ect/
person
1 5 25
2 5 29
3 6 30
4 9 45
etc. etc. etc.
38 7 46
39 6 27
Note that you will need a minimum of 20 observations when using this method and that you should use
columns Y (dependent variable) and X1 (one independent variable) on the data sheet. We used an
example of two continuous variables but one or both of your variables could also be categorical (for
example: variable X = male/female and variable Y = anxiety scores).
HMPYC80/104 109
HMPYC80/104/0/2024
<T-test>
A teacher develops a new method of teaching mathematics and he wants to know how successful this
method is. His hypothesis states that performance in mathematics will improve if the new method of
teaching is used. The independent variable (X variable) is method of teaching and the dependent variable
(Y variable) is performance in mathematics. He decides to divide the independent variable into two
groups and to use a t-test to determine the significance of the difference between the mean performance
of the two groups on the dependent variable. At the beginning of the year he selects two random samples
of 25 pupils each and teaches the one group with the new method of teaching and the other with a more
traditional method of teaching. He obtains the half-year examination score in mathematics for each of
the pupils. This is how he will specify and record his data:

ratio)
teaching 1 to 2 categorical X 1 = new method
2 = traditional method
maths 0 to 100 ratio Y higher score =
better performance in
maths
Data sheet
Observation/subj Y: maths X1: teaching

ect/
person
1 40 1
2 60 1
3 70 1
4 65 1
etc. etc. etc.
49 62 2
50 45 2
110
columns Y (dependent variable) and X1 (one independent variable) on the data sheet.
<One-way analysis of variance>
An industrial psychologist wants to know if different occupational groups differ in their ability to detect
a liar. She formulated the following research hypothesis: Different occupational groups will differ in
their ability to detect a liar. The independent variable (X variable) is occupation and the dependent
variable (Y variable) is ability to detect a liar. She decides to divide the independent variable into three
groups (judges, psychiatrists and secret service personnel), to do a one-way analysis of variance, and to
use an F-test to determine the significance of the difference between the mean performance of the three
groups on the dependent variable. She selects a random sample of 35 persons from each career group and
obtain a percentage of correct responses for each person in a deception accuracy exercise. This is how
she will specify and record her data:

ratio)
career 1 to 3 categorical X 1 = judges
2 = psychiatrists
3 = secret service
accuracy 0 to 100 ratio Y higher score =
better ability to detect a
liar
Data sheet
Observation/ Y: accuracy X1: career

subject/
person
1 50 1
2 60 1
3 50 1
etc. etc. etc.
104 50 3
105 60 3
HMPYC80/104 111
HMPYC80/104/0/2024
<Factorial analysis of variance>
Suppose you want to know if study technique and cognitive style affect scores on a test based on lecture
material and you formulate the following research hypothesis. Both study technique and cognitive style
will affect scores on a test based on lecture material. You are now working with two independent
variables (X variables), namely study technique and cognitive style while the dependent variable (Y
variable) is test performance. You decide to divide both the independent variables into groups and to use
a factorial analysis of variance (and the F-test) to determine the significance of the difference between
the mean performance of the groups for each variable on test performance. This will also help you to
determine the interaction effect of the two independent variables. You work with a sample of 104
first-year students and distinguish between those with a field-independent and those with a
field-dependent cognitive style. Within each of the two cognitive styles students are randomly assigned
to one of four study techniques (no notes, student notes, outline framework, complete outline) and they
have to apply this technique while listening to a taped lecture. You then obtain scores on a test based on
the lecture (minimum score is 0 and maximum score is 20). This is how you will specify and record your
data:
(Do not use more possible (e.g. nominal/ (Y) or nominal/categorical) or
than 8 characters) values categorical, ordinal, Independent the meaning of a higher
interval, or ratio) (X) variable? score (for ordinal, interval,
ratio)
style 1 to 2 categorical X 1 = field independence
2 = field dependence
study 1 to 4 categorical X 1 = technique 1
(no notes)
2 = technique 2
(student notes)
3 = technique 3
(framework outline)
4 = technique 4
(complete outline)
test 0 to 20 ratio Y higher score =
better performance on test
based on lecture material
112
Data sheet
Observation/subject/ Y: test X1: style X2: study

person
1 13 1 1
2 13 1 1
3 10 1 1
4 16 1 1
etc. etc. etc. etc.
103 17 2 4
104 15 2 4
Note that you will need a minimum of 60 observations if you work with two independent variables
divided into two groups each and that you should use columns Y (dependent variable) and X1 and X2
(two independent variables) on the data sheet.
<Simple regression analysis >
A group of students decide to do a study on pain and their hypothesis is that predictions of pain and
memories of pain are related and scores of predicted pain can therefore be used to predict memories of
pain. The independent variable (X variable) is expected pain and the dependent variable (Y variable) is
memory of pain. They decide to do a regression analysis to determine if there is a relationship between
the two variables and to obtain a formula to predict Y values from X values. They obtain a rating
(between 0 and a 100) on expected pain and memories of pain for a sample of 31 volunteers. This is how
they will specify and record their data:
(Do not use more possible (e.g. nominal/ (Y) or nominal/categorical) or
than 8 characters) values categorical, ordinal, Independent the meaning of a higher
interval, or ratio) (X) variable? score (for ordinal, interval,
ratio)
expected 0 to 100 interval X higher score =
worse pain is expected
memory 0 to 100 interval Y higher score =
worse pain is remembered
HMPYC80/104 113
HMPYC80/104/0/2024
Data sheet
Observation/subject/ Y: memory X1: expected

person
1 1 1
2 0 1
3 23 9
4 10 10
etc. etc. etc.
30 6 10
31 11 6
<Multiple regression analysis>
Suppose that the group of students in the previous example add another independent variable, namely
the actual experience of pain. Their problem therefore concerns the prediction of memory of pain
(dependent/Y variable) based on the expectations of pain (independent/X variable) and the actual
experience of pain (independent/X variable). They use a multiple regression analysis to determine the
relationship between each independent variable and the dependent variable as well as the possible effect
of the combination of the two independent variables on the dependent variable (the interaction effect).
They will also obtain a prediction formula for the Y variable based on more than one X variable. This is
how they will specify and record their data for the sample of 31 volunteers:

(Do not use more values (e.g. nominal/ Independent the meaning of a higher
than 8 characters) categorical, (X) score (for ordinal, interval,
ordinal, interval, or variable? ratio)
ratio)
expected 0 to 100 interval X higher score =
worse pain is expected
actual 0 to 100 interval X higher score =
worse pain has been
experienced
memory 0 to 100 interval Y higher score =
worse pain is remembered
114
Data sheet
Observation/subj Y: memory X1: expected X2: actual

ect/
person
1 1 1 1
2 0 1 1
3 23 9 1
4 10 10 2
etc. etc. etc. etc.
30 6 10 5
31 11 6 2
columns Y (dependent variable) and X1 and X2 (two independent variables) on the data sheet.
<Manifest and latent variables> (Intercorrelation matrix and Factor Analysis)
Suppose you want to measure short-term memory and long-term memory and you use three different
tests for each construct. You decide to compute an intercorrelation matrix to determine if the three tests
of short-term memory are related to each other but not to the three tests of long-term memory as well as
if the three tests of long-term memory are related to each other but not to the three tests of short-term
memory. This will indicate that the tests (manifest variables) are good measures of short-term and
long-term memory (latent variables). You obtain scores (on a 10-point scale) on three different tests for
each of the two memory processes for a random sample of 300 children without brain damage. This is
how you will specify and record your data:

ratio)
STM1 0 to 10 interval X higher score =
better short-term memory
(test 1)
HMPYC80/104 115
HMPYC80/104/0/2024
(test 2)
(test 3)
LTM1 0 to 10 interval X higher score =
better long-term memory
(test 1)
(test 2)
(test3)
Data sheet
Observation/subj X1: X2: X3: X4: X5: X6:

ect/ STM1 STM2 STM3 LTM1 LTM2 LTM3
person
1 9 9 8 3 2 2
2 4 4 5 4 5 6
3 5 5 5 5 4 4
4 7 8 7 7 7 8
etc. etc. etc. etc. etc. etc. etc.
349 1 2 2 7 8 7
350 9 9 8 1 1 1
116
T-test
<Hypothesis testing> deals with the relationship between variables X and Y; we test the significance of
the slope of the straight-line graph depicting this relationship, that is the significance of β. The most
straightforward way to determine the significance of the relationship between variables X and Y, is to
calculate the <correlation coefficient> between these variables. Another way to assess the significance
of the slope (β) is to look at the difference between two (or more) groups, or more precisely to look at the
difference between the means of these groups.
The <relationship between variables> can be graphically portrayed by plotting the measures obtained for
the X and Y variables. The range of measures for the X variable can also be divided into two (or more)
groups and the mean Y score for each category is then calculated to plot a “middle road” graph. Instead
of a whole range of, for example, motivation scores we can divide these scores into two groups namely
“poorly motivated” and “highly motivated”. This is what it may look like:
This looks simple enough, but beware of a pothole or two along this way. What we are actually
interested in are the Y scores of these two groups and not the difference in their X scores. We used the X
axis to constitute the two groups – the poorly motivated and the highly motivated groups. Note that we
also call the two groups, two levels or categories of the independent variable. (Do not confuse this with
the levels of measurement!) What we want to know is whether they differ with regard to their Y scores.
Let us assume the Y scores reflect the Research Methodology examination marks. We now determine
the average examination results of the poorly motivated group and the average examination results of the
highly motivated group. Then the situation looks like this:
HMPYC80/104 117
HMPYC80/104/0/2024
We want to know whether the difference in examination marks (the Y scores) is significant. Can you see
that the difference in Y scores is a function of the slope of the line, in other words a function of β? The
steeper the slope the larger the difference on the Y axis, and the less steep the slope the less the
difference in Y scores. So testing for the significance of the difference between groups is simply another
way of testing for the significance of β. But because we ask the question in a different way, the
hypotheses look slightly different. We now suggest, as our null hypothesis, that the average examination
mark for the poorly motivated population is the same as the average examination mark for the highly
motivated population. Or put more formally, that the mean Y scores for group XA (poorly motivated
group of individuals in the population) and XB (highly motivated group) do not differ. We refer to these
mean Y scores as μA and μB, and thus we write:
H0 : μA = μB
Alternatively we can suggest that these means do differ:
H1: μA  μB (nondirectional hypothesis)
Or we can hypothesize that the mean of the poorly motivated population will be lower than the mean of
the highly motivated population:
H1: μA < μB (directional hypothesis)
If we have the X and Y scores of a sample of subjects and we use the X scores to divide them into two
groups, we can determine the significance of the difference between the mean Y scores for groups XA
(poorly motivated group of individuals in the sample) and XB (highly motivated). The statistical test that
is used for testing the significance of the difference between the means of two groups is known as the
t-test. We distinguish between two types of t-tests, one for independent and one for dependent groups.
Independency means the groups have nothing in common. Dependency occurs when a group is tested
and then retested. In this case each member is paired with him-/herself, which means the members of the
two groups (tested and retested) are related. But dependency can also refer to different individuals being
paired off. If the researcher suspects that intelligence plays a role in examination results, he/she can pair
each member of the poorly motivated group with a member of the highly motivated group on the basis of
their intelligence scores. This leads to dependency between the two groups and hence the use of a t-test
for dependent groups.
Here Is an example: Suppose you worked with six persons and the following set of (X,Y) scores was
obtained for motivation (variable X) (measured on a scale from 6 to 17) and examination marks (variable
Y): (17; 80), (6; 30), (7; 40), (15; 70), (9; 50), (14; 60). In other words the first person has a score of 17
for motivation and an examination mark of 80, etcetera. How will you define the two groups to be
compared?
................................................................................................................................................................
................................................................................................................................................................
Have you divided the set of scores in terms of the X variable, motivation? If we have values for
motivation ranging from 6 to 17 we can simply categorise those between 6 and 11 as poorly motivated
and those between 12 and 17 as highly motivated. The poorly motivated group (let us call them Group
118
A) will include (6; 30), (7; 40), (9; 50) and the highly motivated group (Group B) will include (17; 80),
(15; 70), (14; 60). What you would like to know is if there is a difference in the examination marks (Y
variable or dependent variable) of the poorly motivated and the highly motivated groups. Try to
formulate the null hypothesis for this problem.
................................................................................................................................................................
................................................................................................................................................................
The null hypothesis suggests that the mean examination mark for Group A is the same as the mean
examination mark for Group B. Calculate these means for the two groups. Do you think there is a
difference between the groups?
................................................................................................................................................................
................................................................................................................................................................
Remember that you are interested in the examination marks, that is variable Y. The mean examination
mark for Group A is 40 and for Group B it is 70. It seems as if there is a difference in performance
between the poorly motivated group and the highly motivated group, but one will have to test for the
significance of this result. Remember that the statistical test that is used for testing the significance of the
difference between the means of two groups is known as the t-test. This statistic is based on the means of
the two groups and the variability of the scores within the sample. We use a computer to calculate the
t-value. The computer also provides the probability of obtaining this value should the difference
between the mean examination marks for the poorly motivated population and the highly motivated
population actually be equal to zero. If this probability is smaller than the cut-off point (for example
0.05), we reject the idea that the sample reflects a population where the difference in means between the
two motivation groups is equal to zero.
But what does this mean? Remember that we are interested in the relationship between variables X and
Y; the relationship between motivation and examination marks. We determine the significance of this
relationship by looking at the significance of the difference in mean examination marks for a poorly
motivated and a highly motivated group. If this difference is statistically significant (as indicated by the
probability associated with the t statistic), we can conclude that motivation and examination marks are
related.
Applying the theory
A teacher develops a new method of teaching mathematics and he wants to know how successful this
method is. In this case method of teaching is the independent variable and the teacher wants to know if
changes (variations) in this variable has any effect on the other variable, performance in mathematics
(the dependent variable). The teacher will probably hypothesise that performance in mathematics will
improve if the new method of teaching is used – a directional hypothesis.
Can you think of a more specific hypothesis that directly indicates the research design and how the
variables will be measured?
................................................................................................................................................................
................................................................................................................................................................
HMPYC80/104 119
HMPYC80/104/0/2024
The teacher can state that pupils taught by the new method will score higher in the mathematics
examination than pupils taught according to a more traditional method. When planning (designing) your
research, you have to decide on a way of doing the study that will help you to answer your research
question. In this case it makes sense to compare different methods of teaching and to divide the
independent variable into groups rather than working with a range of values or scores for this variable.
The mean performance of these groups on the dependent variable is compared.
The teacher is interested in the difference between groups and more specifically in the difference
between the means of the two groups. Consider the hypothesis and define the X and Y variables and
indicate for which variable he will calculate the mean.
................................................................................................................................................................
................................................................................................................................................................
Do you remember that we use levels of the X variable (the independent variable) to define our groups? In
this case the X variable is the method of teaching and we call the pupils taught with the new method,
Group A, and the pupils taught according to the more traditional method, Group B. The Y variable is the
examination performance and the teacher will calculate the mean examination score for Group A and for
Group B. What will the null hypothesis be in this case?
................................................................................................................................................................
................................................................................................................................................................
The null hypothesis implies that the mean examination score in the population for Group A is the same
as the mean examination score for Group B, in other words, there is no difference between the means.
You can also write this as:
H0 : μA = μB
The alternative hypothesis will be directional because the teacher believes that the pupils taught with the
new method will do better. We can write this as:
H1: μA > μB
At the beginning of the year the teacher selects two samples (25 in each sample) of pupils randomly to
test his hypothesis. They are taught with the new and the more traditional methods respectively. The
half-year examination scores for mathematics for each of the pupils are reported in the following table.
Examination scores for mathematics (Y variable) for the two methods of teaching (X variable)
Group A: New method of teaching Group B: Traditional method of teaching

59 70
74 91
70 80
65 84
66 62
59 53
120
68 34
62 60
73 40
77 43
75 52
80 50
61 70
58 61
58 81
60 75
79 50
80 43
70 52
80 55
65 60
64 56
58 52
72 62
80 45
We can draw a graph of the mean examination scores for Group A (68.50) and Group B (59.24).
Based on these results could you say that the alternative hypothesis is true, and that the new method is
superior?
................................................................................................................................................................
................................................................................................................................................................
HMPYC80/104 121
HMPYC80/104/0/2024
It is clear that the mean for the pupils in Group A is numerically higher than the mean for the pupils in
Group B. You might accept the alternative hypothesis and conclude that the new method is superior.
Another student might have decided that the difference is due to error and that the null hypothesis is true.
Who will be correct? Remember that we said we use a t-test for testing the significance of the difference
between the means of two groups? The t-value associated with the difference in means found for the two
groups is calculated as well as the probability (p-value) of obtaining this result should the population
difference actually be equal to zero. The following is a computer printout using the SAS programme for
doing a t-test for two independent samples. Remember that you do not have to understand all the values.
The results that the teacher is interested in (t-value and p-value) are given in bold type and are tabled.
Mean examination performance for two groups and the t- and p-values for the difference between the
means
Mean for Group A n=25 Mean for Group B Difference t-value p-value
n=25
68.50 59.24 9.26 1.21 0.2360
In the case of a directional alternative hypothesis the probability value has to be divided by two and the
p-value that we should look at is therefore 0.118. The probability is fairly large (0.118 > .05, and
therefore p > the cut-off point or significance level) and the obtained difference between the means is
therefore not significant. The teacher will not reject the null hypothesis and he accepts that there is no
relationship between teaching method and examination performance.
The statistical method of analysis used for this problem is therefore a t-test for testing the significance of
the difference between the means of two independent groups. A <one-way analysis of variance> where
the F-test is calculated could also have been used.
Exercise
When people become physiologically aroused by one stimulus, their emotional response to another
stimulus is heightened. This notion has been applied to the topic of romantic love and sexual attraction.
Some couples will, for example, exhibit greater attraction for each other when their parents produce
arousal by interfering with the relationship. Researchers have proposed that the heightened
122
physiological arousal due to an extraneous source is “misattributed” (misdiagnosed) as an especially
strong romantic attraction toward a member of the opposite sex. They have formulated the following
research question: Is it possible that arousal due to fear might be misattributed as being heightened
sexual attraction to a member of the opposite sex?
Dr Richards decided to conduct a laboratory experiment. She drew a random sample of 40 male college
students and told them that they were going to be electrically shocked as part of a study on learning.
They had to give their consent before the study started. Fear was manipulated by threatening subjects
with different levels of shock. The shock was described as either mild (low fear condition) or strong
(high fear condition). Dr Richards randomly divided the sample into two subgroups and randomly
assigned these subgroups to each condition. To determine attraction to a member of the opposite sex, a
female co-worker of Dr Richards also participated in the study. Immediately after the study the subjects’
attraction to the girl was measured on a 5-point scale and a score was obtained for each subject.
1. Formulate a research hypothesis that applies to this experiment. ..............................................
....................................................................................................................................................
2. Subject the hypothesis you have formulated to the criteria a good hypothesis should adhere to.
....................................................................................................................................................
....................................................................................................................................................
3. Identify the dependent and independent variables, and where appropriate indicate the levels of
these variables.
The dependent variables(s): ....................................................................................................
....................................................................................................................................................
The independent variables(s): ....................................................................................................
....................................................................................................................................................
4. Formulate the alternative hypothesis (if symbols are used they should be defined).
....................................................................................................................................................
....................................................................................................................................................
5. Suppose the results of the statistical analysis are significant. How would you interpret this?
....................................................................................................................................................
....................................................................................................................................................
6. Comment on the external validity of this experiment. ...............................................................
....................................................................................................................................................
HMPYC80/104 123
HMPYC80/104/0/2024
Answers to exercise
1. The research question in this study is: “Is it possible that arousal due to fear might be
misattributed as being heightened sexual attraction to a member of the opposite sex?”
Considering this question we formulated the following research hypothesis: Students who
received a threat of strong electrical shock (high fear condition) will show greater attraction to a
member of the opposite sex than students who received a threat of mild electrical shock (low fear
condition).
Your hypothesis should indicate how the variables are going to be operationalised and it is not
sufficient to simply formulate the research question as a statement. Those of you who started
your hypothesis with “An increase in level of fear ...” or “An increase in arousal due to fear ...”
should note that you have sufficient information to be more specific (see our example).
2. The hypothesis in 1.1 is a statement about the relationship between two variables (level of fear
and level of attraction) and gives clear indications for their measurement and operationalisation
(threatening with different levels of shock and measuring level of attraction).
Note that you should state the criteria for a good hypothesis and apply these to the specific study.
3. The independent variable, in other words that which the researcher manipulates to see what
effect it has, is fear. She manipulates this variable to form two groups by threatening students
with a mild shock/low fear condition or a strong shock/high fear condition. We refer to these
possible values or categories of the independent variable as levels of the variable and in this case
the independent variable has two levels, namely 1) low fear condition and 2) high fear condition.
Please distinguish between levels of the independent variable (used to form our different groups)
and the level of measurement.
The researcher wanted to see what ”he e’fect of different levels of fear (the independent variable)
is on attraction to the opposite sex. The dependent variable is therefore attraction.
You will find it difficult to continue with the study (and the next questions) if you cannot
distinguish between the independent variable (the researcher has control over this variable and
uses it to form groups, etcetera) and the dependent variable (the researcher studies this variable to
see how it changes).
4. The alternative hypothesis should imply that there is a relation between the variables, in other
words between fear and attraction. In this study the researcher not only expects the attraction
score to differ for the two groups, she expects the score for one group to be higher and the
alternative hypothesis should therefore be directional.
The researcher wants to compare two groups (formed on the basis of the independent variable,
fear) in terms of their score on the dependent variable, attraction. She could thus compare the
mean attraction score for the two groups. (Can you see that a t-test would be more suitable for
this study than a correlation?)
124
The alternative hypothesis would be: The mean attraction score of the high fear condition is
higher than the mean attraction score of the low fear condition. (If you said H1 : μA > μB you
should just make sure that you define the symbols.)
5. A significant result implies that the researcher rejects the null hypothesis and accepts the
alternative hypothesis. She can therefore accept that the mean attraction score of the high fear
condition is significantly higher than the mean attraction score of the low fear condition.
It is not sufficient to just repeat the alternative hypothesis as an interpretation and one should be
more specific than just saying that there is a relation between fear and attraction. What is the
nature of this relation and what are the implications thereof? One needs to consider the original
research question and the theoretical background that resulted in the study and the hypotheses. A
possible interpretation of this result is that arousal due to fear is misattributed as being
heightened sexual attraction to a member of the opposite sex.
You should remember that if the hypothesis is directional and the difference in means is in the
other direction, the hypothesis is rejected without having to test it. In other words if the mean of
the low fear group was higher the researcher would have rejected the hypothesis without having
to test it.
6. External validity refers to the extent that the results of this study can be generalised to other
situations, people, etcetera. This experiment does not provide for all the aspects that could affect the
external validity. Some of these aspects are: The fact that it was not done under natural conditions (it is
not the usual context in which you think about attraction to another person and the fear arousing stimulus
differs from what one would find in everyday life); only male students were included in the sample; and
a specific female was used. (A sample size of 40 might also not be sufficient to generalise from this
result to the population.)
HMPYC80/104 125
HMPYC80/104/0/2024
Correlation coefficient
In <hypothesis testing> there are a number of different methods we can apply to determine whether or
not there is a relationship between variables X and Y. All these methods relate to testing the significance
of the slope of the straight-line graph, that is the significance of β. We want to know that if for the larger
population β is in fact equal to zero, what the probability is of obtaining the particular (non-zero) β
value for a sample of this population. The larger this probability the more likely it is that the obtained
non-zero β value is in fact equal to zero. The smaller this probability the more unlikely the chances that
the obtained (non-zero) β value is actually equal to zero.
If there is a relationship between two variables it means that a person’s relative position on one variable
bears a relation to his relative position on the other variable. A positive relationship implies that
relatively high scores on one variable are associated with relatively high scores on the other and
relatively low scores on the first correspond with relatively low scores on the second. An inverse or
negative relationship on the other hand implies that high scores on one variable are associated with low
scores on the other variable. If the variables are not related, changes on one variable do not correspond
with changes on the other.
We refer to the statistical relationship between two variables as a correlation and the statistic used to
describe this is called a correlation coefficient. There are different kinds of correlation coefficients, but
all of them are measures or indices of the relationships between variables. More specifically it indicates
the direction and the strength of these relationships. A correlation coefficient can range in value from
-1.00 to +1.00. These values represent a perfect negative (-1) or a perfect positive correlation (+1). A
value close to 0 indicates a weak relationship while 0 represents no correlation. We can see that the
numerical size of a correlation coefficient indicates the strength of the relationship while the sign
(positive/negative) indicates the direction of the relationship.
The correlation coefficient Is an Index of the extent of the linear relationship between two variables and
as such plays a large role in determining the slope of the graph (or the β value). The correlation
coefficient does not have the same numerical value as β but is part of the β value. If we want to determine
the significance of the relationship between variables X and Y, the most straightforward way is to
calculate the correlation coefficient between these variables. When stating our hypotheses, we indicate
that it is the correlation coefficient (ρ), and not β per se, that we are dealing with.
So we write: H0: ρ = 0
and: H1: ρ  0 (nondirectional hypothesis)
or: H1: ρ > 0 or H1: ρ < 0 (directional hypotheses)
So the scenario is as follows: You have the X and Y scores of a sample of say 10 subjects. You calculate
(or rather you use a computer to calculate) the correlation coefficient expressing the relationship
between the X and Y variables. We normally write this correlation coefficient as rXY. We use r instead of
ρ to indicate that we deal with a sample correlation and not with the correlation in the entire population.
The computer calculates rXY, and provides the probability of obtaining the calculated value should the
population correlation actually be equal to zero – in other words, the probability of obtaining rXY if ρXY =
126
0. If this probability is fairly large we must accept that the obtained rXY value is a fluke, and that it should
actually be equal to zero. On the other hand, if the obtained rXY carries a fairly small probability, we must
reject the hypothesis that ρ = 0, and accept an alternative hypothesis.
You are probably wondering about the terms “a fairly large probability” and “a fairly small probability”.
How large or how small should this probability be? There are no hard and fast rules but psychologists
often work with the cut-off point of 0.05. In other words, when the probability associated with a
sample’s rXY drops below 0.05 (i.e. less than 5%) psychologists tend to reject the idea that the sample
reflects a population where the correlation between variables X and Y is equal to zero (ρXY = 0). By
doing so they accept that the sample in question represents a population where the correlation between X
and Y is not equal to zero (ρXY  0), or perhaps greater or less than zero (ρXY > 0 or ρXY < 0), depending
on the directionality of their alternative hypothesis. Under these circumstances researchers refer to the
obtained correlation coefficient (rXY) as statistically significant. If they do not reject the null hypothesis
they refer to the result as not statistically significant. (Also see <hypothesis testing> where the cut-off
point is discussed.)
Applying the theory

Let us look at a research problem that illustrates this method of data analysis. A researcher is interested
in the relationship between age and anxiety about going to the dentist. He believes that as people get
older, they become less anxious. You need to study this research problem to identify the independent and
dependent <variables> as well as the relationship between them (also refer to <relationship between
variables>). The researcher wants to know if changes (variations) in the variable age has any effect on
the other variable, level of anxiety. In other words he wants to know if level of anxiety “depends” on the
age of a person. Age is therefore the independent variable and the variable, level of anxiety, is called the
dependent variable. The type of relationship is determined by the researcher’s expectations. The
<hypothesis> could state that level of anxiety will change as age changes (nondirectional) but in this
case the researcher will probably state that level of anxiety will decrease with increasing age
(directional). Formulate a hypothesis about the relationship between the dependent and the independent
variable.
................................................................................................................................................................
................................................................................................................................................................
A plausible research hypothesis for this problem is: People become less anxious about going to the
dentist as they become older. Once you have formulated a hypothesis and before you can continue to do
the research and obtain scores that will be used to test the hypothesis, you must define the X and Y
variables for which scores or values will be obtained.
................................................................................................................................................................
................................................................................................................................................................
We have seen that in this case the X variable is age and the Y variable is level of dental anxiety. The
researcher now needs to design a study that will help him to test his hypothesis about the relationship
between variables X and Y. Suppose he uses a questionnaire consisting of four questions to measure
level of anxiety. Scores on each question ranges from 1 to 5. The dental anxiety score is obtained by
summing the responses to the four questions. High numbers mean high anxiety. He is interested in the
linear relationship between age and anxiety scores and decides to calculate a correlation coefficient to
HMPYC80/104 127
HMPYC80/104/0/2024
determine this relationship.
The null hypothesis implies that there is no linear relationship between the two variables, in other words
that the correlation coefficient for the population is zero:
H0 : ρ = 0
The alternative hypothesis is directional because the researcher believes that anxiety scores will decrease
as age increases (implying a negative correlation):
H1 : ρ < 0
He randomly selects a sample of 39 patients to test his hypothesis and obtains their ages and dental
anxiety scores.
Age (X variable) and dental anxiety scores (Y variable)
Age Dental anxiety scores

25 5
29 5
30 6
45 9
26 9
40 10
45 10
71 10
44 5
41 10
33 7
41 8
44 7
27 19
39 10
33 19
21 14
33 8
50 7
39 10
28 9
35 5
41 10
41 11
45 10
30 10
17 9
22 12
128
18 9
32 6
18 12
29 10
29 8
28 10
23 11
18 11
39 6
46 7
27 6
Next he determines whether the relationship between the variables is statistically significant. The
following is a computer printout using the SAS programme for calculating a correlation coefficient for
two continuous variables. You need not understand the complete printout but should focus on the results
that the researcher is interested in, namely the correlation coefficient and the probability (p-value) of
obtaining the calculated value should the population correlation actually be equal to zero. These results
are given in bold type and are also summarised in a table following the printout.
The correlation coefficient (rxy) for age and dental anxiety and the p-value for this result (n=39):
Rxy p-value
-0.18 0.2827
In the case of a directional alternative hypothesis the probability value (p-value) has to be divided by two
(refer to <hypothesis testing>) and the p-value that the researcher should look at is therefore 0.1413. The
HMPYC80/104 129
HMPYC80/104/0/2024
obtained correlation is therefore not significant (0.1413 > .05) and he will not reject the null hypothesis.
He accepts that the sample reflects a population where the correlation is equal to zero. This implies that
there is no linear relationship between age and dental anxiety.
130
One-way analysis of variance
One way of assessing the significance of the relationship (see <hypothesis testing>) between variables X
and Y is to look at the difference between the means of two or more groups. The values of the
independent variable (X) are used to form the groups and the mean score on the dependent variable (Y)
is calculated for each group. Say for example that our Y variable reflects examination marks and our X
variable is motivation. We divide the X values into three groups: a poorly motivated group, an
indifferent group, and a highly motivated group.
The statistical test that Is used for testing the significance of the difference between the means of two
groups is known as the <t-test>. In our example we cannot use the t-test for the difference between
means, because we are working with three groups. Here we use the F-test! This method of analysis is
also known as analysis of variance (because there is only one independent variable we call it a one-way
analysis of variance). The F-test uses variance as an index for the difference between two or more
means. Yes, if you have two groups you can use either the t-test or the F-test, but when we get to more
than two groups you must use the F-test. When you work with three groups the hypotheses look like this:
H0 : μA = μB = μC
H1: μA  μB or μA  μC or μB  μC
In the case of four or more groups you simply expand the hypotheses accordingly. But note the
following: A significant F-value indicates an overall significant β (the slope of the line). It does not tell
us between which means the differences occur, which also means we cannot state directional
hypotheses. For example, we can obtain a significant F-value for our XA (poorly motivated), XB
(indifferent), and XC (highly motivated) groups, but this does not tell us which of the differences
mentioned in the alternative hypothesis (H1) is/are significant. We need ad hoc procedures to determine
where the differences occur. The Scheffé test is an example of such an ad hoc procedure, but in this
course we will not go into the application of such procedures.
HMPYC80/104 131
HMPYC80/104/0/2024
Applying the theory
An industrial psychologist wants to know if different occupational groups differ in their ability to detect
a liar. Suppose the industrial psychologist worked with three occupational groups and formulated the
following research hypothesis: Judges, psychiatrists and secret service personnel will differ in their
ability to detect a liar. Indicate the X and Y variables and define an appropriate method of analysis for
testing this hypothesis.
................................................................................................................................................................
................................................................................................................................................................
The psychologist uses levels of the X variable (occupation) to define the groups (judges, psychiatrists
and secret service personnel). She wants to know if these three groups differ in terms of the Y variable
(ability to detect a liar). Because she is working with only one independent variable divided into more
than two groups and she is interested in the differences between the means of these groups, she will use
a one-way analysis of variance to determine if these differences are significant. (Can you see why a
<t-test> is not suitable?) The hypotheses will be:
H0 : μA = μB = μC
Suppose a random sample of people in each career group (n = 35) were shown 1-minute videotape
segments of ten individuals telling how they felt about a film. Participants were then asked to indicate
whether the person in the videotape was telling the truth. The score on the Y variable for each subject is
the deception accuracy score, which is the percentage of correct responses (in fact five of the ten were
lying). These scores are reported as percentage correct responses.
The means for the three groups and the F-test associated with the differences in the means are calculated
to obtain the p-value (the probability that these differences would have occurred if occupations did not
differ in their ability to detect liars). The SAS computer printout follows the data table. The relevant
results (F- and p-values) are in bold type and are also summarised.
Deception accuracy scores (Y variable) for judges, psychiatrists and secret service personnel (X variable)
Judges Psychiatrists Secret service personnel
50 60 60
60 90 60
50 60 70
60 70 50
70 60 70
50 50 60
60 50 70
50 50 70
40 60 60
70 70 60
50 60 40
70 60 80
40 60 70
90 50 90
132
50 70 80
60 30 60
40 60 80
60 50 90
60 60 50
40 60 40
50 40 50
50 60 40
40 70 50
70 40 70
70 50 70
50 50 70
50 60 60
70 60 60
60 60 70
50 50 60
80 40 80
70 80 60
70 60 80
50 40 50
30 80 60
HMPYC80/104 133
HMPYC80/104/0/2024
Mean deception accuracy score for the occupational groups and the F- and p-values for the differences
between the means
Mean for judges Mean for psychiatrists Mean for secret services F-valu p-value
n=35 n=35 n=35 e
56.57 57.71 64.00 3.39 0.0375
Remember the cut-off point is 0.05. How will you interpret these results and how will you answer the
research question?
................................................................................................................................................................
We reject the null hypothesis because the probability of this result (p-value) under the null hypothesis is
smaller than what we are prepared to tolerate (i.e. 0.0375 < .05). We also say: 0.0375 is smaller than
0.05, the significance level, therefore the result is significant. We can say that the success in detecting
liars differs according to occupation in other words there is a relationship between the occupational
group and the ability to detect a liar. Note that we worked with a nondirectional hypothesis; the result
indicates an overall significant difference and it does not tell us between which means the differences
occur.
134
Factorial analysis of variance
When we work with one independent variable (X variable) we can have two or more groups. These
groups constitute levels of the independent variable and are all part of the one X variable. For example, if
the X variable is motivation, we can distinguish three levels of motivation, namely poorly motivated
(XA), indifferent (X B), and highly motivated (X C). However, your research project may be more
complicated, and you may wish to add more variables. So let’s explore what happens when you add
another variable. The logical thing to do would be to name the new variable Z, but this leaves us with the
problem that after Z we’re out of names. Statisticians have fallen into the habit of calling all additional
variables X, but to add a subscript to identify them as different variables. Following this convention we
will now rename our X variable to X1, and call the newly added variable X2. Here is a drawing where an
additional variable, namely X2, has been added to the familiar Y and X1 axes.
As you can see it’s fairly difficult to represent a three-dimensional picture on a flat surface. So, I am
afraid you will have to use your imagination. Imagine the straight-line graph floating in the
three-dimensional space delimited by the Y, X1 and X2 axes. If you have a problem picturing this, use the
room in which you sit as an example. Look at one corner of the room, where two of the walls meet. That
corner, running from floor to ceiling, is the Y axis. Now look at where the Y axis corner meets the floor.
There are two more corners here. These corners are formed where the two walls meet the floor. Call the
one corner the X1 axis and the other the X2 axis. Here is a picture:
HMPYC80/104 135
HMPYC80/104/0/2024
Now imagine a straight line floating in the middle of the room. If you have a problem imagining a
mathematical line floating in your room, use a broomstick as an example! Now imagine the broomstick
casting two shadows, one on the X1Y wall and one on the X2Y wall. These two shadows contain all the
information we need should we want to reconstruct the original broomstick. Of course we are talking
here of a mathematical reconstruction, but you can easily picture the reconstruction more graphically.
Imagine the two walls as transparent screens, each containing a shadow picture. If we put a light behind
each wall the two shadow pictures are projected into the space of the room, where they meet up to form
a picture of the original broomstick. What we are doing is to construct the line floating in the space
demarcated by Y, X1 and X2 by adding together its X1Y and the X2Y pictures.
There is still one last point to be made. I am sure you will agree that a three-dimensional object has more
to it than simply putting together the flat shadow images on the two walls. The three-dimensional object
is more than the sum of its two-dimensional components. It reminds one of the Gestalt principle, which
says that the whole is more than the sum of its parts. The point we want to make is the following: When
we add the X1Y and the X2Y images we get an interaction effect between the two, which gives an
additional quality to the sum of these images.
Let’s express these ideas more formally. When we deal with the <relationship between variables>, we
can express Y as a function of X. We now express Y as a function of X1 and X2, etc.
Thus Y = f (X1, X2, X3 ..... Xn ).
Now consider the basic straight-line formula:
Y = α + βX
136
We simply expand this formula by adding together the “shadow images” to reconstruct, for example, the
three-dimensional scenario:
Y = α + β1X1 + β2X2+ βX1X2X1X2
Here the term βX1X2X1X2 reflects the interaction effect between variables X1 and X2. For simplicity’s
sake we can consider the interaction effect as a third variable, X3, in which case we write:
Y = α + β1X1 + β2 X2 + β3X3
The above discussion concentrated on three dimensions because we can, albeit with some difficulty, still
draw a three-dimensional graph. Although we cannot picture graphs involving four (or more)
dimensions, nothing stops us from mathematically working with such higher dimensional spaces.
Therefore one can design research projects involving a number of variables. But be warned: although
such multivariate designs can be dealt with mathematically the statistical findings are often very difficult
to interpret and integrate meaningfully.
If we are working with only one independent variable, we determine whether or not there is a
relationship between X and Y by testing the significance of the slope of the straight-line graph, that is the
significance of β (see <hypothesis testing>). Because the <correlation coefficient> (rXY) plays a large
role in determining the slope of the graph, we can calculate this value to determine whether β, the slope
of the line that reflects the relationship between the X and Y variables, deviates significantly from zero.
If one wants to determine the slope of a “broomstick” line floating in a three-dimensional space (or an
even higher order dimensional space) one has to calculate (or rather ask the computer to calculate) a
multiple correlation coefficient, a coefficient indicated by R. Such a coefficient puts together the β
values associated with the different variables to express the slope of the multivariate graph. In other
words, it puts together all the different shadow pictures to determine the slope at which the broomstick
floats in the multidimensional space. The “broomstick” slope is quite complicated because the slope at
which the broomstick floats must be seen in relation to the different “walls” demarcating the space in
which the broomstick finds itself.
Formally the hypotheses can be stated as follows:
Ho: R = 0
H1: R  0
H1: R > 0 or H1: R < 0
We can, however also assess the significance of the relationship between variables by looking at the
difference between the means of groups. If we are working with only one independent variable (X
variable), this variable can be divided into two or more levels, each forming a separate group. Just
imagine how the number of groups increases when one adds more independent variables. In this case
<t-tests> are obviously out of the question, and we can also not get away with a simple <one-way
analysis of variance>. The variance analysis technique we use when working with two or more
independent variables is known as factorial analysis of variance. This technique takes into account not
only the differences between groups associated with specific variables, but also the interaction effects
between variables. We heard of interaction effects, but what does it mean? Let’s hypothesise that
“examination results in research methodology” (variable Y) is a function of motivation (X1) and
intelligence (X2), and suppose we find that intelligence does not play a role (it is not related to, it does
not correlate with) examination results in the poorly motivated group and in the indifferent group.
However, in the highly motivated group we find intelligence to be positively related to examination
results. In other words, for the highly motivated group the more intelligent students fare better than the
HMPYC80/104 137
HMPYC80/104/0/2024
less intelligent students, but this is not the case in the poorly motivated group and in the indifferent
group. Thus intelligence plays a role in examination results, but the role it plays depends on the level of
motivation. In other words, there is an interaction effect between intelligence and motivation.
So, what do the hypotheses look like now? Things are not as complicated as you might think. Keep in
mind the walls and the shadows of the above discussion. To make things easy we do not try to tackle the
multidimensional “broomstick” graph in one go. We limit our hypotheses to individual walls and
shadows. In other words, we split the multiple correlation coefficient I into the different rXY coefficients.
These coefficients, as we know, play a major role in the β values (the slopes of the shadows) on each
wall. In other words, instead of looking at R in one go, we look at the shadow (rXY) it casts on each of the
different walls.
Thus, for variable X1 we write:
Ho: μA = μB = μC
And for variable X2 we write:
Ho: μA = μB
H1: μA  μB
For the interaction effect, X1X2, we write:
Ho: There is no interaction effect between variables X1 and X2.

H1: There is an interaction effect between variables X1 and X2.
We write the interaction hypotheses in words, because the symbolic representations are somewhat
complicated. Also note that variable X1 has three levels (XA, XB and XC), and that two levels were
indicated for variable X2. Dividing motivation into three groups, namely poorly motivated, indifferent,
and highly motivated, implies that it is a three-level variable. Of course variables may have more than
three levels. Furthermore your research problem may require more than two variables, for example
variables X1, X2 and X3. Should this be the case, simply expand the model accordingly. The most
complicated aspect is to keep track of the different interactions. For example, a three variable design
gives four interactions: X1X2, X1X3, X2X3, as well as the overall interaction X1X2X3.
Applying the theory
Suppose you want to know if study technique and cognitive style affect scores on a test based on lecture
material. You are now working with two independent variables, namely study technique and cognitive
style. Relationships between study technique and test performance (the dependent variable) as well as
between cognitive style and test performance are implied. In this case the effect that cognitive style has
on test performance could, however, be influenced by the type of study technique that was used (and
vice versa). We call the effect of the combination of the independent variables on the dependent variable,
the138ffectt.
Suppose you formulate the following research hypothesis from the research question: Both study
technique and cognitive style will affect scores on a test based on lecture material. You decide to work
138
with a sample of first-year students and identify students as having either a field-independent or
field-dependent cognitive style (2 levels) based on their scores on the Hidden Figures Test. Within each
of the two cognitive styles students are randomly assigned to one of four study techniques (4 levels – no
notes, student notes, outline framework, complete outline) and they have to apply this technique while
listening to a taped lecture. They then complete a 20-item multiple-choice test on the lecture material.
Identify an appropriate method of analysis to test the research hypothesis.
................................................................................................................................................................
................................................................................................................................................................
In this problem we have more than one X variable or independent variable, namely study technique and
cognitive style and each variable can be divided into more than one level to form separate groups. We
want to know what the differences are between the groups in terms of their performance on the test based
on the lecture material (the Y variable). We are, however, also interested in the interaction effect of the
two X variables. Stated differently and in terms of the definition of interaction effect: we are interested in
whether the effect of one of the independent variables on the dependent variable, depends on the level of
the other independent variable. Factorial analysis of variance is therefore a suitable method to use.
Because we are dealing with only two independent variables we can be more specific and call it a
two-way analysis of variance.
The null hypothesis about the relation between X1 (cognitive style) and Y (performance on the test)
indicates that field-dependent and field-independent students will on average (the means) perform the
same on the test based on lecture material. The alternative hypothesis indicates that field-dependent and
field-independent students will on average (the means) perform differently on the test based on lecture
material. We can also write:
H0 : μA = μB
H1: μA  μB
Try to formulate the hypotheses about the relation between X2 (study technique) and Y. (Hint: refer to
<one-way analysis of variance> where hypotheses involving more than two groups are discussed.)
................................................................................................................................................................
................................................................................................................................................................
The relationship between each of the independent variables and the dependent variable is similar to the
relationships we find when doing a <t-test> (one independent variable and two groups) or a <one-way
analysis of variance >(one independent variable and two or more groups). However, with two
independent variables we also need to formulate hypotheses about the interaction effect of the two
variables on Y. The null hypothesis states that there is no effect due to the combination of cognitive style
and study technique (or: Differences between the mean scores on the test of field-dependent and
field-independent students do not depend on the study technique used) while the alternative hypothesis
states that there is an interaction effect.
HMPYC80/104 139
HMPYC80/104/0/2024
The following results were found for the sample of students.
Scores on a test based on lecture material (Y variable) for students who used one of four study
techniques for field-independent and field-dependent students (X variables)
Technique 1 Technique 2 Technique 3 Technique 4
Field independence 13 15 19 15
13 19 18 19
10 19 17 16
16 17 19 17
14 19 17 19
11 17 19 15
13 20 17 20
13 17 19 16
11 18 17 19
16 17 15 16
15 18 18 19
16 18 17 19
10 19 15 18
Field dependence 11 12 18 18
14 16 15 19
11 16 15 15
10 17 14 16
15 16 15 19
10 16 18 18
16 16 19 19
16 14 18 19
17 14 18 18
11 16 16 17
16 15 16 16
11 15 18 17
10 15 16 15
The results that we are interested in from our sample are the differences between the means on the test
based on lecture material.
140
Mean performance (n=13 in each cell)
Cognitive style Study technique
1 2 3 4
Field independence 13.15 17.92 17.46 17.54
Field dependence 12.92 15.23 16.62 17.38
The following graph shows how study technique is related to test performance for different cognitive
styles:
Indicate if you think we should accept or reject the null hypothesis that field-independent and
field-dependent students will on average perform the same. Based on your decision, do you think that
there is a relationship between cognitive style and performance on the test?
................................................................................................................................................................
................................................................................................................................................................
It seems as if cognitive style relates to performance with field-independent students performing better.
Now try to determine if there is a relationship between study technique and performance on the test. First
decide if you should accept or reject the null hypothesis that students using different study techniques
will on average perform the same.
................................................................................................................................................................
................................................................................................................................................................
Again we might want to reject the null hypothesis as students seem to do progressively better with
subsequent techniques. This is, however, only true for students with a field-dependent cognitive style.
We see that for field-independent students there is not much of a difference between the means for study
techniques 2, 3 and 4. The size of the effect of cognitive style on test performance varies in accordance
with the type of study technique used (and vice versa!) implying that the combination of cognitive style
and study technique affects performance on the test.
It seems as If we should reject the null hypotheses In all three cases. How should we decide? We want to
know what is the probability that the differences between the means would have occurred if the null
HMPYC80/104 141
HMPYC80/104/0/2024
hypotheses of no differences indicating no relationships were true for the population. As with one-way
analysis of variance the statistician will compute a F-test associated with the results for our samples to
obtain the p-value for each of the two independent variables as well as for the interaction between the
two independent variables.
The F- and p-values for the differences between the mean test performance for different study
techniques and different cognitive styles as well as for the interaction effect (n= 104)
Variable (Type of effect) F-value p-value
Cognitive style 7.78 0.0064
Study technique 33.22 0.0001
Interaction 2.83 0.0426
(Cognitive style * Study
technique)
Interpret the results for each independent variable as well as for the interaction between the variables.
What does this tell us about the factors we should consider when we are planning teaching methods?
................................................................................................................................................................
................................................................................................................................................................
We will reject the null hypotheses for both independent variables separately as well as for the interaction
effect because the p-values are smaller than 0.05. We were correct in assuming that the effect of
cognitive style on performance is significant and that study technique also relates to performance on the
142
test. In the event of a significant interaction effect (as found here), the following statements are
equivalent: 1) The size of the effect of cognitive style on performance varies in accordance with the type
of study technique used (and vice versa); 2) Study technique influences the effect of cognitive style on
performance (and vice versa); 3) Performance depends on the combination of cognitive style and study
technique. Both variables will therefore have to be considered when planning teaching methods.
Exercise
People produce different kinds of free associations from single words. Some associations are obvious
while others are obscure. For example, when someone is presented with the word “hand”, hand-finger
would be an obvious association, hand-ring (because rings are worn on fingers) an indirect yet plausible
association, but hand-tree (because fingers branch from hands) would be a far-fetched association. A
theory known as the network theory was postulated to explain why different kinds of associations exist.
The network theory states that associations exist as memory networks. Direct associations are closely
linked, but the more indirect the association between two terms the greater the distance in the network
between these terms, and the longer it takes to perceive the association.
Professor Modeko conducted the following experiment. He presented subjects with word pairs. For each
pair the subjects had to find an association between the words in question, and press a button to indicate
that they had done so. The time between presenting a word pair and pressing the button was recorded as
reaction time. The word pairs belonged to three categories, namely direct associations, indirect
associations and far-fetched associations. Professor Modeko used two randomly drawn samples – a
group of 90 schizophrenic subjects and a group of 90 normal subjects. He randomly divided each group
in three subgroups, and randomly assigned these groups to each of the three association categories. His
results are summarised below.
Mean reaction time (n = 30 in each cell)

Group Association category
1 2 3
Normal group 6 9 8
Schizophrenic group 1 2 1
1. Identify the dependent and independent variables. Where appropriate indicate the levels of these
variables.
The dependent variable(s): ....................................................................................................
....................................................................................................................................................
The independent variable(s): ....................................................................................................
....................................................................................................................................................
HMPYC80/104 143
HMPYC80/104/0/2024
2. Draw a graph to show how type of association is related to reaction time for different mental
conditions.
To determine if the observed relationships between the independent and dependent variables are
significant, Professor Modeko used the following results.
The F- and p-values for the difference between the mean reaction time for different mental conditions
and different association types as well as for the interaction effect (n = 180)
Variable F-value p-value

Mental condition 196.84 0.0001
Association type 6.54 0.0001
Interaction 2.18 0.100
3. On which method of data analysis are these results based? ....................................................
....................................................................................................................................................
4. The size of the effect of mental condition on reaction time does not vary in accordance with the
association category. Justify this statement.
....................................................................................................................................................
....................................................................................................................................................
5. Interpret the significance of the interaction effect with reference to the difference between
normal and schizophrenic subjects in respect of verbal free associations. Keep in mind that there
is a significant relation between mental condition and reaction time.
....................................................................................................................................................
144
....................................................................................................................................................
Answers to exercise
1. Professor Modeko manipulated different conditions to see what the effect would be on reaction
time and the dependent variable is therefore reaction time.
There are more than one independent variable in this study. Professor Modeko wanted to know if the
type of word-pair (the type of association) affects reaction time and he studied three types of word-pairs.
He also wanted to know if mental state affects reaction time and he worked with subjects classified as
normal and subjects classified as schizophrenic. The one independent variable is therefore type of
word-pair or association and the possible values or categories of this variable are 1) direct,
2) indirect and 3) far-fetched. This independent variable therefore has three levels used to form three
groups. The other independent variable, mental state, has two levels, namely 1) normal and
2) schizophrenic.
3. This study involves two independent variables and each variable is divided into more than one
level to form separate groups. Professor Modeko wants to know what the differences are between
the groups in terms of their performance on the dependent variable. In other words he wants to
know if normal and schizophrenic subjects differ significantly in terms of mean reaction time. He
also wants to know if subjects who received word-pairs involving direct associations, those who
received pairs involving indirect associations and a group who worked with pairs involving
far-fetched associations, differ significantly in terms of mean reaction time.
A one-way analysis of variance for each independent variable separately (mental state; type of
word-pair) would have provided the above information. However, when we are working with
more than one independent variable, we also want to know if the effect of one of the independent
variables on the dependent variable, depends on the level of the other independent variable; is
there an interaction effect? A factorial analysis of variance is therefore a suitable method to use.
HMPYC80/104 145
HMPYC80/104/0/2024
Because we are dealing with two independent variables we can call it a two-way analysis of
variance.
4. Although there might be an interaction effect, this effect is not statistically significant because
the p-value is bigger than 0.05. Association category therefore does not influence the effect of
mental condition on reaction time.
Remember to work with the p-values and to compare these to the cut-off value to
determine significance – you should not compare F-values or p-values directly.
5. The significant relation between mental condition and reaction time implies that normal and
schizophrenic subjects differ in terms of the time it takes them to make a verbal free association. The fact
that the interaction effect is not significant implies that no matter what type of word-pair is presented, the
schizophrenic group will be faster (see your graph). Schizophrenics are therefore more inclined to verbal
free associations.
146
Simple Regression analysis
Often the dependency-independency <relationship between variables> has to do with our point of view
rather than necessity. For example, if we plot examination results on the Y axis, and motivation levels on
the X axis we imply that higher motivation leads to better examination results. But of course high
examination results could also increase motivation levels, in which case we should have plotted
motivation levels on the Y axis and examination results on the X axis. So, we need to analyse our
problem carefully, determine which variable we consider the dependent variable, and then plot this
variable on the Y axis. In other words determine the dependent variable, follow the convention of
plotting the dependent variable on the Y axis, and name the independent variable X.
Correlational studies consider the degree of linear relationship between two variables and there is often
not a separation in time between observations on the two variables. If we, however, want to use the
values of one variable (independent) to predict values on another variable (dependent) there is usually a
separation in time between observations on the two variables. Remember that if you have a range of
scores for two variables you could plot the data on an X axis and a Y axis. If you find the line that best fit
the data you could use this as a prediction line. The method of analysis used to create a mathematical
expression for the prediction line is called regression analysis.
If the Y variable is a function of the X variable and if one knows the relationship between the Y and the
X variables, one can use this information to predict Y values from X values. The prediction formula is Y
= α + βX and we can use this formula to connect any particular X value with a specific Y value. This is
what is looks like if we set β = 2 and α = 1, and we chose X to be equal to 10:
Y = α + βX
Y = 1 + (2)(10)
Y = 21
In this example the prediction is absolute. It does not provide for any possibility of error. We will go one
step further now and include an error component in our prediction formula:
Y = α + βX + e
We therefore obtain the predicted score with the following equation:
Y’ = α + βX
while the actual observed score would be:
Y = Y’ + e
HMPYC80/104 147
HMPYC80/104/0/2024
For this discussion it is sufficient to consider error as distributed normally around zero (no error) (refer
to <latent and manifest variables> for a discussion on the origins of errors). This simply means that
positive and negative errors are present, and that these add up to zero. Should we repeat the prediction of
Y from X several times we would find that on average Y equals 21 (e = 0 and Y = 21 + 0). However, we
would also find Y to be sometimes less (e < 0) and sometimes more than 21 (e > 0). Due to the error
component, the predicted Y scores vary. The amount of variance associated with the predicted Y scores
is a function of the size of the error component. Smaller error scores leads to more accurate predictions,
whereas larger error scores lead to increased deviations from the “true” Y score. If most error scores fall
between -1 and 1 most of the Y predictions fall between 20 (21-1) and 22 (21 +1). On the other hand, if
most error scores fall between -3 and +3, the Y score predictions will vary between 18 and 24, thus
reflecting less reliable predictions.
Applying the theory
Students doing the directed Master’s in Research Psychology at Unisa decide to do a study on pain for
their group project. They want to know if the pain people expect to feel could be used to predict the
memory people at a later stage have of the actual pain. This problem is about a relationship between two
variables, expected pain (independent) and memory of pain (dependent). The hypothesis could be
nondirectional (the one is related to the other) or directional (a person with a higher/lower rating for
expected pain will score higher/lower on memory of pain).
Dental pain data: predicted pain (PREPAIN) (X variable) and memories of pain (REMPAIN) (Y
variable)
Suppose the research psychology students had the following hypothesis: Predictions of pain and
memories of pain are related and scores of predicted pain can therefore be used to predict memories of
pain. Which method of analysis is suitable to test the hypothesis?
................................................................................................................................................................
Remember that we said if there is a relationship between the Y and the X variables, one can use
regression analysis to obtain a formula to predict Y values from X values? If there is no relationship
between the two variables, the β value in the formula is zero (refer to <hypothesis testing>). The students
are therefore interested in the β value and their hypotheses will be:
H0: β = 0
H1: β  0
The alternative hypothesis is nondirectional because the students only indicated a relation and did not
say whether it is positive or negative.
Twenty male and 20 female volunteers with an average age of 33.9 years underwent dental work that
could potentially be painful. Before the treatment the participants answered questions about their
predicted pain. They made a mark on a 100mm line ranging from 0 = not at all painful to 100 = the worst
pain imaginable to rate their feelings. Five months after the treatment, 31 of the participants responded to
a follow-up survey about memories of pain on the same scale. The following table provides the data for
the two variables for the 31 participants who completed the follow-up survey. The data matrix is
followed by the computer printout using the SAS programme for calculating the β value and the
148
probability of finding this value if the null hypothesis is in fact true (the p-value). These results are in
bold type.
Dental pain data: predicted pain (PREPAIN) (X variable) and memories of pain (REMPAIN) (Y
variable)
PREPAIN REMPAIN
1 1
1 0
9 23
10 10
21 22
26 4
26 31
9 9
24 5
50 17
24 23
12 12
79 20
6 12
35 30
2 1
80 6
11 4
7 1
24 15
1 3
20 25
98 49
99 82
7 22
45 22
49 20
3 23
30 24
10 6
6 11
HMPYC80/104 149
HMPYC80/104/0/2024
Note that the estimate for INTERCEPT is the α value while the estimate for PREPAIN is the β value for
this variable. The probability of finding this β value = 0.0001. The null hypothesis can thus be rejected
because the p-value (0.0001) is smaller than 0.05. One can accept that there is a relationship between
predicted pain and memories of pain. The prediction formula based on the data from this study can
therefore be used to predict memories of pain. Reading the α and β values from the printout gives the
following formula:
Y’ = 6.8280 + 0.3895(PREPAIN)
If a person, for example, indicates 70 on the scale for PREPAIN, we would substitute this value in our
formula and his/her predicted REMPAIN would be Y’ = 6.8280 + 0.3895(70) = 34.
150
Multiple regression analysis
When we are working with more than one independent variable we name these variables X1, X2, X3 .......
Xn. Note: if you have only one independent variable it is not necessary to add the subscript to the X. So
you write X instead of X1. If the Y variable is a function of the X variables and if one knows the
relationship between the Y and the X variables, one can expand the prediction formula used in
<regression analysis> to include more than one independent variable:
Y = α + β1X1 + β2X2 + ….... + βnXn + e
Suppose that research methodology examination results (Y) is a function of motivation (X1) and
intelligence (X2). If we know α and the slope of the relationship between motivation and examination
results (β1) and that between intelligence and examination results (β2) we can predict the examination
results from the motivation and intelligence scores. The error component determines the accuracy of our
predictions.
You may feel somewhat confused here and ask: How can we determine the β values if we do not have
the Y scores to begin with (remember we must predict them)? Of course you are right. We have some
homework to do with regard to our β values if we are interested in calculating predictions. A typical
scenario would be as follows: We would use a former class to determine the β values. In other words we
would, for example, use the 1996 Research Methodology class to determine the β values. We would get
their motivation and intelligence scores and also record their examination performance at the end of the
year. From this we would be able to calculate the α and β values, and determine the size of the error
component. This puts us in the position where in 1997 we can measure Johnny’s motivation and
intelligence and feed this information into the prediction formula to determine (within certain limits of
error) whether Johnny could expect to pass at the end of 1997.
Calculating Johnny’s expected examination mark is a practical application of regression analysis, but it
can also be applied on a more theoretical level. We can determine how much the independent variables
contribute to the dependent variable. For example, we can ask: How much do motivation and
intelligence contribute to examination results? This question actually asks two questions: (a) How
much does motivation and how much does intelligence contribute to examination results, and (b) how
much does the combination of motivation and intelligence contribute to examination results? The β
values carry the information about such contributions. But be warned, it is not always easy to “see” the
information by simply inspecting the β values. If we standardise the motivation and intelligence scores
we can compare their β values directly. Remember that motivation and intelligence scores are measured
on different scales. For example, motivation may be measured on a ten-point scale ranging from 1 to 10,
whereas intelligence may be measured on a scale ranging from 20 to 180. It is obviously difficult to
compare motivation and intelligence scores in such a case. Standardising these scores would be to put
them on a scale where motivation and intelligence scores can be compared directly. The z scale is the
scale we most likely would use. Here any score is expressed in terms of a standard deviation from the
mean. For example, if a motivation score of 7 is one standard deviation above the mean of the motivation
scale, and if an intelligence score of 115 is one standard deviation above the mean of the intelligence
scale, then 7 and 115 could be considered equivalent scores. But, unless we knows that both 7 and 115
are one standard deviation above the mean (i.e. that both have a z value of 1), we would not be able to
consider them equivalent scores. In other words, standardising motivation and intelligence scores as z
values enables us to compare the motivation and intelligence scores.
HMPYC80/104 151
HMPYC80/104/0/2024
This was a long explanation to make a simple point: If variables are standardised we can compare their β
values directly. Which means, if the motivation and intelligence scores are standardised we can simply
look at the β values to determine which variable contributes more to examination results. Suppose β1 =
0.5 and β2 = 0.1, then motivation (X1) contributes more to examination results than does intelligence
(X2). But remember we can compare the β values directly only when we deal with standardised
scores.
There is one other point to keep in mind: Remember that independent variables may also interact with
one another and produce interaction effects (also see <factorial analysis of variance>). We therefore
have to include this factor in our formula. If we have two related independent variables (X1 and X2) and
we also add their interaction effect (X1X2), the formula becomes:
Y = α + β1X1 + β2X2 + βX1X2X1X2 + e
βX1X2 indicates the contribution to Y of the interaction effect between variables X1 and X2.
This brings us to the next question (Question (b) above): How much does a combination of motivation
and intelligence contribute to examination results? In other words: To what extent can examination
marks be explained on the basis of motivation and intelligence scores. We can also formulate this
question by asking: How much of the variance of examination marks is explained by motivation and
intelligence scores? If we know the correlation between variables it is easy to calculate the amount of
variance explained – we simply square the correlation coefficient. In other words, if the correlation
between Y and X is 0.8, then a proportion of 0.82 = 0.64 (which is 64%) of the variance in the Y scores is
caused by X. This may sound very mathematical or statistical, but just be logical about this: In our
example it means that 64% of the variance of examination results is due to motivation scores. In other
words, motivation plays a major role in examination results. Other factors that explain examination
results are responsible for only 36% (100% - 64% = 36%) of the examination result variance. In terms of
proportions we would say: The proportion of the differences in examination results explained by
motivation is 0.64. When we deal with more than one independent variable, we have to make use of a
multiple correlation coefficient I, but the principle remains the same. The proportion of the variance
explained is R2. For example, if the correlation of motivation and intelligence (and their interaction
effect) with examination results is 0.9, then a proportion of 0.81 (81%) of the variance of examination
scores is explained by these variables. Thus a proportion of 0.19 (19%) of the examination result
variance is due to other factors. In our formula the error term I represents these factors.
The scenario can be pictured as fo”lows’
152
Most computer programs will provide the following information when you ask for a multiple regression
analysis:
(1) The β values: β1, β2, ....... βn

(2) The two-tailed probability values for these β values: p1, p2 , ........ pn
(3) An estimate of the population multiple correlation coefficient squared: R2
Let’s assume that for our example we get the following values:
(1) β1 = 6 (Motivation); β2 = 20 (Intelligence); βX1X2 = 13 (Interaction effect for X1 and X2)

(2) p1 = 0.003; p2 = 0.04; pX1X2 = 0.02
(3) R2 = 0.81
Now the picture looks like this:
Here all three β values are significant (refer to <hypothesis testing>). Look at the p-values. In all three
cases the probabilities of obtaining the β values in question (given that H0 is true, i.e. that the population
β values = 0), are less than 0.05. In this case we do not have to divide the p-values by two, because the
alternative hypotheses are nondirectional. A second question: Which variable contributes most to
explaining the variance of variable Y? What do you say: Intelligence, perhaps? Maybe, and then maybe
not! Remember, we cannot directly compare the β values because the size of these numbers is a function
of the values of the scales on which they are measured. For example, if intelligence is measured on a
scale ranging from 20 to 180, and motivation on a scale ranging from 1 to 10, then the β value for
intelligence will probably be larger than the motivation β value. We must standardise the scales to
compare the β values directly.
Now for the important question: What does the above result mean? How do we interpret these figures?
The result means that both motivation and intelligence contribute significantly to examination results,
and that the interaction between these variables has a significant effect. It is extremely important to be
able to interpret results like these. You can ask a professional statistician to help you choose the best
statistical techniques, and you can use powerful computers to perform the calculations, but only you can
make psychological sense of the resulting figures. How do we do this? Let’s assume our example
produced the following data:
HMPYC80/104 153
HMPYC80/104/0/2024
Poorly motivated Indifferent Highly motivated
IQ above 100 40 60 80
IQ 100 and below 40 40 40
The numbers in the cells are examination results. For example, the group who is indifferent and has an
IQ above 100 managed an average of 60% in the examination. We use this information to draw shadow
graphs similar to when we do a <factorial analysis of variance>. There are two walls, namely motivation
(X1) and IQ (X2). Let’s start with motivation. The motivation variable has three levels, namely poorly
motivated (XA), indifferent (XB), and highly motivated (XC). Furthermore the numbers in the cells are
Y axis (examination results) numbers. To draw the graph on the motivation wall we simple have to plot
the three motivation levels of the X1 variable against the Y axis. But note how the IQ variable interferes
by giving us two sets of motivation examination results, namely the motivation examination averages
for the high IQ group (40, 60, and 80) and the motivation examination averages for the low IQ group (40,
40, and 40). In other words, the interference of the IQ variable causes a double shadow on our motivation
wall.
This is what the picture sounds like when you put it into words: This plot shows motivation against
examination results, for different levels of IQ. What do we see? We see that the relationship between
motivation and examination results is different for different IQ levels (the slopes of the two graphs
differ). We also see a significant interaction effect between motivation and IQ. Interaction effects are
fairly easy to spot: When the different graphs share a point (here both graphs share point 40) or when the
lines of the graphs cross each other we “see” interaction between the graphs. (Note that sharing a point
can also mean that the two points are not actually touching but are in close proximity to one another.) A
significant interaction effect means that βX1X2 differs significantly from zero, which means the
154
difference between the slopes of the high and low IQ graphs differs significantly from zero (no
difference between the slopes).
Depending on our point of view, we can study the same situation on the other wall, that is the IQ wall.
Now the interference comes from motivation. It splits the IQ graph into three shadows.
If we put this picture into words it sounds as follows: This plot shows IQ against examination results, for
different levels of motivation.
Applying the theory

Suppose that a group of students doing the directed Master’s in Research Psychology at Unisa wants to
know which variables can be used to predict the memory people at a later stage have of the pain they
experienced. They are interested in the effect of the pain people expect to feel as well as the actual
experience of pain. Their problem therefore concerns the prediction of memory of pain based on the
expectations of pain and the actual experience of pain. The relationship between each independent
variable (expected pain and actual experience of pain) and the dependent variable (memory of pain)
could be directional or nondirectional. The possible effect of the combination of the two independent
variables on the dependent variable (the interaction effect) should, however, also be considered.
The students want to know if they can use the values for variable X1 (predictions of pain) and variable
X2 (experience of pain) to predict values of variable Y (memories of pain). Which method of analysis
would be suitable?
................................................................................................................................................................
................................................................................................................................................................
Because they have more than one independent variable they will use multiple regression analysis to
obtain a prediction formula. Their sample consists of 20 male and 20 female volunteers who underwent
dental work that could potentially be painful. Before the treatment the participants answered questions
HMPYC80/104 155
HMPYC80/104/0/2024
about their predicted pain. They made a mark on a 100mm line ranging from 0 = not at all painful to 100
= the worst pain imaginable to rate their feelings. Directly after the treatment they responded to a
follow-up survey about the actual experience of pain. Five months after the treatment, 31 of the
participants responded to a follow-up survey about memories of pain on the same scale. The following
table provides the data for the three variables for the 31 participants who completed the follow-up
survey.
Dental pain data: predicted pain (PREPAIN) (X variable), actual experience of pain (EXPAIN) (X
variable) and memories of pain (REMPAIN) (Y variable)
PREPAIN EXPAIN REMPAIN

1 1 1
1 1 0
9 1 23
10 2 10
21 15 22
26 12 4
26 1 31
9 1 9
24 2 5
50 1 17
24 17 23
12 1 12
79 2 20
6 3 12
35 28 30
2 10 1
80 19 6
11 6 4
7 5 1
24 97 15
1 1 3
20 1 25
98 1 49
99 44 82
7 10 22
45 21 22
49 3 20
3 13 23
30 39 24
10 5 6
6 2 11
156
To compare the variables so that they can determine how much does prediction of pain and how much
does actual experience of pain contribute to memories of pain, they standardise the variables to the same
scale. Multiple regression analysis provides them with the following computer printout – R2 as well as
the β and p-values are in bold type and are also portrayed graphically:
HMPYC80/104 157
HMPYC80/104/0/2024
Which variable contributes more to memories of pain?
................................................................................................................................................................
We can compare the β values directly because we are dealing with standardised scores. It is clear that the
prediction of pain contributes more to the memory of pain than the actual experience of pain. If we look
at the p-values we also see that only in the case of prediction of pain is the result significant (0.0001 <
.05). Note that a prediction formula based on standardised scores does not include an α value.
Look again at the results of the multiple regression analysis. How much does the combination of
prediction of pain and experience of pain contribute to memories of pain and how much of the variance
is due to error?
................................................................................................................................................................
The multiple correlation coefficient squared (R2) = 0.47. This implies that 47% of the variance of Y is
explained by the combination of these two variables while 53% is due to error.
158
Exercise
Researchers are interested in establishing the determinants of depression proneness in middle-aged

women. They included the following variables in their research design:
Y = Depression proneness
X1 = Self-esteem
X2 = Extraversion/introversion
A multiple regression analysis provides the following results (assume that the variables have been
standardised).
1. On which method of data analysis are these results based? .......................................................
....................................................................................................................................................
2. Which variable contributes more to depression proneness? Explain your answer.
....................................................................................................................................................
3. For which variable/s is the result significant if we look at the p-values? Why?
....................................................................................................................................................
4. How would you explain the variance in depression proneness in middle-aged women?
....................................................................................................................................................
HMPYC80/104 159
HMPYC80/104/0/2024
Answers to exercise
1. The researcher is interested in the strength of the relationship between the dependent variable
and the independent variables. If these relationships are significant, regression analysis can be
used to obtain a formula to predict the dependent variable from scores on the independent
variables. Because the study involves more than one independent variable multiple regression
analysis is used to obtain a prediction formula.
2. The variable self-esteem has a larger β value and therefore contributes more to depression
proneness. The β values of the variables can be compared directly because the scores have been
standardised.
3. The result is significant for the variable self-esteem because p < .05.
4. If a result is presented in this format we know that the value indicated on the arrow pointing from
the independent variables towards the dependent variable is the R2 or the proportion of the
variance of the dependent variable explained by the independent variables. The value on the
arrow pointing from the error term represents the proportion of the variance that is due to other
factors.
Sixty-one percent of the variance in depression proneness in middle-aged women is explained by the
combination of self-esteem and introversion/extraversion and 39% is due to error. Note that your
answer would not be complete if you do not name the variables.
160
G. QUANTITATIVE RESEARCH DESIGNS AND STATISTICS.
Latent and manifest variables

Intercorrelation matrix and Factor Analysis
1 Where does error come from?
When we assume that the true <relationship between variables> takes the form of a straight line, we can
express this relationship in terms of a linear model. Your own research results probably do not form a
straight line but look something like the following (and even then you can count yourself lucky!):
The straight line clearly is an approximation. There are different criteria and methods we can use to
position the straight line so that it represents the scores in the best possible way. Representing a set of
scores in the best possible way means to minimise the amount of error incurred when the line substitutes
the actual scores. Remember once we have positioned the line, it is like saying: From now on all scores
in the set fall in this line. But of course we make mistakes when we do this. In reality all scores do not fall
on the line – they form a distribution around the line. To keep things simple we assume (a) that
everywhere on the line the scores are normally distributed around the line, and (b) that the normal
distributions remain the same for any place on the line.
Assumption (a) means that if we have a very large number of scores most scores will fall on the line and
that as one deviates more and more from the line the number of scores at the particular point of deviation
drops. Here is a picture displaying assumption (a) at different points on the line:
HMPYC80/104 161
HMPYC80/104/0/2024
Assumption (b) suggests that the widths of these distributions are the same everywhere on the line. Thus
we can draw two parallel lines, one on each side of the straight line, depicting the region within which
we are likely to find most of the scores:
If you know what you are talking about, you can position the two parallel lines in such a way that a
specific percentage, for example 95% or 99%, of all the scores is included in the demarcated region.
Now look what happens when one predicts Y from X (refer to the discussion on <regression analysis>):
Here we see that any score x on the X axis is associated with a range of scores (y1 to y2) on the Y axis. In
other words, we use the straight-line graph to find the y value associated with a particular x, but in reality
y finds itself anywhere in the interval demarcated by y1 and y2. We have already considered an example
to this effect, but here is another one: Suppose the limits around the straight line are set at 95% and the
error is 5. Given α and β values of 8 and 2 respectively, we have the following prediction formula:
Y = α + βX + e
Y = 8 + 2(X) ± 5
If X = 10, then:
Y = 8 + 2(10) ± 5
Y = 28 ± 5
162
which means that in 95% of all instances (remember the limits were set to include 95% of all scores) Y
finds itself between 23 and 33.
So what is the origin of this error component? Clearly the size of the error component depends on the
width of the distribution of the scores around the straight line. This reflects measurement error plus
error in the structural part of the larger model. Measurement error is a serious issue, and an entire
sub-discipline known as psychometrics has been devoted to it. So far we have simply taken for granted
that we know the values of the X scores. We have not doubted our abilities to measure the independent
variables. Look what happens when we are not certain about the X value in question:
If, because of measurement error x actually lies in the interval between x1 and x2 (two specific values of
the same variable and not two different variables), the predicted y value ranges from y1 to y2.
2 Manifest and latent variables
Measurement error is mainly concerned with validity and reliability issues, and as you have probably
guessed, these are characteristics of our measuring instruments. Reliability refers to the consistency of
measurements. (If I repeat the measurement, do I get the same result?) Validity has to do with the truth
of measurements. (Do I really measure what I think I am measuring?) The concept of latent variables
(also known as hypothetical variables) is important in the understanding of measurement error. Latent
variables are the variables “behind” the manifested Y and X measurements. Latent variables cannot be
observed directly. In measuring these variables we attempt to observe them, but the Y and X scales of
our observations should not be confused with the underlying latent variables themselves. HMPYC80
performance, motivation and intelligence are examples of latent variables. An examination paper, a
motivation scale and an intelligence test are methods by which we attempt to measure these latent
variables. An examination result is an estimate of an individual’s true HMPYC80 performance.
Likewise motivation scores and Iqs are approximations of motivation and intelligence levels. Below is
an example of the relationship between latent variables and measurement scales (manifest variables):
HMPYC80/104 163
HMPYC80/104/0/2024
Now you can see why, instead of a single score we have to consider an interval (x1 – x2) on the X axis.
The X values are estimations of the underlying latent score, often referred to as the true score. The same
argument holds for the Y scores. Consider the following example: Suppose a student receives 49% in the
research methodology (HMPYC80) examination. The examination mark is an estimation of the
student’s true HMPYC80 performance. If the measurement error is 1 (given a specific confidence level,
let’s say 0.95), we are 95% certain that the student’s true performance falls in the 48 to 50 region. If 50 is
the pass mark one immediately thinks: This student should pass! However, the student’s true HMPYC80
performance score could just as well be 48! See how generous lecturers are? They always give the
student the benefit of the doubt by letting him/her pass!
There is a further decrease in the accuracy of latent Y score estimates when we enter the field of
prediction. When we hypothesise that intelligence has an effect on HMPYC80 performance, we actually
deal with latent variables. Taking into account the different sources of error, this is what happens when
we attempt to predict a latent (true) HMPYC80 performance score from a latent (true) intelligence score.
We begin with a single latent (true) intelligence score and we end up with a range of latent (true)
HMPYC80 performance scores. The extent (width) of the range is a function of error involving both
intelligence and HMPYC80 performance scores. Although we have been practising this kind of
presentation you may find the graph somewhat complicated. So let’s summarise the information in a
different way.
164
If you look carefully at this picture you will see that it contains more information than does the previous
graph. The additional information is captured by the directions of the arrows linking the latent variables
with manifest variables (the measured scores). These arrows indicate that the X scores are caused by the
latent variable, intelligence, and that likewise the Y scores are caused by the latent variable, HMPYC80
performance. Here we run into philosophical issues. In pointing the arrows as we have, we adopt a
particular philosophical position. We hold the view that there is an underlying reality of latent variables,
and by applying the right methods we expect to gain access to these hidden truths.
Let’s investigate the rest of the picture: The arrow pointing from intelligence to performance marks
HMPYC80 performance as the dependent latent variable and intelligence as the independent latent
variable. We also see that, apart from intelligence, error I contributes to performance. This source of
error has to do with the fact that intelligence is not the only variable contributing to performance. The δ
and ε symbols reflect measurement error: δ signifies the difference between an observed X score and its
true latent value, and likewise ε indicates the difference between an observed (or sometimes predicted)
Y score and its true latent value. In the linear model (remember Y=α+βX+e) we assume that a latent
variable (e.g., intelligence) can be measured directly as X (IQ) without error. We allow for error in
measuring Y, but the error component in question I does not distinguish between error due to other
important variables being left out and measurement error. We have now refined our thinking to reflect a
more realistic picture: our hypothesis takes the form of the proposition of a causal relation (or relations)
between latent variables, where the latter are imperfectly measured by manifest variables, and where it is
assumed that no proposition of the causal determinants of any latent dependent variable can hope to
include all predictor (independent) variables in the model.
However, we have been working with a somewhat theoretical picture; in practice we cannot determine
the latent values. We cannot get beyond the X and Y measurements in real situations. Therefore we run
into a problem, namely: How do we calculate the extent of error involved if we cannot determine the true
scores? The solution is fairly simple: We use more than one measurement for each latent variable. Thus,
our previous picture expands in the following way:
Here X1, X2 and X3 are three measurements of intelligence, and Y1, Y2 and Y3 are three measurements of
HMPYC80 performance. The measurements (of each latent variable) may refer to different tests or they
may be repetitions of the same test. For example, in measuring HMPYC80 performance we can repeat
an examination paper or we can use three different types of papers, such as a paper consisting of multiple
choice items, a paper requiring practical work, or an oral examination. So far so good. But how do we get
HMPYC80/104 165
HMPYC80/104/0/2024
to the latent variables via these multiple measurements? What do these six measurements in our example
tell us? If we organise them properly, they can tell us quite a lot!
We are really interested in the relationships between these measurements. So let’s organise the
variables in such a way that we can easily see the relationships involved. A grid would give us a clearer
picture. Here is what we could expect:
X1 X2 X3 Y1 Y2 Y3
X1 high
X2 high high
X3 high high high
Y1 low low low high
Y2 low low low high high
Y3 low low low high high high
Note that we need to complete only half the grid because the upper part reflects the same information.
This type of grid is often called an intercorrelation matrix. We usually express the relationships
between variables in terms of <correlation coefficients>, so let’s use “high” and “low” to indicate good
and poor relationships between variables. Inspect the grid. What do you see? Yes, as could be expected
you will see high correlations between manifest variables going together with a particular latent variable
(it is like belonging to the same family), and low correlations between manifest variables going together
with different latent variables.
Of course we have not really been practical here. We are still dealing with the relationship between latent
and manifest variables as if the latent variable values are known to us. Now imagine the opposite
approach. Suppose we obtain the measurements of the six manifest variables, and suppose we calculate
the necessary correlations, and organise the information in the form of an intercorrelation matrix.
Inspecting the patterns of intercorrelations we see three distinct groups: Two groups of high
intercorrelations (one among X variables and one among Y variables) and a group of low
intercorrelations (among X and Y variables). This pattern might suggest (a) that the X variables belong
together, (b) that the Y variables belong together, and (c) that the groups of X and Y variables do not
belong together. It would be logical to conclude that two different latent variables are responsible for this
pattern. The intercorrelation matrix makes us aware of the presence (or perhaps absence) of particular
latent variables.
This seems simple enough, but the scenario may easily become more complicated. An obvious
complication would be to add more measurements, but even in the case of a few variables you must
remember that the intercorrelation matrix usually contains the numerical values of the correlations, and
these are more difficult to read than simply having to deal with “high” and “low” values. Furthermore
the correlations seldom constitute clearly discernible patterns.
For this reason latent variables are unearthed by means of statistical analysis. We can use a technique
such as FACTOR ANALYSIS to see which manifest variables belong together in a family (factor or
latent variable). This technique allows us to look at the manifest variables on the basis of such factors.
166
There are also computer programs such as LISREL, that estimate the β values for each of the linear
relations depicted in our model. The program gives:
(1) The β values (and associated probability values) for the three arrows pointing from intelligence
to the X variables, and it estimates the three δ error variances (thus it describes the
measurement model for intelligence);
(2) The β values (and associated probability values) for the three arrows pointing from HMPYC80
performance to the Y variables, and it estimates the three ε error variances (thus it describes the
measurement model for HMPYC80 performance);
(3) The β values (and associated probability values) for the two arrows from intelligence and e to
HMPYC80 performance (thus it describes the structural part of the larger model);
(4) An estimate of the reliability (consistency) of the measurements of intelligence and HMPYC80
performance.
A program such as LISREL allows us to draw pictures of our models and it then proceed from these
drawings to estimate the parameters in question. Behind the scenes the statistical procedures executed by
such programs can be pretty complicated, but as psychologists we do not have to fully comprehend these
procedures. However it is our responsibility to describe psychological problems and to set up the
necessary “drawings” for the computer. And when the computer gives the statistical values it is our job
to make psychological sense of the data. Remember the computer is a number cruncher, and the
statistician knows the rules whereby numbers are crunched. But only the psychologist can assign proper
meaning to these crunched numbers.
3 A note on manifest and latent variables
You may feel slightly confused: So far we have referred to a latent variable as if such a variable is a
single and solid truth, but I am sure you have an inclination that things are not always this
straightforward. In the examples discussed above we referred to three measurements of intelligence,
thereby implying that each of these measurements is an estimate of the true intelligence score. We can
always include more latent variables in our model. We could, for instance, add motivation and memory
as two additional latent variables. To the extent that these variables contribute significantly to
HMPYC80 performance, error decreases. If we have two measurements of each latent variable the
model looks like this:
HMPYC80/104 167
HMPYC80/104/0/2024
Don’t be confused by these ideas: We use X1 for motivation, X2 for intelligence, and X3 for memory,
and as before we distinguish the different measurements with subscripts. Thus X11 and X1 2 are the two
measurements of the latent variable, X1 (motivation).
Expanding the model in this way is quite logical, but keep in mind that we are still looking at each latent
variable as a singular entity. However, research can also focus on the structure of a latent variable, which
occurs frequently in the developing of new measuring instruments. In this case the latent variable could
be considered in terms of more than one latent variable. Intelligence, for example, can be viewed as
verbal intelligence and non-verbal intelligence. We can now have multiple measurements of verbal
intelligence and multiple measurements of non-verbal intelligence.
There are two points to this discussion:

(1) Don’t get confused when a latent variable is divided into multiple latent variables. Simply think
of these “splinter” latent variables as latent variables in their own right.
(2) However, note the theoretical implications for the relationship between manifest variables:
Suppose we view intelligence as a singular latent variable, and suppose we have two
measurements of intelligence. Suppose further that we are unaware that these are measurements
of verbal and non-verbal intelligence. What would the intercorrelation matrix show? On the one
hand the two measurements have intelligence in common. Therefore, they should be correlated.
On the other hand, they measure different aspects of intelligence and should therefore not be
correlated. To see the differential effects of verbal and non-verbal intelligence, we need to view
them as separate latent variables and obtain multiple measurements for each. This kind of study
is known as research into the construct validity of a latent variable (in this case, intelligence).
Applying the theory
The following problems usually occur when one is constructing an instrument to measure certain
constructs. Suppose you want to measure memory. Memory has, however, been defined in terms of three
different processes, namely short-term memory, intermediate memory and long-term memory. As
memory deficits are critical indicators of brain damage it is important to evaluate all three these memory
processes. These constructs can, however, not be directly observed (we call them latent variables). You
therefore use a test to measure short-term memory (we call the score on the test the manifest variable).
Because of measurement error you decide to play it safe and to use three different tests to obtain an
indication of each person’s short-term memory. You do the same for intermediate and for long-term
memory. You expect a relationship between the underlying constructs and the three measures of each
construct. You also expect the tests that measure the same process to be related (to correlate) but you do
not expect high correlations between tests that measure different constructs.
The following are scores (on a 10-point scale) on three different tests for each of the three memory
processes for a random sample of 10 children who do not have brain damage. These tests are: STM-1,
STM-2 and STM-3 for short term memory; IM-1, IM-2, and IM-3 for intermediate memory; and
LTM-1, LTM-2, and LTM-3 for long term memory.
168
Memory data: scores on STM-1, STM-2, STM-3, IM-1, IM-2, IM-3, LTM-1, LTM-2, and LTM-3
STM-1 STM-2 STM-3 IM-1 IM-2 IM-3 LTM-1 LTM-2 LTM-3
9 9 8 5 6 8 3 2 2
4 4 5 6 5 5 4 5 6
5 5 5 6 5 5 5 4 4
7 8 7 9 10 7 7 7 8
5 3 4 8 9 3 5 5 5
2 1 2 3 2 2 4 3 3
5 6 6 9 10 5 4 3 3
5 5 4 4 4 5 9 10 10
1 2 2 6 7 1 7 8 7
9 9 8 4 4 9 1 1 1
The following is a computer printout using the SAS programme for calculating correlations between
variables. The results are also displayed in an intercorrelation matrix reflecting only half the grid to
make it easier to interpret.
HMPYC80/104 169
HMPYC80/104/0/2024
170
Intercorrelation matrix for the three sets of memory tests (n=10)
STM- STM-2 STM-3 IM-1 IM-2 IM-3 LTM-1 LTM-2 LTM-3

1
STM-1 1.00
STM-2 .94 1.00
STM-3 .95 .97 1.00
IM-1 .06 .15 .22 1.00
IM-2 .13 .22 .26 .96 1.00
IM-3 .95 .96 .96 .00 .03 1.00
LTM-1 -.44 -.34 -.49 .18 .20 -.44 1.00
LTM-2 -.45 -.34 -.48 .13 .15 -.42 .96 1.00
LTM-3 -.37 -.27 -.39 .19 .17 -.32 .94 .98 1.00
Remember that you would expect high correlations between the manifest variables (tests) going together
with a particular latent variable (memory process), and low correlations between manifest variables
going together with different latent variables. Indicate if you think the following tests are good measures
of the particular latent variables and justify your answers.
STM-1: ..............................................................................................................................................
IM-3: ....................................................................................................................................................
Did you see that the manifest variable STM-1 is a good measure of the latent variable short-term
memory because it has high correlations with STM-2 (0.94) and STM-3 (0.95) and low correlations with
most of the other tests that measure different latent variables, namely intermediate and long-term
memory? The test IM-3, however, is not a good measure of intermediate memory. The test correlates
poorly with IM-1 (0.00) and IM-2 (0.03) and it has high correlations with STM-1 (0.95), STM-2 (0.96)
and STM-3 (0.96) that measure a different latent variable, namely short-term memory. Did you notice
that there is a good relationship between IM-1 and IM-2 and these tests can therefore be regarded as
measures of intermediate memory?
---------------------------------------
Let us consider another problem involving the measurement of latent variables. Suppose you designed a
test battery (more than one test) to measure a certain construct. You would like to know if your manifest
variables (test scores) are in fact all measuring the same construct (latent variable). On the other hand,
you might have a battery of tests measuring different things but you would like to know if there is
something similar about these tests (or some of them) and if so what this similarity is. In other words:
What is the underlying construct or latent variable that these tests are measuring?
A researcher at Unisa’s Institute for Behavioural Sciences administers six tests of cognitive abilities to a
large number of randomly selected students. Each of these tests can be regarded as a variable. Based on
theory, he suspects that the six tests are not measuring six but a smaller number of variables. He will
HMPYC80/104 171
HMPYC80/104/0/2024
have to determine between which tests (or groups of tests) relationships exists and then he has to
interpret these relationships in terms of the construct(s) that the related tests have in common.
The tests and their labels are: vocabulary (V), reading (R), synonyms (S), numbers (N), arithmetic
(standardised test) (AS), and arithmetic (teacher-made test) (AT). After the tests were administered and
scored, he represented the data in a data matrix with a score on each test for each student. We will not
report it here as this data set is very large. The table will look similar to the data matrix in the previous
problem. He then calculated the correlation coefficients between the tests. The correlations between the
six tests have been summarised in a matrix.
Intercorrelation matrix
Variables V R S N AS AT
V 1.00
R .72 1.00
S .63 .57 1.00
N .09 .15 .14 1.00
AS .09 .16 .15 .57 1.00
AT .00 .09 .09 .63 .72 1.00
In the previous problem we said that those tests that correlate have something in common, they measure
the same underlying construct or latent variable. Correlation matrices are however often large and it
becomes impossible to interpret the pattern of correlations just from an inspection of the correlations.
The researcher will use a statistical technique such as FACTOR ANALYSIS which indicates how
many factors (latent variables) the battery of tests measure and also which tests belong with which factor
(latent variable). The following is the solution of the FACTOR ANALYSIS .
Factor solution matrix
Tests Factor 1 Factor 2

V .83 .01
R .79 .10
S .70 .10
N .10 .70
AS .10 .79
AT .01 .83
This means that two factors or latent variables have been extracted. The values in the table are called
factor loadings. The idea is to identify for each factor the tests (manifest variables) with high loadings on
the particular factor.
172
To help you identify variables loading highly on factors, mark loadings with an absolute value higher
than 0.3 with “*”, while leaving all other loadings blank. Do this in the following table: (Please note that
0.45 and -0.45, for example, both have an absolute value higher than 0.3.)
Exercise: Simplified factor solution matrix

V
R
S
N
AS
AT
Did you complete the table with an “*” under factor 1 for V, R and S and under factor 2 for N, AS and
AT? It is even more important to give a name to each factor, in other words, to identify the two latent
variables measured by the six tests. The argument is that the variables (tests) that have a high loading on
the same factor must have something in common which caused them to cluster together in the correlation
matrix in the first place. It is this common source which makes up the derived factor and the name we
give the factor represents a psychological construct. In fact we are giving theoretical meaning to this
factor. Giving theoretical meaning to a factor is so important that even if a variable loads highly on a
factor, but one cannot explain why it belongs there because it appears to be quite different from the other
items with high loadings on that factor, one should rather exclude it in the definition (derivation) of the
factor.
Have a look at the tests that load highly on the first factor. They are vocabulary, reading and synonyms
tests. In what way are these tests similar? What do you think is the common source that makes up factor
1? Try to give theoretical meaning to this factor.
................................................................................................................................................................
................................................................................................................................................................
Can you see that the three tests all involve “words” in some way? The common source might therefore
be verbal ability. Name the three tests that load highly on the second factor and try to give theoretical
meaning to this factor.
................................................................................................................................................................
................................................................................................................................................................
These tests are numbers and the two arithmetic tests. We might conclude that the common source for
these three tests is numerical ability. These tests can therefore be regarded as measures of the latent
variable “numerical ability”.
HMPYC80/104 173
HMPYC80/104/0/2024
Exercises
Practitioners at Bubba’s Mental Health Clinic routinely administer five psychological tests to patients
who are admitted with complaints of depression: The Jeff Abdul Depression Inventory, the Kierkegaard
Existential Anxiety Inventory, the La Pew Personal Hygiene Scale, the Midas Greed Scale, and Wundt’s
Inventory of Introvertive Inclination. Dr Bubba, hoping to cut costs associated with administering all of
the tests, has requested a FACTOR ANALYTIC study of the five measures. It is his hope that this study
will identify redundancies between the tests, leading to the elimination of some of them. The results are
summarised in the following two tables.
Variables Depression Anxiety Hygiene Greed Introversion
Depression 1.00
Anxiety -.13 1.00
Hygiene -.17 -.45 1.00
Greed .54 .05 -.61 1.00
Introversion -.13 .97 -.52 .14 1.00

Depression .12 .82
Anxiety .85 -.14
Hygiene -.81 -.23
Greed .30 .77
Introversion .90 -.20
1.1 Identify the manifest variable(s) and the latent variable(s) associated with this study. Justify your
answer briefly.
....................................................................................................................................................
....................................................................................................................................................
1.2 Interpret factor 2, in other words give theoretical meaning to this factor.
....................................................................................................................................................
....................................................................................................................................................
174
---------------------------------------
The SAT (Stanford Achievement Test) consists of a battery of seven tests, namely: (1) Vocabulary, (2)
Reading, Part A, (3) Reading, Part B, (4) Word Study Skills (5) Mathematical Concepts, (6)
Mathematical Computations and Application, and (7) Listening Comprehension. A psychologist
analyses the intercorrelations between these seven tests, and obtains the following intercorrelation
matrix.
Variables Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7
Test 1 1.00 .63 .62 .63 .64 .61 .61
Test 2 1.00 .87 .86 .61 .62 .50
Test 3 1.00 .83 .59 .62 .51
Test 4 1.00 .64 .63 .52
Test 5 1.00 .73 .61
Test 6 1.00 .57
Test 7 1.00
The psychologist then performs a FACTOR ANALYSIS, which yields the following factor solution
matrix.

Test 1 .76 .17
Test 2 .89 -.33
Test 3 .87 -.28
Test 4 .87 -.24
Test 5 .80 .30
Test 6 .78 .24
Test 7 .68 .30
Answer the following questions on these results:
2.1 Consider the factor solution matrix and then give meaning (i.e. a name) to factor 1. Justify your
answer briefly.
....................................................................................................................................................
....................................................................................................................................................
2.2 Interpret the results obtained by explaining briefly what conclusions the psychologist might draw
from this study. Base your answer on the data in the intercorrelation matrix and the factor
HMPYC80/104 175
HMPYC80/104/0/2024
solution matrix (i.e. refer to both matrices in your answer).
....................................................................................................................................................
Answers to exercises
1.1 Five psychological tests were administered and scores were obtained on these tests. The results
on these five tests are the manifest variables because these results were not just implied but they
were actually observed.
Underlying each test result there is a personality characteristic that is “causing” the test result.
Tests that have the same underlying characteristic will correlate. One can study a intercorrelation
matrix to see which tests are related. To make this pattern of relationships clear, FACTOR
ANALYSIS is based on the matrix and tests that have the same underlying characteristic will
now also load on the same factor. The two factors identified in the factor solution matrix are
therefore the latent variables or the underlying psychological constructs believed to “cause” the
results on the five test.
1.2 The tests Depression and Greed load on factor 2 and a psychological construct that explains
performance on both these tests should be identified. In other words: What does these two tests
have in common? The characteristic measure by this factor could be a somewhat more worldly
form of depression than that represented by factor 1 and it is linked to insatiable greed.
It is not sufficient to say that this factor consists of depression and greed. Some of the innovative
answers linking “depression” and “greed” that were received from students are: a feeling of
insecurity; lack of satisfaction; and concerned with the self.
2.1 All the tests load highly on factor 1 and we need to identify a construct that probably underlie or
“cause” performance in all these tests. If we look at the type of tests included in this battery we
see that they all deal with abilities learned at school and we thought that this factor can be called
“academic ability”.
Please note that we obtain “loadings” on a factor, not “scores”, and that tests that correlate will
load on the same factor (factors and tests do not correlate).
2.2 The intercorrelation matrix indicates the relation between the various tests. In other words each
correlation coefficient indicates the extent to which two tests are related. A higher correlation
between two tests indicates that these two tests have the same underlying characteristic. All the
values in the intercorrelation matrix are average to high (above 0.50) indicating that the tests are
all related and therefore all measure the same underlying characteristic. (Please note that a
correlation coefficient indicates the strength of a relationship between two variables and does not
refer to how well subjects performed on a test.)
In the factor solution matrix tests that have the same underlying characteristic will load on the
same factor. Because all the tests have high loadings on factor 1 and relatively low loadings on
factor 2, they seem to measure the same common factor (see question 2.1). (Remember that we
work with the absolute value of the loading of a test on a factor.)
176
The psychologist could conclude that the tests measure the same common factor because a) based on the
values in the intercorrelation matrix (0.50) the tests are all related and b) all the tests have high loadings
on factor 1 and relatively low loadings on factor 2.
HMPYC80/104 177
HMPYC80/104/0/2024
Qualitative analysis
There are many different traditions of qualitative data analysis (e.g. phenomenology and grounded
theory) each of which typically suggest quite detailed analytic procedures that may be followed. There
are also some general principles and techniques that apply to most forms of qualitative data analysis.
The purpose of qualitative analysis is usually to obtain a “rich” understanding of the material. In other
words, qualitative analysis places more emphasis on “sense meaning” and understanding a phenomenon
or experience rather than predicting its occurrence. There is no objective, external instrument that will
do the work of generating such an understanding for you - you have to bring your own subjective
capacities to bear. Data analysis can either be inductive or deductive in nature. According to Kawulichi
and Holland (2012), you can follow an inductive approach when you want to describe or explore some
phenomenon, while a deductive approach can be used when you are already familiar with the setting and
want to confirm or explain a phenomenon, based on several cases. Inductive approach works well if you
want to go from specific to general, which means approaching the data analysis from data level, then
looking at them more broadly. In this case, you would identify codes from the raw data rather than
relying on pre-conceived codes from existing literature or from your own beliefs and knowledge of the
topic. Deductive approach would involve looking at the study from a general standpoint, and move to the
specific data. In other words, you may use codes that have already been used from other studies or even
apply codes based on your own specific knowledge. Your choice of which approach to use, will be
influenced by the theoretical framework (theory supporting your study) and the purpose of the study
(what does the study aim to achieve, investigate or highlight).
The methods of qualitative analysis are therefore not so much procedures for manipulating data, as
techniques for disciplining your thinking to help you move closer to your subject and beyond
common-sense interpretations. Here are some suggestions:
I. Immerse yourself in the material
You need to become thoroughly familiar with your data, and the only way of doing so is literally to
live with it. Immersing yourself in the material will help you to obtain a general impression, feeling
and an overview of the material. examples of qualitative approaches which emphasise immersing
yourself in the data are Ethnography which depends on becoming a part of the setting being studied
and Grounded Theory (Glaser & Strauss, 1967) which talks of working from the “bottom up” rather
than the “top down” - that is to allow your analysis to be guided by the data rather than by theoretical
considerations. If you're working with interview transcripts or some other textual material, a good
first step is always to simply read through the material several times until you are thoroughly
acquainted with it. If you feel you need more structure at this stage, you could consider using the
“grand tour” and sub-questions you generated when formulating your research project as a reading
guide.
II. Bracket your preconceived ideas
178
III. <Bracketing> (temporarily putting one’s preconceptions to one side) is a good technique for
cultivating an open and empathic attitude towards one’s data. Don't wait until the end.
Analysis works best if you start doing it while you're collecting the data. Yes, this will “bias” the data
collection process - leading to more useful and interpretable material. One of the distinctions that
separate qualitative from quantitative analysis is that during the data collection, the researcher might
have a sense of common experience which may later be reported as themes. This does not mean that
you will merely identify a theme and then collect data to support or justify it, but rather as you collect
data, you need to be consciously aware of what could be the major findings. The following questions
can help you to deal with this process:
IV. What I am hearing, seeing, sensing, feeling, thinking and intuiting with regard to the
participants?
V. What am I hearing, seeing, sensing, feeling, thinking and intuiting with regard to the data?
VI. What is common among participant’s experiences? (What is unique?)
VII. Do I have enough data to justify it as a common experience?
Follow a cyclical analytic process
Unlike some forms of quantitative research which move in a linear fashion from an hypotheses to a
formal test of the hypothesis, qualitative research typically cycles through this process many times in
quick succession. For instance, after reading through a set of interview transcripts with people in an
old-age home, a researcher may form the impression that although most of their complaints focus on
material difficulties they seem more genuinely concerned about no longer being considered
productive members of society. She then re-reads the material to see if such an interpretation can be
supported - carefully noting passages that seem to confirm this as well as counter-examples. While
doing this, she notices that men appear to complain more than women, and again checks carefully to
see if this is the case. She now starts to wonder if homing in on complaints in this manner gives a fair
reflection of these old people's lives and decides to try and see what, if any, positive elements are
mentioned by the old people. In this way she gradually forms a rounded picture of their life-worlds.
VIII. Ask: “what is missing?”
Silences are often as important as positive content. What is not present? Who does not speak? Why
not? (For example, if your material comes from community meetings where everybody supposedly
has a voice, you may notice that in fact the meetings tend to be dominated by middle-aged and older
men.)
Look for the obvious
What aspects of the material are so obvious that they go-without-saying? What are the basic
assumptions which imbue the material with common-sense meanings?
______________________________________________________________________
Looking for the obvious
It has been said that the purpose of qualitative research is to make the strange familiar
and the familiar strange. Try the following exercise to get a feel for what this could mean
HMPYC80/104 179
HMPYC80/104/0/2024
in your own life: Next time you have a meal with your family (or with people you know
well) observe them closely while imagining that they are from another planet and that
although you understand their language you know nothing else about their ways. Is the
food dished into separate bowls or plates or does everybody eat from the same pot? What
kinds of food do they eat and in what sequence? What are the seating arrangements, or do
they stand while they eat? If there is conversation, see if you are able to understand what
they're saying to each other - remember, you know the language, but nothing else about
these people. If you put some serious effort into this exercise, always looking for what
would ordinarily seem obvious, you may find that you've been living with some rather
strange people all these years.
______________________________________________________________________
IX. Look for the unique
The purpose of the analysis is not only to find commonalities across different accounts, but also to
highlight contradictions. You aren't necessarily expected to produce a single, definitive version of
what is going on but rather you are also expected to highlight contrasting experiences.
X. Throw away some of the data
You will in any case always have more than you need. Do not turn into a data fetishist.
XI. Free associate
This normally works best if you do it with friends rather than on your own. You could also try doing
it with people noted for their complete lack of official expertise. The trick is not to be overly critical
at this stage. Try and generate as many ideas about what is going on in your data as possible -
whether they appear non-sensical or not. You can make time to be critical later.
XII. Restructure the material
Qualitative data is usually quite unstructured, and novice researchers often fall into the trap of
immediately wanting to impose some structure, for instance by assigning codes to recurring
phenomena and then counting how many times they occur. Such restructuring can be very useful, but
it is important to remember that there are always numerous different ways in which the data can be
restructured, and simply performing one arbitrary restructuring does not constitute an analysis.
Remember, the main aim of qualitative analysis is to break the whole into separate parts in the
material and then re-assemble these parts to form a coherent meaning. Thus, purpose is not to find
the best (or correct) structuring of the material, but to help you gain new insights by presenting things
in different ways. Restructuring works by reordering material and grouping similar things together,
for example by cutting and pasting, using marker pens, or assigning codes in a computer program for
qualitative analysis. Try to categorise the material using a variety of different criteria and levels of
analysis - e.g. different kinds of content, different themes, different linguistic categories. Don't fixate
on one coding scheme too quickly. Play around with simple and complicated coding schemes,
cross-tabulations, diagrams and hierarchical orderings of the material. If you're feeling creative, try
re-writing the material as a play or a story. By re-casting the material, particular features that may at
first have seemed natural and inevitable may become highly significant.
180
Below are three extracts from much longer interviews with South African women being treated for
breast cancer (Berger, 1990). First read through the extracts in sequence:
Fathima: Most of the friends I have that's got a bust removed it's all people that hasn't got
children... If they can't bring children in the world maybe it's something that's
affecting your bust because the first thing the doctor asked me is have you got
children... If somebody could've told me how it happens that you get breast cancer,
you know if you could've been on the alert hey... You could've looked more after your
breast, and I wouldn't have let it go on for so long when I went to --- Hospital.
Gertrude: She must be careful... because they say a woman, I did read it in a book, must have a
checkup regularly by a doctor for breast cancer... because now Patty is my daughter
and I don't know if it will travel, you know what I mean, if it will affect her also.
Ann: I was reading about it, if a mother don't breastfeed her babies, now my babies were all
bottle babies. Now I don't get the, I don't know what you call it, the baby must now
drink on you. Say the exercise inside your breast, where the milk comes from... maybe
there was always something, a clot, whatever's in your breast that came together after
all these years. I thought that. It's worthwhile that the baby drinks on you, for your
own health.
There are of course many ways in which you could restructure such data. Berger (1990) used three levels
of illness meaning - symptoms, explanatory models and life world. The above extracts have all been
identified as relating to explanatory models, and in a sense this material has therefore already been
restructured. However, we would like you to try some further restructuring. Could you identify three
explanations for breast cancer given by these women?
Explanation 1:
................................................................................................................................................................
Explanation 2:
................................................................................................................................................................
Explanation 3:
................................................................................................................................................................
Now go back to the extracts and underline and number sections of talk that relate to each of the
explanations. (Hint: Each woman may give more than one explanation.) Does this restructuring enhance
your insight into the women's understanding of their illness?
The three explanations we identify are not having children (Fathima and Ann), not having regular
check-ups (Fathima and Gertrude), and hereditary causes (Gertrude). You may have different categories,
e.g. not having children and not letting your children breastfeed could be two separate categories. If you
had more material you could at this point cross-tabulate explanations with sources of information (some
of the sources mentioned are inferences from doctors' questions, observation of friends' illnesses, books,
HMPYC80/104 181
HMPYC80/104/0/2024
and unspecified other reading material). Alternately, you could perhaps decide to restructure the material
in a different way, for example. by coding all the instances where women appear to take the blame for
their illness. You could also try and re-write each interview as a story told by the woman involved, or
you could try and write one story that incorporates as much as possible from the different women's
experiences. Remember that you are not trying to prove something about the material, but to rearrange it
so as to understand it better.
Involve the “subjects” in the analysis
Take partial analyses back to the research participants and get their opinions. (Imagine how Berger could
have enhanced her analyses if she asked the women to comment on them.) Consider engaging in
participatory action research, that is involving the “subjects” in all aspects of the research, including the
analysis, and feeding the results directly back into social action.
Be reflexive
Keep a diary. Concentrate on the process. Rework your analysis in such a way that it explicitly includes
a description of your own role in the production and transformation of the data.
Treat your analysis as data
Revisit the suggestions listed above, but this time think of your analysis as the data to work on. Repeat.
Additional sources
Miles, M.B., & Huberman, A.M. (1994). Qualitative data analysis. An expanded sourcebook. Sage.
HMPYC80 (Tutorial Letter 102) Resource 02: Research Review Inventory. Unisa.
References
Berger, S. (1990). An "inside story" - the illness experiences of women with breast cancer. Unpublished
MA dissertation, University of Cape Town.
Glaser, B.G., & Strauss, A.L. (1967). The discovery of grounded theory. Strategies for qualitative
research. Weidenfeld & Nicolson.
Ryan, G. W., & Bernard, H. R. (2003). Techniques to Identify Themes. Field Methods, 15(1), 85–109.
https://doi.org/10.1177/1525822X0223956
182
Text: What is it?

What is a text? Text is one of those ordinary words that has gained a heavy load of philosophical
meaning in recent years. Text usually refers to a piece of written material. But have a close look at the
page you are reading at this moment. What do you see? Sentences and words? Look closer. Yes, what
you see are nothing but marks on paper, but these marks are not simply randomly drawn images. They
mean something. They are symbols, chaining together to form words and sentences. Now imagine
somebody speaking to you. You hear the words. What are those words? Listen closely: they are only
sounds. They are marks (sound images), impressions made on the air that surrounds us. Do you get the
picture? A text is a collection of marks that have meaning. The meaning emerges from the text in a kind
of chain action. Letters or sounds are stringed into words and words are stringed into sentences.
Let’s make it real difficult now. Watch two people who are walking down the street. If you watch them
for a while you can tell whether they know each other, from the way they interact. Do you realise that
you have just been reading a text? For the observer, the persons' movements are marks, stringed into a
meaningful text.
If we look at text in this way we come to realise that the world surrounding us is a collection of texts.
This is the first lesson: Research projects lurk everywhere in the world around you. You only have to
learn to read the text. But before we become more practical about producing and reading texts a brief
note on different philosophical meanings of text. There are essentially three ways of making sense of the
texts we see around us:
· As referring (however inaccurately) to the real world (this is a positivist position)
· As reflecting (however indirectly) the intentions, experiences, thoughts and feelings of the
people involved (the hermeneutic position)
· As producing and sustaining certain truths about the people and the world in which they find
themselves (the semiotic position).
Do not be too concerned if these statements do not make much sense at this stage. Simply keep in mind
the common notion of a text, namely text as a piece of written material. Then remember that such texts
can be treated as if they represent (refer to) a real world, or they can be treated as if they capture the
intentions and experiences of the author of the text, or in the third place they can be treated as if they
produce and sustain the author and his world.
From raw texts to research texts

It is one thing to say that we are surrounded by texts, but to capture those texts is a different kettle of fish.
Where and how does one begin? One can begin anywhere: an interesting book, newspaper articles, radio
talk shows, television news, soap operas or sitcoms, people at work, or on the sports field, or in the park.
Once you have become aware of the many texts surrounding you, you can start taking notes, file
clippings, save recordings and keep artefacts. The notes and stuff you collect are raw texts. From these
texts you can produce a research text. Progressing from raw texts to a research text means the material
has to be interpreted and analysed. There are no prescribed guidelines for producing research texts, but
an important prerequisite is that one has to get to know the raw material. One has to get involved, reflect
about the material, and work with it. The more familiar the material the easier it is to analyse and
interpret.
HMPYC80/104 183
HMPYC80/104/0/2024
Observation
Observation as a method of collecting material is often used in sociology and anthropology - although as
yet less often in psychology. Introductory research methods textbooks are rich sources of advice on the
practicalities involved. Should you observe as an outsider or attempt to blend in with the surroundings
and become a “participant observer”? How can you best gain access to the research setting? What should
your “cover story” be? Should you take notes while “in the field” or write down your impressions
afterwards?
Additional sources
Breakwell, G.M., Hammond, S., & Fife-Schaw, C. (1995). Research methods in psychology. Sage.
See chapters 14 (Direct Observation) and 20 (Ethnographic Research).
184
Interviews
One-on-one interviewing is a very popular way of collecting research material, and there is a wealth of
advice in introductory textbooks on how to go about doing interviews. There are also some techniques
from counselling psychology (such as reflecting content and feelings, using minimal encouragers,
making eye-contact, etc.) which you may want to try out. Interviewing techniques vary enormously
depending on the kind of information you're after and your theoretical orientation. A phenomenologist
may want to ask only a single broad question at the beginning of an interview, steering the rest of the
interview as little as possible. Somebody who is after people's ideas and opinions on a number of quite
specific issues, on the other hand, may opt for a semi-structured interview with predetermined questions
and a fairly fixed sequence. Group interviews around a particular topic (often called focus groups) is
another alternative.
Novice researchers almost invariably tape-record or video their interviews. This could be useful if you're
interested in people's exact words and the ways in which they say them, but also creates a lot of work. It
can easily take five hours to transcribe one hour of tape-recorded conversation. For many kinds of
research questions, note-taking (while perhaps using a tape-recorder as a back-up) is equally
appropriate. It also works well to do interviews in pairs, with one researcher doing the interviewing,
while the other takes notes. If you do opt for full transcription, you should consider what transcription
scheme is most useful. For most social science research, ordinary writing is sufficient, but if you're
interested in detailed conversational analysis, you may want to use one of a number of more complicated
schemes showing tone of voice, timed pauses, overlapped speech and so on.
______________________________________________________________________
How much work is transcription really?
If you have access to a tape- or video recorder, try the following exercise: Ask a family
member or friend to tell you what they think should be done to combat AIDS in South
Africa. Record exactly 5 minutes of the interview. Now transcribe the recording,
carefully noting how long it takes to do the transcription. A figure commonly mentioned
is that transcription on average takes five times as long as the original recording. How
many minutes of transcription did you require for every minute of recording? (And how
accurately does your transcription reflect what was actually said?)
______________________________________________________________________
If, even after having done the above exercise, you're still keen on the idea of recording interviews (or
spontaneous talk), here are three practical hints that can make life a bit easier: One, for recording, get a
condenser microphone (the sort that takes a tiny battery) - this is not very expensive, fairly easy to get
hold of, and far more important than having a high quality tape-recorder. Two, place the tape recorder on
a soft surface such as a cushion - it tends to improve the sound quality. Three, for transcription, try and
obtain a foot control so that you can stop and start the tape-recorder without having to stop writing or
typing.
Additional sources
HMPYC80/104 185
HMPYC80/104/0/2024
Content analysis: An example
There are numerous methods for analysing texts. Some of these methods require complex and intricate
analyses. For the moment, however, we will not play clever, assuming any prior knowledge of textual
analysis. The approach suggested here is simple and well structured. The reader should not experience
any problems, although it is only natural to sometimes doubt one's abilities and to wonder whether one
really is on the right track. However, do not worry about possible errors in your approach. We all make
mistakes, but the mistakes we make tend to show themselves. The important thing is not to avoid
mistakes at all costs. It is the ability to go back and correct these mistakes.
The idea behind the kind of analysis demonstrated here is to translate your raw text into a psychological
text. The present example uses a diary text. However, the raw text need not necessarily be diary entries,
any text can be submitted to this kind of analysis. Each diary entry is analysed separately. The advantage
of a diary is that it provides a number of entries (different texts). The more texts one has available for
analysis the easier it becomes to recognise consistent themes.
Here is an example of a run of the mill kind of diary entry that we can use for analysis.
Starting time: 17:50 Date: 1995 - 10 - 04
Dogs were particularly friendly/playful this morning. Brian phoned - asked for lift to university - his
car still at garage. Had 3rd lecture with master students - on philosophy of science - they’re
forming group cohesion, but still complaining about the amount of work in course. I can empathise,
but they are at master’s level and they were warned about work-load in selection interviews. Sure
they’ll make it! Had tea with colleagues from law - debate whether CV’s are private or public
documents. Still don’t know. Debate made me late for meeting. Somewhat boring - People need to
vent their feelings, but must it be at expense of others? Cleaning lady phoned - out of washing
powder - rushed to shops - got groceries - dogs still friendly. Returned to university - finished
correspondence - students turned up without appointment. Had to see them, came far - from deep
platteland. Someone (didn’t get the name) phoned, waiting for contribution to congress
proceedings - Late again. Day was getting stressful - Remembered friendly dogs at home. Why
this writing? I feel I am wasting time!
Signed off: 18:08 Number of words: 200

(approximately)
The first step in analysis is to code the text. Coding refers to labelling sections of the text according to a
specific system. In this case, we will use three sets of labels, the first concerns the level at which events
occur, the second the way in which events are regulated, and the third set of labels refers the kind of
mind process involved.
Here are the labels for event level:
A1. Physical: This label refers to issues concerning the physical environment, as well as to any
186
interaction between an individual/group/community and the physical
environment
A2. Intrapersonal: This label refers to matters concerning the individual, such as an individual's
plans, emotions, thoughts, etcetera.
A3. Interpersonal: This label refers to interaction between two or more individuals.
A4. Societal: This label refers to the larger community, for example, all inhabitants of a town,
all citizens of a country, the population of the world, etcetera.
A5. Metaphysical: This refers to the domain beyond empirical (perceivable) reality.
And here are the labels for regulatory processes:
B1. Factual: This label refers to events regulated by perceivable incidents or generally
accepted truths
B2. Ideological: This label refers to events regulated by generally accepted norms (e.g. ethical
norms)
B3. Political: This label refers to events regulated by power differences among individuals,
groups, societies, etcetera.
B4. Philosophical: This label refers to events regulated by abstract ideas the truth of which cannot be
factually determined.
Here is set of labels for mind processes:
C1. Cognition: This label refers to perception, thought, decision making, etcetera.
C2. Affect: This label refers to feelings, mood, emotions, sentiments, desire etcetera.
C3. Conation: This label refers to will, drive, purposefulness, motivation, goal-directedness,
etcetera.
Let’s code the example diary entry with these labels. You could perhaps try coding before you look at
the answers. Do not worry if you find the coding process difficult or if you come up with answers that
differ from those in the example analysis. The aim is not to get it absolutely right. Simply try to be
consistent in how you assign the labels. People do not always agree about how labels should be assigned.
This is why researchers often use more than one evaluator in coding texts. The evaluators are supposed
to do their coding individually and then to meet to discuss their differences.
Having coded the text it becomes possible to read the text on a more generalised level. See whether you
can jot down some ideas about the text, and organise them into a general and perhaps psychological
reading of the text. Remember, the point is not to come up with an absolutely correct interpretation. The
text affords a number of different readings. Trust your intuition and your knowledge of psychology. Do
this exercise now, before you look at my coding and the interpretation of the text.
.....................................................................................................................................................................
.....................................................................................................................................................................
.....................................................................................................................................................................
.....................................................................................................................................................................
.....................................................................................................................................................................
.....................................................................................................................................................................
HMPYC80/104 187
HMPYC80/104/0/2024
Upon inspection of the codes (see next page) the following picture emerges: Most events happen on the
interpersonal level, are factual in nature and are accompanied by affect. It is interesting these factual
events are accompanied by affect and not so much by cognition or conation. Here is my psychological
reading of the text:
The day starts off with an interaction between the individual and her dogs. This interaction is
complementary, but the next interpersonal event is not. The friend’s request for a lift is a demanding
event. Note how the demands keep building during the day. There are student complaints, a meeting
where people demand attention, a cleaning lady who needs washing powder, students who want to be
seen, correspondence that needs to be dealt with, and an editor who wants an article for publication. And
note the emotional reactions: Initially there is no sign of affect (the friend’s “demand” for a lift is taken
as matter of fact), but one by one signs of emotional reaction appear (empathy, boredom, frustration
(“late again”), and finally an overall emotional reaction to all these demands surfaces and shows itself
clearly: “Day was getting stressful”. Note how this feeling is contrasted with the complementary
interaction with the dogs.
TEXT PHRASES CODING
Dogs were particularly friendly/playful this morning. A3 B1 C2
Brian phoned - asked for lift to university - his car still at garage. A3 B1 C3
Had 3rd lecture with master students - on philosophy of science - they’re A3 B1 C2

forming group cohesion, but still complaining about the amount of work in
course. I can empathise, but they are at master’s level and they were warned
about work-load in selection interviews. Sure they’ll make it!
Had tea with colleagues from law - debate whether CV’s are private or public A3 B2 C2
documents. Still don’t know.
Debate made me late for meeting. Somewhat boring - People need to vent A3 B1 C2
their feelings, but must it be at expense of others?
Cleaning lady phoned - out of washing powder - rushed to shops - got A3 B1 C3

groceries -
A3 B1 C2
dogs still friendly.
A3 B1 C1
Returned to university - finished correspondence - students turned up without
appointment. Had to see them, came far - from deep platteland. Someone
(didn’t get the name) phoned, waiting for contribution to congress
proceedings - Late again.
A2 B1 C2
Day was getting stressful - Remembered friendly dogs at home. Why this
writing? I feel I am wasting time!
Do not discard your own attempt simply because it may differ from mine. Rather check on the points
(labels) where we differ, and decide for yourself which interpretation are more useful for your
understanding of the text. Do you agree that there is a lot of psychology woven into a seemingly innocent
text? What would you see as a prominent motif emerging from this diary entry? Write down your ideas
188
before you look at the next section.
....................................................................................................................................................................................
Here is the motif I selected: This individual experiences stress on an interpersonal level, due to many
demands. Note that the motif refers to an event level (interpersonal) and to a mind process (affect). But
keep in mind that this analysis is based on a single diary entry. It may be necessary to analyse a number
of entries before one can see whether the motif in question emerges in more entries and whether it grows
into a theme that runs through the diary. Also this does not mean that the text should be reduced to a
single theme. A number of different themes may emerge from a text.
Additional resource:
HMPYC80/104 189
HMPYC80/104/0/2024
Thematic Analysis
Analysis of qualitative data is usually is guided by the purpose and theoretical framework of the study
Wegner et al., 2012). The purpose of thematic analysis is to understand the meaning of the data by
identifying themes. Thematic analysis is the most common way of analysing qualitative data, as it is
embedded across most methods of searching for themes in the data. In other words, thematic analysis is
a generic approach used in analysing data, by looking for patterns, such as similarities, common and
different experiences in the data.
How do we find themes?
Qualitative research is based on the pursuit of themes. Themes are patterns, ideas subject, meaning or
common experience that emerges from the data. These are features of participants’ accounts,
experiences, attitudes or even perceptions that emerge from the raw material. Ryan & Bernard (2000)
provides eight observational techniques that can help you to identify themes from the data:
1) Repetitions
Repetitions in the text can often be a clue of a possible theme. Participants often repeat ideas, thoughts
and feelings which are important to them and therefore in most cases, these repetitions sum up their life
experiences. When analysing a raw material, it is important that you become aware of the repetitions
since these are often building blocks of themes.
2) Indigenous Typologies and Categories
One of the ways in which you can themes is by looking at the unfamiliar, local words and for familiar
words that are used in unfamiliar ways Ryan & Bernard (2003). For example in a study conducted by
Spradley (1972), and recorded conversations among tramps at informal gatherings, meals and card
games. In these recordings, participants kept mentioning the idea of “making a flop,” which turned out to
be a local term for finding a place to sleep for the night (Ryan & Bernard, 2003). When analysing
qualitative data, you need to familiarise yourself with the local language that is being used in the data.
3) Metaphors and analogies
A study by Lakoff and Johnson (1983), observed that people often represent their thoughts, behaviours,
experiences with metaphors and analogies (Ryan & Bernard, 2003). You need to be constantly aware of
the metaphors, analogies that are used in the data since in qualitative data analysis we are also interested
in picking up metaphors which may tell us more about the meanings behind a certain phenomenon or
experience.
4) Transitions
When analysing the data, you may observe shifts that occur as sentences progress within the text. There
may be different ideas or certain “build-up” of knowledge as the text progresses, which may indicate
shifts in topics. Similarly, in speech, pauses, changes tone of voice or the presence of particular phrases
190
may indicate transitions and themes (Ryan & Bernard, 2003).
5) Similarities and differences
Similarities and difference are common ways of identifying themes or patterns in the data. The process
of identifying similarities and differences is done systematically by comparing different “units” or parts
of data. In most instances, similarities are “ building blocks” of themes. In other words, the motive
behind finding themes is to bring forth meanings which are similar across the participants or artefacts.
The idea is about highlighting what is a common experience while also observing the differences across
the observations within the data.
6) Linguistic connections
One of the ways to identify themes is by looking for words and phrases that indicate attributes and
various kinds of causal or conditional relations in the data. This method is applicable to studies that are
interested in studying the significance of language and text to support a specific observation.
7) Missing data
This method is often done at advanced stages of doing thematic analysis. The goal is to ask yourself what
is missing here? what is here? The usage of this method can be prompted when you start identifying gaps
in the data, in other words as you read a text, you might realise that what you already know may not be
complete and that there might be gaps in your data. The gaps might be as a result of participants not
mentioning a certain experience as often as they mention others. As you first read the text, you might
have a general overview of themes that may emanate from the data. The next step could be to mark them
with highlighters and then read the text again. As you read the text again, look for themes that remain
unmarked. This tactic will force you to search for the new and less obvious themes, which ultimately
will help you to identify the missing data.
8) Theory-related to Material
Generally, primary data contains rich information on the characteristics of the informants. However, it is
also important to understand how data can illuminate questions of theoretical importance (Ryan &
Bernard, 2003). It is also important to observe the setting, context of the participants and ways of
thinking about people, processes, objects and events since these have an impact on how we understand,
apply and adapt certain theories.
Coding?
This process is called “coding”, which means assigning labels or phrases to text from the data. As you
code, you are assigning meaning to the text that you are reading. Code is often a short phrase or a word
that captures the essence of the text, sentence or paragraph. Below is an example of coding taken from
Saldaña (2013).
HMPYC80/104 191
HMPYC80/104/0/2024
The method prescribed by Braun and Clarke (2006) is among the common ways of doing thematic
analysis. The following steps can be taken to find themes from the raw material data:
Table 06
Common ways of doing thematic analysis
Phases Description of the analysis process
1. Familiarise yourself with i. Narrative preparation (i.e. transcribing data).

data ii. ( Re-) reading and noting down initial ideas.
2. Generating initial codes i. Coding interesting features of the data in a
systematic fashion across data set.
ii. Collating data relevant to each code.
3. Searching for themes i. Collating codes into potential themes.
ii. Gathering all data relevant to each potential theme.
4. Reviewing themes i. Cheeking whether themes work in relation to the
coded extracts.
ii. Checking if themes work in relation to the entire data
set.
iii. Reviewing data to search for additional themes
iv. Generating a thematic “map” of the analysis.
5. Defining and naming i. On-going analysis to refine the specifics of each
themes theme and the overall story that the analysis tells.
ii. Generating clear definitions and names of each
theme.
6. Producing the report i. Selection of vivid, compelling extract examples.
ii. Final analysis of selected extracts.
iii. Relating the analysis back to the research question,
objectives and previous literature reviewed.
Adapted from Braun and Clarke (2006)
192
References
Braun, V. and Clarke, V. (2006) Using thematic analysis in psychology. Qualitative Research in
Psychology, 3(2), 77-101. http://eprints.uwe.ac.uk/11735
Lawler, J. M., Lakoff, G., & Johnson, M. (1983). Metaphors We Live by. Language, 59(1), 201.
https://doi.org/10.2307/414069
Spradley, J. P., & Phillips, M. (1972). Culture and stress: A quantitative analysis. American
Anthropologist, 74(3), 518-529. https://doi.org/10.1525/aa.1972.74.3.02a00190
Saldaña, J. (2013). The Coding Manual for Qualitative Researchers (2nd Ed.). SAGE.
https://doi.org/10.1017/CBO9781107415324.004
HMPYC80/104 193
HMPYC80/104/0/2024
Discourse analysis
Discourse analysis is a <constructionist> approach to analysing research material. It can be defined as
the act of showing how particular discourses are deployed to achieve particular effects in particular
contexts. To do discourse analysis one needs to have some level of 'cultural competence'. This does not
involve knowledge of particular ethnic cultures (such as say Zulu culture or Portuguese culture),
although this could be useful too. Rather, it involves the much more difficult task of being able to strike
a critical distance from the everyday culture that dominates most of our lives, whatever our ethnic
background. Amongst other things, one could label this the culture of modernity (with its emphasis on
science, technology and progress); capitalist culture (which sets up a dichotomy between individual and
society); Euro-American culture (which establishes the rational, white male as an implicit standard); or
consumer culture (which reifies and homogenises our desires). The challenge is to render this familiar
cultural landscape of the modern world 'strange'. For the most part, textbooks and journal articles in the
field of psychology are of little use in this, as they are themselves very much embedded in the modernist,
capitalist, Euro-American cultural milieu. However, there is a growing number of discourse analytic
studies in psychology which are starting to develop a critical appreciation of our culture.
Outside psychology the pickings are richer, both in established university departments such as history,
philosophy and anthropology, and in the newer interdisciplinary fields such as media and gender studies.
While economic pressures therefore continue to force students into greater specialisation, competence as
a discourse analytic researcher in psychology (and, we would argue, as a psychologist per se) requires
less rather than more specialisation.
The concept of cultural competence could of course easily give rise to various forms of abuse. Many of
us have encountered examples of a certain kind of pompous academic who uses his or her scholarship
(his or her superior cultural competence) as an indiscriminate weapon to devalue opponents'
contributions. Others become so enamoured with their insights into social theory that they take notice of
specific facts only if they can make them fit their pet theory. We nevertheless believe that cultural
competence, seen as an ongoing process and as a willingness to engage with critical debates about
modern culture (rather than as a once-off intellectual achievement), is a prerequisite for doing useful
discourse analytic research. Another prerequisite is gaining practice in performing discourse analytic
analyses.
There are also very many forms of discourse analysis outside of psychology, some of which are closely
related to the kind of constructionist research we have been talking about here. However, if you do a
library search for discourse analysis you'll find that the label also encompasses other, not necessarily
related approaches, such as some highly technical work in linguistics.
In psychology most discourse analytic work has been done by a group of British social psychologists
including leading figures such as Margaret Whetherell, Ian Parker, Erica Burman and John Potter, but
you will notice it seeping into increasingly more branches of the discipline, including developmental and
cognitive psychology. Much of the work done thus far has taken the form of a critique of <positivist>
(and sometimes <interpretivist>) work in these fields, and one often hears the complaint that discourse
oriented researchers have not (yet?) built up a sufficient stock of their own 'findings' and are overly
194
parasitic on established work.
In South Africa there are discourse analytic researchers in psychology at many of the leading
universities, including UCT, UWC, Wits, University of Natal, Rhodes University, University of Venda,
Vista, University of Durban-Westville and Unisa. Several editions of the South African Journal of
Psychology have featured discourse analytic work, and a volume (Levett et.al., 1997) of semiotic studies
by South African psychologists has recently appeared.
What is a discourse?
One of the main frustrations for people trying to form an impression of what discourse analysis is about
(as well as for experienced analysts) is the pervasive vagueness regarding what exactly constitutes a
discourse. Although we are also sometimes frustrated by this state of affairs, our feeling is that there are
certain advantages to the situation. For example, because discourse analysts cannot rely on hard and fast
definitions of what constitutes a discourse, they are continually reminded that they themselves are the
principal instruments of the analysis, and that they can never fall back on a set of technical procedures or
definitions.
One very useful distinction made by Parker (1992), and echoed in Coyle, is between discourse and text.
Texts are particular spoken or written utterances (or, by extension, pictures or actions which can be 'read'
as texts). Discourses are the general systems of signification which help to give texts meaning. Here is a
text:
Anne: What do you think - should I wear this outfit to the party?
Tumelo: Uh, no... maybe not.
Lerato: It's a bit revealing, and you know what guys can be like.
We know just what Lerato means because she draws on what has been termed the Male Sexual Drive
discourse (Hollway, 1989), a culturally-sanctioned understanding of men as having strong sexual desires
which are difficult to control once they have become aroused. There may also be many other discourses
(especially gender-related discourses) in operation, all helping to make the text appear meaningful and
unremarkable - for example, the idea that women are particularly interested in outward appearance, but
that women who take the sexual initiative are sluttish. The whole pre-party scene - girlfriends intimately
together in a bedroom, sharing each others' secrets and anxieties, giving support and advice, both fearing
and looking forward to the anticipated contact with boys - is in fact a very familiar script, rehearsed over
and over again in American movies and television soap operas until we have come to think of it as
entirely natural and unexceptional.
Here is another example text from a treatise on "How immunology became a science":
Fodar, in 1887, found that if anthrax bacilli were added to blood and the mixture was kept
at body temperature, the number of bacilli was greatly reduced. This was confirmed by
Nuttall in the following year. Nuttall also found that, if blood were kept for some time or
heated to 55 C., it lost the power of killing bacteria. Since the blood used contained
phagocytes, these results tended to confirm Metchnikoff's theories. (Bigger, 1941, p.
171)
This text draws on the discursive patterns associated with a certain kind of historical writing (sometimes
called 'Great Men' or 'Whig' history) which presents science as an inexorable progression based on key
empirical discoveries.
HMPYC80/104 195
HMPYC80/104/0/2024
Even the simplest text can be seen as an amazing balancing act perched on top of an intricate scaffolding
of discourses, and the aim of discourse analysis is to restore to us the sense of vertigo which such a
performance should properly elicit. Every utterance relies on a wealth of implicit meaning. A classic
example is the phrase:
The king of France is bald.
Of course this little utterance is trying to slip something past us, since France in fact does not have a
king. So, to avoid such sleights of hand we have to make all implicit meanings explicit, for example by
re-writing the sentence as: There is a king of France and the king of France is bald. But that still leaves
France and baldness implicit, so we would have to say There is a country called France, the country has
a king, kings can suffer from baldness and the king of France is bald. But what about the idea of
country? We should really also make it clear that the statement implies that there is such a thing as a
country. And what about the idea that countries have kings - shouldn't that be made explicit too? So we
have: There is such a thing as a country, there are such things as kings, there is such a thing as baldness,
some countries have kings, there is a country called France, France has a king, kings can suffer from
baldness and the king of France is bald. You can see that this little exercise is never-ending - that soon
we'll have to start making explicit what is meant by concepts like 'being' and 'having' as well.
Discourse analysis is about making implicit meanings in texts explicit and, like the king of France
example, is in principle never-ending. Thus different analysts prefer to focus on different levels and
kinds of implicit meaning, and what is important analytic work to one analyst may appear to be a trivial
re-writing exercise to another.
One way of organising the different kinds of implicit meaning looked for by discourse analysts is in
terms of its political import. Those who are concerned to expose false political ideas may use terms like
ideology, myth and mystification to label the systems of meaning they detect operating in particular texts.
Those who are less politically oriented may use more neutral labels such as cultural meaning systems,
cultural schemas, communicative arrays, tacit meaning systems, recurrent discursive patterns,
linguistic repertoires, interpretive repertoires, narratives, grand narratives and common forms of
understanding. The term discourse is becoming perhaps the most commonly used, and can be seen as a
kind of half-way station between explicitly political terms such as ideology and bland technical terms
such as interpretive repertoire.
Another way of distinguishing among the different kinds of implicit meanings that discourse analysts
look for in texts is in terms of meaning vs form. Some analysts are more interested in the meaningful
content of discourses, for example seeing how stereotypical ideas about masculinity are reproduced in
South African men's drinking talk (Kaminer & Dixon, 1995), or how a positivist epistemology
permeates most academic work in psychology, while others are more interested in linguistic forms such
as rhetorical devices, metaphors and scripts that are repetitively used in texts.
Examples of rhetorical devices are extreme case formulations (using words such as always, never,
nobody, everyone) and pre-emptive denials (such as "I'm not a racist but...). Examples of metaphors are
the key metaphors found by Potter and Reicher (1987) in talk about communities - spatial metaphors
(e.g. talking about the community as close knit, integrated, tight), organic metaphors (the community
matures, grows, evolves) and metaphors of agency (the community is said to act, know or feel). With
repeated use these metaphors have entrenched a certain view of what a community is, so that it is now
196
difficult for us to imagine that it is possible to think of communities in any other way. Examples of
scripts are the kinds of scripts we almost unconsciously use when engaging in conversation, which in
fact follow quite strict rules of turn taking, topic changes and so on. Perhaps the primary rule of
middle-class Western conversation is the so-called 'no gap, no overlap' rule, that is there shouldn't be
long silences and people shouldn't talk at the same time. Many other acts we engage in every day (such
as signalling a minibus taxi or answering the phone) can also be seen as following more-or-less set
scripts, with people behaving as if they were acting out a part in a play.
Although we have not provided a direct answer to the question What is a discourse? we hope that you at
least now have an idea of the range of things discourse analysts may be talking about when they refer to
discourses, interpretive repertoires, scripts and the like. For some purposes it can be useful to be quite
specific about what exactly we mean by discourse (e.g. are we talking about an ideology, an
epistemology or a rhetorical device?), but it is equally important simply to be aware of the range of
linguistic phenomena which could be called discourses and to be able to recognise examples of them.
What do discourses do?
Foucault's and Parker's definitions of discourse, respectively "Practices that systematically form the
objects of which they speak" and "A system of statements which constructs an object" give a good
indication of what discourses are about - they construct particular realities. Another definition (Collins,
1997) also captures this idea rather succinctly: "Narratives that organise meaning so as to produce what
then show up as facts."
This does not mean that discourses are magical pronouncements which make things true simply by
saying them, for example you (unfortunately) can't become a millionaire simply by saying "I'm a
millionaire". Rather, discourses create social objects on which you can (consciously or otherwise) draw
to make your explicit statements seem natural, reasonable and true. This also works the other way round,
in that every time you invoke a discourse, you also help to further consolidate the status of the object it
refers to as unproblematic and real. Of course, some discourses refer to objects that do have a physical
reality, but as Parker (1992) argues, although "the object that a discourse refers to may have an
independent reality outside discourse, [it] is given another reality by discourse" (p. 9). (Think, for
example, of different map projections, which all refer to the actual globe, but each also give it another
reality.)
This is therefore the principal function of discourses: To make social objects seem natural, reasonable
and real; to set up a consensual world within which particular texts may flourish. Thus, whatever texts
may appear to be doing (e.g., describe, argue, explain), they also always re-present a universe of
assumed knowledge - for the most part without us even being aware that it is happening.
Apart from creating a particular kind of world and particular subjects to live in it (as the alien language
did), language of course also performs various other less grandiose functions, and often discourse
analysts are interested in tracing how this is achieved. For instance, particular rhetorical strategies may
be used to justify, question, accuse or warrant the trustworthiness of the speaker. In doing an analysis
you should therefore have two questions foremost in your mind: What does this text do? and How does it
do it?
HMPYC80/104 197
HMPYC80/104/0/2024
Table 07
Steps to conduct a discourse analysis
1. Determine the research questions.
2. Select your sample of data.
3. Collect records, including transcripts, videotapes, and so on.
4. Conduct interviews or conversations.
5. Transcribe in detail.
6. Code the data (as the first pass).
7. Analyse the data, looking for any patterns of variability or inconsistency in the discourse; Then
go through the data again, looking for those same patterns. Validate by looking at how the
discourses assist in your understanding, what you still want to know, and what you learned.
8. Compile the results section, and remember to present your argument clearly.
Note: From Potter and Wetherell (2010) eights steps to conduct a discourse analysis.
Additional sources
Breakwell, G.M., Hammond, S., & Fife-Schaw, C. (1995). Research methods in psychology. Sage. See
Chapter 16 (Discourse analysis).
Burman, E. (1996). Psychology discourse practice. From regulation to resistance. Taylor & Francis.
Parker, I., & the Bolton Discourse Network (1999). Critical Textwork. An introduction to varieties of
discourse and analysis. Open University Press.
References
Garner, M., Wagner, C., & Kawulich, B. (2009). Teaching Research Methods in the Social Sciences.
Ashgate.
Hollway, W. (1989). Subjectivity and method in psychology: gender, meaning and science. Sage.
Levett, A., Kottler, A., Burman, E., & Parker, I. (1997). Culture, power and difference: Discourse
analysis in South Africa. Zed Books.
Parker, I. (1992). Discourse dynamics. Critical analysis for social and individual psychology.
Routledge.
Potter, J., & Reicher, S. (1987). Discourse of community and conflict: The organization of social
categories in accounts of a 'riot'. British Journal of Social Psychology, 26, 25-40.
Potter, J., & Wetherell, M. (2010). Discourse and social psychology. Sage.
198
Ensuring Trustworthiness of the

Data
In the research world, scientist or researchers are constantly encouraged to “prove” that their results
“valid”. In order to ensure that the results are accepted by other researchers, they need to detail the
trustworthiness of the data. In quantitative research the terms reliability and validity are used; in
qualitative studies, the term trustworthiness of the data may be used. Qualitative researchers are
encouraged to describe their studies in rich and thick detail to allow the reader to determine whether the
methods used in the study are appropriate and whether the findings are trustworthy. Lincoln and Guba
(1985) proposed four criteria that may be used to ensure the trustworthiness of the data:
1. Credibility- this term is used when addressing the activities that make it more credible than the
findings were derived from the data ( prolong engagement, persistent observation, triangulation,
peer debriefing and negative case sampling) (Wegner et al., 2012).
2. Transferability- this term refers to findings that have applicability in other contexts. One of the
ways to ensure transferability is by having a thick description of the interpretation of the data.
The basis for making similarity judgements; it is achieved by maintaining all versions of the
data in their original forms and by the presentation of thick description (Wegner et al., 2012).
3. Dependability- this terms refers to activities that are aimed at showing that the findings are
consistent and could be repeated, should the researcher conduct the study again. Dependability
can be achieved by triangulation of methods and providing an audit trail; this audit trail provides
evidence for the accuracy of translation of information from various data sources and provides
the means for ensuring the confirmability of the findings, allowing reconstruction of events and
processes that led to the conclusion in the research (Wegner et al.,2012).
4. Confirmability- this term refers to the degree of neutrality or the extent to which the findings of
a study are shaped by the respondents and not the researcher bias, motivation, agenda or
interest. The aim is to ensure that the findings are grounded in the data and to demonstrate that
the data and findings were derived from events, rather than the researcher’s perspective.
As mentioned in the opening passage, when evaluating qualitative research the appropriate term to use
is “trustworthiness”, while quantitative research seeks to ensure reliability and validity. Although
quantitative and qualitative research have different goals, Lincoln and Guba (1985) and Corley (2004)
tabled a comparison between traditional criteria which are mostly used for quantitative research and
trustworthiness criteria, as provided on the next page.
HMPYC80/104 199
HMPYC80/104/0/2024
Table 08
Techniques to Ensure Trustworthiness
Traditional Criteria Trustworthiness criteria Examples
Internal Validity Credibility Member checks

• Transcripts returned to interviewees
to ensure their accuracy i.e., did they
mean what they said?)
• Clarifying questions via email (if
required)
Triangulation of data types
Multivocality
Independent/external peer debriefings
• Review and evaluation of data
interpretations, categorizations and
analysis
External Validity Transferability Detailed (thick) description of the case
or context
Reliability Dependability Purposive and snowball sampling

Participants’ confidentiality protected
Processes reported in detail
Objectivity Confirmability Acknowledgement of study’s

limitations Meticulous data
management and recording
• Verbatim transcription of interviews •
Clear notes on theoretical and
methodological decisions
• Accurate records of contacts and
interviews
Based on Lincoln and Guba (1985) and Corley (2004).
References
Garner, M., Wagner, C., & Kawulich, B. (2009). Teaching Research Methods in the Social Sciences.
Ashgate.
Lincoln, Y.S. and Guba, E.G. (1985). Naturalistic inquiry. Sage.
200
Bracketing (phenomenology)
<Qualitative analysis> involves understanding material as an insider and therefore requires that you
cultivate an open and empathetic attitude. Phenomenology, for instance, involves trying to understand a
person's life-world from their own perspective. As Dreyer Kruger (a leading South African
phenomenologist) puts it: "It is necessary to give up manipulation of the phenomenon in favour of
allowing this to show itself by an intimate communion with it" (1990, p. 404). Often your preconceived
notions can get in the way of empathetic understanding and you therefore need to try and 'bracket' them,
or put them to one side. Imagine what it would have been like if you knew absolutely nothing about the
topic being studied and your only source of information were the data you have collected.
____________________________________________________________________
Here is a short extract from an interview with James, a self-confessed paedophile
(someone who engages in sexual activities with children) who operates in the Gauteng
area:
Are you a violent paedophile?

No, I've never hurt a child in my life. The things I do to children they like.
Like what?
Touching them, making love to them.
And if a child says 'please stop' would you stop?
Yes, of course, immediately. I'll never force myself on a child.
Burrows (1995)
What pre-suppositions would one have to bracket to understand James' world from his
perspective? Does bracketing pre-suppositions mean the same as giving them up
entirely? Is it possible (or desirable) to have an 'intimate communion' with the
phenomenon of paedophilia?
__________________________________________________________________________
References
Burrows, S. (1995, October). Found object. Proceedings of the 1st Annual South African Qualitative
Methods Conference (A spanner in the works of the factory of truth). University of the
Witwatersrand, Johannesburg.
Kruger, D. (1990). The question of knowing and truth in psychology and psychotherapy. In J. Mouton
& D. Joubert (Eds.) Knowledge and method in the human sciences. Human Sciences Research
Council.
Additional resource:
HMPYC80/104 201
HMPYC80/104/0/2024
APPLICATION OF THE RESEARCH REVIEW INVENTORY (R.R.I. revised)

Examples and applications of the R.R.I.:
Application 01
This section links to “Action 01: Act as peer reviewer of (a) an article and (b) proposals” in Tutorial
Letter 102. Here you are expected to review an article (and later on proposals) with the aid of the
Research Review Inventory (R.R.I.3) as provided in the Resource Material in Tutorial Letter 102.
You are expected to perform the following self-assessment task. This practice run will help you to
prepare for the article review as required for Assignment 01. In this example the previous R.R.I.
(Revised) was used and therefore the answers on 2 items will differ from the R.R.I.3.
Download and study the following article:

Dekel, Bianca & Andipatin, Michelle (2016). Abused Women's Understandings of Intimate
Partner Violence and the Link to Intimate Femicide [85 paragraphs]. Forum Qualitative
Sozialforschung / Forum: Qualitative Social Research, 17(1), Art. 9,
http://nbn-resolving.de/urn:nbn:de:0114-fqs160196. DOI:
http://dx.doi.org/10.17169/fqs-17.1.2394
[ Use the following URL web address to access the above article:
http://www.qualitative-research.net/index.php/fqs/article/view/2394/3939
or, http://www.qualitative-research.net/index.php/fqs/article/view/2394 ]
If there are words in the article that you are unfamiliar with you should consult a dictionary or look up
the meaning on the internet. There may also be theoretical references that may be unfamiliar. Use
Wikipedia or other internet resources to gain an overview of what is being referred to. Example:
discourse analysis.
Review the article. Then, make an informed decision of what scale rating applies for each relevant item.
For a published research article all the rating items must be used! Keep track of your reasoning and
justification in a R.R.I. Review Log.
You have to apply the information contained in the review criteria to the article you have to rate. So, take
the first criterion which pertains to the title of the study. After you have studied the article, look at the
title and then decide if it meets the first set of criteria. If it meets the criteria sufficiently you give a rating
of 3. If it doesn't meet the criteria at all, you give a rating of 1. If it is somewhere in the middle, but
insufficient, give a rating of 2.
Please document your reasoning for each rating carefully. Use the Review Log for the Research Review
Inventory (R.R.I. revised) in the Resource Materials (Tutorial letter 102) to keep a personal log of your
answers and comments.
As discussed in Tutorial Letter 102 the Research Review Inventory (R.R.I. revised) is used to conduct
the review. The R.R.I. (revised) Rating Scale ratings are as follows:
1 = Absent, incorrect, or completely wrong.
2 = Only partially addressed. It is does not attain the item requirement(s) or criterion/criteria.
202
3 = Sufficiently up to standard in terms of the item criteria.
Every item should be reviewed and rated on the three-point scale (1 to 3).
Important note:
All of the items in the Research Review Inventory (Resource material 2) can be read as questions.
For example, when rating items 01, 02 and 03 are converted into a question it will read as follows:
Question 01.
Does the title contain (at least) one central psychological construct from the research proposal or
study?
Question 02.
Does the title indicate the research setting or demographic information of the sample population?
Question 03.
Does the title point to, or refer to the research method, design or paradigm? This may be
signalled by key words such as: “lived experiences”, “perceptions”, “attitudes”, “constructions”,
“discourses”, “quasi-experimental”, “critical study”, “psychometric properties”, “correlational
study”, and “inferential study”.
Answer these questions by selecting a rating, i.e.: 1 = Absent or incorrect; 2 = Only partially addressed;
or 3 = Sufficiently up to standard.
Remember to use the Review Log for the Research Review Inventory (R.R.I. revised) in Resource 3
(Tutorial letter 102) to keep a detailed log of your answers and comments.
--------------------------------------------------------------------------------------------------------------------
After you have completed the review you can compare your answers with the memorandum that is
provided on the next page.
HMPYC80/104 203
HMPYC80/104/0/2024
Application of the Research Review Inventory: Example 1 with answers
The correct answers for a review of the following article are now provided and discussed:
Dekel, Bianca & Andipatin, Michelle (2016). Abused Women's Understandings of Intimate
Partner Violence and the Link to Intimate Femicide [85 paragraphs]. Forum Qualitative
Sozialforschung / Forum: Qualitative Social Research, 17(1), Art. 9,
http://nbn-resolving.de/urn:nbn:de:0114-fqs160196. DOI:
http://dx.doi.org/10.17169/fqs-17.1.2394
This review is based on the previous version of the Research Review Inventory (R.R.I. revised).
Therefore 2 of the items will differ from the R.R.I.3 that ius provided in Tutorial Letter 102.
First read the entire article to get an overview of the situation. Analyse the scenario in terms of research
interest, statement, problem or question, and identify the constructs, themes or variables.
RATING SCALE OF THE RESEARCH REVIEW INVENTORY (R.R.I. revised)

For articles and completed research reports all 58 items are used in the review.
Items should be reviewed and rated on a three-point rating scale from 1 to 3. The ratings are defined as
follows:
1 or a: Absent, incorrect, or completely wrong.
2 or b: Only partially addressed. It is does not attain the item requirement(s) or
criterion/criteria.
3 or c: Sufficiently up to standard in terms of the item criteria.
Remember, if the item criterion is not explicitly addressed then allocate a low rating. Good academic
writing is clear and transparent. If a reader needs to guess at whether an issue has been implemented and
addressed or not, then the writing is neither clear nor transparent. Only give a high rating if the item
criteria have been explicitly addressed.
Q. 1 2 3 Rating aspect, and comment:

01 1 2 3 Title: one psychological construct:
Up to standard: “understandings” is a key psychological construct.
02 1 2 3 Title: sample setting or demographic:
Up to standard: “Abused Women” (in the context of intimate partner violence and
intimate femicide)
03 1 2 3 Title: research method/design/paradigm:
Only partially addressed by “understandings” as this point to a some likely qualitative
research approaches. But, neither the specific method of “discourse analysis” nor the
approach/paradigm of the “feminist poststructuralist perspective” are indicated.
04 1 2 3 Abstract: research issue/interest/problem:
Up to standard: “(how do) women survivors of intimate partner violence understand
the abuse they endured and the possible link to intimate femicide”
05 1 2 3 Abstract: setting/demographic:
204
Up to standard: “Seven South African women, aged 23 to 50 years, with a history of
different manifestations of intimate partner violence” and “women who live in
abusive relationships”
06 1 2 3 Abstract: research question/statement:
Up to standard: “how women survivors of intimate partner violence understand (and
explain) the abuse they endured”, and “the possible link to intimate femicide”
07 1 2 3 Abstract: key theory:
Up to standard: “feminist poststructuralist perspective”
08 1 2 3 Abstract: research method/design:
Up to standard: “discourse analysis”
09 1 2 3 Abstract: data collection method:
Up to standard: “open-ended interviews.”
10 1 2 3 Abstract: sample & strategy:
Seven South African women, aged 23 to 50 years, with a
history of different manifestations of Intimate Partner Violence (IPV) participated
Not indicated.
11 1 2 3 Abstract: data analysis method/technique:
Up to standard: “The data was analyzed by means of discourse analysis”
12 1 2 3 Abstract: actual findings [of article]:
Up to standard: “The women's understandings were filtered through the particular
social context in which their abusive experiences occurred. The findings highlighted
that contemplating femicide was too threatening, and consequently participants drew
on discourses of femininity, romantic love, and others to justify their remaining in their
violence-ridden relationships.”
13 1 2 3 Keywords: (a) one psychological construct, (b) one psychological theory:
Both are up to standard: (a) “abuse” and “Psychology”, note that without Psychology
qualifier this could study be located in other disciplines; (b) “feminism” is a paradigm
and theory that also to psychology, and the social sciences.
14 1 2 3 Keywords: setting /demographic (any two):
Only partially addressed: “domestic” and “South Africa” provide some information,
but this is incomplete. The following would have been relevant: Abused woman,
intimate relationships, femicide.
15 1 2 3 Keywords: research method/design:
Up to standard: “discourse analysis”
16 1 2 3 The research context:
Up to standard: The context (who, what, where, how, how come, and when) of the
research issue is clearly described throughout the text.
17 1 2 3 Role-players/interested/affected parties:
Up to standard: The relevance of the research issue or problem for a specific group(s)
or stakeholder(s) is clearly specified and explained throughout the discussion.
Stakeholders and groups include: individuals - survivors of Intimate Partner Violence,
their families, their households, the affected communities Families, schools,
universities, workplaces, and the media (see 6. Future Research). Lastly, researchers
who wish to gain a rich perspective that in turn will enable the development of new
understandings that have the potential to be responsive to complex problems in society;
and relevant practitioners, professionals, organisations and institutions who wish to
inform their strategies and programs to reduce the risk level of intimate femicide, and
to guide prevention policy.
18 1 2 3 A psychological study (issue/interest/problem):
Up to standard, for example: “Intimate partner violence (IPV) is recognized as a severe
HMPYC80/104 205
HMPYC80/104/0/2024
medical and public health concern for women (BOONZAIER, 2008; TOWNS &
ADAMS, 2009)” and “A dominant factor contributing to the high IPV rate in South
Africa is the fact that South African culture is to a large extent still regarded as
patriarchal and
hierarchical, where women are expected to be obedient and men are viewed as
the disciplinarians in the family. Although these traditions are beginning to shift,
socially dictated gender roles are pervasive and relate to the widespread
manifestation of IPV in South Africa. Violence for men is normalized and thus
asserting their masculinity through violence is seen as a socially acceptable
means of exercising power over women (JEWKES et al., 2009).”
Other psychology terminology and constructs include: perceptions, patriarchy, gender
and sex roles, identity, family dynamics, social norms and expectations. The study is
firmly located within Psychology.
19 1 2 3 Clarification of interests to role-players/interested/affected parties:
Up to standard: The aim is to better understand intimate femicide in order to inform
families, schools, universities, workplaces, and the media (see 6. Future Research); and
researchers who wish to gain a rich perspective that in turn will enable the development
of new understandings that have the potential to be responsive to complex problems in
society; and relevant practitioners, professionals, organisations and institutions who
wish to inform their strategies and programs to reduce the risk level of intimate
femicide, and to guide prevention policy.
20 1 2 3 Literature review (12 sources for proposal/ 24 for article):
Up to standard: literature from multiple sources is used to highlight applicable theories
that are relevant to understand and describe (or explain) the psychological complexities
and dynamics of the research issue or phenomenon in its various contexts. Note that
new literature is only introduced later in the discussion (see 5. Data analysis).
Although this is not a conventional sequencing of the discussion, the relevant theories
are presented throughout the article. (This integrated discussion of the literature in the
results section is probably due to limited space in journals for peer-reviewed articles.)
21 1 2 3 Three theoretical perspectives:
Up to standard: One meta-theoretical approach and viewpoint was deployed (feminist
poststructuralism). Other theoretical views include: Traditional patriarchal gender
discourses position and a dominant patriarchal family discourse. Dominant
prescriptions of femininity, such as traditional feminine practices where nurturance
and selflessness, and construction of the "good wife”, and a sense of responsibility for
the man's well-being, the limitations inherent in religious discourses and religious
constructions of the sanctity of marriage. / Dual construction or "splitting" into (1) the
fairy tale romance discourse, which involved an idealized form of romantic discourse,
and (2) The second was the dark romance discourse – that can hamper women's ability
to understand the many complex and often manipulative factors affecting abusive
relationships that serve to keep both parties trapped in the cycle of violence.
The intergenerational transmission of abuse is frequently explained from a social
learning perspective (see 5.9.1.4 Intergenerational transmission of abuse.) / According
to JEWKES et al. (2009), social norms support the use of violence and in doing so,
desensitize us to the use of violence and render the use of violence in many
circumstances as legitimate. / In order for women to understand the complexities of
intimate femicide, it is crucial that they first learn to deal with and overcome
self-blame, and realize that they are not responsible for the abuse, and that the abuse is
206
out of their control (JACKSON & MANNIX, 2004; SMITH & RANDALL, 2007).
22 1 2 3 Two existing empirical research findings:

Up to standard: There is a reasonably exhaustive discussion of empirical literature
findings, see 1. Introduction, 5. Data analysis, 6. Future Research). Examples:
Research has found that men are more likely to be violent if they abuse substances.
having witnessed or experienced violence in childhood might result in more
acceptance of violence (ABRAHAMS et al., 2013). This finding is similar to a study
conducted by HYDÉN who found that women often have mothers and mothers-in-law
who appear to be "co-offenders" of the abuse, as they convince the women to forgive
the violence and
to remain with their abusive partners. Unemployment and low income have
consistently been identified as risk factors for intimate femicide (ARETAKIS, 2008;
CAMPBELL,
2004; MATHEWS et al., 2004; WEIZMANN-HENELIUS et al., 2012).
23 1 2 3 Literature review is in dialogue with the body of published works:
Up to standard: The findings are synthesised and presented as a continuous integrated
discussion in dialogue with the existing bodies of literature. (see 1. Introduction, 5.
Data analysis, 6. Future Research). Note that much of the literature review is integrated
in the Discussion section.
24 1 2 3 No digressions/detours:
Up to standard
25 1 2 3 Academic integrity: plagiarism-free literature review:
Up to standard. The entire manuscript from beginning to end appears to be
plagiarism-free. All works, ideas, theories, facts and statements are appropriately
referenced, acknowledged and cited. As stated in Question 12: The article appears to
be plagiarism-free, as the style of writing is consistent and is of high academic quality.
As mentioned before, in this case we are dealing with a peer-reviewed article; this
means that you can assume that the manuscript is plagiarism-free. In general,
peer-review processes of academic journals include various plagiarism-checking
mechanisms and procedures.
26 1 2 3 Prescribed in-text citation style:
Up to standard. The criterion states that APA7 or Other can be used. In this case
another referencing style is prescribed by the journal. It is used correctly and
consistently. (Consult the FQS journal citation style as provided on the FQS website.)
27 1 2 3 Research ontology/paradigm/approach:
Up to standard. The stated approach/paradigm is the “feminist poststructuralist
perspective”. This is explained, for example “A significant feature of post structural
feminism is that it attempts to unpack how women's understandings of
experiences and phenomena which affect them, are constructed within
discourses and power relations (BEASLEY, 1999).”
28 1 2 3 Research question/statement: (a) setting/demographic, (b) psychological construct
Up to standard. The research question or research statement is evident and adheres to
the expected criteria:
“this study sought to explore how women survivors of IPV understand intimate
femicide and their level of risk in a violent relationship.”
“The objective of this study was to investigate the discourses that women survivors of
IPV draw on to understand intimate femicide.”
“The discourse analysis in this study translated into analyzing the texts of women
survivors of IPV, and unpacking the discourses they drew on to understand intimate
HMPYC80/104 207
HMPYC80/104/0/2024
femicide. It involved examining how these discourses assisted or hampered their

understanding of intimate femicide, specifically in the context of an abusive
relationship.”
Demographic information: women from a shelter from a middle-class suburb of Cape
Town, and the second shelter is located on the Cape Flats; a sample of women who had
experienced IPV at the hands of their intimate male partners; only women who were
comfortable with the English language would be considered; The mean age of the
participants was 38 years. Only one woman was married to her abusive male partner at
the time of the interview, one woman was currently in the process of getting a divorce,
and one woman was already divorced from her abusive ex-husband, while four women
were at the shelter as a result of abuse experienced at the hands of ex-boyfriends. On
average, the women in this study had three children. Only two of the participants were
employed at the time of the interview. Six of the participants were ‘Colored’ (see Note
1) and one participant was ‘Black’.”
29 1 2 3 Apt research method/ design:
Up to standard. An appropriate method(s) and research design(s) was chosen to
research the question or issue. An exploratory qualitative research design was used.
“The data in this study lent itself to discourse analysis, which is in keeping with the
theoretical framework of the study, namely feminist poststructuralism.” “In discourse
analysis, the focus is on language and how an individual's understanding of experience
is constructed through language.” (See 4.1 Sampling and data collection, and 4.2 Data
analysis.) The discussion of the method is detailed and comprehensive.
30 1 2 3 Apt data/evidence to be gathered:
Up to standard. Open-ended semi-structured interviews.
“MORGAN (2007) posits that individual interviews can encourage
participants to share more information about sensitive issues because of the
confidential nature of the conversation. FARQUHAR and DAS (1999) point out
that confidentiality can also become an issue when conducting focus groups on
personal and sensitive topics if the location of the participants is not taken into
consideration. Therefore, because the women whom we interviewed lived with
each other, we felt that the best way to ensure confidentiality was to conduct
individual interviews. Our concern and decision to conduct individual interviews
was later confirmed in an interview, when one of the women commented that she
would not speak about her abuse in front of the other women, because they all
lived together and she would have to see them every day. Furthermore, the use
of a flexible interviewing approach afforded the first author the opportunity to be
able to listen to the women's understandings from their own vantage point, using
their own words, as the purpose of open-ended questions is to encourage
participants to express their way of understanding in their own words
(BOONZAIER & DE LA REY, 2003).”
31 1 2 3 Apt sampling population, strategy and size:
Up to standard: The sampling selection criteria were: only women in abusive
heterosexual relationships were selected (women who experienced abuse experienced
from fathers or brothers or others were excluded). The second criterion was related to
the issue of language for this study (English speaking). To this end purposive sampling
was used. The sampling strategy and size are discussed in sufficient detail.
32 1 2 3 Data collection, code, analysis: Reliable or dependable/credible/consistent:
Up to standard: The qualitative, discourse analysis, data collection and coding
208
techniques appear to be sufficiently dependable and credible. “Discourse analysis was
conducted in the following way, based on the suggestions proposed by WILLIG: The
first author immersed herself in the data by transcribing each interview and then
thoroughly reading and re-reading transcripts. She also listened to the tapes while
reading the transcripts, which led to getting a better grasp of the participants' accounts
as well as to the words that were spoken, as well as to how they were spoken.” A
detailed approach to deconstructing the dominant discourses in the analysis is
provided. The 2 researchers coded (and analysed) the interview data together
(inter-coder agreement, reasoned consensus). “The coding of the transcripts was
guided by the objectives of this study. In addition, the coding was guided by what
emerged from the data (crystalisation), as opposed to previous literature indicating
what should emerge. This was done as we wanted new categories of understanding to
emerge (ibid.). We thus listened to what the women themselves highlighted as
important.” The researchers used “recordings and transcribed the interviews. The
recorded interviews were checked against each transcribed interview a few times to
ensure that it was done without error.” (transparency/verifiability). The processes and
procedures of the study are sequentially and logically presented and this renders the
expected results credible and believable. The criteria for this item are met.
33 1 2 3 Data collection, code, analysis: Valid or trustworthy/ believable/ transferable/
confirmable
Up to standard:
The processes used to deconstructing the dominant discourses are clearly described.
The findings of the research can be transferred to other contexts by the readers.
The processes and procedures of the study are sequentially and logically presented and
this renders the expected results believable.
Confirmability as intersubjective agreement refers to the degree to which the results
could be confirmed or corroborated by other researchers, in this case through
agreement of findings and interpretations in the literature.
The non-verbal communications in the interviewed were also attended to and used as
data: “She (the researcher) also listened to the tapes while reading the transcripts,
which led to getting a better grasp of the participants' accounts as well as to the words
that were spoken, as well as to how they were spoken.” The criteria for this item are
met.
34 1 2 3 Data management:
Procedures whereby the data was managed are described, although not in sufficient
detail. It appears that the qualitative coding and analysis was done without the aid of
analysis software. The data was recorded and transcribed. Pseudonyms were used.
Questions that are not answered include: How the recordings and transcriptions (data)
were stored, and was it done securely? This important omission of safe and secure data
storage informs the lower rating of 2. The criteria are only partially addressed.
35 1 2 3 Ethics: non-malevolence and avoidance of harm:
This aspect is not explicitly discussed.
However, (a) the research is clearly described and appears to be non-malevolent and
avoidant of harm. And (b) the article mentions that ethical oversight and approval was
granted by the Senate Higher Degrees Research Ethics Committee at the University of
the Western Cape. The two shelters also approved all components of the study. The
research article reportedly adheres to institutional research policies, guidelines and
procedures. The criterion is only partially addressed.
36 1 2 3 Ethics: informed consent:
HMPYC80/104 209
HMPYC80/104/0/2024
However, the article mentions that ethical oversight and approval was granted by the
Senate Higher Degrees Research Ethics Committee at the University of the Western
Cape. The two shelters also approved all components of the study.) The research article
reportedly adheres to institutional research policies, guidelines and procedures. The
criterion is only partially addressed.
37 1 2 3 Ethics: (a) voluntary participation, (b) right to withdrawal:
However, the article mentions that ethical oversight and approval was granted by the
Senate Higher Degrees Research Ethics Committee at the University of the Western
Cape. The two shelters also approved all components of the study.) The research article
reportedly adheres to institutional research policies, guidelines and procedures. The
criterion is only partially addressed.
38 1 2 3 Ethics: privacy, confidentiality and anonymity:
Up to standard: This aspect is explicitly discussed.
“Great emphasis was placed on confidentiality and anonymity and each participant was
informed that counselling would be provided should they feel a need for support after
the interview. This was done through the social worker at each shelter.” And
“Pseudo-names are used.”
39 1 2 3 Ethics: ethical oversight and permission:
Up to standard: This aspect is explicitly discussed.
The article mentions that ethical oversight and approval was granted by the Senate
Higher Degrees Research Ethics Committee at the University of the Western Cape.
The two shelters also approved all components of the study.
40 1 2 3 Grammar and spelling (95% correct):
Up to standard: The grammar and spelling of the manuscript are of high quality.
American spelling rules were used, with very few errors.
41 1 2 3 Document in toto is plagiarism-free:
Up to standard: The entire manuscript from beginning to end appears to be
plagiarism-free. All works, ideas, theories, facts and statements are appropriately
referenced, acknowledged and cited. The article appears to be plagiarism-free, as the
style of writing is consistent and is of high academic quality. As mentioned before, in
this case we are dealing with a peer-reviewed article; this means that you can assume
that the manuscript is plagiarism-free. In general, peer-review processes of academic
journals include various plagiarism-checking mechanisms and procedures.
42 1 2 3 Correct prescribed in-text citation style of manuscript:
Up to standard: The criterion states that APA7/Other can be used. In this case another
referencing style is prescribed by the journal. It is used correctly and consistently.
43 1 2 3 A complete list of references is presented:
Up to standard: A reference list is provided. The reference list is complete
44 1 2 3 List of references in correct prescribed format:
Up to standard: The reference list adheres to the required standard and format for the
FQS journal.
45 1 2 3 Acceptable academic/scientific standard and style:
Up to standard: The work is of high academic standard, and an appropriate
academic and scientific style is used that suits the feminist poststructuralist paradigm
and epistemology, and the discourse analysis method.
.... End of proposal review. (Continue in the case of an article review.)
46 1 2 3 Data collection, coding & analysis: reliable/dependable/credible (see Q32):
210
Up to standard: As discussed in question 32.
47 1 2 3 Data analysis and interpretation procedure comprehensible/ able to follow:
Up to standard: The processes and procedures of the study are sequentially and
logically presented.
48 1 2 3 Data collection, coding & analysis: Valid or trustworthy/ believable/ transferable/
confirmable (Q33):
Up to standard: As discussed in question 33.
49 1 2 3 Ethical data analysis and interpretation ethical (Q35-39):
This aspect is not explicitly discussed for all ethical criteria, due to omissions as
previously discussed in Questions 35 to 39.
50 1 2 3 Results: presentation appropriately and correctly:
Up to standard
51 1 2 3 Results: presentation well structured, can follow:
Up to standard
52 1 2 3 Discussion: interprets/explains and implications:
Up to standard. A discussion is presented which interprets or explains the results and
their implications for the research question/issue. This is interspersed within the
discussion of Section 5. Discourse Analysis of the Understanding of Intimate
Femicide, and elaborated in Section 6. Future Research.
53 1 2 3 Discussion: links to literature theories (and empirical)
Up to standard: The discussion links the results to the existing bodies of literature of:
(a) existing theoretical literature and (b) existing empirical literature findings. Again,
this is interspersed within the discussion of Section 5. Discourse Analysis of the
Understanding of Intimate Femicide, and elaborated in Section 6. Future Research.
54 1 2 3 Golden research thread throughout manuscript:
Up to standard. The research paradigm, theoretical perspectives in the literature,
research methods, data gathering techniques, data coding and analysis, and
interpretations that were employed are in line with the research question and aim.
There is a ‘golden thread’ that runs through the manuscript and connects each
subsection and section to the whole.
55 1 2 3 Alternative explanations/interpretations:
Up to standard. The possibility of alternative explanations/interpretations of the
phenomenon is considered. Almost every section has more than one discussion of
meanings or discourses that are evident in the participant’s interviews. This is
particularly evident in Section 5.10.1 Romantic discourses, where more than just a
single discourse is presented. There is sufficient variations in the discourses and
multiple co-contributing factors are identified and discussed in each sub section of
Section 5. Discourse Analysis of the Understanding of Intimate Femicide.
56 1 2 3 Further research indicated:
Up to standard. The possibility or need for further research is raised in Section 6.
Future Research.
57 1 2 3 Strengths, weaknesses & limitations:
Absent: The strengths, weaknesses and limitations of the study are not discussed, not
even in Sections 6 or 7.
58 1 2 3 Final conclusion:
Only partially addressed. There is a conclusion, but it is very brief and does not
succinctly reiterate the main research finding or result.
The review ratings are generally very high. This is to be expected as the article was published in a
peer-reviewed journal where there are strict quality control measures and quality-enhancing feedback
HMPYC80/104 211
HMPYC80/104/0/2024
mechanisms.
212
Application 02
The Research Review Inventory (R.R.I) is provided in Resource Material 02 of Tutorial Letter 102. The
58 items of the R.R.I. were used to review the article:
Van der Walt, F. (2018). Workplace spirituality, work engagement and thriving at
work. SA Journal of Industrial Psychology/SA Tydskrif vir Bedryfsielkunde,
44(0), a1457. https://doi.org/10.4102/sajip.v44i0.1457
URL web address: https://sajip.co.za/index.php/sajip/article/view/1457 (access option one)

or https://sajip.co.za/index.php/sajip/article/view/1457/2196 (second access option)
(Note that there are two ways to access the journal article and that the first option does not include the
keywords. Therefore there are two correct ratings for the keyword ratings of items 13 and 14.)
First read the entire article to get an overview of the situation. Analyse the scenario in terms of research
interest, statement, problem or question, and identify the constructs, themes or variables.
This review is based on the previous version of the Research Review Inventory (R.R.I. revised).
Therefore 2 of the items will differ from the R.R.I.3 that ius provided in Tutorial Letter 102.

follows:
1: Absent, incorrect, or completely wrong.
2: Only partially addressed. It is does not attain the item requirement(s) or
criterion/criteria.
3: Sufficiently up to standard in terms of the item criteria.
Remember, if the item criterion is not explicitly addressed then allocate a low rating. Good academic
writing is clear and transparent. If a reader needs to guess at whether an issue has been implemented and
addressed or not, then the writing is neither clear nor transparent. Only give a high rating if the item
criteria have been explicitly addressed.
HMPYC80/104 213
HMPYC80/104/0/2024
Item Rate Rating item (R.R.I) Justification

01 3 01. The title contains at least one Yes, the item criterion is achieved and a rating of 3
central psychological construct from (sufficiently up to standard) is allocated.
the research proposal or study. “Workplace spirituality", "work engagement" and
"thriving at work" are key psychological
constructs.
02 2 02. The title indicates the research The item criterion is not completely satisfied and a
setting or demographic information of rating of 2 (partially addressed) is allocated.
the sample population. Any one of the "Work" and "workplace" are indicated. This is a
following is sufficient: where, who, generic description of the research setting and is
when, under what circumstances, or insufficient. The title should have included the
the phenomenon that is being work context: "small, medium and macro
investigated. enterprises (SMMEs)" or the locality of: "the Free
State province" or "South Africa" for a complete
description.
03 1 03. The title points to, or refers to the No, the item criterion is not addressed and a rating
research method, design or paradigm. of 1 (absent) is allocated.
This may be signalled by a keyword or The title does not provide any indication of the
phrase such as: “lived experiences”, research method, design, or paradigm that was
“perceptions”, “attitudes”, followed.
“constructions”, “discourses”,
“quasi-experimental”, “critical study”,
“psychometric properties”,
“correlational study”, “inferential
study”, and so on.
04 3 04. The abstract indicates the research The abstract or summary is provided in the text
interest, issue or problem. box at the beginning of the article. The item
criterion is achieved and a rating of 3 (sufficiently
up to standard) is allocated.
The research orientation is to investigate how
employees can sustain high performance over the
long term. The research purpose was to establish
whether spiritual workplaces
will enhance employees’ work engagement and
thriving at work.
05 3 05. The abstract indicates the setting, Yes, the item criterion is achieved and a rating of 3
demographic or circumstances (sufficiently up to standard) is allocated.
pertaining to the research interest or The study was quantitative in nature, and data were
question. Any one of the following is collected from
sufficient: where, who, when, under Setting: employees working at small, medium and
what circumstances, or the macro enterprises (SMMEs) in one geographical
phenomenon that is being investigated. area in South Africa.
Phenomenon: workplace spirituality, work
214
engagement and thriving at work.
06 3 06. The abstract contains the specific Rate: 3
research question or research The purpose of the study was to establish whether
statement. spiritual workplaces will enhance employees’
work engagement and thriving at work.
07 1 07. The abstract indicates the key Rate: 1
theoretical perspective(s) that is (are) The abstract does not mention theoretical
used in the literature. perspective.
08 3 08. The abstract specifies the Rate: 3
(proposed) research method, study The abstract states that a quantitative research
type or research design. approach is used with a cross sectional survey
design.
09 3 09. The abstract indicates the Rate: 3
(proposed) data collection method. The abstract indicates that a survey
was conducted by means of a self-administered
questionnaire.
10 2 10. The abstract describes the Rate: 2
(proposed) population sample, and A sample is mentioned but sampling strategy is not
sampling strategy or approach. mentioned. "Data were
collected from employees working at Small,
Medium and Macro Enterprises (SMMEs) in one
geographical area in South Africa. The final
sample consisted of 259 employees." It is not
exactly clear how the sample(s) was/were drawn.
(proposed) data analysis method or No data analysis method is mentioned in the
technique. abstract.
12 3 12. For completed research projects Rate: 3
only -- the abstract briefly mentions the It is stated that: "The findings of the study show
actual findings of the study, and states that there is a positive and significant
how the study answers, addresses or relationship between workplace spirituality, work
illuminates the research issue or engagement and thriving at work.
question. Furthermore, workplace spirituality significantly
(Proposals will receive a “1” for this influences the variance in both work engagement
item as the actual research has not yet and thriving at work." Practical or managerial
been performed.) implications and the contribution or value-add of
the study are also indicated.
13 1 13. The list of keywords contains at * The keywords are not indicated in the adobe
or least (a) one key psychological acrobat (pdf) document. Some students accessed
2* construct and (b) one key reference to the article directly and did not note the keywords
a psychological theory. on the main webpage. In this case a rating of 1 is
correct.
------------------------------------------------------
* The keywords of the study are indicated on the
article webpage
https://sajip.co.za/index.php/sajip/article/view/145
HMPYC80/104 215
HMPYC80/104/0/2024
7. A psychological construct is indicated, but there

is no mention of a psychological theory. In this
instance a rating of 2 is correct.
14 1 14. The list of keywords indicates the Rate: 1 or 3 *
or research setting and demographic * The keywords are not indicated in the adobe
3* information of the sample. acrobat (pdf) document. Some students accessed
Any two of the following is sufficient: the article directly and did not note the keywords
where, when, who, under what on the main webpage. In this case a rating of 1 is
circumstances, or the phenomenon that correct.
is being investigated. ------------------------------------------------------
* The keywords of the study are indicated on the
article webpage
https://sajip.co.za/index.php/sajip/article/view/145
7. In this instance the research setting of SMME’s
(Small, Medium and Macro Enterprises) is
indicated. The phenomenon that is being
investigated is also stated: thriving at work, work
engagement. In this instance a rating of 3 is
correct.
15 1 15. The list of keywords indicates the Rate: 1
research method, study type or The keywords are not indicated in the adobe
research design. (This may be acrobat (pdf) document. Some students accessed
indicated directly or indirectly by the article directly and did not note the keywords
phrases such as: “lived experiences”, on the main webpage. In this case a rating of 1 is
“perceptions”, “attitudes”,
correct.
“quasi-experimental”, “critical study”, ------------------------------------------------------
“psychometric properties”, The list of keywords of the study are indicated on
“correlational study”, “inferential the article webpage
study”, and so on.) https://sajip.co.za/index.php/sajip/article/view/145
7. However, the research method, study type or
research design is not indicated. A rating of 1 is
correct.
16 3 16. The context (who, what, where, Rate: 3
how, how come, when) of the research The research context is provided in the
issue is clearly described. Introduction.
(Note that the rating issue can be
addressed anywhere in the proposal or
article.)
17 3 17. The relevant role-players, affected Rate: 3
parties or interest groups with regard to The relevant role-players, affected parties or
the research issue or problem are interest groups mentioned include: employees,
specified or mentioned. Role-players, organisation, SMME’s, people
affected parties and interest groups can (academics/researchers) who work with the body
include: the researcher, individuals,
of knowledge regarding workplace spirituality,
clients, students, beneficiaries,
216
patients, families, households, thriving at work and work engagement.
communities, practitioners,
professionals, businesses,
organisations, institutions, funders,
legislators, the environment (ecologies
such as a river, watershed,
plant/tree/insect/animal species, or a
natural landmark); and cultural,
ceremonial or historic practices and
sites.
18 3 18. The research issue or problem is Rate: 3
firmly located within the discipline of This is psychological study since it focuses on the
psychology as indicated by appropriate themes of: affective-motivational states, thriving at
psychological terms and constructs. work, work engagement, and sense of
(For example: perception, attitude, connectedness.
cognition, emotion, experience,
behaviour, discourse, identity,
community, ideology, etcetera.)
19 3 19. The interest in the problem, issue Rate: 3
or phenomenon is clarified. Why, how The article provides information on the affected
or how come is the research topic of individuals who are employers, employees and
interest to the relevant role-players, organisations with regard to productivity to
affected parties or interest groups? promote sustainable performance. The authors
states that emphasis needs to be placed on the work
experience, rather than on work outcomes. It is
also important that SMMEs develop employees
holistically, and that they create spiritually based
organisational cultures and that they pay more
attention to relationship management and
networking.
20 3 20. For research proposals at least 12 Rate: 3
references from relevant scientific and More than 24 references from relevant scientific
disciplinary literature are provided in and disciplinary literature are used in the article --
the literature overview. For research in the introduction; the research purpose and
articles at least 24 references are used objectives; and the literature review (workplace
in the literature review and results spirituality, work engagement, thriving at work,
discussion. and theoretical framework).
21 3 At least three relevant theoretical Rate: 3
perspectives are discussed, compared, The main theoretical framework used is the
evaluated and integrated -- to self-determination theory of Ryan and Deci
understand, describe or explain the (2000). The self-determination theory explains that
psychological dynamics of the individuals have three psychological needs,
research issue, phenomenon and namely autonomy, competence and relatedness
context. (Theoretical triangulation). needs.
Note that older work of seminal This is discussed in relation to various scholarly
theorists may still be relevant. findings on workplace spirituality (Saks's theory of
However, more recent and current the connection between workplace spirituality and
updates of theories should also be employee engagement), work engagement,
HMPYC80/104 217
HMPYC80/104/0/2024
presented. thriving at work (the socially embedded theory of

thriving at work; the social cognitive theory of
Bandura), and in relation to previous findings. The
in-depth introduction and literature review
sections are more than sufficient to explain the
psychological dynamics of the research issue,
phenomenon and context (theoretical
triangulation).
22 3 22. At least two relevant empirical Rate: 3
research results are incorporated and Ample empirical studies are referenced in the
discussed -- to further understand, article.
describe or explain the psychological
dynamics, occurrence, incidence or
prevalence (i.e. census, survey,
epidemiological or client data) of the
research issue or phenomenon.
23 3 23. The literature discussion continues Rate: 3
the scientific/ academic/ disciplinary The literature discussion continues the disciplinary
dialogue with published bodies of dialogue with published bodies of knowledge and
knowledge and expertise. (For expertise.
proposals a preliminary overview of
literature is provided, with indications
of key studies, theories and theorists
that will be used and elaborated on.)
24 3 24. The literature review is relevant to Rate:3
the topic, without unnecessary The literature review speaks directly to the
digressions or detours. The literature research issue that is investigated or explored in
review speaks directly to the research this study.
issue that is investigated or explored in
this study.
25 3 25. The author(s) (apparently) wrote Rate: 3
the text with academic integrity. It The article appears to be plagiarism-free, as the
appears to be their own and unique style of writing is (a) consistent and (b) of good
work, written in their own words and academic quality. In this case we are dealing with a
style. The literature review appears to peer-reviewed article; this means that you can
be plagiarism-free. All authors and safely assume that the manuscript is
literature sources are acknowledged in plagiarism-free. In general, peer-review processes
some way (even if the citation style is of academic journals include various
incorrect or imperfectly used.) plagiarism-checking mechanisms and procedures.
The work is published by a reputable journal and is
deemed to be plagiarism-free.
26 3 26. All authors and sources in the Rate: 3
literature review are cited correctly and Yes, all authors and sources in the literature review
follow the prescribed in-text citation appear to be cited correctly, and the prescribed
218
style. (APA7 style for HMPYC80, or in-text citation style ("APA or Other"), in this case
in the case of published journal articles the prescribed journal standard for citations is used
-- the prescribed referencing style of consistently.
the specific journal was adhered to.) The South African Journal of Industrial
Psychology used the APA6 style in 2018 when the
article was published.
27 3 27. The research ontology or paradigm Rate: 3
or approach is clearly demarcated and In the section "Research design - Research
is substantiated. It is made clear in the approach" it is stated that the study was conducted
discussion why the approach, in the positivist paradigm, and it was quantitative
paradigm or ontology was selected. in nature. (This upfront approach is noteworthy
and refreshing.)
28 3 28. The research question or research Rate: 3
statement is clearly formulated and In the "research purpose and objectives" on page 3
demarcated. The following two a discussion is provided on the aim of this article.
constructs are indicated in the research The aim (and main objective) of the research is to
question or statement: (a) demographic report on the relationship between thriving at
or contextual information of the work, work engagement and workplace
population or setting of the inquiry; spirituality.
and (b) a relevant psychological The secondary objective was to establish the
keyword or construct. degree to which the variance in thriving at work
and work engagement can be attributed to
workplace spirituality.
The (a) demographic or contextual information of
the population or setting of the inquiry is indicated
as "Hence, it is appropriate to consider thriving at
work in South African SMMEs, because it has not
been considered in this kind of work environment".
(b) Relevant psychological keywords or constructs
are used. Appropriate research hypotheses were
developed.
29 3 29. An appropriate (proposed) method Rate: 3
and (proposed) research design (will The Quantitative research study aimed to
be) were chosen to research the determine whether there is a statistically
question or issue. The research method significant relationship between the independent
and design are indicated. A rationale variable, namely workplace spirituality, and the
for the selected method is provided. dependent variables, namely work engagement
(Research design refers to a plan and and thriving at work. The survey research design
structure of the investigation that is that was used was cross-sectional in nature, as the
used to obtain evidence in order to research was carried out at a specific point in time.
answer or illuminate the research A self-administered questionnaires were used to
question). collect the data.
(Note that hypothesis testing applies to Research hypotheses were developed and tested.
randomised clinical trials, Statistical Package for the Social Sciences (SPSS)
experimental and quasi-experimental (version 20) was used to carry out the statistical
designs. Hypothesis testing may analysis. Descriptive statistics consisted of
HMPYC80/104 219
HMPYC80/104/0/2024
sometimes apply to correlational measures of central tendency, including means,

studies with the use of correlation or medians, standard deviations, skewness and
regression; or designs that test for kurtosis. Inferential statistical analysis was
group or sub-group difference with the performed using the Pearson’s
aid of t-tests, Anova, F-tests, Manova, product-moment correlation to determine
and other inferential statistics.) statistically significant relationships between
workplace spirituality, thriving at work and work
engagement, and their respective facets. A
multiple regression analysis was performed to
establish the degree to which the variance in work
engagement and thriving at work can be attributed
to workplace spirituality.
30 3 30. The data or evidence that was (will Rate: 3
be) collected to investigate the Data were collected through self-administered
research themes or variables appears questionnaires in a survey. The questionnaires
appropriate. included: a self-constructed biographical
questionnaire, a questionnaire to measure thriving
at work (learning and vitality), a questionnaire
measuring workplace spirituality (meaning and
purpose, awareness of life, and compassion), and a
questionnaire measuring work engagement
(vigour, dedication, absorption).
31 3 31. The (proposed) sampling strategy, Rate: 3
sample population and sample size are A total of 35 SMMEs operate within various
discussed in detail, and appear to be sectors of the economy were
appropriate. (The strengths and included in the study. It was not possible to obtain
limitations of the sampling strategy official statistics regarding the number of
must be indicated for academic employees that are working at SMMEs in the town
research projects or proposals, but is situated in the Free State province of South Africa.
not required for peer reviewed A proxy statistic was sued by using a sample
articles.) drawn from the number of residents in the town in
the national census data.
The Raosoft sample size calculator recommended
a sample size of 379 for this population size at a
confidence level of 95%. A non-probability
sampling technique of convenience sampling was
used.
However, over-sampling (in anticipation of low
questionnaire return rates) was not done. Of the
379 questionnaires that were distributed only 259
were returned and usable. The study had a
response rate of 68.3%.
32 3 32. The (proposed) data collection, Rate: 3
coding and analysis method(s) are Data Collection:
220
reliable; or dependable and credible. The self-administered questionnaires were
* For a study that uses quantitative personally distributed by the researcher and
research techniques, reliability is trained fieldworkers. The questionnaires were
demonstrated through consistency, collected personally by the researcher and the
replicability, and accuracy. Any one of fieldworkers at a prearranged time and day.
the following must be reported: Cronbach’s alpha coefficient (scale reliability
correlation statistics provide evidence coefficients) are reported for all scales and
of reliability of an instrument or compared to previous research findings.
technique, such as: internal item-test
and scale reliability coefficients,
test-retest correlations, and parallel
test-form correlation, inter-coder or
inter-observer correlations.
The data collection and coding
techniques are clearly explained, and
sufficient information is provided to
the reader to follow the procedures.
33 3 33. The (proposed) data collection, Rate: 3
coding and analysis method(s) are
trustworthy and valid. The factor analysis of thriving at work, workplace
* For a study that uses quantitative spiritualty
research techniques, any two of the and work engagement confirmed the factor
following sub-aspects must be structure of the
reported. Internal validity of measures constructs.
refer to: Prior to the distribution of the final
(i) face validity or content validity questionnaire, the constructed questionnaire was
(questions or statements should be distributed
appropriate to the research question); to two experts in the field of HR management, in
(ii) criterion validity (appropriate order to
criterion variables or instruments are ensure face validity.
selected); and The convergent validity of all the constructs
(iii) construct validity (constructs appearing in the measurement model is good (i.e.
selected for the study should be the AVE scores were all above 0.5). The
congruent with each other; this also discriminant validity is also supported which
includes convergent validity and indicates that the constructs are significantly
discriminating validity. different from
External validity relates to the degree each other. This implies that the instruments are
that the findings of the research can be reliable and
transferred to other contexts. The valid for this study
results can be generalisable and
applied to other similar settings,
populations, and situations. The
statistical power and adequacy of the
sampling strategy influences the
generalisablity of findings. The
researcher demonstrates objectivity by
HMPYC80/104 221
HMPYC80/104/0/2024
remaining neutral, uninvolved and

scientifically objective. (Human
experiences in the world are assumed
to be observable and measurable in the
form of a selection of variables and
their statistical inter-relations via
standardised tests, psychometric
instruments and procedures in the
positivistic/ post-positivist paradigm.)
34 1 34. The (proposed) procedures Rate: 1
whereby the data are (to be) There is no indication of procedures whereby the
professionally and ethically managed data were professionally and ethically managed.
are explained in sufficient detail. Any Examples could have included: data were captured
one of the following is sufficient: the on a password locked spreadsheet, and
information is appropriately recorded, questionnaires were stored under lock and key.
stored, and secured with a password or
lock.
35 1 35. The ethical prescription of Rate: 1
non-malevolence and avoidance of The author states that "The purpose of the study as
harm are complied with. (The rights well as ethical considerations, such as informed
and needs of vulnerable and minority consent, anonymity and confidentiality, was
groups are adhered to by prescribed explained in a cover letter. It was indicated in the
ethical and legal frameworks.) (If cover letter that participation was voluntary and
deception is a legitimate part of the that participants may withdraw from the study at
study design, this is dealt with in an any time" (page 6).
ethical and professional manner.) Non-malevolence and avoidance of harm are not
mentioned in the article, nor (apparently) referred
to in the cover letter.
(The item criterion is not explicitly and clearly
addressed.)
36 2 36. The ethical directive of informed Rate: 2
consent is adhered to. (Vulnerable and The author states that "The purpose of the study as
minority groups are treated within the well as ethical considerations, such as informed
prescribed parameters of relevant consent, anonymity and confidentiality, was
ethical and legal frameworks.) explained in a cover letter. It was indicated in the
cover letter that participation was voluntary and
that participants may withdraw from the study at
any time" (page 6).
The phrase informed consent is mentioned.
However, there is no further explanation or
explication of this concept or how it was
operationalised. Ideally the procedure how this
was acquired must be stated. Signed informed
consent forms as standard requirements for ethical
222
research. The criterion is only partially addressed
but not fully implemented/explained.
(Remember, if the item criterion is not
explicitly addressed then allocate a low rating.
Good academic writing is clear and
transparent. If a reader needs to guess at
whether an issue has been implemented and
addressed or not, then the writing is neither
clear nor transparent. Only give a high rating if
the item criteria have been explicitly
addressed.)
37 3 37. The ethical mandate of voluntary Rate: 3

participation is adhered to, and The author states that "The purpose of the study as
participants have the right to withdraw well as ethical considerations, such as informed
from the research study. consent, anonymity and confidentiality, was
explained in a cover letter. It was indicated in the
any time" (page 6).
It was indicated in the cover letter that
participation was voluntary and that participants
may withdraw from the study at any time. This is
the only ethical mandate that is clear and
self-evident.
38 2 38. The ethical imperatives of privacy, Rate: 2
confidentiality and anonymity are not The author states that "The purpose of the study as
violated. Ethics: privacy, well as ethical considerations, such as informed
confidentiality and anonymity: consent, anonymity and confidentiality, was
explained in a cover letter. It was indicated in the
any time" (page 6).
It was indicated in the cover letter that
participation was voluntary and that participants
may withdraw from the study at any time.
Anonymity and confidentiality are mentioned,
however, there is no clear discussion or
explanation how this was implemented. How was
privacy ensured? How were confidentiality and
anonymity ensured? The criterion is partially
addressed.
39 1 39. The ethical imperative of ethical Rate: 1
oversight is fulfilled. The research Ethical oversight and permission by relevant
proposal or project adheres to authorities and institutions are not mentioned in
institutional research policies, the article.
guidelines and procedures. The
HMPYC80/104 223
HMPYC80/104/0/2024
research proposal was (or will be)

submitted to all appropriate
institutional ethics review panels for
ethical oversight, and official written
permission was (will be) obtained to
perform the research.
Ethics: ethical oversight and
permission:
40 3 40. The grammar and spelling in the Rate: 3
research document are at least 95% The grammar and spelling in the article are of high
correct. quality.
41 3 41. The entire document from Rate: 3
beginning to end is plagiarism-free. All The article appears to be plagiarism-free, as the
important works, ideas, theories, style of writing is (a) consistent and (b) of good
authors, facts and statements are academic quality. In this case we are dealing with a
acknowledged in some way (even if peer-reviewed article; this means that you can
the citation style is incorrect or safely assume that the manuscript is
imperfectly used.) The authors(s) plagiarism-free. In general, peer-review processes
wrote the text with academic integrity; of academic journals include various
and it appears to be their own and plagiarism-checking mechanisms and procedures.
unique work. The work is published by a reputable journal and is
deemed to be plagiarism-free.
42 3 42. The various authors and sources in Rate: 3
the proposal or article (the entire Authors and sources in the article (the entire
document) are cited correctly and document) are cited correctly and follow the
follow the prescribed in-text citation prescribed in-text citation style. The South African
style. This should be correct for the Journal of Industrial Psychology used the APA6
entire document. (APA7 style for style in 2018 when the article was published.
HMPYC80 proposals, or the
prescribed referencing style of the
specific journal.) Was the correct
citation method used?
43 3 43. A complete list of all references Rate: 3
that were cited is provided at the end of A complete list of all references that were cited is
the document. Is it complete? provided at the end of the document.
44 3 44. The reference list at the end of the Rate: 3
document adheres to the required The reference list at the end of the document
standard and format. (APA7 style for adheres to the required standard and format. The
HMPYC80 for proposals, or in the South African Journal of Industrial Psychology
case of published journal articles -- the used the APA6 style in 2018 when the article was
prescribed referencing style of the published.
specific journal was adhered to.)
45 3 45. The work is of overall acceptable Rate: 3
academic standard, and it is written The work is of overall acceptable academic
224
using an appropriate standard, and it is written using an appropriate
academic/scientific style suitable to the scientific style suitable to the research paradigm,
research paradigm, approach and approach and journal.
journal (in the case of published
articles).
46 3 46. The data collection, coding and Rate: 3
analysis method(s) were performed Data collection: the self-administered
reliably; or dependably and credibly. questionnaires were personally distributed by the
(Apply the criteria as discussed in item researcher and trained fieldworkers. The
32.) questionnaires were collected personally by the
researcher and the fieldworkers at a prearranged
time and day.
Scale reliability coefficients: Cronbach’s alpha
coefficients are reported for all scales and are
compared to previous research findings. (The
criteria as discussed in item 32 were suitably
fulfilled.)
47 3 47. The way in which the data were Rate: 3
analysed and interpreted is explained The way in which the data were analysed and
in sufficient detail and provides interpreted is explained in sufficient detail and
sufficient information for a reader to be provides sufficient information for a reader to be
able to follow the procedures. able to follow the procedures.
48 3 48. The data collection, coding and Rate: 3
analysis method(s) were applied in a The author reports that the factor analysis of
correct and substantiated manner and thriving at work, workplace spiritualty and work
appear trustworthy and valid. (Apply engagement confirmed the factor structure of the
the criteria as discussed in item 33.) constructs (construct validity).
Prior to the distribution of the final questionnaire,
the constructed questionnaire was distributed
to two experts in the field of HR management, in
order to ensure face validity.
The convergent validity of all the constructs
appearing in the measurement model is good (i.e.
the scores were all above 0.5). The discriminant
validity is also supported which indicates that the
constructs are significantly different from
each other. This implies that the instruments are
reliable and valid for this study. (The criteria as
discussed in item 33 were suitably fulfilled.)
49 2 49. The data analysis and interpretation Rate: 2
were performed ethically. (Refer to It is unclear whether and how the data analysis and
RESEARCH ETHICS: items 35 to 39.) interpretation were performed ethically with
regard to non-malevolence, avoidance of harm,
informed consent
privacy, confidentiality and anonymity. Most of
HMPYC80/104 225
HMPYC80/104/0/2024
the issues are mentioned but not unpacked and

discussed in sufficient detail. The information is
not wrong, nor is it completely absent. The
criterion is partially attained, albeit in a very
limited way. A rating of 2 is allocated.
50 3 50. The results are presented Rate: 3
appropriately and correctly. The results are presented appropriately and
correctly.
51 3 51. The results are clearly structured, Rate: 3
well presented and easy to follow. The results are clearly structured, well presented
and easy to follow.
52 3 52. A discussion is presented which Rate: 3
interprets or explains the results and A discussion is presented which interprets or
their implications for the research explains the results and their implications for the
question or issue. research question or issue.
53 3 53. The discussion links and explains Rate: 3
the results to the existing bodies of The discussion clearly links and explains the
literature of as discussed in the results to the existing bodies of literature of as
literature review. It links to existing discussed in the literature review and to the
theoretical perspectives, and where hypotheses.
relevant to existing empirical literature
findings.
54 3 54. The research paradigm, theoretical Rate: 3
perspectives in the literature, research There is a “golden thread” that runs through the
methods, data gathering techniques, document and connects each subsection and
data coding and analysis, and section to the whole.
interpretations that were employed are
in line with the research question and
aim of the study. There is a “golden
thread” that runs through the document
and connects each subsection and
section to the whole.
55 1 55. The possibility of alternative Rate: 1
explanations or interpretations of the No alternative explanations or interpretations of
results is considered. the results is considered.
56 3 56. The possibility or need for further Rate: 3
research is raised. "This study focused on only a contextual factor
(i.e. workplace spirituality) to promote work
engagement and thriving, and, as such, it is
recommended that future studies should include
both contextual and individual factors, in order to
establish how the interaction between these factors
influences self-directed behaviour (Wallace et al.,
2013)" (p. 9).
226
And "Although work engagement has been
the focus of many empirical studies, thriving at
work is a newer construct, and, as such, more
research should be conducted regarding the
individual and contextual variables that can
promote thriving at work" (p. 9).
57 3 57. The strengths, weaknesses and Rate: 3
limitations of the study are presented. Strengths, weaknesses & limitations are included
(see: Limitations and Recommendations, p. 9.)
Firstly, a non-probability sampling technique was
used, and data were gathered at a specific location.
This may compromise the external validity of the
research reported on. Caution needs to be taken
when generalising the findings to the larger
population.
Secondly, common-method variance may have
influenced the results, because a cross-sectional
study was conducted, and self-administered
questionnaires were used to collect the data, which
contained more than one construct that
respondents had to respond to at the same time.
Thirdly, the sample was not representative of the
population, and therefore, the findings cannot be
generalised to the larger population.
However, an acceptable response rate was
obtained, and the final sample consisted of an
adequate number of respondents in order to
achieve the research objectives of the study.
Despite the above limitations, the study holds
important implications for SMMEs that wish to
promote sustainable performance.
58 3 58. A final conclusion is presented that Rate: 3
succinctly reiterates the main research A final conclusion is presented that succinctly
findings or results. reiterates the main research finding or result.
HMPYC80/104 227
HMPYC80/104/0/2024

Application 03
The Research Review Inventory (R.R.I) is provided in Resource Material 02 of Tutorial Letter 102. The
58 items of the R.R.I. are used to review the article:
Langa, M. (2017). Boys to men: Narrating life stories of fatherhood and work life amongst young
black men. PINS, 55, 61 - 83. http://dx.doi.org/10.17159/2309-8708/2017/n55a5
https://www.pins.org.za/pins/pins55/langa.pdf
Use the Research Review Inventory (See Resource material 01, 02 and 03 in Tutorial Letter 102) to
review the article. This review is based on the previous version of the Research Review Inventory
(R.R.I. revised). Therefore 2 of the items will differ from the R.R.I.3 that ius provided in Tutorial Letter
102.

follows:
1 or a: Absent, incorrect, or completely wrong.
2 or b: Only partially addressed. It is does not attain the item requirement(s) or
criterion/criteria.
3 or c: Sufficiently up to standard in terms of the item criteria.
Remember, if the item criterion is not explicitly addressed then allocate a low rating. Good
academic writing is clear and transparent. If a reader needs to guess at whether an issue has been
implemented and addressed or not, then the writing is neither clear nor transparent. Only give a high
rating if the item criteria have been explicitly addressed.
Item Rate Rating item (R.R.I) Justification

01 3 01. The title contains at least one Yes, the item criterion is achieved and a rating of 3
central psychological construct from (sufficiently up to standard) is allocated.
the research proposal or study. "fatherhood" and "work life" are key psychological
constructs.
02 3 02. The title indicates the research Yes, the item criterion is achieved and a rating of 3
setting or demographic information (sufficiently up to standard) is allocated.
228
of the sample population. Any one of The who: "Boys to men", "young black men", and
the following is sufficient: where, the what: "fatherhood and work life", "life stories"
who, when, under what are indicated.
circumstances, or the phenomenon
that is being investigated.
03 2 03. The title points to, or refers to the
2 The item criterion is not fully achieved. A rating of
research method, design or paradigm. 2 (only partially addressed) is allocated.
This may be signalled by a keyword Although the research method is signalled, i.e.
or phrase such as: “lived "narrating" points to a "narrative approach" that
experiences”, “perceptions”, places the story of each participant at centre stage in
“attitudes”, “constructions”, the data analysis. It is also mentioned that "Life
“discourses”, “quasi-experimental”, stories" were used and this point to a qualitative
“critical study”, “psychometric approach.
properties”, “correlational study”, However, photovoice as the main method or
“inferential study”, and so on. (auto-)photography is not expressly mentioned.
The method or design that was used in the study is
"photo voice" or auto-photography followed by
semi-structured individual interviews
("Auto-photography" by Noland, 2006;
"photo-biographical research" by Blackbeard and
Lindegger, 2007; and "Photovoice" by Kessi,
2011). The different names for the research method
are not discussed, nor unpacked in the article. The
approach or paradigm is also not unpacked. It is not
stated whether this participatory action research, or
liberatory psychology, or interpretive research, or
narrative research, or psychohistory/
psychobiography, or other. This omission makes the
exact approach and research design unclear.
04 3 04. The abstract indicates the The abstract (or summary/ executive summary) is
research interest, issue or problem. provided at the beginning of the article.
Yes, the item criterion is achieved and a rating of 3
(sufficiently up to standard) is allocated.
The research interest, issue or problem is indicated:
" the life stories of young black men, specifically
how they negotiate their masculine identities over
time." And "tracked a group of young black men
over a period of nine years, from when they were
adolescent boys (between the ages of 13 and 18
years), until they were young adults (between the
ages of 23 to 26 years at the time of the writing)."
05 3 05. The abstract indicates the setting, Yes, the item criterion is achieved and a rating of 3
demographic or circumstances (sufficiently up to standard) is allocated.
pertaining to the research interest or "Young black men over a period of nine years, from
question. Any one of the following is when they were adolescent boys (between the ages
sufficient: where, who, when, under of 13 and 18 years), until they were young adults
HMPYC80/104 229
HMPYC80/104/0/2024
what circumstances, or the (between the ages of 23 to 26 years at the time of the
phenomenon that is being writing)." And "The aim of the study was to explore
investigated. how the participants spoke about their relationships
with their fathers as young adolescent boys and how
they were now fathers to their own children as
young men."
06 3 06. The abstract contains the specific Rate: 3
research question or research The aim of the study was to examine the life stories
statement. of young black men, specifically how they negotiate
their masculine identities over time. And, to explore
how the participants spoke about their relationships
with their fathers as young adolescent boys and how
they were now fathers to their own children as
young men. The study looks at participant's
experience of: "My life as a young black man in the
new South Africa”. (This is sufficient as a research
statement.)
07 1 07. The abstract indicates the key Rate: 1
theoretical perspective(s) that is (are) The abstract does not mention any theoretical
used in the literature. perspectives.
08 2 08. The abstract specifies the Rate: 2. A rating of 2 (only partially addressed) is
(proposed) research method, study allocated.
type or research design. Life stories were researched by using cameras to
take photos on the theme under the theme, “My life
as a young black man in the new South Africa”.
These were used to facilitate semi-structured
interviews. Follow-up interviews were conducted
with some of the participants.
Note that the exact method/ design that was used in
the study is not indicated -- i.e. "photo voice" or
auto-photography followed by semi-structured
individual interviews ("Auto-photography" by
Noland, 2006; "photo-biographical research" by
Blackbeard and Lindegger, 2007; and "Photovoice"
by Kessi, 2011). The different names for the
research method are not discussed, nor unpacked in
the article. The approach or paradigm is also not
unpacked. It is not stated whether this participatory
action research, or liberatory psychology, or
interpretive research, or narrative research, or
psychohistory/ psychobiography, or other. This
omission makes the exact approach and research
design unclear.
230
(proposed) data collection method. Photographs and semi-structured interviews were
used. This is sufficient and correct.
10 2 10. The abstract describes the Rate: 2
(proposed) population sample, and This discussion is incomplete and does not attain the
sampling strategy or approach. item requirement(s) or criterion/criteria.
The sample population is South African black men.
This is a large population, and it is not clear where
the sample was drawn from (Alexandra,
Johannesburg).
It is not mentioned where the research sample was
specifically accessed (at a school).
The method of sampling and inclusion is also not
mentioned (random or non-random, convenience,
etc.).
11 2 11. The abstract indicates the Rate: 2. A rating of 2 (only partially addressed) is
(proposed) data analysis method or allocated.
technique. No specific data analysis method is mentioned.
However, it is stated that "Key themes in the life
stories included…" and "narratives of being a young
black man".
This indicates that a narrative (or qualitative) data
analysis method was used regarding life stories in
the context of semi-structured interviews.
12 3 12. For completed research projects Rate: 3
only -- the abstract briefly mentions The findings are listed: " Key themes in the life
the actual findings of the study, and stories included relationships with mothers,
states how the study answers, experiences of growing up without fathers, entering
addresses or illuminates the research the world of work, and being fathers themselves
issue or question. which encouraged them to also reflect about their
(Proposals will receive a “1” for this own relationships with their fathers. It is clear from
item as the actual research has not yet their experiences that narratives of being a young
been performed.) black man are not static, but continuously change
depending on the context, and time."
13 2 13. The list of keywords contains at Rate: 2.
least (a) one key psychological The list of keywords include:
construct and (b) one key reference (a) Fatherhood and masculinities as key
to a psychological theory. psychological constructs.
(b) -
(There is no keyword that references a
psychological theory. Note that fatherhood,
masculinities, and narratives do not refer to
specific psychological theories.)
The criteria are only partially addressed.
research setting and demographic Fatherhood, masculinities, narratives (the
information of the sample. phenomenon that is being investigated).
HMPYC80/104 231
HMPYC80/104/0/2024
Any two of the following is Young black men in “My life as a young
sufficient: where, when, who, under black man in the new South Africa” (who are
what circumstances, or the investigated). The context and setting are
phenomenon that is being sufficiently described for a rating of 3.
investigated.
research method, study type or Narratives and photography are mentioned.
research design. (This may be However, there is no mention of photo voice, life
indicated directly or indirectly by stories, or interviews. The criterion is only partially
phrases such as: “lived experiences”, addressed.
“perceptions”, “attitudes”,
“quasi-experimental”, “critical The method or design that was used is "photo voice"
study”, “psychometric properties”, or auto-photography followed by semi-structured
“correlational study”, “inferential individual interviews ("Auto-photography" by
study”, and so on.) Noland, 2006; "photo-biographical research" by
psychohistory/ psychobiography, or other.
(See the textbook chapter on Participatory methods).
This omission makes the exact approach and
research design unclear.
16 3 16. The context (who, what, where, Rate: 3
how, how come, when) of the The research context is provided in the Introduction.
research issue is clearly described. Township masculinities and " The dominant
(Note that the rating issue can be discourse that has emerged out of these studies is
addressed anywhere in the proposal that young black men are more likely [than young
or article.) men of other races] to engage in risky and violent
behaviours. The construction of hegemonic
masculinity has been found to be a key element in
these risk-taking behaviours. According to Connell,
(1995), hegemonic masculinity is the term used to
refer to the dominant cultural stereotype of
masculinity. This includes, for example, the
dominant social construction of men as brave,
strong, aggressive and resilient, in many societies.
The other dominant cultural stereotypes associated
with hegemonic masculinity also suggest that a real
man is evidenced in the ability to demand sexual
intercourse with multiple partners. Accordingly,
232
hegemonic masculinity is predominantly
encapsulated in being a heterosexual, healthy,
competitive male who sees women as sexual objects
and competes against other males for access to
women (Connell, 1995; Niehaus, 2005)." …
" Later, Connell (2000; 2005) accepted some of
these criticisms that hegemonic masculinity is
characterized by the constant process of negotiation,
translation and reconfiguration in order to adapt to
new historical times and contexts. This shows that
the identity of being a male is not static, which is
also evident in the current study in how young men
spoke about their relationships with their fathers and
how they were fathering their own children. Morrell,
Jewkes and Lindegger (2012) raised a concern about
how the term hegemonic masculinity is often used in
the existing scholarship in South Africa. The
concept is given a specific, fixed content and
purpose. As a result, it has lost its fluidity and
flexibility as some researchers use the term rigidly
(ibid)."
17 3 17. The relevant role-players, Rate: 3
affected parties or interest groups The relevant role-players, affected parties or interest
with regard to the research issue or groups mentioned include: researchers who study
problem are specified or mentioned. masculinity, growing up without fathers, the role of
Role-players, affected parties and mothers, boyhood, and fatherhood;
interest groups can include: the
the interviewees themselves,
researcher, individuals, clients,
students, beneficiaries, patients, and boys, young men and their mothers;
families, households, communities, young fathers in general;
practitioners, professionals,
businesses, organisations,
institutions, funders, legislators, the
environment (ecologies such as a
river, watershed,
plant/tree/insect/animal species, or a
natural landmark); and cultural,
ceremonial or historic practices and
sites.
18 3 18. The research issue or problem is Rate: 3
firmly located within the discipline of "Research to explore how the participants spoke
psychology as indicated by about their relationships with their fathers as young
appropriate psychological terms and adolescent boys and how they were now fathers to
constructs. (For example: perception, their own children as young men. The aim of the
attitude, cognition, emotion,
study was to examine the life stories of young black
experience, behaviour, discourse,
identity, community, ideology, men, specifically how they negotiate their masculine
etcetera.) identities over time." This is a psychological study.
HMPYC80/104 233
HMPYC80/104/0/2024
19 3 19. The interest in the problem, issue Rate: 3

or phenomenon is clarified. Why,
how or how come is the research Yes, the research is located in current research
topic of interest to the relevant studies on township masculinities (Chili, 2013;
role-players, affected parties or Langa, 2010; Ratele, 2013) and relate to various
interest groups? (Why or how is the discourses and stereotypes of masculinity. The
research of benefit to the role-players relevance of the research for various role-players,
mentioned in Question17?) including the interviewees, is evident.
20 3 20. For research proposals at least 12 Rate: 3
references from relevant scientific More than 24 references from relevant scientific and
and disciplinary literature are disciplinary literature are used in the article -- in the
provided in the literature overview. introduction, literature review, and the discussion.
For research articles at least 24 (Refer to the texts in the references.)
references are used in the literature
review and results discussion.
21 3 At least three relevant theoretical Rate: 3
perspectives are discussed, There is an in-depth discussion of township
compared, evaluated and integrated -- masculinities, problems focussed research
to understand, describe or explain the associated with young black men, The dominant
psychological dynamics of the discourse that has emerged out of these studies is
research issue, phenomenon and that young black men are more likely [than young
context. (Theoretical triangulation). men of other races] to engage in risky and violent
Note that older work of seminal behaviours. hegemonic versions of masculinity, and
theorists may still be relevant. criticisms have been levelled against Connell’s
However, more recent and current conception of hegemonic masculinity -- that
updates of theories should also be hegemonic masculinity is often used in the existing
presented. scholarship in South Africa. The concept is given a
specific, fixed content and purpose.
The next discussion of research on fatherhood in
South Africa with a main focused on absent
fathering and meanings made by those who grew
without fathers is challenged. For example, the
definition of a father in this article is not limited to
the biological relationship, but to “any older man
who fulfils a role of care, protection and provision in
relation to children” (Richter & Morrell, 2006: 2).
The literature, in general, does not acknowledge the
positive role that female-headed households play in
raising boy children. And, many studies into
“alternative” constructions of masculinity are based
on the constructivist premise that masculinity is
fluid, multiple and flexible (Frosh et al, 2002; Shefer
et al, 2007).
22 3 22. At least two relevant empirical Rate: 3
234
research results are incorporated and A number of empirical research studies are
discussed -- to further understand, referenced in the article (more than 2).
describe or explain the psychological
dynamics, occurrence, incidence or
prevalence (i.e. census, survey,
epidemiological or client data) of the
research issue or phenomenon.
23 3 23. The literature discussion Rate: 3
continues the scientific/ academic/ The literature discussion continues the disciplinary
disciplinary dialogue with published dialogue with published bodies of knowledge and
bodies of knowledge and expertise. expertise.
(For proposals a preliminary
overview of literature is provided,
with indications of key studies,
theories and theorists that will be
used and elaborated on.)
24 3 24. The literature review is relevant Rate: 3
to the topic, without unnecessary The literature review speaks directly to the research
digressions or detours. The literature issue that is investigated or explored in this study.
review speaks directly to the research
issue that is investigated or explored
in this study.
25 3 25. The author(s) (apparently) wrote Rate: 3
the text with academic integrity. It The article appears to be plagiarism-free, as the style
appears to be their own and unique of writing is (a) consistent and (b) of good academic
work, written in their own words and quality. In this case we are dealing with a
style. The literature review appears to peer-reviewed article; this means that you can safely
be plagiarism-free. All authors and assume that the manuscript is plagiarism-free. In
literature sources are acknowledged general, peer-review processes of academic journals
in some way (even if the citation style include various plagiarism-checking mechanisms
is incorrect or imperfectly used.) and procedures. The work is published by a
reputable journal and is deemed to be
plagiarism-free.
26 3 26. All authors and sources in the Rate: 3
literature review are cited correctly Yes, all authors and sources in the literature review
and follow the prescribed in-text appear to be cited correctly, and the prescribed
citation style. (APA7 style for in-text citation style ("APA or Other"), in this case
HMPYC80, or in the case of the prescribed journal standard for citations is used
published journal articles -- the consistently.
prescribed referencing style of the The Psychology in Society (PINS) journal uses the
specific journal was adhered to.) Harvard referencing style - refer to the style-guide
on the website.
(Note that there are 2 references that appear in the
references but are not in the text. Also, one of the
references in the text does not have the same date in
the reference list. These are minor issues and are
HMPYC80/104 235
HMPYC80/104/0/2024
negligible -- the overall standard is high.)

27 2 27. The research ontology or Rate: 2
paradigm or approach is clearly (a) It is stated that "The photo voice method was used to
demarcated/defined and is (b) capture the life stories of young black men." "This
substantiated -- it is made clear in the study focused on exploring the narratives of how
discussion why the approach, young black men in the context of Johannesburg’s
paradigm or ontology was selected. Alexandra Township. A narrative approach was
used."
In the concluding remarks the author states that:
"Photography provided the participants with the
opportunity to share their detailed life stories. This
research methodology has proved to be a novel way
of allowing participants to be champions of their
own narratives in terms of events that they chose to
focus on and share in the interviews." However, the
research approach and paradigm is not elaborated
nor explained in detail.
The method or design that was used is "photo voice"

or auto-photography followed by semi-structured
individual interviews ("Auto-photography" by
Noland, 2006; "photo-biographical research" by
psychohistory/ psychobiography, or other.
(See the textbook chapter on Participatory methods).
This omission makes the exact approach and
research design unclear.
28 3 28. The research question or research Rate: 3
statement is clearly formulated and The aim of the study was to examine the life stories
demarcated. The following two of young black men, specifically how they negotiate
constructs are indicated in the their masculine identities over time from
research question or statement: (a) adolescence to young adults.
demographic or contextual Another aim was to explore how the participants
information of the population or spoke about their relationships with their fathers as
setting of the inquiry; and (b) a young adolescent boys and how they were now
relevant psychological keyword or fathers to their own children as young men.
construct. The study looks at participant's experience of: "My
life as a young black man in the new South Africa”.
236
The research question is clear.
29 2 29. An appropriate (proposed) Rate: 2
method and (proposed) research No clear and in-depth rationale for the chosen
design (will be) were chosen to research method is provided.
research the question or issue. The
research method and design are The method or design that was used is "photo voice"
indicated. A rationale for the selected or auto-photography followed by semi-structured
method is provided. individual interviews ("Auto-photography" by
(Research design refers to a plan and Noland, 2006; "photo-biographical research" by
structure of the investigation that is Blackbeard and Lindegger, 2007; and "Photovoice"
used to obtain evidence in order to by Kessi, 2011). The different names for the
answer or illuminate the research research method are not discussed, nor unpacked in
question). the article. The approach or paradigm is also not
(Note that hypothesis testing applies unpacked. It is not stated whether this participatory
to randomised clinical trials, action research, or liberatory psychology, or
experimental and quasi-experimental interpretive research, or narrative research, or
designs. Hypothesis testing may psychohistory/ psychobiography, or other.
sometimes apply to correlational (See the textbook chapter on Participatory methods).
studies with the use of correlation or The author mentions that a longitudinal study
regression; or designs that test for design was used.
group or sub-group difference with
the aid of t-tests, Anova, F-tests,
Manova, and other inferential
statistics.)
30 3 30. The data or evidence that was Rate: 3
(will be) collected to investigate the Data were collected through photographs (first on
research themes or variables appears disposable cameras and later on cell phones) and
appropriate. subsequent semi-structured interviews over a period
of time in a longitudinal study. The questions or
themes explored in the interviews are discussed.
31 2 31. The (proposed) sampling Rate: 2
strategy, sample population and The researcher initially went to two high schools in
sample size are discussed in detail, Alexandra and recruited 32 Grade 10 and 11 boys
and appear to be appropriate. (The between the ages of 13 and 18 years in 2007. The
strengths and limitations of the researcher advertised the study by visiting classes of
sampling strategy must be indicated Grade 10 and 11 to meet with boy learners to explain
for academic research projects or the nature of the research project and to provide
proposals, but is not required for peer them with an information sheet, parental consent
reviewed articles.) and participant assent forms which required the
appropriate signatures.
Only boys who returned the signed forms were
included in the study. The number of participants
exceeded 18, the initial number that the researcher
wanted.
Following the first phase of data collection in 2007,
the researcher followed 12 of the participants over a
HMPYC80/104 237
HMPYC80/104/0/2024
period of nine years (2007 to 2017). At the time of

writing this article, these participants were between
the ages of 23 and 27 years. Eleven of them had
completed high school, one had dropped out in
Grade 11 and four had completed tertiary-level
diplomas. Ten of the participants were working, one
was unemployed and one was in prison. I conducted
between four and fourteen individual follow-up
interviews with each of the 12 participants since
2007. During this second longitudinal phase of the
study, the participants used their cell phones rather
than disposable cameras to take photos representing
their lives.
The method of sampling can be deduced from the
above discussion (convenience sampling with
follow-up over time).
The exact method or reason or method for selecting
(a) these 2 schools and (b) the 12 participants for
follow up are not clearly explained. (Was it due to
convenience or access, what led to the inclusion of
the final 12 participants? These issues were not
addressed.) A rating of 2 (only partially addressed)
is allocated.
32 2 32. The (proposed) data collection, Rate: 2 (partially addressed)
coding and analysis method(s) are The step-wise data collection method was discussed
reliable; or dependable and credible. as indicated in the previous item. All interviews
(* For a study that uses quantitative were transcribed.
research techniques, reliability is The author states that "a narrative approach was
demonstrated through consistency, used" (p. 65) under the heading data analysis.
replicability, and accuracy. Any one "The initial stage of data analysis involved reading
of the following must be reported: each transcript many times with the aim of
correlation statistics provide determining the prevailing themes relating to
evidence of reliability of an masculinity and its contestation in terms of voices to
instrument or technique, such as: which the participants publicly or privately
internal item-test and scale reliability subscribe. It was in the interest of this study, given
coefficients, test-retest correlations, its narrative approach
and parallel test-form correlation, to place the story of each participant at centre stage
inter-coder or inter-observer in the data analysis." P.65
correlations. " This involved organising the analysis of each
The data collection and coding narrative coherently in terms of meanings that each
techniques are clearly explained, and participant made about his upbringing, relationship
sufficient information is provided to with mothers, absent fathers, search for work,
the reader to follow the procedures.) working and becoming fathers. Particular attention
was also paid to the manner in which personal
238
* In qualitative research techniques, stories were shared. The photos taken by the
credibility, dependability, and participants were also central in the analysis as
consistency can be achieved through many of the narratives were linked to back to the
various practices. Any two of the photos. The photos were seen and interpreted by the
following must be reported: thick participants as an important representation of
rich descriptions (in-depth significant events in their personal lives." p.65
illustration that explicates culturally Thick rich descriptions are provided in the
situated meanings and abundant discussion.
concrete detail), triangulation or Transparency and verifiability - there were
crystallization, reasoned consensus, "member checks" by way of repeated interactions
multi-vocality (different perspectives, over a number of years with participants (a
nuances and voices are articulated), longitudinal study design was used).
and via inter-observer or inter-coder (There is no report of an audit trail.)
agreement. Transparency and The author reports that "Reflections and personal
verifiability are established via: dynamics about the relationship that the researcher
extensive written field research notes, formed with the participants is beyond the scope of
member checks by participants, peer this article." (p.64) This points to reflexivity to make
review and an audit trail. Reflexivity the biases, values, and experiences of the researcher
is used by the researchers to make the explicit -- but is not discussed.
biases, values, and experiences that The exact method of data coding and analysis is not
they bring to the research explicit. described in any detail, and this raises questions
Any changes to the study or about the method used.
unexpected occurrences are For this reason a rating of 2 (only partially
documented. (For more detailed addressed) is allocated.
method-specific criteria see
Creswell, 2007; and other sources).
33 2 33. The (proposed) data collection, Rate: 2 (partially addressed)
coding and analysis method(s) are
trustworthy and valid. As the discussion of the previous item shows there
(* For a study that uses quantitative are some methodological issues and steps that have
research techniques, any two of the been omitted from the discussion and reporting of
following sub-aspects must be results.
reported. Internal validity of
measures refer to: The believability and trustworthiness of the findings
(i) face validity or content validity should be supported by a sequentially and logically
(questions or statements should be presentation of the processes and procedures of the
appropriate to the research question); research method and technique. There is evidence of
(ii) criterion validity (appropriate this, but it is not fully clarified.
criterion variables or instruments are
selected); and Transferability refers to the degree that the findings
(iii) construct validity (constructs of the research can be transferred to other contexts
selected for the study should be by the readers (called “external validity” in
congruent with each other; this also quantitative approaches). The results appear to be
includes convergent validity and transferable to a large extent.
discriminating validity. Confirmability is the inter-subjective agreement
External validity relates to the degree refers to the degree to which the results could be
HMPYC80/104 239
HMPYC80/104/0/2024
that the findings of the research can confirmed or corroborated by other researchers. If
be transferred to other contexts. The findings are corroborated or confirmed by others
results can be generalisable and who examine the data, then no inappropriate biases
applied to other similar settings, impacted the data analysis. This cannot be directly
populations, and situations. The established by studying this article and is lacking.
statistical power and adequacy of the
sampling strategy influences the "For the purpose of this
generalisablity of findings. The article, only four case studies were chosen to
researcher demonstrates objectivity highlight the participants’ narratives of
by remaining neutral, uninvolved and growing up without fathers, the positive role that
scientifically objective. (Human their mothers played in their lives,
experiences in the world are assumed becoming fathers themselves, finding job
to be observable and measurable in opportunities and reflecting on challenges
the form of a selection of variables associated with being a young black man in the new
and their statistical inter-relations via South Africa." (p.64) However, insufficient support
standardised tests, psychometric is provided to show why only 4 participants' stories
instruments and procedures in the were selected and how they link to the results from
positivistic/ post-positivist all 12 participants.
paradigm.) )
The (proposed) data collection, coding and analysis
* In qualitative research techniques the method(s) appears to be trustworthy and valid (with
concept of “validity” (from a few questions). However, more information on the
quantitative designs) is not a single, methodological steps and actions would have raised
fixed or universal concept, but the reporting of results to a next level.
“rather a contingent construct,
inescapably grounded in the
(Note that the results are authentic, coherent and
processes and intentions of particular
research methodologies and projects” cogent -- and link well to the theories and findings
(Winter, 2000, p.1). Any two of the that were discussed in the literature review. The
following concepts that are either finding that "the present study confirms
underlined or in italics must be Bozalek’s (1997) South African based finding that
reported, even if there is overlap with female-headed households can raise well-adjusted
the previous rating item: adolescent boys. This argument debunks the notion
that boy children always need their fathers to
The believability and develop healthy masculine identities." is drawn from
trustworthiness of the findings a very small sample -- and it is doubtful whether this
(called “internal validity” in study debunks the notion, however it goes some way
quantitative approaches) is dependent to dispel the ubiquity of that notion. More research
on the richness of the data. The is needed to debunk the above notion (to some
participants of the study are the only extent the research claims exceed the "power of
ones that decide if the results actually generalisibility" or range and scope of the research
reflect the phenomena being studied. design).
Triangulation is a commonly used
method to verify accuracy, and A counter argument is that the 12 participants out of
involves cross-checking information the initial 32 may have had a form of self-selection
240
from multiple perspectives and which made them continue -- as opposed to those
multiple methods of data collection. who did not continue or were not selected to
Triangulation may involve the continue with the research. It is not clear how the 12
involvement of multiple researchers, participants came to be in the longitudinal study.
multiple data sources or types of data,
multiple methods of design and
analysis, and multiple theoretical
lenses. The processes and procedures
of the method and technique are
sequentially and logically presented.
This done so that findings and
conclusions about data
interpretations are made explicit and
rendered credible and believable.
Disconfirming cases are used to test
the findings and interpretations.
Transferability refers to the degree

that the findings of the research can
be transferred to other contexts by the
readers (called “external validity” in
quantitative approaches). However,
researchers bring their own unique
perspectives to the research process
and data interpretation as a subjective
or inter-subjective aspect in
qualitative research. Reflexivity is
used to make the subjective biases,
values, and experiences of the
researcher explicit. Transferability
therefore refers to dense descriptions
of processes and procedures used in
the study to try and answer the
research question.
Confirmability as inter-subjective
agreement refers to the degree to
which the results could be confirmed
or corroborated by other researchers.
If findings are corroborated or
confirmed by others who examine the
data, then no inappropriate biases
impacted the data analysis.
Qualitative researchers seek
illumination, understanding, and
extrapolation to similar situations.
HMPYC80/104 241
HMPYC80/104/0/2024
(For more detailed method-specific

criteria see Creswell, 2007; and other
sources)
34 1 34. The (proposed) procedures Rate: 1
whereby the data are (to be) There is no indication of procedures whereby the
professionally and ethically managed data were professionally and ethically managed.
are explained in sufficient detail. Any Examples could have included: data were captured
one of the following is sufficient: the on a password locked spreadsheet, and
information is appropriately questionnaires were stored under lock and key.
recorded, stored, and secured with a
password or lock.
35 2 35. The ethical prescription of Rate: 2
non-malevolence and avoidance of The ethical prescription of non-malevolence and
harm are complied with. (The rights avoidance of harm are not directly discussed. There
and needs of vulnerable and minority is no extended discussion of ethics apart from the
groups are adhered to by prescribed discussion of informed consent (p.65).
ethical and legal frameworks.) (If [For this review with the R.R.I. it must be discussed
deception is a legitimate part of the in detail for a rating of 3.]
study design, this is dealt with in an
ethical and professional manner.) In terms of ethical behaviour it is clear that the
researcher went the extra mile to support the
participants.
" Furthermore, I used my clinical skills as a
counselling psychologist to contain some
participants and help them deal with feelings of
disappointment and other challenges of life."
"Follow-up interviews provided an opportunity for
them to reflect and “think beyond the confines of
their limited worlds and to envision the range of
possible paths their lives could follow”. Follow-up
interviews were used to offer some guidance about
possibilities that the participants could follow in
realising their career goals." And
"Other follow-up individual sessions were held with
Simon to debrief him regarding the death of his
brother. It is important to note that the researcher is a
qualified Counselling Psychologist, which helped
him to contain the participants’ feelings and
emotions. However, it is also important to note that
it was not always easy to respond to the participants’
emotions as a Psychologist and still fulfil the role of
the researcher. Overall, debriefing and making
follow-ups provided useful data on the emotional
world of the participants, which also facilitated a
242
deeper understanding about their feelings,
relationships and difficulties in their everyday
lives". The commitment to, and demonstration of
non-malevolence and avoidance of harm, in a
context of low socio-economic status and resources,
is admirable and praise-worthy.
Another aspect, involving ethical reflexivity is

mentioned on page 76: " My questions in the above
extract reveal that
I was explicitly pushing Marcus to say something
about his sexual identity. With
hindsight and self-reflection now, I was also able to
uncover my own prejudices
in associating certain occupations (for example,
hairdressing) with being gay (as the other boys had
indicated)."
This is a commendable reflection and shows that the
researcher dealt with evolving and emerging
dynamics within himself and with the participants in
a mature and honest way. A solid relational ethical
stance is reflected.
It also emphasises a key point/argument in this

article that: "This reinforces the argument that
Ratele (2013) makes about the need to create safe
spaces for young black men to talk about themselves
and to reflect on their identities. Such safe spaces
may contribute immensely to the development of
positive masculinities that are non-violent,
non-aggressive and future-oriented in an egalitarian
way. It is also through initiatives such as this study
that young black men may begin to imagine new
lives that are characterised by positive futures."
(p.77). An appropriate and splendid ethical stance
toward the wider community is reflected in this
quote that embodies non-malevolence and
avoidance of harm at all appropriate levels
(individual, familial and community).
36 2 36. The ethical directive of informed Rate: 2

consent is adhered to. (Vulnerable Informed consent was gained as follows:
and minority groups are treated "After receiving ethical approval from the
within the prescribed parameters of Department of Education, I went to two high schools
relevant ethical and legal in Alexandra and recruited 32 Grade 10 and 11 boys
frameworks.) between the ages of 13 and 18 years in 2007. The
HMPYC80/104 243
HMPYC80/104/0/2024
researcher advertised the study by visiting classes of

Grade 10 and 11 to meet with boy learners to explain
the nature of the research project and to provide
them with an information sheet, parental consent
and participant assent forms which required the
appropriate signatures. Only boys who returned the
signed forms were included in the study."
This appears to be in order.
It is not clear whether the research was part of an

academic qualification or under the auspices of a
university, and whether a university ethical
committee provided oversight. (This might be the
case as the author is/was a member of staff at a
university, as reported on the first page of the
article.)
Informed consent from the follow-up group of 12

boys/men is not mentioned. This might have been
appropriate for the longitudinal study.
[For this review with the R.R.I. it must be discussed

in detail for a rating of 3.] The criterion is partially
achieved, rate 2.
37 1 37. The ethical mandate of voluntary Rate: 1
participation is adhered to, and There is no mentioned of voluntary consent and
participants have the right to right to withdrawal; except for the actions implied in
withdraw from the research study. the letter of consent. As this is not explicitly and
clearly discussed a rating of 1 is allocated.
38 1 38. The ethical imperatives of Rate: 1

privacy, confidentiality and
anonymity are not violated. Ethics: Privacy is not discussed in the article - and is
privacy, confidentiality and potentially thwarted by the use of photographs with
anonymity: identifiable persons or characteristics.
Confidentiality is not explicitly discussed in the

article.
Anonymity is not explicitly discussed in the article.
It is also not clear that there was explicit consent for

the use of photographs of the participants and their
family members. This should be clearly determined
244
and reported. (Rules on the privacy and consent
issues with regard to photographs have become
more explicit in the last few years.)
The criterion is not achieved. Rate 1.

39 2 39. The ethical imperative of ethical Rate: 2
oversight is fulfilled. The research
proposal or project adheres to Ethical oversight and permission by the Department
institutional research policies, of Education was reportedly obtained.
guidelines and procedures. The
research proposal was (or will be) It is not clear whether the research was part of an
submitted to all appropriate academic qualification or under the auspices of a
institutional ethics review panels for university, and whether a university ethical
ethical oversight, and official written committee provided oversight. (This might be the
permission was (will be) obtained to case as the author is/was a member of staff at a
perform the research. university, as reported on the first page of the
Ethics: ethical oversight and article.)
permission:
There are some doubts with regard to adequate and
appropriate oversight. Rate 2.
40 3 40. The grammar and spelling in the Rate: 3
research document are at least 95% The grammar and spelling in the article are of high
correct. quality.
41 3 41. The entire document from Rate: 3
beginning to end is plagiarism-free. The article appears to be plagiarism-free, as the style
All important works, ideas, theories, of writing is (a) consistent and (b) of good academic
authors, facts and statements are quality. In this case we are dealing with a
acknowledged in some way (even if peer-reviewed article; this means that you can safely
the citation style is incorrect or assume that the manuscript is plagiarism-free. In
imperfectly used.) The authors(s) general, peer-review processes of academic journals
wrote the text with academic include various plagiarism-checking mechanisms
integrity; and it appears to be their and procedures. The work is published by a
own and unique work. reputable journal and is deemed to be
plagiarism-free.
42 3 42. The various authors and sources Rate: 3
in the proposal or article (the entire Authors and sources in the literature review appear
document) are cited correctly and to be cited correctly, and the prescribed in-text
follow the prescribed in-text citation citation style ("APA or Other"), in this case the
style. This should be correct for the prescribed journal standard for citations is used
entire document. (APA7 style for consistently.
HMPYC80 proposals, or the The Psychology in Society (PINS) journal uses the
prescribed referencing style of the Harvard referencing style - refer to the style-guide
specific journal.) Was the correct on the website.
citation method used? (One of the references in the text does not have the
same date in the reference list. This is a minor
mistake and is negligible -- the overall standard is
HMPYC80/104 245
HMPYC80/104/0/2024
high.)
43 3 43. A complete list of all references Rate: 3
that were cited is provided at the end A complete list of all references that were cited is
of the document. Is it complete? provided at the end of the document. (Note that there
are 2 references that appear in the references but are
not in the text. These are minor issues and are
negligible -- the overall standard is high.)
44 3 44. The reference list at the end of the Rate: 3
document adheres to the required The reference list at the end of the document adheres
standard and format. (APA7 style for to the required standard and format.
HMPYC80 for proposals, or in the
case of published journal articles --
the prescribed referencing style of the
specific journal was adhered to.)
45 3 45. The work is of overall acceptable Rate: 3
academic standard, and it is written The work is of overall acceptable academic
using an appropriate standard, and it is written using an appropriate
academic/scientific style suitable to scientific style suitable to the research paradigm,
the research paradigm, approach and approach and journal.
journal (in the case of published
articles).
46 2 46. The data collection, coding and Rate: 2 (partially addressed).
analysis method(s) were performed The step-wise and longitudinal data collection
reliably; or dependably and credibly. method was discussed as indicated in the previous
(Apply the criteria as discussed in item. All interviews were transcribed.
item 32.) The author states that "a narrative approach was
used" (p. 65) under the heading data analysis.
"The initial stage of data analysis involved reading
each transcript many times with the aim of
determining the prevailing themes relating to
masculinity and its contestation in terms of voices to
which the participants publicly or privately
subscribe. It was in the interest of this study, given
its narrative approach
to place the story of each participant at centre stage
in the data analysis." (p.65)
" This involved organising the analysis of each
narrative coherently in terms of meanings that each
participant made about his upbringing, relationship
with mothers, absent fathers,
search for work, working and becoming fathers.
Particular attention was also paid to the manner in
which personal stories were shared. The photos
246
taken by the participants were also central in the
analysis as many of the narratives were linked to
back to the
photos. The photos were seen and interpreted by the
participants as an important representation of
significant events in their personal lives." (p.65)
Thick rich descriptions are provided in the
discussion.
Transparency and verifiability - there were
"member checks" by way of repeated interactions
over a number of years with participants (a
longitudinal study design was used).
(There is no report of an audit trail.)
The author reports that "Reflections and personal
dynamics about the relationship that the researcher
formed with the participants is beyond the scope of
this article." (p.64) This points to reflexivity to make
the biases, values, and experiences of the researcher
explicit -- but is not discussed.
The exact method of data coding and analysis is not
described in any detail, and this raises questions
about the method used.
For this reason a rating of 2 (only partially
addressed) is allocated.
47 2 47. The way in which the data were Rate: 2 (partially addressed).
analysed and interpreted is explained The way in which the data were analysed and
in sufficient detail and provides interpreted is explained was not done with sufficient
sufficient information for a reader to detail and did not fully provide sufficient
be able to follow the procedures. information for a reader to be able to follow the
procedures.
48 2 48. The data collection, coding and Rate: 2 (partially addressed).
analysis method(s) were applied in a This follows from the discussion for item 33 and the
correct and substantiated manner and 2 previous items (46 and 47).
appear trustworthy and valid. (Apply
the criteria as discussed in item 33.)
49 2 49. The data analysis and Rate: 2
interpretation were performed Most of the ethical issues are not discussed at all,
ethically. (Refer to RESEARCH and in some places with insufficient detail. It is not
ETHICS: items 35 to 39.) possible to state with certainty that the data analysis
and interpretation were performed ethically.
However, the researcher embodied an ethical stance
in supporting the participants and reporting on
difficult interpersonal dynamics. Some of the ethical
issues are discussed in detail. A rating of 2 is
allocated.
50 3 50. The results are presented Rate: 3
HMPYC80/104 247
HMPYC80/104/0/2024
appropriately and correctly. The results are presented appropriately and

correctly.
51 3 51. The results are clearly structured, Rate: 3
well presented and easy to follow. The results are clearly structured, well presented and
easy to follow.
52 3 52. A discussion is presented which Rate: 3
interprets or explains the results and A discussion is presented which interprets or
their implications for the research explains the results and their implications for the
question or issue. research question or issue.
53 3 53. The discussion links and explains Rate: 3
the results to the existing bodies of The discussion clearly links and explains the results
literature of as discussed in the to the existing bodies of literature of as discussed in
literature review. It links to existing the literature review and to the hypotheses.
theoretical perspectives, and where
relevant to existing empirical
literature findings.
54 2 54. The research paradigm, Rate: 2
theoretical perspectives in the The research paradigm and approach were not
literature, research methods, data clarified - but apart from this there is a “golden
gathering techniques, data coding and thread” that runs through the document and connects
analysis, and interpretations that were each subsection and section to the whole. The
employed are in line with the criterion is partially achieved (rate 2).
research question and aim of the
study. There is a “golden thread” that
runs through the document and
connects each subsection and section
to the whole.
55 2 55. The possibility of alternative Rate: 2
explanations or interpretations of the The landscapes in which the young men grow up is
results is considered. described in a complex, nuanced and layered way by
the author. The multiple lenses and meanings are an
excellent way of hosting many narratives and
multiple (sometimes contradictory or evolving)
meaning making by participants. In this sense there
are many realities that are described on the article,
and not just one reality, via thick and rich
descriptions. The "participants are champions of
their own narratives" (p.78).
On the other hand, there is no explicit mention of
alternative or contrarian explanations or
interpretations at the close of the discussion. The
criterion is partially met (rate 2)
56 1 56. The possibility or need for further Rate: 1
research is raised. Further research is not discussed.
248
57 1 57. The strengths, weaknesses and Rate: 1
limitations of the study are presented. Strengths, weaknesses & limitations are not
discussed in any detail.
58 3 58. A final conclusion is presented Rate: 3
that succinctly reiterates the main A final conclusion is presented that succinctly
research findings or results. reiterates the main research finding or result.
A sample report of the Turnitin software “plagiarism and similary check” follows on the next few pages.
This is the final resource in this tutorial letter.
2024 © Unisa
HMPYC80/104 249

HMPYC80_2024_TL_104_0_B

Uploaded by

Copyright:

Available Formats

You might also like

HMPYC80_2024_TL_104_0_B

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HMPYC80_2024_TL_104_0_B

Uploaded by

Copyright:

Available Formats

HMPYC80/104/0/2024

Tutorial Letter 104/0/2024

A. RESEARCH AND LEARNING TOOLS

C. ETHICS AND RESEARCH

D. REFERENCING AND FORMATTING

E. PROPOSALS, ARTICLES AND PRESENTATIONS

F. SOCIAL SCIENCE RESEARCH PARADIGMS

G. QUANTITATIVE RESEARCH DESIGNS AND STATISTICS

Factorial analysis of variance 135

H. QUALITATIVE RESEARCH DESIGNS AND METHODS

I. APPLICATION OF THE RESEARCH REVIEW INVENTORY (R.R.I. revised)

J. Example research proposal with Turnitin report

About this resource

A. Research and learning tools

C. Ethics and research

D. Referencing and formatting

E. Research proposals, articles and presentations

F. Social science research paradigms

G. Quantitative research design and statistics

H. Qualitative research designs and methods

I. Application of the Research Review Inventory (R.R.I. revised).

J. Example research proposal with Turnitin report

A. RESEARCH AND LEARNING TOOLS

Learning activity: What is it?

The following is a generic description of a learning activity:

A. RESEARCH AND LEARNING TOOLS

Project planning and management

PROJECT PLAN STRUCTURE

A good research project management plan includes the following:

JOE'S PROJECT PLAN (example):

ACTIVITIES NEEDED TO COMPLETE THE PROJECT

IMPLEMENT THE PLAN

MONITOR, CONTROL AND EVALUATE THE PROJECT

TIME SCHEDULING AND PLANNING

SNAFU'S and Murphy's law:

How to make Murphy's reverse law work for you:

Research and course journal

AIM (What did you want to achieve?)

(1) Reflection - Research problem

Reflection - I'm Stuck

(4) Reflection - aims and doing

WHY PORTFOLIO ASSESSMENT?

PUTTING A PORTFOLIO TOGETHER

4. Your own organising principle

What items to include

Such evidence could

Goals and objectives

Some important aims with portfolio work should be to:

Reactions to feedback on assignments and activities

Other items to include

Boud, D. (ed.). (1981). Developing student autonomy in learning. Kogan Page.

Dewey, J. (1938). Experience and education. Collier.

Research: What is it?

So, what is “Research”?

Smith, L. T. (1999). Decolonizing methodologies: Research and indigenous peoples.

The basic elements of “Research”

Chilisa, B. (2012). Indigenous research methodologies. Sage.

Smith, L. T. (1999). Decolonizing methodologies: Research and indigenous peoples.

C. ETHICS AND RESEARCH