CHARTERED INSTITUTE OF RISK AND SECURITY MANAGEMENT

PROFESSIONAL DIPLOMA IN RISK & SECURITY MANAGEMENT

RESEARCH METHODS AND PROJECT MANAGEMENT

Module RRM 400

• Published by: Chartered Institute of Risk and Security Management

12 Highfield Road
Southerton
Harare, Zimbabwe

Chartered Institute of Risk and Security Management is a distance teaching and open learning
institution.

© Chartered Institute of Risk and Security Management: All rights reserved. No part of this publication
may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior permission of the CIRSM.
 To the student
This module contains information obtained from authentic and highly regarded
sources. Reasonable efforts have been made to publish reliable data and
information, but the compiler and institution cannot assume responsibility for
the validity of all materials or the consequences of their use. The compiler and
institution have attempted to acknowledge the authors of all material
reproduced in this module.
 CHARTERED INSTITUTE OF RISK AND SECURITY MANAGEMENT

"The Six Hour Tutorial Sessions"

As you embark on your studies with the Chartered Institute of Risk and Security Management
(CIRSM) by open and distance learning, we need to advise our students so that they can make
the best use of our modules and other learning materials, your time and the lecturers who attend
to you virtually. The most important point that students need to note is that in e-learning, there
are no lectures like those found in conventional face-to-face learning. Instead, there are learning
packages that may comprise written modules, audio recordings, video recordings and other
referral materials for extra reading. All these including WhatsApp, Telegram, Twits, Blogs,
Skype, telephone and email can be used to deliver learning to students. As such, at CIRSM, we
expect the lecturers to lecture to students virtually on the stipulated six-hour tutorials designed
to give students robust introductory knowledge to their programmes. We believe that the
teaching and learning task is accomplished by the learning package that students receive at
registration.

What then is the purpose of the six-hour tutorial for each course on offer?
At CIRSM, like any other e-learning programmes, the students are at the core of learning. After
they receive the learning packages and other learning materials, it is obvious that they will come
across concepts/ideas that may not be that easy to understand or that are not so clearly explained.
They may also come across issues that they do not agree with, that actually conflict with the
practice that they are familiar with. Through interaction and discussion groups, friends can bring
ideas that are totally different and new and arguments may begin. Students may also find that
an idea is not clearly explained and they may remain with more questions than answers. They
need someone to help them in such matters. This is where the six-hour tutorial comes in.
For it to work, you need to know that:
 This is one requirement in e-learning
 The lecturer has to introduce the course adequately for students to progress on
their own.
 The student should prepare questions, queries, clarifications, for the topics to the
 discussed. For the lecturer to help you effectively, give him/her the concerns
beforehand so that in cases where information has to be gathered, there is sufficient
time to do so. If the questions can get to the lecturer at least two weeks before the
tutorial, that will create enough time for thorough preparation.

In the tutorial, the students are expected and required to take part all the time through
contributing in every way possible. They can give their views, even if they are wrong, (many
students may hold the same wrong views and the discussion will help correct the errors), they
still help them learn the correct thing as much as the correct ideas.

There is also need for both students and the lecturer to be open-minded, frank, inquisitive and
should leave no stone unturned as they analyze ideas and seek clarification on any issues. It has
been found that if tutorials are done correctly, students do better in assignments and
examinations because their ideas are streamlined. By introducing the six-hour tutorial, CIRSM
hopes to help students come in touch with the lecturers who mark their assignments, assess
them, and guide them in preparing for writing examinations and assignments and who run
students’ general academic affairs. This helps students to settle down in their course having
been advised on how to go about their learning.
Professional networking with students is, therefore, upheld by CIRSM.

The six-hour tutorials should be so structured that the tasks for each session
are very clear. Work for each session, as much as possible, follows the
structure given below.

Session I (Two Hours)

Session I should be held at the beginning of the semester. The main aim of this session is to
guide you, the student, on how you are going to approach the course. During the session, you
will be given the overview of the course, how to tackle the assignments, how to organize the
logistics of the course and formation of study groups that you will belong to. It is also during
this session that you will be advised on how to use your learning materials effectively.

Session II (Two Hours)

This session comes in the middle of the semester to respond to the challenges, queries,
experiences, uncertainties, and ideas that students are facing as they go through the course. In this
session, difficult areas in the module are explained through thecombined effort of the students and the
lecturer. It should also givedirection and feedback where students have not done well in thefirst
assignments as well as reinforce those areas where performance in the first assignments is good.

Session III (Two Hours)

The final session, Session III, comes towards the end of the semester. In this session, students
polish up any areas that they still need clarification on. The lecturer gives students feedback on
the assignments so that they can use the experiences for preparation for the end of semester
examination.

Note that in all the three sessions, students identify the areas that their lecturer should
give help. They also take a very important part in finding answers to the problems
posed. As students, you are the most important part of the solutions to your learning
challenges.

Conclusion
In conclusion, we should be very clear that six hours is too little for lectures and this does not
limit, in view of the provision of fully self-contained learning materials in the package, to look
for supplementary sources to augment this module. We, therefore, urge students not only to
attend the six-hour tutorials for this course, but also to prepare oneself to contribute in the best
way possible so as to maximize beneficiation.

BEST WISHES IN YOUR STUDIES.

CIRSM
 RRM 400

Writing a research Proposal

Research is a systematic collection, analysis and interpretation of data to answer a certain question
or solve a problem.
Purposes of Research
1. Exploration

This is an to attempt to develop an initial rough understanding of some phenomenon.

2. Description

This is the precise reporting and/or measurement of the characteristics of some population or
phenomenon under study addressing the ‗What, How, Where, How‘ questions

3. Explanation

It is the discovery and reporting of relationships among different aspects of the phenomenon
under study addressing the ‗Why‘ question. Research Proposal

The meaning of proposal … suggests looking forward, to what the researcher plans to do in the
future. It is a concise and coherent summary of your proposed research. It is a planned tool of
action which sets out the central issues or questions that you intend to address, logical mapping of
activities to be executed, how data will be collected and analysed, factual and instrumental
resources to be used and the time frame in order to achieve the desired result. It outlines the general
area of study within which your research falls, referring to the current state of knowledge and any
recent debates on the topic. By so doing, your research proposal becomes an evaluation instrument
for stakeholders.

Page 0 of 136 CIRSM

 RRM 400

Functions of a Research Proposal

• A research proposal gives you an opportunity to demonstrate that you have the aptitude for
graduate level research, for example, by demonstrating that you have the ability to
communicate complex ideas clearly, concisely and critically – your expertise in a particular
area.

• To convince others that one has a worthwhile research project and that one has the
competence and work plan to complete it.

• It becomes a basis of decision for Stakeholders researchers, research funding agencies,

educational institutions, and supervisors that the research is worth spending scarce
resources on.

• An academic proposal also helps us to match your research interest with an appropriate
supervisor.

Different Formats of Research Proposals

Research Proposals come in different formats. However, the formats depend on the needs of the
institutions, i.e. Business, Academic and Funding Agencies. It should be noted that there are
nonetheless shared features to all the proposals. These features are reflected mainly in the
Academic Oriented Proposals.

General Features/Steps to the Proposal Writing Process Research Topic

A research topic is a cumulative outcome of your general knowledge of issues and your personal
experience in the world. Generally the topic is a broad area of interest, for example, Determinants
of criminal activities in activities in Mabvuku, Harare, Zimbabwe. However, your topic should
reflect that your study will a ground breaking one. Avoid over-research issues. In addition, it must
be Specific, Measurable, Attainable, Reliable and Time framed (SMART).

Key issues when choosing a research topic

• Should reflect topical issues as covered in Public Media, Donor interests and interests of
organizations/institutions.

• Topic must be precise and researchable: not too broad-focused.

• Must be grounded in literature

• Not redundant, must fill a gap and must reflect that the study a ground breaking one.

Page 1 of 136 CIRSM

 RRM 400

Introduction

An introduction is the teaser for your background summary. It is meant to be short and attention
grabbing, and make the reader actually want to read further into the background summary. One
opens up (Introduces) one‘s area of interest for discussion or/and problematises one‘s area of
research inquiry. The aim of an introduction is to open up the subject for discussion, attention, etc.
The introduction also presupposes that the audience are fairly familiar with the subject of one‘s
study. However, among other issues, the introduction should highlight a description of the problem
that will be addressed. In this section the researcher might discuss the nature of the research and
the purpose of the research. Note that on the purpose of the research, you should highlight the
objectives of the study and what one intends to discover from the study. You should also briefly
state the rationale of the research problem. This is where you briefly explain why your research
topic is worthy of study and may make a significant contribution to the existing body of literature.
In fact, an introduction is a mental road map that answers the following questions:

1. What am I studying?

2. Why is this topic important to investigate?

3. What do we already know about this topic or what have other experts discovered about the
topic?

4. How will my research advance new knowledge or new ways of understanding?

The introduction leads the reader from a general research issue or problem to your specific area of
research. It puts your research question in context by explaining the significance of the research
being conducted. This is usually done by summarizing current understanding (research to date)
and background information about the topic. This is followed by a statement of the purpose of your
research issue or problem. This is sometimes followed by a hypothesis or a set of questions you
attempt to answer in your research. You may also explain your methodology (how you will
research this issue) and explain what your study can reveal. It also may contain a summary of the
structure of the rest of the paper.

Components of the Research introduction One

• Establish the problem or issue you want to research

• Highlight the importance of the problem/issue, and/or

Page 2 of 136 CIRSM
 RRM 400

• Make general statements about the problem/issue, and/or

• Present an overview on current research on the issue or problem.

Two

Provide an overview of existing thinking about and/or research into your research problem.

Three

Identify a gap, problems or unresolved issues in the existing knowledge/research that your research
can fill or identify a research focus that will be useful:

• ―The previous research has mistakenly assumed that….‖ or ―Although most experts in the field
believe …., they have overlooked …‖

• ―None of the previous research has examined ……‖

• ―Despite prior observations of voter behavior in local elections in Harare, it remains unclear
why some single mothers choose to avoid....‖ and/or

• ―Consequently, these factors need to examined in more detail....‖ or ―Evidence suggests an

interesting correlation, therefore, it is desirable to survey different respondents....‖ Four:

State your research question, your hypothesis and your knowledge claim, making sure to place
your research within the gap in existing knowledge.

• State the intent of your study, including the research question and your hypothesis,

• Outline the key characteristics of your study,

• Describe important results that you have found or hope to find

Page 3 of 136 CIRSM

 RRM 400

Background

It says more than a mere Introduction to the problem.

It assumes that lack of familiarity of your audience with the problem and the need to provide its
―backstage; could be a spatial (where it happened) or time (historical evolution) context. It relates
to the problem one wants to study to a larger problem.

The background study for your research includes a review of the area being researched, current
information surrounding the issue, previous studies on the issue, and relevant history on the issue.

You should explain the broad background against which you will conduct your research. You
should include a brief overview of the general area of study within which your proposed research
falls, summarising the current state of knowledge and recent debates on the topic. This will allow
you to demonstrate a familiarity with the relevant field as well as the ability to communicate clearly
and concisely.

Use the first paragraph or two of the Background to explain what is known generally in the area
of your study. Cite key references, but do not write an extensive review of literature; instead, direct
the reader to a recent review.
Then focus in on the problem that your study addresses.

Literature Review

Literature review (previous studies) can be defined as the process of surveying what has been done
on the topic or problem one wants to work on. The materials can be purely be theoretical or
empirical. Note that a literature review is not a mere list or summary of books and articles.
However, it is a combination and synthesis in the form of an integrated description of the field of
study which include inter alia: theories, models and studies. Literature review shows the originality
of your proposed study. Sources of relevant literature include: Electronic; references of references,
policy documents, academic books and journals.

Rationale for Literature Review

- Literature provides a conceptual framework for the reader so that the research question and
methodology can be better understood.

Page 4 of 136 CIRSM

 RRM 400

-Literature review demonstrates to the expert reader, notably funding agencies and academics that
the researcher is aware of the breadth and diversity of literature that relates to the research question.

Functions of Literature Review

- Ensures the production of original work.

- Acknowledges the contribution of previous works.

- It situates your research focus within the context of the wider academic community in your field;
reports your critical review of the relevant literature.

- Identifies a gap within that literature that your research will attempt to address.

- It showcases your talents of: understanding, interpretation, analysis, clarity of thought, synthesis,
and development of argument.

- The process of conducting and reporting your literature review can help you clarify your own
thoughts about your study. It can also establish a framework within which to present and analyse
the findings.

- Enables a reader to appreciate that you have up to-date awareness of the relevant work of others,
and that the research question you are asking is relevant.

- Enables you identify your research questions and their importance;

- Establishes the relevance of your research/practice/theory in your field;

- Establishes what is already known or understood about your topic;

- Establishes how your research might add to this understanding or challenge existing literature,
theories, methodology and beliefs.

Before you commence literature review, make sure you ask yourself the following questions:
What research question(s) of my study which should be answered by my literature review?
Has anyone else done anything similar to my study?
Is my research relevant to research/practice/theory in my field?
What is already known or understood about this topic?

Page 5 of 136 CIRSM

 RRM 400

How might my research add to this understanding, or challenge existing theories and beliefs?
Steps and Process in Literature Review

a) Decide on what one may want to derive from the literature.

b) Provide indicative issues one wants to derive from the materials one would have collected.
These could be:

i. Conceptual theoretical issues surrounding the subject/project; ii. Demonstrating

one‘s critical ability to evaluate literal materials; and, iii. Rationalizes the importance of
one‘s problem of study.
Literature review assists the research design, establishes connection between the specific topic and
larger theme/context evaluate empirical findings from previous works.

iii. Shows gaps in methodological approaches.

c). Prepare a notebook appropriately classified to reflect the type of information that one is
looking for from the materials collects. d). Compilation of the information collected. e). Make
an evaluation of the materials.

Output from the Literature Survey

-Theoretical and methodological discourses and rebuttals surrounding the subject to date.

-A summary of study areas and subjects that have been handled through different initiatives.

The data that has been generated along the way.

-Significant benchmarks and trends in the area and topic handled.

-Icons of caps in the trend

-A synthesized write up of the review pinpointing the caps which the project/research would
like to focus on.

Research Framework

A research Framework is a structure that provides ―guidance for the researcher as study
questions are fine tuned, methods for measuring variables are selected and analyses are
Page 6 of 136 CIRSM
 RRM 400

planned‖ (Liehr and Smith, 1999: 13). Once data are collected and analysed, the framework is
used as a mirror to check whether the findings agree with the framework or whether there are
some discrepancies; where discrepancies exist, a question is asked as to whether or not the
framework can be used to explain them.

Research Theory

A theory is ―a set of interrelated concepts, which structure a systematic view of phenomena for
the purpose of explaining or predicting. A theory is like a blueprint, a guide for modeling a
structure‖ (Imenda 2014). On the other hand, ―A scientific theory is a well-substantiated
explanation of some aspect of the natural world, based on a body of facts that have been repeatedly
confirmed through observation and experiment. Such factsupported theories are not ―guesses‖ but
reliable accounts of the real world. (American Academy for the Advancement of Science, 2010:
1).

Theoretical Framework

It refers to the theory that a researcher chooses to guide him/her in his/her research. ―A theoretical
framework is the application of a theory, or a set of concepts drawn from one and the same theory,
to offer an explanation of an event, or shed some light on a particular phenomenon or research
problem‖ (Imenda, 2014). Note that a theoretical framework consists of theories that seem to be
related and the theory provides a point of focus for attacking the unknown in a specific area. A
theoretical framework is more formal and used for studies based on existing theories. It refers to
the concepts found in Theories/Models. Note that the theories would have been tested and
confirmed to explain the phenomena being studied.

A theoretical framework is formulated to explain why the relationship between two or more
variables exists. However, it should be noted that a theoretical framework has been tested and
proven to test relationships e.g The Theory of Reasoned Action in predicting human behaviuor
(Figure 1). It is grounded in and based on scientific theory. It is derived from specific concepts and
propositions that have been tested and proven to explain the relationship/phenomena.

Page 7 of 136 CIRSM

 RRM 400

Figure 1: Theory of Reasoned Action

Ajzen and Fishbein (1980:8)

Functions of a Theoretical Framework

- To determine what theories and ideas exist in relation to the chosen subject.

- Helps to ‗frame‘ the research and show that one is knowledgeable about the key concepts,
theories, and models that relate to the topic.

- Provides the definitions and models which help in guiding the research.

- Provides scientific justification for the investigation: it shows that the research is not just
coming ―from the blues,‖ but that it is both grounded in and based on scientific theory.

Conceptual Framework

It is you modified theoretical framework. A conceptual framework is based on concepts adopted

from theories related to the research topic, empirical observation/studies and existing theories. It
can be used to explain the relationships between the dependent and independent variables or to
develop a theory. It is also a system of concepts, assumptions, expectations, beliefs, and theories
that supports and informs your research—is a key part of your design (Miles & Huberman, 1994;

Page 8 of 136 CIRSM

 RRM 400

Robson, 2011). Miles and Huberman (1994) defined a conceptual framework as a visual or written
product, one that ―explains, either graphically or in narrative form, the main things to be studied—
the key factors, concepts, or variables—and the presumed relationships among them‖ (p. 18).

It is primarily a conception or model of what is out there that you plan to study, and of what is
going on with these things and why—a tentative theory of the phenomena that you are
investigating.

It incorporates pieces that are borrowed from elsewhere, but the structure, the overall coherence,
is something that you build, not something that exists ready-made. It is important for you to pay
attention to the existing theories and research that are relevant to what you plan to study, because
these are often key sources for understanding what is going on with these phenomena.

Functions of a Conceptual Framework

- It informs the rest of your design— to help you to assess and refine your goals and objectives,
develop realistic and relevant research questions, select appropriate methods, and identify
potential validity threats to your conclusion.

- It helps you justify your research.

-It clarify concepts and propose relationships among the concepts in a study.

-It provides a context for interpreting the study findings.

- It explains observations.

- Encourages theory development that is useful to practice.

Research objectives

These are concrete statements which describe what the research is trying to achieve at the end of
the study. Research objectives should be closely related to the statement of the problem. In other
words, research objectives are a re-working of the statement of the problem. A well-worded
objective will be SMART, i.e Specific, Measurable, Attainable, Realistic, and Timebound given
the available resources. In addition, a research objective should be Relevant, Feasible, Logical,

Page 9 of 136 CIRSM

 RRM 400

Observable, Unequivocal and Measurable. Objective is a purpose that can be reasonably

achieved within the expected timeframe &with the available resources

Overall Objective (AIM)

It refers to broad developmental goal within which the project/research proposal falls. It provides
a short statement of the development goal being pursued by the research, e.g. ―improve security
issues in programmes‖

Governments or Development partners often express the broad development goal(s) in general
development strategies of broad policy agendas.

Specific objectives

These state precisely what you intend to do. Specific objectives include obtaining answers to
research questions or testing the research hypothesis, contributing factors in a coherent way in a
logical sequence. They must be SMART. In addition, they start with action verbs such as to: assess;
explore; identify; determine; establish; compare; verify; calculate; describe; and, analyse.

Specific objective should also be presented briefly and concisely. They cover the different aspects
of the problem. They should be clearly phrased in operational terms, specifying exactly what
researcher is going to do, where, & for what purpose. They should be realistic when one considers
their local conditions.

The formulation of research objectives helps the researcher to avoid the collection of data which
are not strictly necessary for understanding and solving problem that he or/she has

Statement of the problem

A problem statement is the description of an issue currently existing which needs to be addressed.
It provides the context for the research study and generates the questions which the research
aims to answer.
The statement of the problem is the focal point of any research.

It is also an analysis of the major factors that may influence the problem and discussion of why
certain factors need more investigation.

It provides a brief description of solutions to problem that have been tried in the past, how well
they have worked, and why further research is needed.

Page 10 of 136 CIRSM

 RRM 400

Hint when writing a statement of the problem:

a) Highlight the ideal situation of your issue under investigation.

b) The current or prevailing situation vis-a- vis what should be ideal.

c) Briefly state the factors underlying the prevailing situation

d) Conclude by stating that unless policies and programmes give priority to address the current
situation, the achievement of an ideal situation will always remain a pipe-dream.

Justification

A justification explains the importance of your study; why it is worthwhile undertaking; an agent
spend funds on the study/project; why should you spend so much time on the study/project.
Justification of study stems solely from the outcome of your literature review. From the literature
review, you can establish the following:

- That the methodologies of previous studies employed did not adequately explain the
phenomenon;

- May be that new methodology in other field of studies may contradict the existing
knowledge about the phenomenon and offer fresh insight which you may want to apply;

- May be that the way the problem and its associated concepts were approached and defined
were problematic; and,

- May be that contemporary problem have falsified at the previous claims about the issue
you are investigating.

So justification can show the theoretical, methodological scientific significance of the research and
practical relevance of your study. In addition, by the time you do thorough review, any of the
aforementioned points will come out and that will help provide a good justification for your
research.

Hint on writing the justification of your study.

You can use your statement of the problem to guide you when writing the justification of your
study. For instance, if you ask the question ―Is there a significant relationship between criminal
activities and education, age, sex, religion, place of residence, region of residence, income?‘, then

Page 11 of 136 CIRSM

 RRM 400

the contribution of your research would probably be coming up with technologically advanced
investigating techniques to curb criminal activities performed for instance by people highly
educated people (the white collar jobs). Note that your research in advanced criminal investigating
techniques would be important. This could be a groundbreaking approach that will change the way
criminal investigation is conducted.

You can also use the deductive reasoning approach to explain the significance of your study. This
means that you can start by highlighting the contribution of your study broadly and then taper off
gradually to a specific group or person. That is, the contribution of your study to a specific group
e.g. the society and organizations as a whole, then its contribution to
individuals

Methodology

It is a single paragraph where you would briefly state the target population, study area, research
design, the data collection methods and tools (justifying the selection of each method in brief) and
how you are going to analyse your data. Note that you link the justification to your research
objectives.

From my presentation, note the various sub-topics of the methodology section.

Target Population

The sampled population whose characteristic (knowledge, attitudes, beliefs and practices) you are
interested in. Justify the selection of the target population e.g. why you are only interested in
understanding the criminal behaviour of women aged 15 to 69. Study area

This would be the exact location from which you would carry out your study from. Also remember
you have to justify the selection of your study area. Research Design

A research design is the plan and procedure for research. It is the overall scheme or program of
research. It is the plan, structure and strategy of investigation, conceived to obtain answers to
research questions. You would briefly present the format of your study, highlighting and justifying
each of the data collection methods and tools you intend to use in your research. You would state
Page 12 of 136 CIRSM
 RRM 400

whether or not the study was purely qualitative, quantitative or whether it triangulated the
qualitative and/or quantitative research methods. You would also briefly how you intend to have
the data recorded, managed and analysed.

Data collection methods and tools

In this section you would explicitly present in detail the respective methods and tools which you
intend to use in your study. Remember for each method, you should clearly state the method, how
many, tool, with who, why, sampling, selection of the respondents and how data were collected
and recorded.

How data is going to be recorded and analysed (Data manangement and analysis- briefly)

Ethical issues

Basically ethical issues are meant to enforce good morals and correct the bad ones. Ethical issues
refer to conforming to standards of conduct of a given profession or group.

Key ethical agreements in research

1. Voluntary participation

- People must not be forced to participate.

- People have right to withdraw at any moment.

- Have right not to discuss certain questions e.g. experiments on ….. - Refusal bearing no
consequences.
2. Risks

There should be no harm to participants, e.g Physical (mostly associated with medical research) and
Psychological (reminding teenagers of events they do not want to e.g rape).

3. Benefits

-Will the participants have any direct benefit by participants?

-State the indirect benefits and their levels of probability (don‘t overstate anything).

Page 13 of 136 CIRSM

 RRM 400

4. Anonymity

- Protection of participants‘ identity.

- Research must not make it possible to link certain responses with specific respondents.

- For adolescents to admit being sexually active anonymity must be assured.

5. Confidentiality

- The researcher identifies given person‘s responses but promises not to do so publicly.

6. Informed Consent

- Culmination of all ethical considerations

Budget

• List all activities you will carry out

• Cost each item

• EG. Questionnaire designing (optional)

• Pretesting: data collectors : per diems, transport, accommodation Data

collection: data collectors, per diems, salaries, transport
• Data entry: cost per questionnaire

• Data analysis

• Report writing

• Equipment

• Dissemination

Activity Plan

• Allocate time to each activity

Page 14 of 136 CIRSM

 RRM 400

• Better to allocate more time than less

• If time not specified by donor, allocate reasonable time for the study average one year.

• Best Illustrated with a Gantt Chat below.

Referencing

It is a method used to demonstrate to your readers that you have conducted a thorough and
appropriate literature search, and reading. Equally, referencing is an acknowledgement that you
have used the ideas and written material belonging to other authors in your own work.

There several referencing styles which include inter alia: The Harvard; Chicago; Vancouver
Convention; Oxford or footnoting system; American Psychological Association (APA); Modern
Language Association (MLA); American Chemical Society (ACS); and, American Institute of
Physics (AIP). A referencing style sets rules of telling you how to acknowledge the thoughts, ideas
and works of others in a particular way. In addition, it is not only a crucial part of successful
academic writing, but also a means to avoid plagiarism.

It should be noted that each referencing style has its own different origins and features. However,
Referencing style requirements cover the two elements of a referencing system: a. in-text citations
such as author-date citations or footnotes b. reference lists or bibliographies. What is of paramount
importance, however, is that you make sure you strictly adhere to the referencing style prescribed
by lecturer/tutor, assignment instructions guide, college handbook. This is largely because
referencing styles are established systems of referencing with consistent rules.

When referencing your work, you should know that consistency is very important. In addition, do
not mix referencing styles within one piece of writing.

Page 15 of 136 CIRSM

 RRM 400

Common mistakes in writing research proposals

• Failure to provide the proper context to frame the research question.

• Failure to delimit the boundary conditions for your research.

• Failure to cite landmark studies.

• Failure to accurately present the theoretical and empirical contributions by other

researchers.

• Failure to stay focused on the research question.

• Failure to develop a coherent and persuasive argument for the proposed research.

• Too much rambling– going ―all over the map‖ without a clear sense of direction. (The best
proposals move forward with ease and grace like a seamless river.)

• Too many citation lapses and incorrect references.

• Too long or too short.

• Failing to follow the APA style.

• Sloppy writing

Page 16 of 136 CIRSM

 RRM 400

Introduction to Qualitative Research

Essence of Qualitative Research

Qualitative research is an approach in which a researcher often makes knowledge claims based
primarily on a constructivist perspective. This perspective uses multiple meanings of individual
experiences, meanings socially and historically constructed, with an intent of developing a theory,
a pattern, or basis of advocacy (Creswell, 2003). Qualitative researchers study things in their
natural settings, attempting to make sense of, or interpret phenomenon in terms of the meaning
people bring to them.

Qualitative Research emphasizes understanding of verbal narratives using flexible designs. It is a

means for exploring, describing, interpreting and understanding the meaning of individuals, groups
or human problem. Data analysis is inductively built (from particular observations in order make
generalisations).

Qualitative methodology is highly descriptive, explanatory and exploratory (investigative). This

type of methodology answers the questions; ‗what, who, how and why‘.

Qualitative Research Designs

1. Phenomenology

This is a design which deals with the understanding of the essence of a phenomenon and
how individuals experience the world. The concern of the research is to understand social
and psychological phenomena from the perspectives of the people involved. The operative
word in phenomenology is ‗describe‘.

A phenomenologist views human behaviour as a product of how people interpret their

world. S/he attempts to see things from the participant‘s point of view. In the process of
Page 17 of 136 CIRSM
 RRM 400

the phenomenological research design, the researcher set aside his/her own experiences in
order to understand those of the participants understudy. In-depth interviews, documentary
review and films/videos are used for the phenomenological research design. Note that in-
depth interview questions will focus on participants‘ experiences, feelings, and beliefs
about the theme.

Examples of questions when using the phenomenological research design

a) What have been your experiences of sexual rape?

(b) How do you feel to be a married woman but experiencing sexual rape?

(c) What do you think are the root causes of your problem(s) of sexual rape?

The procedure in phenomenology involves studying a ―small‖ number of participants to develop

patterns and relationships.

2. Grounded theory

Whereas a phenomenological study aims at describing the essence of an activity or event,

grounded theory aims at providing an explanation or theory behind the events. You use
primarily in-depth and key informant interviews and existing documents to build a theory
based on the data. You identify themes from the coded qualitative data to build the theory.
Sample sizes with these studies are often also larger, between 20 and 60 respondents. This
is meant to better establish a theory. Grounded theory can help inform design decisions by
better understanding how a community of users currently use a product or perform tasks.
For example, the factors underlying criminal activities in any given community.

3. Ethnography

It is a strategy of inquiry in which the researcher studies an intact cultural group in a natural
setting over a ‗long‘period of time by collecting primarily observational and interview
data. In ethnography, you immerse yourself in the target participants ‘environment to
understand the goals, cultures, challenges, motivations, and themes that emerge. You will
be able to represent a view of a world as it is structured by the participants under
observation. You will also be able to describe and interpret a group‘s system with regards
to its sociocultural, religious, political and economic organization and also examine a
Page 18 of 136 CIRSM
 RRM 400

group‘s observable and learned patterns of behaviour, customs and way of life. Rather than
relying on interviews or surveys, you experience the environment first hand, and sometimes
as a

―participant observer.‖

Ethnographic study involves:

• Participation in the life of a group.

• Research in a natural setting.

• Face to face interaction with participants.

• Accurate reflection of participants‘ perspectives and behaviour.

• Utilisation of multiple data sources.

• Framing of behaviour and action in socio-political and historical context

• Noting differences among people with regards to, age, residential situation, and
ethnicity.

• Noting different behaviours, cultural beliefs, attitudes, norms.

Focus of Ethnographic Questions

Ethnographic questions take note of the following:

• Differences among people with regards to, age, residential situation, ethnicity;
• Different behaviours, cultural beliefs, attitudes, norms;
• The structures (policies, rules);
• The processes (community organization, problem solving, decision making); and,
• The statuses (health, economic, educational, social and political).

4. Narrative Inquiry

Strategy of inquiry in which the researcher studies the lives of individuals and asks one or
more individuals to provide cohesive stories about their lives (Clandinin and Connelly,
2000:18). Subjects recount the events of their lives and narrate them into temporal order
and meaning. Such stories would be subsequently presented to the reader(s).

Page 19 of 136 CIRSM

 RRM 400

Narrative inquiry is also a social science which represents and understands an individual‘s
experiences. The inquiry serves the researcher who wishes to understand a phenomenon as
an experience rather than to formulate a logical or scientific explanation. (Kramp,
2004:104). As a qualitative research design, narrative inquiry allows the research to provide
a rich description, contextual (background or appropriate) understanding and explanation
of the place or event under observation.

You conduct in-depth interviews, read documents, and look for themes; in other words,
how does an individual story illustrate the larger life influences that created it. Often in-
depth interviews are conducted over weeks, months, or even years, but the final narrative
does not need to be in chronological order. It can be presented as a story (or narrative) with
themes, and can reconcile conflicting stories and highlight tensions and challenges which
can be opportunities for innovation.

5. Case Study

A case study is an exploration of a case over time (activity, process, one or more individuals
or organisations) through detailed in-depth data collection involving multiple sources of
information rich in context (Creswell, 1998:61). It is used when one is interested in
exploring, explaining and describing an organization, entity, company or event.
A case study is exploratory (investigative), explanatory or descriptive.

Case studies are to be used when:

• The researcher wants to focus on the answers to the ‗Why‘ and ‗How‘ questions;

• The behavior is to be observed, not manipulated;

• You want to further understand a given phenomenon; and,

• If the boundaries between the context and phenomena are not clear.

Multiple methods can be used to gather data case study data. These include: in-depth
interviews, key informant interviews; observation; and, historical documentation.

Page 20 of 136 CIRSM

 RRM 400

Qualitative Data Collection Methods

Data for qualitative research can be collected through a wide range of data sources.
However, a research should always remember to match research objectives with the
appropriate data collection methods.

When ever you are writing any qualitative research method, always explicitly state the
following: method, how many, tool, with who, why, sampling, selection of the
respondents and how data were collected and recorded.

1. Observation

Observational techniques are methods by which a researcher can gather first-hand data
on programmes, processes, or behaviours being studied.

Observations provide researchers with an opportunity to collect data on a wide range

of behaviours, to capture a great variety of interactions, and to explore particular
research questions.

By directly observing operations and activities, the researcher can develop a holistic
perspective (i.e. an understanding of the context within which the issues under
investigation operate).

Successful observation requires background information on what is being observed.

Observation is the basic technique of qualitative data collection which should continue
throughout the field work. There is need for careful observation of events and behaviour
that provides valuable non-verbal clues as to what is actually happening.

Whatever the purpose of the study, four questions face the observer:

a) What should be observed – setting, participants, social behaviour and activities?

b) How should the observations be recorded?

c) What procedures should be used to ensure the accuracy of the observations?

d) What relationship should exist between the observer and the observed, and how can such a
relationship be established?

Page 21 of 136 CIRSM

 RRM 400

Types of Observations

Covert Participatory Observation

It involves a researcher joining the group he or she is studying, without pronouncing his/her status
to the group. The researcher becomes ‗part of, or an actor‘ in the events he is studying.

The aim of covert participant observation is as far as possible to attain an entirely familiar presence
in the lives of the people being studied i.e., to become invisible and not affect feelings, attitudes
and behaviours of the subjects as much as possible. The researcher participates in and observes the
socio-cultural content of a household, organisation, community and gains important insights into
everyday life.

Overt Participatory observation

Involves the researcher being open with the group he is going to study. Before joining a group the
researcher is likely to inform the group's members (either personally or through the agency of a
sponsor) the following: his/her presence; the purpose of the research; its scope; how long the
research will last; and any other relevant information about the study. It is advisable for researchers
to keep a field diary where day-to-day observations, along with the researcher‘s ideas, thoughts
and feelings are recorded.

Merits of observations

• Gain direct ‗first hand ‗ information.

• May reveal issues not anticipated by researcher.

• Useful for methods regarding behaviour.

• Useful for recording contextual (background) factors.

Disadvantages of observations

• Observation may alter learning environment, inhibiting programme participants.

• Inter-rater reliability (observer consistency) required before data collection occurs.

Page 22 of 136 CIRSM

 RRM 400

• Structured schedules may not capture critical factors Unstructured observations are
difficult to carry out.
• Costly-labour intensive and involves analysis costs

2. Focus Group Discussions (FGDs)

It is a research technique that collects data through group interaction on a topic determined by the
researcher. FGDs provide normative information or a group generated perspective. In that group,
concepts will be defined broadly. Given that FGDs capitalise on group dynamics, they allow a
small group of respondents to be guided by a skilled moderator into increasing levels of focus and
depth on the key issues of the research topic.

FGDs are a qualitative technique involving between six and twelve people talking spontaneously
and freely on themes considered important to an investigation/research. It has a moderator and a
note taker. Each FGD should last between one and one and half hours.

Reasons for using FGDs

1. Group interaction which generally stimulate richer responses and allow new and valuable
thoughts to emerge

2. The researcher can observe the discussion and gain first hand insights into the respondents‘
behaviours, attitudes, language (including body language) and feelings.

3. FGDs can be completed more quickly and generally less expensive than a series of in-depth
interviews.

Setting up FGDs

The number of groups for FGD depends on the purpose of the research and the nature of data
required. However, the following general guidelines should be followed:

a) Conduct at least two FGDs for each independent variable considered relevant to the topic.
e.g. if sex is a relevant variable, then you should have 4 groups, two for men and two for
women. If the researcher considers marital status to be important as well, the number of
groups should be eight.

Page 23 of 136 CIRSM

 RRM 400

b) Conduct FGDs until information obtained is no longer new or seems repetitive. If during
FGD you find that the information obtained is being repeated and no new information is
forthcoming, then most probably you have done enough FGDs.

c) If responses would vary according to geographical location , e.g. urban areas versus rural
respondents, then separate FGDs should be conducted in every relevant location

d) Economic conditions determine relevant variables and hence the characteristics of each
FGD must reflect the underlying differences.

e) Consider also race when conducting FGDs. Separate FGDs should be conducted with a
specific race.

General Guidelines when designing FGD guides

An FGD guide is the tool used to collect data from FGDs. However, either a structured or a
semi structured FGD guide can be used to collect qualitative data. What you need to appreciate
are the merits and demerits of each type of guide. Remember also, the care with which you
plan and design your FGD guide will influence the quality of the information you obtain.

Hints when designing a FGD guide

1. Do not begin with difficult/sensitive questions. Put these near the end so that if respondents
decide not to answer these, you do not lose their willingness to answer earlier questions.

2. Do not make respondents feel they ought to know the answers. Help them by saying,
‗Perhaps you have not had time to give this matter much thought?‘ May be they can find
the answer latter.

3. Decide carefully whether you should avoid emotional or sensitive words. Using words like
greedy, oppressed or immoral may seem to imply a judgment. Such words can cause bias
in answers. Respondents may be reluctant or nervous to give answers.

4. Avoid making assumptions. Why are men alone in this community undertaking criminal
activities?

5. Do not use confusing questions. Avoid questions like, ‗Why would you prefer and at the
same time not prefer your child to be vaccinated?‘ Keep it simple and positive. Ask, ‗Why
would you wish your child to be vaccinated?‘
Page 24 of 136 CIRSM
 RRM 400

6. Use both direct and indirect questioning. Direct questions are those ―normal‖ questions
that we can ask to friends, family members, and people who we know well. On the other
hand, indirect questions are little more formal and polite. They are used when talking to a
person we do not know very well, or in professional situations.

Whenever you design your FGD guide make sure the following information is
available:

a) Identification information eg. province, city, the district, ward, enumeration area,
suburb (if data was collected in an urban area), village (rural area), date;

b) Instructions regarding the following:

• The number of FGDs to be conducted and specifying the age and sex
composition for each FGD;

• When invitations should be sent;

• Who should facilitate and record the data;

• General introduction of the researchers and the objectives of the study should
be done before the discussion begin;

• Clearly stated issues of discussion;

• Duration of the discussion (between 1 and 1 and half hours;

• How protocol should be formulated and observed according to cultural setting;

and,

• How questions and suggestions from the participants about the research process
should be recorded.

Recoding and Analysis of FGD data

At the end of each day, the team should summarise the discussion. Cards with different
numbers and colours should be used for recording responses to each item discussed. Cluster
reports should be written combining all the FGD sessions for each respective cluster. At
district level, a cluster report combining all the sites should be written.

Page 25 of 136 CIRSM

 RRM 400

3. Interviews

This technique is characterised by extensive probing and open-ended questions. The emphasis is
on obtaining answers to carefully phrased questions. Interviewers are trained to deviate only
minimally from the question wording to ensure uniformity of interview administration. Interviews
are conducted on one-on-one basis between a respondent and a highly skilled interviewer.

Whenever conducting interviews, the following questions should face the interviewer:

• What should be asked?

• How should the questions be asked?

• How should the information obtained be recorded?

• What procedures should be used to assure the accuracy of the interview?

• What relationship should exist between the interviewer and the respondent and how can
such a relationship be established.

Types of interviews

3.1 Directive Interviews

3.1.1 In-depth interviews

These interviews are conducted with the actual people who are experiencing or experienced
a given phenomenon. An in-depth interview guide is the instrument used to collect the data.
The indepth interview guide and the interview itself could be structured, semi structured
and unstructured, but must be uniformly administered to the same group of people.
Unstructured interviews follow the form of conversations more closely and the interviewer
does not follow a rigid form (take care of the trigger question and provide impetus to the
respondent to keep on track). When to use in-depth interviews

• Complex subject matter with knowledgeable respondents - Interviews with the

actual (key) people who were involved, e.g. struggle for independence – war
veterans

Page 26 of 136 CIRSM

 RRM 400

• Highly sensitive subject matter - A study among women who have had an abortion
regarding their feelings about sexuality and family planning.

3.1.2. Key informant interviews

These are interviews with knowledgeable community experts who deal/interact with the target
population on daily basis. They provide valuable insights to the target population‘s behaviour,
attitudes, beliefs, important statistics pertaining to the target population and also key
recommendations for the study.

3.2 Non-Directive Interviews

3.2.1 Autobiography

A written story about someone‘s life (in the form of a book), where the writer is the subject. It is
the subject‘s own expression of his/her personal experience given within a chronological
framework. The term ‗life history‘ is also used in this connection. The autobiography concept
appears to be more precise in the sense that it emphasises the fact that it is the subject who is
expressing her/himself. Many autobiographies are written by people who were involved in
historical events that transcends their personal lives. So reading an autobiography may educate a
reader about the history and events above and beyond the subject of the book.

3.2.2. Biography A story of someone‘s life (written in the form of a book) by a different person.
Note that in biography, the informant (interviewee) is in control of the exploration of the topic.
However, the issues to be included in the biography are suggested by the interviewer. Biographers
often spend ‗years‘ studying their subjects, allowing them to provide insights into motivations,
contexts, as well as events. A biography maintains a more objective tone, thus the reader is more
likely to get the story from several angles (the interviewee‘s and the writer‘s), rather than only
from the angle of the subject of the book. The general rule in biography is that the respondent is
free to explore any field that appears to be related to the topic, even if it is a distant relationship. It
is therefore important for the interviewer to make it clear to the respondent, the purpose of the
study and what information the researcher is seeking. The interviewer should be a partner in the
dialogue, and should avoid putting the informant under any constraints as much as possible. The
role of the interviewer is to provide incentive and encouragement by being a listener.

Page 27 of 136 CIRSM

 RRM 400

Issues to consider when carrying out interviews

a) Sex - Stimulating or blocking the interaction during the interview. b) Age

• Everything depends on the subject, topics and social images.

• In ordinary social environment of the subject, with whom would the interviewer be
identified with.

c). Length of the interview.

While length of the interview varies in relation to the objectives of the study, an
autobiography or a biography of a single individual may be a study in itself.

d). Availability of the subject.

It is important to choose the times when the subject availability permits the most open
dialogue. It should be understood that some interviews require a high level of
concentration from both the interviewer and the informant and that the fatigue factor
must be taken into account.

In the case of a biography, the basic material is furnished in 20, 30, or even 40 hours of
interviews. It is better to come back a second time rather than to rush the interview.

e). Social setting and social class of the protagonist (hero being interviewed)

This is not only important in terms of language, but also in relation to inter-personal
attitudes, which influence the form and content of the communication.

4. Individual Case Histories

This method is used to capture the status of those people who must have been identified as
appropriate to the research problem. E.g. in the process of participatory methodology, the
community members may identify individuals considered to be very poor (or very rich)
and who may have engaged in activities/projects designed to improve their standard of
living.

The strategy here is to capture a wide spectrum of such individuals in a given community.
The approach can involve the interviewing of both successful and unsuccessful cases. The
idea is to capture the perceptions of the respondents as best as possible in their own words
and language.

Page 28 of 136 CIRSM

 RRM 400

5. Documentary Review

The method provides a record of events over a period of time, and a useful overview where
data are ready and available. It may also provide an indication of the state of development
uninfluenced by the evaluator‘s presence. Documentary review assists in-depth evaluation
and assists evaluator's understanding of the context. Public documents (policy documents,
minutes of meetings, newspapers) and private documents (journals, diary or letters) can be
used for documentary review.

6. Audio-visual material

This is an unobtrusive method of data collection which collects data using photographs,
video-tapes and art/objects. Provides an opportunity for the informant to share directly his
or her own ‗reality‘. The methods is creativity because it captures attention visually.

Sampling Methods in Qualitative Methodology

In order to collect these types of data for a study, a target population, community, or study
area must be identified first. It is not possible for researchers to collect data from everyone
in a sample area or community. Therefore, the researcher must gather data from a sample,
or subset, of the population in the study. In quantitative research, the goal would be to
conduct a random sampling that ensured the sample group would be representative of the
entire population, and therefore, the results could be generalized to the entire population.
The goal of qualitative research is to provide indepth understanding and therefore, targets
a specific group, type of individual, event or process. To accomplish this goal, qualitative
research focus on criterion-based sampling techniques to reach their target group. There
are three main types of qualitative sampling: purposeful sampling, quota sampling, and
snowballing sampling. The following descriptions describe the reasons for choosing a
particular method.

1. Convenient sampling

It relies on the availability the respondents (anyone in the target population whom the
researcher comes across). It is also known as accidental sampling (interviewing whoever

Page 29 of 136 CIRSM

 RRM 400

happens to stroll by BUT in the target population). The assumption is that by simply being
in a given locality, one can get a reasonably representative cross-section of the target
population. The method is fast, saves time & energy, easy to employ and is cost effective.
It is useful for pre-testing a questionnaire. However, there is no control over representation
of sample, hence results may not be generalized.

2. Purposive sampling

Also known as judgmental sampling. In this type of sampling, participants are selected or
sought after based on pre-selected criteria based on the research question. A researcher
purposively selects certain groups of people for their relevance to the issues being studied
and on the basis of his/her knowledge of the population. This sampling procedure is often
used when studying social phenomena that are rare to be studied using a representative
sample of the population. It ensures that researchers at least get some information from
respondents hard to locate e.g criminal activities among adolescents with dwarfism
syndrome or lymphoma patients. However, there is no assurance that those selected are in
any way representative of some population of more general interest. The sample size may
be predetermined or based on theoretical saturation, which is the point at which the newly
collected no longer provides additional insights.

3. Snow balling

Also known as reputational referral sampling. Each person (with desired attributes)
interviewed is asked to suggest additional people (with the same attributes) for
interviewing. It is often used in the absence of a complete or reliable sampling frame or
when access to subjects for interviewing is difficult. The sampling technique is often used
for studying special populations e.g. drug users, homeless people and members of secret
societies. It is used primarily for explorative research (breaking new grounds). The
interviewer may locate an individual with desired attributes and after interviews, may
request to be referred to other respondents of similar attributes. Technique is good for
gaining access to knowledgeable people who may provide indepth knowledge on a given
phenomenon. However, there are problems with representation & generalizability.

Page 30 of 136 CIRSM

 RRM 400

4. Quota sampling

People are selected into a sample on the basis of pre-specified background characteristics
e.g. age, sex, race, marital status, place of residence and religion. A quota is produced in
accordance with percentage of the population composed of a particular stratum e.g. 49%
male and 51% female, 75% rural and 25% urban, 4% white and 96% black. Accidental
sampling is done to identify respondents in each stratum. There is assurance that certain
strata of the population will be included. Hence, resembles probability sampling and strives
for representativeness.

Hint on sampling

Note that once a sampling method has been determined, the researcher must consider the
sample size. In qualitative studies, sampling typically continues until information
redundancy or saturation occurs. This is the point at which no new information is emerging
in the data. Therefore, in qualitative studies is it critical that data collection and analysis
are occurring simultaneously so that the researcher will know when the saturation point is
reached. It is important to understand that the saturation point may occur prematurely if the
researcher has a narrow sampling frame, a skewed analysis of the data, or poor
methodology. Because of this, the researcher must carefully create the research question,
select an appropriate target group, eliminate his or her own biases and analyze data
continuously and thoroughly throughout the process to bring validity to the data collected.

Methods of recording qualitative primary data a) Observation Data

In recording unstructured observation, two questions are important:

1. When should the observer make notes? The best time for recording observation is
immediately, as this will result in the minimum amount

of selective bias and distortion through memory. However, this is not always possible
as the observer may lose relevant aspects of the observation if s/he is writing and
observing simultaneously. The observer must write their impressions as soon as possible
after and observation.

Page 31 of 136 CIRSM

 RRM 400

2. How should notes be made?

Observations can be recorded in a number of ways:

(a) The most comprehensive of these being video or audio equipment.

(b) More commonly though, researchers record their observations through field
notes. b) Interview Data

• Tape or digital recording

• Note taking

• Writing up after the event

Controlling errors in data collection a) Observation

The following are procedure which should be used to ensure the accuracy of the observation:

(i) Reflexivity. This involves the researcher monitoring, and reflecting on his/her
own influence and on the effect of the social context, on data production. (Foster, 1994:91).

(ii) Triangulation. It involves checking the validity of findings and or/observational

data themselves by cross-checking with other sources of data.

(iii) Respondent validation. Involves checking the validity of the researcher's

observations by making reference to the subjects‘ perceptions.

b) Interview

The following are ways for controlling for these potential sources of error:

• To compare one respondent‘s account with other accounts i.e. crosschecking information;

• Complementing interview data with other forms of data i.e. triangulation;

• Carrying out repeated interviews with the same respondent; and,

• To exclude interviewer bias from recording of responses, some mechanical method should
be used.
Page 32 of 136 CIRSM
 RRM 400

Attributes of a good interviewer

1. Accurately receiving the information.

Fatigue, boredom, bias and expectation of a specific answer, among other things, can affect
this ability. If the interviewer keeps on taking notes during the interview, s/he might miss
something that is said, in addition to disrupting the flow of the discussion.

2. Accurately recalling the information

This can be affected by: a confusion of content between interviews; selective retention on
the part of the interviewer; the interviewer‘s attempt to retain too much information.

3. Critically evaluating information

During the interview, it is the function of the interviewer‘s ability to identify the actual
level of richness of the content being provided. It is also important that the interviewer
steers the respondent away from irrelevant information, and induces richness when
superficial answers are being provided. The research must act upon the information being
received and altering the interview as it takes place. This is really an issue of probing,
focusing and staying on track with respect to the interview objective. A researcher should
display some familiarity with the phenomenon and setting under study and strong
conceptual (theoretical) interests. Use also a multi-disciplinary approach, as opposed to a
narrow grounding or focus in a single discipline. There is also need for good ‗investigative‘
skills, including doggedness (perseverance or determination), the ability to draw people
out, and the ability to ward off premature closure.

Having presented the qualitative data collection methods and the sampling strategies, let me give
you an example of how you would present the data collection methods whenever you are writing
the methodology section of your research/ dissertation. Let me assume that you have been asked
to carry out a research on the topic, ‗Gender based violence (GBV). A case study of Gweru,
Zimbabwe‘. The objectives the study are: to identify the forms of GBV in Gweru; to establish the
underlying factors to the forms of GBV so identified; and, to assess the availability, accessibility
and acceptability of reporting structures in Gweru.

Page 33 of 136 CIRSM

 RRM 400

Always remember that you should explicitly state the following: the method; how many; tool;
with who; why (you are justifying the the selection of the method by linking to your objectives);
selection of the respondents; sampling; and how data were collected and recorded – an approach
I have earlier on alluded to.

Data collection methods a) Focus Group Discussions (FGDs) (State the method first)

Fourteen FGDs were undertaken with males and females comprising both the younger and older
generation using a semi-structured FGD guide. (Note that in the first sentence, I have stated the
number of FGDs, participants and the tool). Seven FGDs were conducted with males and seven
others with women. Females were selected so as to test and validate information obtained from
male FGDs. Each FGD had between 6 and 12 participants. Since FGDs are largely exploratory, it
was necessary to get a group generated perception of the forms of violence common in Gweru.
(Note that I am now justifying the use of FGDs but however, linking to my objectives). Moreover,
FGDs provided data on what the community perceived as the underlying causes to GBV. (Another
justification that has been linked to my second objective). The community‘s perceptions and
evaluative analysis of the availability, accessibility, and acceptability of reporting structure for
GBV was also drawn upon in the FGDs (Another justification that has been linked to my third
objective). Selection of respondents for each FGD was based on the following age cohorts: 15-19;
20-29; 30-39; 40-49; 50-59; 60-69; and 70-79 years. Each age cohort had its own FGD. (The
sentences above are showing how I have selected my FGD participants). Convenient sampling
and willingness to participate were the basis of selection of the
FGDs‘ participants. (I am now articulating the sampling strategy). One of the researchers
facilitated the FGDs using the local Shona language. The other researcher recorded the responses
for the participants using a tape recorder and extensive note-taking. (I am highlighting how the
FGDs were conducted and how data were collected).

b) In-depth interviews (State the sub-heading of the method)

Twelve face-to face in-depth interviews were undertaken with selected women who suffered from
the identified forms of GBV such as physical violence, spousal rape, rape by a stranger and forced
marriage. These interviews were conducted using an in-depth interview guide. In-depth interviews
Page 34 of 136 CIRSM
 RRM 400

helped solicit information about personal experiences of each form of GBV. They also provided
personal regarding the factors underlying each form of GBV. (Note that at times you can use in-
depth interviews to follow up on issues raised from other methods of collecting data such as a
survey, documentary review and key informant interviews). The GBV victims were first identified
through the survey. However, snow-balling was further utilised for the identification of other in-
depth interviewees. The author conducted the in-depth interviews, while she also recorded the data
through notes and tape recordings.

c) Key informant interviews (State the method)

Six key informant interviews were conducted with the Victim Friendly Unit (VFU) police officers,
Musasa Project and Ministry of Health and Child Welfare (MoHCW) focal persons. A semi-
structured key informant interview guide was used to solicit data. Both males and females were
selected from each category. Two officials from each organization were purposively selected. Key
informant interviewees were community experts who provided statistics on the prevalence of the
various forms of GBV in Gweru. They provided data on how they investigate and deal with each
of the identified form of GBV. The diversity of key informants provided varying perspectives on
the socio-cultural, economic, religious, legislative, political and environmental factors underlying
GBV. They also provided information regarding the availability, accessibility and acceptability of
the reporting structures for GBV. Key informants provided confidential incidences, statistics and
recommendations pertaining to GBV in Gweru. Their particular community based knowledge and
understanding of GBV issues provided a logical basis for the development of more practical,
locally oriented and detailed plans, recommendations and solutions. Key informant interviews
were conducted by the author. The same author recorded data through note taking and tape
recording.

Qualitative data analysis using the thematic approach

Qualitative data from FGDs, the key informant and in-depth interviews should be transcribed,
translated and typed. Data can be analysed using the critical realism approach. This is a scenario
where data is firstly demarcated into thematic segments which inform the objectives of the
research. For instance, let us assume that you analyzing qualitative data for the topic, ‘Factors
influencing criminal activities in Mabvuku, Harare‘ and the objectives of the study are: to identify
the commonly reported criminal activities in Mabvuku; to explore the factors underlying the

Page 35 of 136 CIRSM

 RRM 400

criminal activities so identified; and, assess the effectiveness of law enforcing agents in cubing
criminal activities. You should identify and code the themes for each objective. For instance for
the first objective, you must be able identify and code each of the criminal activities that have been
reported from the qualitative data. For the second objective, you should also identify and code
separately the factors underlying criminal activities. Note that you can identify the following
factors which should be coded separately: social; cultural; economic; environmental; political and
religious. Among the social factors, you might be also able to identify separate subthemes which
underlie the social factors. These subthemes could be peer pressure, double and single orphanhood,
divorces, and lack of parental guidance. So basically, you should code the main theme. Within
each segment of the main theme, you should also code the sub segments representing different
aspects of the main theme. Subsequently, assess the prevalence of codes in order to describe the
commanality of views while at the same time discussing similarities and differences in related
codes across distinct original sources/contexts. In addition, compare also the relationships between
codes, that is sub-themes.You can display the output of your analsyis uin the form of a structure
e.g a tree. Note that the stem of each tree would represent a particular theme (code), while the
branches represent the identified issues around the particular theme. You can also attach common
quaotations to the respective thematic general findings.

Presentation and interpretation of qualitative data

As you present your qualitative findings, make sure the findings are answering/responding to the
research objectives, research questions, research problem and also showing the research methods
that were use. In addition, remember you are presenting the findings you noted from the identified
themes and sub-themes. Firstly, start with a general statement which highlights a particular theme.
Then explain the theme or the subtheme also highlighting the method(s) which revealed the
finding. You should then substantiate your explanations with selected quotation(s) that are
poignant and/or most representative of the theme or research finding. Below is my example of how
you may present the findings for the topic on Gender based Violence in Gweru:

Qualitative data revealed that employment status of a woman is can influence physical abuse.
Women who were not formally employed were noted to be economically challenged. As a result,
they highly dependent on their husbands for survival. Such dependency was noted as a facilitating

Page 36 of 136 CIRSM

 RRM 400

factor to physical abuse during key and in-depth interviews. One woman aged 34 years in an in-
depth interview supported the view that unemployed women are always on the writhe of physical
abuse as she reported that:

Normally unemployed women are mostly subject to physical abuse because of their 100%
dependency on the husband. How I wish I was employed. My husband always beat me with
clinched fists each time I tell him that we no longer have grocery, and more so at the
monthend.
He always accuses me of being a parasite, waiting to suck his energy. Surprisingly, he does
not allow me to sell even pre-paid cell phone top up cards to earn extra cash so that I cease
to be a parasite. I cannot report my case to police because if he is to be locked in cells, that
will be a double tragedy to me. I do not have resources to look after myself and children.
Key informant interviews with Victim Friendly officers also highlighted the view that unemployed
and informally employed women suffer the greatest quandary of physical abuse. One of the officers
stipulated that:
More than 60% of women who have ever reported cases of GBV to this station are either
unemployed or informally employed. They usually say that poverty is disadvantaging them
and it has become the key determinant of GBV. However, it is surprising to note that before
the cases are taken to court, the majority of them withdraw the cases, a development that
likely perpetuates other cases of GBV.

Page 37 of 136 CIRSM

 RRM 400

Introduction to Quantitative Research Definition and Essence

Lisa (2010) asserts that quantitative research is the systematic empirical investigation of observable
phenomena via statistical, mathematical or computational techniques. Babbie (2010) asserts that
the goal in conducting quantitative research study is to determine the relationship between one
thing [an independent variable] and another [a dependent or outcome variable] within a population.
The objective of quantitative research is to develop and employ mathematical models, theories and
hypotheses pertaining to phenomena. The process of measurement is central to quantitative
research because it provides the fundamental connection between empirical observation and
mathematical expression of quantitative relationships.

Quantitative data is any data that is in numerical form such as statistics, and percentages. The
researcher analyses the data with the help of statistics and hopes the numbers will yield an unbiased
result that can be generalized to some larger population. Quantitative data collection methods
emphasize objective measurements and the statistical, mathematical, or numerical analysis of data
collected through polls, questionnaires, and surveys, or by manipulating pre-existing statistical data
using computational techniques. Quantitative research focuses on gathering numerical data and
generalizing it across groups of people or to explain a particular phenomenon. Lisa (2008) notes
that quantitative research also focuses on numeric and unchanging data and detailed, convergent
reasoning rather than divergent reasoning [i.e., the generation of a variety of ideas about a research
problem in a spontaneous, free-flowing manner]

Characteristics of Quantitative Research

• The data is usually gathered using structured research instruments.
• The results are based on larger sample sizes that are representative of the population.
• The research study can usually be replicated or repeated, given its high reliability.
• Researcher has a clearly defined research question to which objective answers are sought.

Page 38 of 136 CIRSM

 RRM 400

• Data are in the form of numbers and statistics, often arranged in tables, charts, figures, or
other non-textual forms.
• The research can be used to generalize concepts more widely, predict future results, or
investigate causal relationships.
• Researcher uses tools, such as questionnaires or computer software, to collect numerical
data.

The overarching aim of a quantitative research study is to classify features, count them, and
construct statistical models in an attempt to explain what is observed.

Quantitative Research Concepts

Target Population

This is the entire population or group that a researcher is interested in studying and analyzing, e.g
all men between 18 and 79 years. The target population defines those units for which the findings
of the survey are meant to generalize.

Sampling

The selection of a small group of individuals from whom data is collected for generalizations to a
larger group of people.

Sampling frame

A list of all the elements from which a sample is drawn e.g. all youths (males and females) aged
15 to 24 in Seke Communal lands.

A Representative Sample

• This is a subset/ ‗cross section‘ of a population that closely matches the characteristics of
its population as a whole. A representative sample is one that has strong external validity
in relationship to the target population the sample is meant to represent. In other words, the
sample is a fairly accurate reflection of the population from which the sample is drawn.

• A representative sample is intended to make some inferences and generalisations about the
whole population.

Page 39 of 136 CIRSM

 RRM 400

Let me use Figure 3.1 to show the difference between a population and a sample. Note that the
population is the sampling frame from which the target population can be drawn.

Figure 3.1: The relationship between target population and a sample

Unrepresentative sample

This is a biased sample which does not reflect the distribution of characteristics of the target group. A
researcher intentionally or unintentionally seeks out respondents who support his theory or views.
Researchers commonly use biased samples because of laziness/sloppiness. They
take samples from what happens to be easily available rather than taking their time and effort to generate
an adequate sample and drawing a justification. Potential consequences of unrepresentative sampling

- Data collected from the unrepresentative sample cannot be generalised to the target
population, and is therefore biased.

- Data collected from the unrepresentative sample does not represent the population from
which it is drawn.

- Biased samples are less likely to contain numbers proportional to the whole population.

Page 40 of 136 CIRSM

 RRM 400

Sampling error

Unavoidable, but measures the differences between sample statistics and population
parameters usually resulting from probability sampling methods.

Non-sampling error

It emanates from other factors that reduce the accuracy of estimates based on sample results and
these include:

1. Non-response

• Problem inherent in all interviewing methodologies (door – door, mail, telephone, internet
etc)

Affects representativeness of the sample and generalisability of sample results

2. Inherent sampling bias (Internet polling)

How representative is the sample of internet users to the general population?

Serious problems of generalizability.

3. Poor sampling frame

Affects the generalisability of results. E.g Focusing only on the army in a study on HIV & AIDS
among the uniformed forces or focusing only on the youths when studying the prevalence of
criminal activities in an area.

Over Representative Subgroups/ Weighting

Weighting is a correction technique that is used by survey researchers. It refers to statistical

adjustments that are made to survey data after they have been collected in order to improve the
accuracy of the survey estimates. There are two basic reasons that survey researchers weight their
data. One is to correct for unequal probabilities of selection that often have occurred during
sampling. The other is to try to help compensate for survey nonresponse. This entry addresses
weighting as it relates to the second of these purposes. Data collected from survey respondents are
called raw data, or unweighted data. Often, data collected from surveys is not exactly
representative of the target population even in those instances where quotas have been applied.
However, when data collected from survey respondents are adjusted to represent the population
Page 41 of 136 CIRSM
 RRM 400

from which the sample was drawn, the resulting data are called weighted data. Imagine you have
a target population that is evenly split by gender. If you then interview a sample of 400 people
within this population, 300 of whom are male and 100 female then you would know that you
sample overrepresents men.

Weighting the resulting data can help you to correct this imbalance. Let us assume that in your
study, the target proportions for both men and women are supposed to be 50%. The proportion of
men would therefore need to be
―down weighted‖ from 75% (300 out of 400 interviews) to 50% while the proportion of women
needs to be ―up weighted‖ from 25% to 50%.

In this case, weighting would be multiply the existing female interviews by 2, so that the female
response is amplified in the data. For example, on the gender question we have 100 people
answering female but after weighting this becomes 200 as the "female" data is counted twice. The
male interviews need to be correspondingly down weighted. In this instance we need to get 300
responses to effectively count as 200 so we multiply the male responses by two-thirds (or 0.67).
Before weighting we have 300 males coded on the gender question. Multiplying by two-thirds
gives us 200 males, equaling the number of female responses after weighting. The numbers used
to multiply the responses from each proportion of the sample are called weighting factors. A
summary of the weighting factors for this example is shown in Table 3.1.

Table 1: The weighting procedure

Group Before weighting After weighting
Number % of Number of % of target Weighting
of interviews sample interviews population factor

Males 300 75 200 50 0.67

Females 100 25 200 50 2

Note that typically weighting is used to match the population profile on more than 1 variable to get
as representative a sample as possible. For example, to get a representative sample of a country‘s
population we might weight on a number of demographic variables such as gender, age, region
and social grade.

Page 42 of 136 CIRSM

 RRM 400

Uses of weighting

a). Adjust for probability

One of the basic principles of probability samples is that every respondent must have a known,
non-zero chance of being selected. In the typical media poll, the sample is designed to represent
all adults 18 years of age and older.
In practice, this also means ―among those residing in telephone households‖ because that is the
way the data are collected. So survey data is weighted for the probability of selecting a respondent.
b). Adjusting for sample design

In a stratified sampling design, groups of potential respondents with an identifiable characteristic

are selected with unequal weights in order to increase the sample size. This is sometimes called
an oversample. It is used to generate more precise estimates of how that group feels or is reacting
to the campaign. For example, in some surveys in the 2008 election cycle, African Americans, who
represent about 12 percent of the U.S. population, are being oversampled to assess their reactions
to Barack Obama and Hillary Clinton. That means that the probability of selecting an
AfricanAmerican respondent is increased and the probability of selecting a white respondent is
decreased, in relation to their relative proportions in the adult U.S. population. Data can be used
straightforwardly to compare the reactions of whites and African Americans. But if you want to
produce a national estimate of support for Clinton and Obama, a weight will have to be created to
―reduce‖ the contribution of each African American respondent and ―increase‖ the contribution
of each white respondent. Otherwise, the estimate of support for each candidate will not reflect
their actual support. c). Adjusting for demographics

After all of the data have been collected, some simple frequencies are run on selected demographic
variables so they can be compared to known characteristics of the population obtained from an
external source, such as the Zimbabwe Census. This adjustment is made because some
demographic groups tend to be overrepresented or underrepresented in the sample. For example,
young men are considerably harder to reach at home than older women. So unweighted data
frequently include a larger proportion of older women and a smaller proportion of younger men
than what the Zimbabwe Census reports. A pollster typically makes small adjustments -- called
post-stratification weights -- to bring the sample into line with known population characteristics
such as age, sex, region and education.

Page 43 of 136 CIRSM

 RRM 400

Design effect

This is an adjustment that should be used to determine survey sample size. According to Henry
(1990), Frongillo (1996) and United States Census Bureau (2001), design effect is basically the
ratio of the actual variance, under the sampling method actually used, to the variance computed
under the assumption of simple random sampling. For an example, ―The interpretation of a value
of (the design effect) of, say, 3.0, is that the sample variance is 3 times bigger than it would be if
the survey were based on the same sample size but selected randomly. An alternative interpretation
is that only one-third as many sample cases would be needed to measure the given statistic if a
simple random sample were used instead of the cluster sample with its (design effect) of 3.0‖. The
designer effect is used to determine how much larger the sample size or confidence interval needs
to be. In general, for a well-designed study, the design effect usually ranges from 1 to 3. It is not
uncommon, however, for the design effect to be much larger, up to 7 or 8, or even up to 30. Note
that the main components of the design effect are the intra class correlation, and the cluster sample
size.

Page 44 of 136 CIRSM

 RRM 400

Quantitative Data collection

Quantitative Research Designs

Remember in Chapter 1, l defined a research design as a format of the study. Similar to qualitative
research, quantitative research has also different designs.

1. Experimental designs:
The main focus of experimental designs is to determine causation. Manipulation of one or more
variables under carefully controlled conditions is typically common. They are better suited for
explanatory studies. Among the experimental designs, there are true experimental designs and
quasiexperimental designs.

2. Non-experimental designs:
There is no manipulation of variables by the researcher. Variables are observed/studied as they
exist, e.g surveys.

Quantitative data collection method

Survey
It is a systematic quantitative method of collecting data. It collects information from a sample of
individuals to quantify their beliefs, attitudes, values and behaviour. Surveys are the most
frequently used mode of observation. The value of surveys depends on the representativeness of
the sample. They normally use large numbers of people when compared to experiments and
qualitative designs. Survey are used to describe, explain and explore situations or events. They are
also normally used in studies that have individual people as units of analysis (respondents). A
standardized questionnaire is the instrument use when collecting data using a survey. It is
administered to respondents to solicit for responses to the ‗what‘ question.
Surveys can be used to test existing theory and hypothesis rather than to discover theory.

Page 45 of 136 CIRSM

 RRM 400

Types of surveys
1. Cross-sectional
 Observations are made at one point in time (snapshot approach)
Figure 4.1.
 Single unrepeated survey
 Suited to descriptive and exploratory studies

Figure 4.1: A Cross-sectional survey

Page 46 of 136 CIRSM

 RRM 400

3. Longitudinal survey
Conducting a study at different times (Figure 4.2).

Page 47 of 136 CIRSM

 RRM 400

Figure 4.2: Example of a longitudinal survey

Types of Longitudinal studies

Trend Studies
Different samples of the same population are studied at different times. Each sample represents the
population at the time of study.
Cohort Studies
The same (specific) population studied at different times. Different samples of the same population
are studied.
Panel Studies
Data are collected over time from the same sample of the population. Same respondents over a
time
Sample size determination
Results based on a total count are more reliable than those based on a sample.
The bigger the sample is the smaller the standard error and the better the estimates. Quantitative
sample sizes tend to be bigger compared to qualitative ones.

How large may a sample be or how big is a ‗representative sample? There are several methods of
determining the sample size. However, you can use the formula n = Z²pq/e². Where: n = sample
size; z = standard normal deviation set at 1.96 corresponding to 95% confidence interval; p =
Page 48 of 136 CIRSM
 RRM 400

proportion of the target population to the entire population; q = proportion of the entire population
excluding the target population (1-p); and e = maximum allowable error margin set at 0.05.
Rule of thumb in sample size determination
In the absence of statistical guidelines, a sample size of between 5% - 10% of the population is big
enough to be representative.

Quantitative sampling design

Probability sampling

Selection of samples typically involve some random selection mechanism. A sample will be
representative of the population from which it is selected if all members of the population have an
equal chance of being selected. Probability sampling is normally used in large scale surveys. It
provides an efficient method for selecting a sample that adequately reflects variations existing in a
population. Probability sampling is more representative when compared to other sampling
methods. Random selection of respondents reduces conscious and unconscious bias by the
researcher. However, note that probability sampling can be costly and demanding given that it may
be technical.

Types of probability sampling techniques a. Simple

Random Sampling
This is the most basic method of probability sampling. Each element in the population has an equal
chance of being sampled (1/N chances). When using the simple random sampling technique, make
sure a complete list of all elements in the population is available. Each element is assigned a
random number which can be randomly selected. Simple random sampling ensures that there is no
bias in the selection process. However, it may be costly in carrying out given that it requires the
technical know-how. In addition, the technique does not guarantee perfect representation
(sampling error)

b. Systematic random sampling

The researcher must know his/her total population and the sample size. With these two parameters
you firstly calculate the sampling interval (n) = Population size/ sample size. eg. if the population
size 100000, sample size 4000, then the (n) would be 25.

Page 49 of 136 CIRSM

 RRM 400

Randomness is in the selection of the first case in the first sampling interval. However, the random
number should be between 1 and 25 which you would call x or any letter. Suppose you select 5,
the number you will select in the next sampling interval is x + 25. This would be 5 + 25 = 30, then
30+n = 35; 35+n =40, etc. Systematic random sampling is simpler to use when compared to simple
random sampling. Randomization is done only once. Nonetheless, the technique may suffer from
sampling bias. c. Stratified Random Sampling

Involves dividing the population into 2 or more strata and then taking either a simple random
sample or a systematic sample from each stratum. Note that the strata must be internally
homogenous with heterogeneity between or among them. Strata are selected on the basis of
variables relevant to the study. They maybe age, sex, economic status, place of residence etc. You
need to sample each stratum according to its percentage in the total population i.e. use the
probability proportional to size to allocate the sample proportionately to each strata.
Disproportionate sampling is when strata are not sampled according to their percentage in the
population. It is done when proportionate sampling yields a very small number of elements in a
population and this does not allow detailed data analysis to be done. It should be noted that
stratification increases precision by reducing sampling error. The main challenge with stratification
can be on deciding which variables to govern stratification e. Cluster sampling

It is used when it is not practical or very difficult to obtain a complete list of all elements in a
population because of the massiveness of the sampling frames. E.g. of massive sampling frames
include all adolescents in Zimbabwe, all university students in Zimbabwe, the population of
Zimbabwe, number of church members in the country etc.

Cluster sampling involves:

1). Randomly selecting existing clusters in a population through simple random, systematic or
stratified random sampling

OR, creating clusters and randomly selecting some from among them.

2. Randomly selecting individual elements from randomly selected clusters and interviewing them.

Multi-stage cluster sampling is one form of cluster sampling involving various stages at which
sampling is done.

For example, How would you sample adolescents for participation in a national baseline survey
on cyber security?
Page 50 of 136 CIRSM
 RRM 400

NB. Obtaining a complete list of all adolescents in Zimbabwe is next to being impossible.

Stage 1: Obtain sampling clusters at national level. List all provinces in Zimbabwe & randomly
select those to collect data from.

Stage 2: Obtain sampling clusters at provincial level. List all districts in the selected provinces &
randomly select some for data collection

Stage 3: Obtain sampling clusters at district level. List all wards in selected districts and randomly
select some for data collection

Stage 4: Obtain sampling clusters at ward level. List all villages in selected wards and randomly
select some for data collection

Stage 5: Obtain sampling clusters at village level. List all households in the villages and randomly
select some for data collection

Stage 6. Randomly select adolescents from selected households and interview them.

It should be noted that cluster sampling is highly efficient and cost effective when compared to
other probability sampling methods. It enables researchers to conduct studies that would otherwise
be impossible. Nonetheless, the technique has more than one sampling error given that it has more
than 1 cluster sample.

NB. In discussing methodology, you discuss sampling down to the selection of individual
participants.

Quantitative research tool - Questionnaire

This is the tool or instrument used when collecting survey data which will be subsequently anaysed.
It is mainly characterized by close-ended questions e.g. (yes/no), (strongly, agree, disagree, and
strongly disagree).

Questions in a questionnaire must answer the broad and specific objectives of the study. In
designing the questionnaire, you should be aware of the indicators to be measured. The
questionnaire format should be spread out and orderly. You should choose appropriate response
format and this should be well-spelt out at the beginning of the questionnaire. You should indicate
how the respeonse should be shown (circling, ticking or putting an X where appropriate). The
common features of a questionnaire are shown below. Introduction;
Page 51 of 136 CIRSM
 RRM 400

Every questionnaire to have an introduction which explains what the study is about.

Instructions and introductory comments;

Questionnaire to contain clear, specific instructions & introductory comments where appropriate.

The relevant questions;

- Questions must be relevant to the study (appropriate indicators to answer objectives)

- Questions must be relevant to respondents

- Field of study must also be familiar to respondents

- Mostly has close-ended questions, open-ended questions should be few and avoided.

- Response category for closed-ended questions should be exhaustive & mutually exclusive
(except for multiple responses).

- Questions may also be made up of statements measuring attitudes, beliefs and knowledge
(Matrix questions).

- Indicate (by ticking) in the appropriate space provided, whether you Strongly Agree (SA),
Agree (A), Disagree (D), Strongly Disagree (SD), Undecided (U).

- Multiple response question(s). Note that in a multiple response question, a respondent has
an option to give more than one answer. Eg.

Which form(s) of gender based violence have you ever heard?

1. Yes 2. No
a) Physical abuse 1 2

b).Spousal rape 1 2

c). Rape by a stranger 1 2

d). Emotional abuse 1 2

e). Economic abuse 1 2

Other… (Specify)…

Page 52 of 136 CIRSM

 RRM 400

In designing a questionnaire, make sure you create short and concise questions. Avoid long,
winding and complicated questions as they would confuse the respondents. Also avoid double-
barreled questions. This is a single question or statement that may produce 2 different answers or
combining questions, e.g.

- Please circle your most appropriate response to the following statement?

Sex education must be taught at primary schools and the schools must be allowed to provide
condoms to sexually active pupils.

1. Yes 2. No. 3. Don‘t know.

When designing a questionnaire also remember to ask for one piece of information at a time. Avoid
biased and leading questions e.g.

- Don‘t you agree with the notion that----

- Wouldn‘t you say that ----

- Isn‘t it fair to say ----

Make sure you avoid asking questions with names of prestigious persons attached to them e.g. Do
you agree with the Prime Minister that -----.Take note of Contingency questions and questions that
do not apply to every respondent (skip patterns). E.g. Have you ever had sex? 1. Yes 2. No (skip
to Question 5)

Sequence of questions in a questionnaire:

- Ask related questions in an ordered way.

- Start with most interesting questions if questionnaire is

selfadministered.

- For interviewer-administered questionnaire, start with general questions, e.g. demographic

data.

- Do not start with awkward or embarrassing questions – respondents may just give up.

Page 53 of 136 CIRSM

 RRM 400

- Start with easy and non-threatening questions. This encourages respondents to carry on
with the questionnaire. - Ensure options are mutually exclusive

– e.g. ―How many years have you worked in academia: 0-5, 6-10, 11- 15, over 15.‖ Not,
―0-5, 5-10, 10-15…‖

Pre-testing a questionnaire

Involves administering a questionnaire to a very small sample (not in your area of study)
and noting all the problems that arise for the interviewer and subject. This approach
highlights any ambiguity in the wording of your questions. It would enable you to note the
necessary and desirable changes in wording, format and ineffective questions. Methods of
administering a questionnaire

1. Self- administered

Respondents are asked to complete the questionnaire by themselves. The questionnaire

can be mailed or delivered at home. Advantage
- Cheaper and easily done

Disadvantage

- Poor response rate

- Telephone interviews

2. Telephone interviews

Advantage

- Cheaper and convenient

Disadvantage

- Respondents might be suspicious

- Limited to people with telephones

- Potential respondents can easily hang up

- Use of answering machines

Page 54 of 136 CIRSM

 RRM 400

3. Interviewer administered

Interviewer administered (face to face / interview survey)

Advantage

Higher response rate than the other two

Interviewers will explain or probe where necessary

Interviewers can make observations relevant to the study Disadvantage

- Expensive to conduct

Guidelines for interviewer administered questionnaires

- Sensitize the police/ traditional leaders in the area

- Dress appropriately

- Establish rapport

- Be neutral

- Familiarity with the questionnaire

- Read questions and record answers correctly

4. Computerised methods

Computer Assisted Telephone Interviewing (CATI)

- An interactive front-end computer system that aids interviewers to ask questions over the
telephone.

- Answers are then keyed into the computer system immediately by the interviewer

- Reduces manpower needs

- Lower costs

- Speeds up the process of data collection

Quality control

The aim is to ensure that quality data are collected. Measures include:
Page 55 of 136 CIRSM
 RRM 400

- Pre-testing questionnaire

- Training enumerators on the tool

- Debriefing during data collection

- Field editing

- Double entry of data

- Data cleaning

Let me show you how would write the survey method using the method, how many, tool, with
who, why, sampling, selection of the respondents and how data were collected and recorded
approach I had use in qualitative methodology. My example will still use the topic on Gender based
violence in Gweru.

Survey (Explicitly state the sub-heading of the method)

A face-to-face survey, using a questionnaire was carried out with males and female between 15
and 79 years of age. The survey was undertaken in order to quantify and determine the magnitude
of the commonly reported forms of GBV in Gweru, as well as the factors underlying the forms of
GBV so identified. The survey also quantified the extent of the availability, accessibility, and
acceptability of the reporting structures by the target population.

Sample Size Determination

The sample size was calculated using the formula; n = (Z2pq)/e2. Where n = the desired sample
size; Z2 = the standard normal deviate usually set at 1.96 which corresponds with 95% confidence
interval; p = the proportion of males and females (the target population) between 15 and 79 years
to the total population in Senga Gweru; q = the proportion of total population in Senga excluding
15 to 79 year old males and females (1- p); and e2 = desired level of precision set at 0.05. In order
to calculate the sample size, a 95% confidence interval was used. In addition, 0.05 error margin
was assumed to be the degree of accuracy. This means that deviations of sample estimates from
the true population were only considered accurate if the significant level was pegged at 5%. The
test statistics for Z scores was also used.

Page 56 of 136 CIRSM

 RRM 400

Senga has a total population of 4710 of whom men and women between exact age 15 and 79 years
constitute 1297 (Note that this is my assumption. You have to use the correct figures and cite the
source). Therefore, the sample size calculation was (1.962*0.28*0.72)/0.052. The calculated
sample size was 303 male and female respondents. However, for robust analysis, the sample size
was set at 500 respondents.

Sampling procedures

You state all the sampling procedures you will have used. It can be a multistage sampling where
you would start off by selecting a ward and enumeration areas, household and then the actual
respondent. You will also need real or the actual figures of the distribution of the population in
your study area to assist you in explaining the sampling procedures.

Page 57 of 136 CIRSM

 RRM 400

Descriptive Statistics

Measures of central tendency

Central tendency is a score that indicates where the center of the distribution tends to be located.
It informs you about the shape and nature of the distribution.
Mean
This is the score located at the mathematical center of a distribution. It is used to summarize
interval or ratio data in situations when the distribution is symmetrical and unimodal.
Xˉ ∑x
n

Where Xˉ (x bar) = mean;

∑x = sum of all data values; n = number of all data values

Example.
The following data shows the weight in kilograms (kgs) of 9 robbers: 123; 79; 62; 145; 87; 91; 84;
133; and, 87.
Mean = 123 + 79 + 62 + 145+ 87 + 91 + 84 + 133 + 87
9 = 99kgs

Median
The median is the middle value of rank ordered data. Value that separates the higher half of a data
set from the lower half. It can be calculated by arranging all values from lowest to highest and
determining the value in the middle.
If there is an odd number of values in the data set, then the median is the middle value. Assuming
that you are given the data set 3, 2, 1, 6, 7, 4, 9, you must first arrange the data set from the lowest
to the highest. Thus, 1, 2, 3, 4, 6, 7, 9. From this arrangement, the median is 4.

Page 58 of 136 CIRSM

 RRM 400

If there is an even number of values in the data set, 3, 2, 1, 6, 7, 4, 9 and 8, then the median is the
mean of the two middle values. Firstly arrange the data set from the lowest to the highest, 1, 2, 3,
4, 6, 7, 8, 9. From this data set, the median is the average of two middle values, 4 and 6 = 5.
Mode
The most frequent measurement. Examples:

a) 132, 139, 131, 138, 132, 139, 133, 137, 139. Mode = 139.

b) 3, 3, 3, 5, 10, 11, 3, 6, 4, 8, 14, 21, 2, 17, 3, 15. Mode = 3

c) 56, 23, 48, 78, 94, 35, 88, 69, 44, 53, 27. There is no mode

Measures of Dispersion a).

Standard deviation

This is a function of sampling error which is meant to measure how much sample statistics are
deviating from the true population parameters. The bigger the sample size, the smaller the standard
error and the more representative the sample is.

Calculating the standard deviation

The formula for the sample standard deviation of a data set (s) is:

Where:
xi is each value is the data set; x-bar is the
mean; and, n is the number of values in the data set.
Steps in calculation the standard deviation.

1. Calculate the average of the data set,

2. Take each value in the data set (x) and subtract the mean from it find the square of its
distance to the mean,

3. Square each of the differences,

4. Add up all of the results from Step 3 to get the sum of

Page 59 of 136 CIRSM

 RRM 400

squares,

5. Divide the sum of squares (found in Step 4) by the number of numbers in the data set minus
one; that is, (n – 1) to get the sample variance,

Note that the sample variance is the average of the squared differences from the mean.
Example.
You have the following data set: 1, 2, 4, 6, 7, and you are asked to calculate the standard deviation.
Solution.
Calculate:

1. The mean of the data set is: 1+2+4+6+7/5 = 4

2. Subtract each vale of the data set from the mean: (1-4) =-3; (24) =-2; (4-4) = 0;
and, (6-4) = 2; (7-4) = 3.

3. Square each of the differences and you get: 9, 4, 0, 4, 16.

4. Sum of squares = 33.
5. Sample variance = 8.25
6. Standard deviation is the square root of 8.25 = 2.87. b). Range
- Numbers between the lowest and highest.
- Minimum is the lowest figure, while maximum is the highest figure in a sample.
- Note that any figures outside the defined range are called outliers.

Distribution of a sample/ Skewness

It is basically the distribution of the data. Skewness helps you to understand the relationship
between three measures namely: the mean; median; and, mode.
a). Normal distribution
A normal distribution of data is one in which the majority of data points are relatively similar,
occurring within a small range of values, while there are fewer outliers on the higher and lower
ends of the range of data (Figure 5.1).

Page 60 of 136 CIRSM

 RRM 400

When data are normally distributed, plotting them on a graph results in an image that is bell-shaped
and symmetrical, like the one shown here. In such a distribution of data, the mean, median, and
mode are all the same value, and coincide with the peak of the curve.
The normal distribution is also often called the bell curve because of its shape. Note that in a normal
distribution, skewness = 0. This means that the mean = median= mode.

Figure 5.1: The normal distribution curve

a) Bimodal distribution
This is when the data has two peaks (Figure 5.2) Picture.

b) Left and right skewed data

Data can be skewed either to the left or to the right, Figure 5.3). Note that when data is skewed to
the left, the majority of the data values fall to right of the mean, skewness < 0. However, when
data is skewed to the right, the majority of the data values fall to the left of the mean, skewness >
0.

Page 61 of 136 CIRSM

 RRM 400

Figure 5.3: Skewness of data

Kurtosis

It is the sharpness of the peak of a frequency-distribution curve. Kurtosis is typically measured

with respect to the normal distribution Kurtosis of a distribution is classified into three categories
namely: Mesokurtic; Leptokurtic; and, Platykurtic (Figure 5.4).

Mesokurtic

A mesokurtic distribution is one with tails shaped in roughly the same way as any normal
distribution. The kurtosis of a mesokurtic distribution is neither high nor low, rather it is considered
to be a baseline for the two other classifications. Binomial distributions for which the p value is
close to ½ are also considered to be mesokurtic. Mesokurtic distributions have excess kurtosis of
zero.

Leptokurtic

It is a distribution with a kurtosis greater than a mesokurtic distribution.

Leptokurtic distributions are sometimes identified by peaks that are thin and tall. The tails of these
distributions, to both the right and the left, are thick and heavy. Leptokurtic distributions are named
by the prefix "lepto" meaning "skinny." One famous example of a leptokurtic distributions is the
Student‘s t distribution. Leptokurtic distributions have positive excess kurtosis >3.

Platykurtic

These are distributions with slender tails. Many times they possess a peak lower than a mesokurtic
distribution. The name of these types of distributions come from the meaning of the prefix "platy"

Page 62 of 136 CIRSM

 RRM 400

meaning "broad." Examples of platykurtic distributions are all uniform distributions and discrete
probability distributions. Platykurtic distributions have negative excess kurtosis<3.

Figure 5.4: Types of Kurtosis distributions

References

Page 63 of 136 CIRSM

 RRM 400

Inferential Statistics

Inferential statistics refers to the analysis of data that helps to describe, show or summarise data in
a meaningful way such that patterns may be acknowledged. However it does not allow us to make
conclusions beyond the available data. Inferential statistics is the main rationale for the extension
of concepts of descriptive statistics to inferential statistics. Inferential statistics techniques allow
researchers to use samples to make generalisations about the populations from which the samples
were drawn.

Two important techniques when drawing inferences from a set of data are:

• Estimation of parameters; and,

• Testing of statistical hypothesis

Drawing inferences from statistics

• A sample is made from a population.

• Parameters that can be calculated for that sample e.g. mean, variance, standard deviation
etc.

• The estimates are used to generalise the whole population taking into consideration some
uncertainties and errors.

• Adjustments are made to obtain better estimates for the population.

• Sometimes you want to check if the postulations you made in your conceptual framework
are correct for a certain population. You also use these samples to check for the accuracy
of your postulations. E.g. an expert may claim that 20% of students at all tertiary institutions
in Zimbabwe are more likely to be thieves.

Page 64 of 136 CIRSM

 RRM 400

• This is called a hypothesis and samples of students may be analysed to validate this
hypothesis.

Modelling assumptions

• When making inferences, four main modelling assumptions are made

• Normality of the data

• Homogeneity of variances

• Linearity

• Independence Normality

of data

• The data used in statistical inference is assumed to come from a normal distribution.
Remember, this is a distribution that is symmetric about the mean.

• This is important when carrying out hypothesis tests. Homogeneity of variances

• Data from different samples obtained from the same population should have the same
variance.

• The Analysis of Variance (ANOVA) depends on this assumption.

Independence

• The random variable should produce independent observations.

• This means there is a choice on choosing the members of the sample since none will depend
on another.

• This is important when carrying out Chi Square tests. Point Estimates
• If a parameter for a population is not known, you may estimate a value for the parameter
by using a sample.

• If a random sample of size n is taken then you have:

• Unbiased estimate for the population mean

Page 65 of 136 CIRSM

 RRM 400

- Unbiased estimate of the population variance

Example
A railway enthusiast simulates train journeys and records the number, x, to the nearest minute,
trains are late according to the schedule being used. A random sample of 50 journeys gave the
following times.
4 3 10 5 2 14
21 9 22 36 14 34
8 15 41 23 13 7
5 34 26 17 8 43
19 5 23 13 12 10

Given that ∑χ=738 and ∑χ2=16 526, calculate to two decimal places, unbiased estimates of the
mean and the variance of the population from which this sample was drawn.

Solution

Interval estimates

• Information obtained from samples only give a rough idea about the population itself.

Page 66 of 136 CIRSM

 RRM 400

• Sometimes we need to know the range of values where we would expect the true value of
the parameter to lie.

• These are called confidence intervals.

• Usually statisticians use a 95% confidence interval for general purposes.
The 95% Confidence Interval

• It shows the range of values where you are 95% confident that the true mean will lie.
• The interval should be central, hence it leaves 5% shared on either side of the interval as
shown in Figure 6.1.
Figure 6.1: The 95% confidence interval

• The 95% confidence interval for the mean is given by:

The 99% Confidence Interval

• It shows the range of values where we are 99% confident that the true mean will lie.
• The interval should be central, hence it leaves 1% shared on either side of the interval.

Page 67 of 136 CIRSM

 RRM 400

• The 99% confidence interval for the mean is given by:

Example
A plant produces steel sheets whose weights are known to be normally distributed with a standard
deviation of 2.4kg. A random sample of 36 sheets had a mean weight of 31.4kg. Find the 99%
confidence interval for the mean.
Solution 1.
Data: n
= 36

= 31.4

The 99% confidence interval can be calculated from:

Solution 2

Page 68 of 136 CIRSM

 RRM 400

The T distribution for the Confidence interval

• If the sample size is small i.e. n<30, and the population standard deviation is unknown, you
use the T-distribution.

• This is a symmetric distribution which has a shape that depends on the degrees of freedom
v.

• The degrees of freedom can be calculated as v = n-1.

• If the standard deviation is not known, and the sample size is small, we use the parameter
T ~t(n-1).

The % confidence interval can be obtained from the formula:

Table 6.1 shows the values that can be obtained from the T distribution table.

Page 69 of 136 CIRSM

 RRM 400

Table 6.1: Values that can be obtained from the T distribution table

Finding the t value:

To find the t value to use in the formula, we find the degrees of freedom (v) from the formula: n=
n-1; and the value of p from the formula.

Where, is the interval required.

Example 3
The mass, in grams, of a packet of biscuits of a particular brand, follows a normal distribution with
mean μ. Ten packets of biscuits are chosen at random and their masses noted. The results, in grams,
are:
397.3; 399.6; 401.0; 392.9; 396.8; 400.0; 397.6; 392.1; 400.8; 400.6.

These can be summarized as follows: ∑x = 3978.8, ∑x2 – 1 583 098, 3.

Calculate a 95% confidence interval for μ.

Solution Stage
1.

Page 70 of 136 CIRSM

 RRM 400

Stage 2.
Calculate also the degrees of freedom:
v = n-1
=10 -1
=9

The critical value of t from the table is t = 2.262

Stage 3.

Page 71 of 136 CIRSM

 RRM 400

Hypothesis Testing
This is a systematic way of validating an assumption or idea based on expertise or postulated value.
The modelling assumption of normality is used to carry out all hypothesis tests.
Steps in testing for a hypothesis.
1. Defining a variable
The variable is the property whose values are being analysed e.g. mass of a tomato, length of a pod
e.t.c. There are two types of hypothesis: a). Null hypothesis
This is a hypothesis which assumes that no change has occurred from the assumption made. It is
expressed as:

:
,

Where, is the postulated population mean.

b). Alternative hypothesis
This is a hypothesis that suggests a change in the system. It is modelled in three main ways:

a).

• This is a one tailed test which suggests that the mean has increased.

We reject the null hypothesis if the value is in greater than the critical value.

b).

Page 72 of 136 CIRSM

 RRM 400

• We reject the null hypothesis if the value of the test statistic is less than the critical value.
c)

We reject the null hypothesis if the test statistic is in one of the critical regions.

2. Define the distribution for hypothesis testing

You state the distribution that will be used to compute the test statistic.

• If the variance is known, or the sample size is large(>30), you define the distribution of the
mean as

• If the variance is unknown and the sample size is small(<30), we define the distribution of
the mean using a t distribution

Page 73 of 136 CIRSM

 RRM 400

3. Compute the test statistic

• The test statistic Z(for a large sample) or T (for a small sample) can be calculated from the
formula:

Or

4. Criteria for rejection

• Define the region where the null hypothesis will be rejected depending on the distribution
being used.
5. Conclusion

• If the test statistic is in the critical region, reject the null hypothesis and accept the
alternative hypothesis.

• Otherwise you accept the null hypothesis and conclude that the conditions expected have
not changed.
Summary of hypothesis (slide 55)
One-tailed test One-tailed test Two-tailed test
(lower tail) (upper tail)
H0: µ= µo H0: µ= µo H0: µ= µo H1: µ≠ µo
H1: µ< µo H1: µ> µo

10% significance level Reject H0 if z<- Reject H0 if z>1.282 Reject H0 if z<1.645

1.282 Or

z>1.645

5% significance level Reject H0 if z<- Reject H0 if z>1.645

1.645
Or z>1.96
(written |z|>1.96)

Page 74 of 136 CIRSM

 RRM 400

1% significance level Reject H0 if z<- Reject H0 if z>2.326 Reject H0 if z<2.576

2.326 Or
z>2.576
(written
|z|>2.576)

Example of hypothesis testing

Each year trainees throughout the country sit for a test. Over a period of time it has been established
that the marks can be modelled by a normal distribution with mean 70 and standard deviation 6.
This year it was thought that trainees from a particular country did not perform as well as expected.
The marks of a random sample of 25 trainees from the country were scrutinized and found that
their mean mark was 67.3.
Does this provide evidence, at the significance level, that trainees from the country did not perform
well as expected?

Solution
Let X be the mark of a trainee from the particular county
𝐻_0: 𝜇= 70
𝐻_1: 𝜇<70
One tailed test in the lower tail at the 5% significance level.
If 𝐻_0 is true, 𝑋 ~ 𝑁 (70 ,6^2/25)
Test statistic 𝑍= (𝑥 𝜇)/√(𝜎^2/𝑛)

Reject 𝐻_0 if z < -1.645 (from table above)

Conclusion
Since the test statistic Z=" "2.25< -1.645, we reject 𝐻_0 and conclude that at the 5% significance
level, the trainees did not perform as well as expected (the mean is significantly less than 70)

Page 75 of 136 CIRSM

 RRM 400

Chi square tests

This is a statistical test used to compare two groups of data, each with independent parameters.
There are two types of chi square tests:

• Chi square goodness of fit test; and,

• Chi square test for independence.
Chi square goodness of fit test
This is used to test whether some data can be modelled by a postulated distribution. For example
a researcher may postulate that the green and red colour strands on bean pods are in the ratio 3:2.
o test if this is true, a chi square test is executed by using a test statistic that assesses the differences
between the observed frequencies and expected frequencies.
Chi square test for independence
• This is used to compare two different samples or population under various conditions, to
check whether each condition is affected by the group from which the observations are
made. For example, a researcher may want to check if the opinion of people (in favour or
against) on hanging prisoners is dependent on whether they are males or females. A
contingency table will be used in such a case to check for independence.
Stages in carrying out a chi square test

a) Define the variable under test e.g mass, weight, length e.t.c.
b) Formulate the hypothesis

• : X follows the given distribution

• : X does not follow the given distribution

c) Define the distribution to calculate the expected frequencies from.

d) Compute the test statistic
from the formula:

Where: O is the observed

frequency; and, E is the expected frequency.

e) Create the criteria for rejection

State the values of the test statistic that will be rejected according to the chi square
distribution.

Page 76 of 136 CIRSM

 RRM 400

f) Below is a diagram showing the chi square distribution and some values of the distribution.

Example
The table below shows the number of committed crimes for one day during a particular period of
time.
Day of Monday Tuesday Wednesday Thursday Friday
the week
Number 121 87 87 91 114
of crimes

(a) Find the frequencies expected from the hypothesis that the number of crimes does not
depend on the day of the week.

(b) Test at the 5% significance level whether the differences in the observed and expected
frequencies are significant.
Solution
a). Let X be number of absentees on any day.

Page 77 of 136 CIRSM

 RRM 400

- : The number of absentees does not depend on the day of the week.

- : The number of absentees depends on the day of the week.

- If is true, the frequencies expected are uniformly distributed as shown

below.
Day of Monday Tuesday Wednesday Thursday Friday Total
the week
500

Number of 121 87 87 91 114

crimes

b). Calculate the test statistic using the following table format:

O E (O - E)2
E
121 100 100 4.41 1.69
87 87 100 100 1.69 0.81
91 100 1.96
114

∑O= 500 ∑E= 500 10.56

From the table:

c). Criteria for rejection:

Page 78 of 136 CIRSM

 RRM 400

• Degrees of freedom

d). If the test statistic

> the critical value (9.488 from table), you

e). Since the test statistic

> 9.488, we reject and
reject . conclude that the number
of crimes depend on the
day of the week
(it is not
independent of the day).

Chi square test for

independence

• The steps of
formulating the
hypothesis and defining a
distribution to calculate
the expected frequencies
will differ.

• All other stages will remain the same.

i). Formulate the hypothesis:

Page 79 of 136 CIRSM

 RRM 400

• the observations from the two groups are independent.

• the observations from the two groups are dependent on each other. ii). Define the
distribution for expected frequencies:
A contingency table which contains row totals and column totals is used to calculate the expected
frequencies on the basis that is two observations

are independent, then

c). Calculate the expected frequencies (E):

Page 80 of 136 CIRSM

 RRM 400

Where: Row total is the total in the horizontal line containing the observation;
Column total is the total in the vertical line containing the observation; and Grand total is the total of
all the observations.

Example
Members of a forensic investigation team are investigating whether the weather has an effect on their
evidence.
Their results are as follows:
Outcome Weather

Good Bad Total

Conviction 12 4 16

Acquittal 5 8 13

Further investigation 7 14 21

Total 24 26 50

Formulate a suitable null and alternative hypothesis and use a chi square distribution to test the claim
at the 1% level of significance.
Solution there is independence between the weather and the results

obtained.

there is dependence between the weather and the results obtained.

is true, there is independence and the expected frequencies are calculated
from:
If

Page 81 of 136 CIRSM

 RRM 400

The table below shows the expected frequencies:

Outcome Weather

Good Bad Total

Win 7.8 8.32. 16

Draw 6.24 6.76 13

Lose 10.08 10.92 21

Total 24 26 50

Calculate the test statistic:

𝑑𝑒𝑔𝑟𝑒𝑒𝑠 𝑜𝑓 𝑓𝑟𝑒𝑒𝑑𝑜𝑚= (𝑟𝑜𝑤𝑠 1) (𝑐𝑜𝑙𝑢𝑚𝑛𝑠 1)

= (3 1) (2 1) =2.
The table below shows the calculations from the test statistic.
O E (O - E)2
E
12 7.68 2.43
5 6.24 0.246…
7 4 10.08 0.941… 2.243…
8 8.32 0.227…
14 6.76 6.956…
10.92

Page 82 of 136 CIRSM

 RRM 400

∑O= 50 ∑E= 50 6.956…

• From the table,

Criteria for rejection

• If the test statistic (2), we reject

• From tables of the distribution

Conclusion: Since the tests statistic is less that the critical value, we do not reject 𝑯0 and conclude
that there is independence between the performance of the team and the weather.

Analysis of variance
Analysis of variance (abbreviated as ANOVA) is used to test for when multiple sample cases are
involved.
The ANOVA technique enables us to perform a simultaneous test to draw inferences about whether
samples have been drawn from populations having the same mean.

Page 83 of 136 CIRSM

 RRM 400

The basic principle of ANOVA is to test for differences among the means of the populations by
examining the amount of variation within each of these samples, relative to the amount of variation
between the samples.
Then the said two estimates of population variance are compared with Ftest, wherein you work out.
F = Estimate of population variance based on between samples variance
Estimate of population variance based on within samples variance

Hypothesis for ANOVA

𝐻_0:𝜇_1=𝜇_2=𝜇_3=…=𝜇_𝑛
_1: at least one of the means is significantly different from the others.
There are two types of ANOVA tests
One way ANOVA (comparing means from different classes under one condition)
Two way ANOVA (comparing means from different classes under many conditions)
The F distribution can be used to assess the test statistic calculated as shown above.

Linear regression and Correlation

• Linear regression is the relationship between two or more variables, or the response of one
variable with respect to another.

• Linear correlation is the measure of the strength of the relationship between the variables.
• There are three main types of correlations as shown on the next page:

Page 84 of 136 CIRSM

 RRM 400

Types of correlations

Types of variables in a linear relationship

• The independent variable is the one that you can control, or is the stimuli.
• The dependent variable is the variable which shows the response due to an input of the
independent variable.

• The independent variable is x and is plotted on the horizontal axis.

• The dependent variable is y and is plotted on the vertical axis. Scatter diagram
• This is the diagram showing the values of x plotted against the corresponding y values.
• From a scatter diagram, two important results can be obtained:
• The regression line of y on x (line of best fit); and,
• The correlation coefficient (strength of the linear relationship).

Line of best fit

Page 85 of 136 CIRSM

 RRM 400

Regression line y on x

• This is the equation of the line of best fit given in the form:

Where: a is the y intercept (the value of y where the line cuts the y axis); and, b is the gradient of the
line.
The values of a and b can be calculated from some analysed scatter lot using the method of least squares.

and,

• The linear regression correlation coefficient r is the value between -1 and 1 which shows the
linear degree of scatter.

Page 86 of 136 CIRSM

 RRM 400

r = 1 indicates positive linear correlation r = -1

indicates negative linear correlation r = 0 indicates
no relationship
The nearer the value of r is to 1 or -1, the closer on the scatter diagram are to the regression line.

Correlation on scatter plot

Calculating r
The value of r can be obtained from the formula:

Interpreting r

• If 0<r<0.5 there is weak positive correlation

• If 0.5<r<1.0 there is strong positive correlation
• If -0.5<r<0 there is weak negative correlation
• If -1.0<r<-0.5 there is strong negative correlation
Example

The following data represents the lengths (χ) and breadths (y) of twelve cuckoos‘ eggs measured in
millimeters (mm).
Page 87 of 136 CIRSM
 RRM 400

χ 22.3 23.6 24.2 22.6 22.3 22.3 22.1 23.3 22.2 22.2 21.8 23.2
y 16.5 17.1 17.3 17 16.8 16.4 17.2 16.8 16.7 16.2 16.6 16.4

Draw a scatter diagram for the data.

Obtain the least squares regression line of y on x and plot this on the scatter diagram.

Page 88 of 136 CIRSM

 RRM 400

Solution

To find the equation of the regression line y on x, we use the equations to find a and b.

• Obtain

The summary of data therefore is as follows:

x Y x2 y2 Xy

Page 89 of 136 CIRSM

 RRM 400

18 20 42 54 324 400 1764 2916 756

30 40 60 54 900 3600 2916 1080
46 54 62 68 1600 2116 3844 4624 1800 2160
60 80 2916 6400 2852 3672
80 88 66 80 3600 4356 6400 4800
92 100 6400 7744 10 000 5280 7040
8464 9200

∑x=528 ∑y=666 ∑x2=34 464 ∑y2=46 820 ∑xy=38 640

Use the formula for r and get r = 0.86 (strong positive correlation)

References
Crawshaw, J. and Chambers, J.(2001).Concise Course for Advanced Level Statistics(4th edition).
United Kingdom: Nelson Thornes.
Kothari ,C.R. (2004). Research Methodology: Methods and Techiques (7th edition). New Delhi:New
Age International (P) Ltd., Publishers

Page 90 of 136 CIRSM

 RRM 400

QUANTITATIVE DATA ENTRY AND CLEANING

In this chapter, I will explain how you would analyse data using the Statistical Package of Social Scientists
(SPSS).
SPSS was specifically designed to analyse data from the social sciences (as its title indicates) but it is
now widely used by a variety of fields and individuals. With SPSS you can perform highly complex data
manipulation and analysis with simple and straight forward instructions/commands. SPSS can take data
from almost any type of file and use them to generate tabulated reports, e.g. from Microsoft Excel. The
SPSS language consists of selfexplanatory commands e.g. the command to perform regression analysis
is ―Regression‖. Starting SPSS

- For you to be able to access SPSS, its software has to be first installed on to your computer.
- Open the SPSS window by double clicking its icon on the ‗desktop’ of your machine or click
Start, Program, then choose SPSS for windows.

- When SPSS first opens, a dialogue box appears asking, ‘what would you like to do?’ and a
number of options are provided (Figure 7.1);

- To create a new data file click on the box for type in data and then click ok.

Page 91 of 136 CIRSM

 RRM 400

Figure 7.1: SPSS dialogue box which asks you on ‘what would you like to do?’ in SPSS

To open an existing data file check on open an existing data source and then click OK. This will open
the data View spreadsheet (Figure 7.2).

Page 92 of 136 CIRSM

 RRM 400

Figure 7.2: The SPSS data View spreadsheet

SPSS Windows/Interfaces
There are three basic windows in SPSS:
a). Data Editor – this displays the contents of the data file. You can have only one data file open at a
time. The data editor has the following windows:
i). Variable view window – this is the dictionary window where variables are defined, characterised and
specified; and,
ii). Data view window – the window used for entering data in the data editor. b). Output Navigator
– this is where all SPSS results are displayed. Once a statistical procedure is run, an output view
window is created that holds results.
c). Syntax window – this is where the choices selected in the dialogue boxes can be saved. These can
then be edited and executed in future. Such operations will automatically appear upon clicking the paste
functions. Defining variables/Creating a dictionary
A good practice in data entry is defining the variables first before entering the data. Defining a
variable means converting questionnaire language to SPSS language.
- To define a variable:
Click on the Variable view tag at the bottom of the screen (Figure 6.3). Figure 6.3 Variable view tag

Page 93 of 136 CIRSM

 RRM 400

I will then take you step by step in defining each variable in each column. The first step is to enter
the variable name in the appropriate column (name) correctly following the rules:

 Place the arrow in the cell of Row 1 of the NAME column and click.
Then type the variable name.

The characteristics of the chosen variable name.

 - Can be up to eight (8) alphanumeric characters.

 - Must begin with a letter.
 - Hyphens (-), ampersands (&), and space can NOT be used but underscore (_ ) can be used.
Variable Type column

- This refers to the type of data.

- The type automatically appears once the variable name is entered.
- To change the variable type, highlight in the second column and the three period symbols on the
right side of the cell.
Page 94 of 136 CIRSM
 RRM 400

- The variable type dialogue box opens and then choose the appropriate
- There are eight variable types: numeric; comma; dot; scientific notation; date; dollar; custom
currency; and, string.
Width column

- Is the width of actual data entries that is determined by the number of characters contained in
the variable code.

- The default width of numeric variable entries is eight (8). String variables may be up to fifty
(50) characters.

- Employing the up/down-ward arrows (inside the cell) on the right hand side increases or reduces
the width of the variable entries. You may simply type a new number in the cell.
Decimals column

- The number of digits to the right of the decimal place to be displayed for data entries.
- Does not apply to string data.
- Decimal places are only used when dealing with continuous data e.g height, shoe sizes, or
income

- The values can be altered in the same way as for the width
Label column

- This is the label attached to the variable name. It is the broad definition of a variable.
- It is not confined to eight characters like variable name.
- The label remind you of the meanings of variables.
- This variable label is also displayed in the output from statistical analysis.
Values column

- These are labels attached to the category codes. These show the variable categories represented
by the numeric codes.

- Once an interger code is assigned to each response category (coding) the variable automatically
becomes ‘numeric’.
Clicking on the three period symbols under the respective sixth column of the window opens the
value labels dialogue box as shown in Figure 7.4. Figure 7.4: The dialogue box for the value
labels column

Page 95 of 136 CIRSM

 RRM 400

Missing column

- These are missing value codes.

- SPSS recognises a period symbol (...) as indicating a missing value.
- If other codes have been used, (e.g. 99,999) these have to be declared to represent missing values
by highlighting the respective cell in the seventh column.

- This is done by clicking the three period symbol and filling the resulting missing values dialogue
box accordingly.
Column

- This is the width of the variable column.

- The default cell width for numerical variables is eight.
- When width value is larger than the column values, only part of the data entries might be seen
in the data view.
Alignment column

- Refers to the alignment of variable entries.

- By default, SPSS aligns numerical variables to the right and string variables to the left hand side
of the cell. It is generally good to adhere to this default.

- The column can be adjusted in the same way as the width (third column)
Measure column

- It refers to the scaling of the data.

Page 96 of 136 CIRSM
 RRM 400

- Ordinal scale distinguishes variables basing on the element of rank (agree, disagree, neutral)
- Nominal scale distinguishes variables on the basis of a name e.g sex of the respondent, religion
and level of education.

- Scale refers to interval data e.g age groups, ranges of monthly income.

Data entry
You shall enter your data (the responses from the questionnaire) in the Data view window. Note
that if you point your cursor above any one of the columns of the data view you shall see each of
the coded question from the hard copy of your questionnaire. All you need to do is to enter the
respective code (response). In other words, you will be transferring the responses from the hard
copy of your questionnaire to the Data view window of SPSS.

Data cleaning
Data cleaning is the process of ensuring that data has been correctly and accurately entered in
SPSS. It is part of quality control. It is very unusual for real data to arrive and be analysed without
problems and/or errors. In large studies the process of data cleaning consumes considerably more
time and effort than the primary analysis of interest.
Main sources of error

- Wrong codes may have been entered.

- Skip patterns may not have been respected.
- Data may not have been captured.
- Missing data may have been treated as available data or vice versa

The data cleaning process

Run the frequency distributions of all variables in order to identify the following:
d) Outliers;
These are values outside the reasonable range for a variable, for instance, on a study about adolescents
you find age of 40 years.
b) Whether all the questions and responses have been captured; and
- Inspecting questionnaire by questionnaire to ensure that all data has been captured.
c) Whether the skip patterns have been observed.

Page 97 of 136 CIRSM

 RRM 400

QUANTITATIVE DATA PRESENTATION AND ANALYSIS

The most interesting part in research is the analysis, interpretation and presentation of findings. It
is equally important to note that when interpreting your findings you should remain guided by your
objectives and conceptual framework. The purpose of analysing data is to obtain usable and useful
information. The analysis, irrespective of whether the data is qualitative or quantitative, may:
describe and summarise the data; identify relationships between variables; compare variables;
identify the difference between variables; and, forecast outcomes
Quantitative data analysis
Having cleaned your data in SPSS, the next stage is to analyse the data. Remember, the choice of
a statistical procedure depends on what you want to present. For instance, a frequency distribution
shows the number of the respondents who completed a particular question; percentages indicate
the proportion of the respondents who answered in a particular way; the mean is the average
number of a score or age; median shows the middle value in a range; and a cross tabulation shows
the relationship between the dependent and independent variable(s). Three types of quantitative
data analysis
Univariate analysis
This is an analysis of the distribution of cases, however, using one variable at a time, e.g. marital
status. The following are steps you should follow in SPSS if you want to analyse one variable e.g.
marital status.
Click "Analyze" from the toll bar-> "Descriptive Statistics" -> "Frequencies" (Figure 8.1)

Page 98 of 136 CIRSM

 RRM 400

Figure 8.1: Commands in univariate anaylsis

Once you select frequency distributions, a dialogue box appears, Figure 8.2. You shall see that
from the left side of the dialogue box are the coded variables and the right box (which is usually
empty) is where you are supposed to put the selected variable. In the Figure 8.2, I have scrolled
through the variables until I have found ‗marital status‘ as my variable of interest. So I have clicked
the arrow to move that variable into the "Variable(s):" box. Should you wish to command SPSS to
calculate the measures of central tendency, dispersion, distribution and percentile values, select
the Statistics command which is at the top right side of the opened dialogue box. . You will also
realise that there are icons where you can also select the charts, format, style and Bootsrap of the
selected variable if need be.

Page 99 of 136 CIRSM

 RRM 400

Figure 8.2: A dialogue box that would appear upon clicking the ‘frequency distributions’
command

Upon selecting the variable (marital status) you can either click the Paste or
OK command. Note that if you are to click Paste, SPSS takes you to the Syntax window where the
selected option has been saved. If you are to select (highlight) the saved option(s) and click ‗Run‘,
SPSS takes you to the Output window. However, if you are to select the variable and just click
OK, SPSS immediately takes you to the Output Window (Figure 8.3). Remember the Output
window displays the distributions of the selected variable(s).

Figure 8.3: SPSS Out-put window

Page 100 of 136 CIRSM

 RRM 400

Given that I had selected marital status as my variable of interest, the Table 8.1 will be displayed
in the SPSS Output window. Note that there are two tables in this output. The first breaks down
the frequency of valid region data versus missing. The second table shows frequency distribution
with four labeled columns.

Table 8.1: SPSS Output window Frequencies Statistics

(q101) Marital status

N Valid 500
Missin g
0

Page 101 of 136 CIRSM

 RRM 400

(q101) Marital status

Valid Cumulative
Percent Percent
Frequency Percent
Vali Single/Never d 38.8
married 194 38.8 38.8
Married/cohabi 93.0
ting
54.2 54.2
Divorced/separ ated 271
97.0
Widowed
Total 100.0
20 4.0 4.0

15 3.0 3.0

500 100.0 100.0

What is important however is for you to be able to interpret the meaning of the figures in the SPSS
Table 8.1 above. You will note that the first column of the table is showing the coded marital
categories (never married, married/cohabiting, divorced/separated and widowed) and ‘Total’
(representing the sample size). The second column, ‗Frequency,’ shows the frequency distribution
of the respondents in the various marital categories and the sample size of 500. The third column,
‗Percent’, depicts the proportion of the respondents in each marital category to the entire sample
size. Note that the proportion adds up to 100 %. The fourth column, ‘Valid Percent,’ is the percent
when missing data are excluded from the calculations. In other words, these are the percentages of
the respondents who selected each response after you disregarded missing responses. Use the valid
percent column when analyzing your data because it shows the frequency distributions after
disregarding any missing responses. The fourth column, "Cumulative Percent", adds the
percentages of each of the proportions in the third ‗Column‘ from the top of the table to the bottom,
culminating in 100%. This is more useful when the variable of analysis is ranked or ordinal, as it
makes it easy to get a sense of what percentage of cases fall below each rank.

You can analyse and interpret the meaning of the figures displayed from the SPSS output window
using your own table or graph. Note that a table or graph (and not both) should have the following:
title (which should reflect the meaning of the figures) and the sample size (n, for a representative
sample and N for a national study). Make sure you draw in Excel your graphs, e.g. pie chart,
column, line or bar graphs for quality presentation of results. For column, line and bar graphs,
make sure you label the X and Y axis, insert data labels, remove grid lines and format the Y axis
such that the percentage axis is up to 100. Note that percentage distributions in a pie-chart should
add up to 100.
Page 102 of 136 CIRSM
 RRM 400

When analyzing and interpreting data, you should start with the text. Within the text, make
reference to the table or figure. In your text, the first statement should be a general statement which
points to the general observation(s) of the frequency distributions in the table or graph. The general
statement should then be substantiated with the presentation of the frequency distributions in the
table. The first time you present the first frequency distribution, you should make reference to your
table or graph in that particular sentence. In presenting the frequency distributions, words which
show the magnitude of the proportion, such as the majority, a significant proportion, a sizeable
proportion, a small proportion should be used.
The following is an example of univariate analysis using data in Table 8.2.
Table 8.2: SPSS Output window showing the distribution of the respondents’ level of
education Statistics (q100) Level of education
N Valid 500
0
Missing

(q100) Level of education

Valid
Perce
Frequenc nt Cumulative
y Percent Percent
Vali Never been d to
school 4 0.8 0.8 0.8
Primary 80 16.0 16.0 16.8 87.2
94.6
Secondary 352 70.4 70.4 100.0

A Level 37 7.4 7.4

Tertiary 27 5.4 5.4
Total 500 100.0 100.0

The sample was highly educated. (Note that this is my general observation after noting that only
1% of the sample had never been to school). The majority of the respondents, 99%, reported that
they had attained some form of education. Over two-thirds of the respondents, 70%, reported that
they had attained secondary education (Table 8.3). (Note that if you had drawn a graph then you
say Figure 1). A sizeable proportion of the respondents, 16%, reported that had completed primary
school. Small proportions of the respondents, 7% and 5%, reported that they had completed

Page 103 of 136 CIRSM

 RRM 400

secondary and tertiary education, respectively. Only an insignificant proportion of the respondents,
1%, reported that they had never been to school.

Table 8.3: Percentage distribution of the respondents’ level of education

Level of Education Percentage
Never been to school 0.8
Primary 16.0
Secondary 70.4
A‘ Level 7.4
Tertiary 5.4
Total 100.0
n = 500
OR you can make use of a figure - pie chart

Figure 8.4: Percentage distribution of the respondents’ level of education

n = 500
OR a column graph

Page 104 of 136 CIRSM

 RRM 400

Figure 8.4: Percentage distribution of the respondents’ level of education

n = 500

Analysing a Multiple response question

Remember when the multiple response question was defined during the creation of a dictionary in
SPSS, each response was created as a separate variable. Subsequently, separate responses for each
defined variable were entered during data entry. The SPSS commands for analyzing a multiple
response question are similar to the ones I highlighted above. However, what is now different is
that you select all the coded variables of the multiple response question at once and then click the
OK or Paste command. Below are Tables (8.4) from the SPSS Output after selecting the variables
of a multiple response question.

Table 8.4: The SPSS output for a multiple response question

Q301. Forms of gender based violence which you have ever heard of in Gweru, Zimbabwe.

Page 105 of 136 CIRSM

 RRM 400

Statistics
(q301.6)
Voluntari
ly
accompa
nying (q301.7)
a sexual Agreeing to
partner assist with
(q301.2) (q301.4)
(q301.Physi Forced (q301.3) Willful visit (q301.5) Rape domestic
cal abuse marriage Spousal rape to in-laws by a stranger chores

500 500
NValid 500 500 500 500 500
0
0
Missin g 0 0 0 0 0

Frequency Table

(q301.1) Forms of gender based violence which you have ever heard of_
Physical abuse
Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 498 2 99.6 99.6 99.6

No
500 .4 .4 100.0
Tota
l
100.0 100.0

(q301.2) Forced marriage

Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 423 84.6 84.6 84.6

No 77
15.4 15.4 100.0

Page 106 of 136 CIRSM

 RRM 400

Tota
l 500 100.0 100.0

(q301.3) Spousal rape

Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 411 89 82.2 82.2 17.8

No 17.8 17.8 100.0
500
Tota
l
100.0 100.0

(q301.4) Willful visit to in-laws

Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 96 19.2 19.2 19.2

No 404
80.8 80.8 100.0
Tota 500
l
100.0 100.0

(q301.5) Rape by a stranger

Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 420 80 84.0 84.0 84.0

No
500 16.0 16.0 100.0
Tota
l
100.0 100.0

Page 107 of 136 CIRSM

 RRM 400

(q301.6) Voluntarily accompanying a sexual partner

Freque Perce Valid Cumulativ e
ncy nt Percent Percent

Vali Yes d 75 15.0 15.0 15.0

No 425 85.0 85.0 100.0
Tota 500 100.0 100.0
l

(q301.7) Agreeing to assist with domestic chores

Frequency Percent Valid Cumulative
Percent Percent

Vali Yes d 86 17.2 17.2 17.2

No 100.0
414 82.8 82.8
Total
500
100.0 100.0

When analyzing the multiple response question, you draw one table or graph to show the frequency
distributions of all the variables. From each table of the SPSS output, you only select the frequency
distribution of your interest. For instance, in the above SPSS Output, one is interested in knowing
the respondents‘ knowledge about the forms of gender based violence. Thus, you should select the
percentage distributions for those who indicated a
Yes‘response in each variable. Note that the presentation technique is similar to the one I have
highlighted when I gave my example of the univariate analysis of the variable ‗Level of education.
What is however different is that the total percent of a multiple response question does not add up
to 100. Remember, you would have chosen the respondents who would have answered ‗Yes‘ to
each of the variables. I would use the above SPSS Output to show you how a data from a multiple
response question is presented and analysed.

Page 108 of 136 CIRSM

 RRM 400

Example
Awareness of the various forms of gender based violence was high although misconceptions were
common. (Note that this is my observation after considering the frequency distributions of
knowledge of the various forms of gender-based violence in Table 8.5.) Physical abuse was the
most commonly known type of gender-based violence as reported by 100% of the respondents
(Table 8.5). Force marriage, as a form of gender-based violence, was reported by 85% of the
respondents. Rape by a stranger was reported by 84% of the respondents. About 82% of the
respondents reported spousal rape. As noted earlier, the study noted misconceptions regarding the
forms of gender-based violence in Gweru. A sizeable proportion of the respondents, 19%, reported
that willful visit to in-laws is a form of genderbased violence. Similarly, 17% of the respondents
reported agreeing to assist with domestic chores. Further, a reasonable proportion of the
respondents, 15%, reported voluntarily accompanying a sexual partner as a form of gender-based
violence.

Table 8.5: Percentage distribution of the respondents’ knowledge about the various forms of
gender-based violence
Form of gender based violence Percentage
Physical abuse 99.6
Forced marriage 85.6
Spousal rape 82.2
Rape by a stranger 84.0
Voluntarily accompanying a sexual partner 15.0

Willful visit to in-laws 19.2

Agreeing to assist with domestic chores 17.2

n = 500

Bivariate analysis
It is used when you are interested in establishing the statistical and direction of relationship
between the dependent and independent variable(s). The following are the SPSS procedures for
bivariate analysis.
Click "Analyze" -> "Descriptive Statistics" -> "Frequencies"-> ―Crosstabs‖ (Figure 8.5).

Page 109 of 136 CIRSM

 RRM 400

Figure 8.5: SPSS commands for bivariate analysis

Once you click the above command, a dialogue box (Figure 8.5) opens. To the left side of the
dialogue box are the defined variable, and to the right are options ‗Row‘, ‗Column‘, ‗Statistics‘,
‗Cells‘, ‗Format‘.

Page 110 of 136 CIRSM

 RRM 400

Figure 8.6: A dialogue box that would open after clicking the crosstab command

Use the upper and lower arrows to select the dependent and independent variables, respectively.
If you are to click ‗Cells‘, note that another dialogue box opens (Figure 8.7). This box has options
to select the ‘Counts’,
‘Percentages’, and ‗Residuals‘. Then click, ‗Observed‘ counts, and all the ‘Row, Colunm and Total’
percentages and then click ‘Continue’.

Page 111 of 136 CIRSM

 RRM 400

Figure 8.7: A dialogue box that would open after clicking the ‘Cells’ command

If you are to click ‗Statistics‘, SPSS takes to another dialogue box (Figure 8.8) which has various
options of statistical analysis. You select the statistical analyse which best suit your interest then
click ‘Continue’.
Figure 8.8: A dialogue box that would open after selecitng the ‘Statisctics’ command

Once you are satisfied with your selected option, click either Paste or OK. Remember, if you click
Paste, SPSS takes you to the Synax Window where the selected options are saved. Highlight the

Page 112 of 136 CIRSM

 RRM 400

saved selected options and click ‘Run’. This command will take you to the SPSS Output Window
(Table
8.6). You can also get to the SPSS Output Window by just clicking ‗OK‘ after selecting all your
options. What is important to note is that you are interested in finding out the extent of committing
any crime, however, by marital status. So you can see the number of males who committed any
crime by marital status, the ‗Yes‘ row (the row wise interpretation). You can note that 226 males
ever committed a crime out of the 401 males. However, among the males who ever committed a
crime, 71 had never married/single, 142 were married/cohabiting, 7 were divorced/separated,
while 6 were widowed.
The second row is showing the proportions (in percentages) of males by marital status who ever
committed a crime. For instance for the never married = 71/226 *100 =31.4%. Note that the
denominator is 226 (representing all males who ever committed a crime) and also that all the
proportions add up to 100%. So basically the figures are showing the distributions of all males who
ever committed a crime by marital status. It is equally important to also bear in mind that the
proportion of male respondents who ever committed a crime to the entire sample is 56.4%
(226/401)*100.
The third row is now showing the proportions of males who ever committed a crime within a
particular marital group (the column wise interpretation). For instance, all single/ never married
males were 101 [71 who said yes, + 30 who said No]. So within the never married males alone,
70.3% (71/101*100) reported that they had ever committed a crime. Similarly, within the
married/cohabiting males alone who constituted 268 respondents of the sample 53.0%
(142/268*100), reported that they had ever committed a crime. You can use the two examples to
interpret the proportions within the divorced and widowed marital groups. What you also need to
understand is that all column percentages add up to 100. This is largely because if we say 53.0 of
the never married males reported that they ever committed a crime, then 47%, as shown in the
‗No‘ column reported that they had never committed a crime. However, you are comparing the
commitment of a crime within marital groups, note that the sample size is now 401.
The last table of the cross tabulation shows the results of your statistical tests. It should be noted
that any association (correlation) whether positive or negative can be statistically significant if the
chi-square p-value is less than or equal to 0.05. This is largely because you would have used a 95%
confidence interval (error margin). So the SPSS cross tabulation below, the p- value is 0.006. This
shows that the association between committing a criminal activity is associated with marital status.
Unlike the presentation of the univariate analysis, you have to state the strength of the association
when analyzing and presenting findings from a cross tabulation.
Table 8.6: SPSS Tables of a cross tabulation Crosstabs Case Processing Summary
Cases

Valid Missing

Page 113 of 136 CIRSM

 RRM 400

Perce Perce Total Perce

nt nt nt

N N N
(q402a) Have you ever
contracted an STD? *
100.0
(q101) Marital status 401 80.2% 99 19.8% 500
%

Page 114 of 136 CIRSM

 RRM 400

(q402a) Have you ever committed any crime? * (q101) Marital status Cross tabulat

(q101) Marital status

Single/Ne Married/ Divorce

ver married cohabitin d/separ
g ated Widow
ed Total
(q402a) Have you Yes Count ever committed any 71 142 7 6 226

62.8%
crime? % within 3.1% 2.7% %
(q402a) Have you 100.0
ever committed any 31.4%
crime?

53.0% 56.4%
% within (q101) 38.9% 42.9%
Marital status 70.3% 56.4%
35.4%

% of Total 17.7% 126 1.7% 1.5% 175

No
Count 30 72.0% 11 8 %

100.0 %
% within 6.3% 4.6%
(q402a) Have you
ever 17.1%
committed any
crime %

Page 115 of 136 CIRSM

 RRM 400

% within (q101) 47.0% 61.1% 57.1%

Marital status 29.7% 43.6%

% of Total 31.4%
7.5% 2.7% 2.0% 43.6%

Total Count 101 268 18 14 401

% within 66.8% 4.5%

(q402a) Have you
ever committed any 3.5% 100.0
crime 25.2%

% within (q101)
Marital status
100.0% 100.0% 100.0
100.0%
% of Total
100.0%
100.0
66.8%
25.2% 4.5% 3.5%

Asymptoti Chi-Square Tests

c
Significan
ce
(2sided
Value Df )
Pearson Chi- 12.489
Square a 3 .006

Likelihood Ratio 12.769 3 .005

Linear-by-Linear
10.683
Association
1 .001
N of Valid Cases 401 a. 0 cells
(0.0%)
have
expected count less than 5. The minimum expected count is 6.11.

Consistent with my 2 different interpretations of the figures from the SPSS output above (the
column and row interpretations which depend on how you would have selected your dependent
and independent variables), there are also 2 different ways of presenting and interpreting the
findings from a cross tabulation. Remember the techniques of starting with a general statement,

Page 116 of 136 CIRSM

 RRM 400

present the percentage distributions to substantiate your assertion, state the direction of the
relationship (inverse, positive and the group which is either more or less likely to respond in a
particular way when compared to other categories), making reference to the table or figure and
indicating the sample size (n).

Example
1. When drawing comparison within males who ever committed a crime.

Criminal activities are common in Gweru. The majority of the respondents, 56%, reported that
they had ever committed a crime (Figure 8.9). Figure 8.9: Percentage distribution of the
respondents who ever committed a crime

n = 226; p = 0.006
Further analysis demonstrated that committing a criminal activity is associated with marital status
(p = 0.006). Married respondents were more likely to report that they had ever committed a crime
when compared to all marital groups. For instance, while, 63% of the married respondents reported
that they had ever committed a crime, only 3% of the widowed respondents reported the same
(Table 8.7).

Table 8.7: Percentage distribution of the respondents who ever committed a crime by marital
status
Marital status Percentage
Never married 31.4
Married/cohabiting 62.8
Divorced/separated 3.1
Widowed 2.7

Page 117 of 136 CIRSM

 RRM 400

Total 100.0
n = 226
N.B. From the first presentation and analysis, it seems as if married respondents are more likely
to commit criminal activities when compared to the rest of the marital groups. However, note that
the first presentation could be biased given that the cross tabulation output shows that married
respondents dominated the sample. They constituted 66.8% of the entire sample. Thus, when
interpreting data, one should be mindful of the aforementioned observation. This is primarily
because if you are to design policies and programmes aimed at curbing criminal activities and use
the first approach to analyse your data, you might end up designing programme targeting married
men. Yet in actual fact you should also remember that married men constituted more than two
thirds of the sample size.

2. When drawing comparison of males who ever committed a crime within the different marital
groups.

Criminal activities are common in Mutare. The majority of the respondents, 56%, reported that
they had ever committed a crime (Table 8.8). Further analysis also demonstrated that committing
a crime is associated with marital status (p=0.006). Never married respondents were more likely
to report that they had ever committed a crime when compared to all marital groups. For instance,
while 70% of the never married respondents reported that they had ever committed a crime, only
39% of the divorced respondents reported the same.

Table 8.8: Percentage distribution of the respondents who ever committed a crime by marital
status
Marital status Percentage P value
Never married 70.3 0.006
Married/cohabiting 53.0
Divorced/separated 38.9
Widowed 42.9
Total 56.4

Page 118 of 136 CIRSM

 RRM 400

n = 401
N.B. The second presentation and analysis of the cross tabulation seem to be more realistic. This
is primarily because one is comparing the commitment of criminal activities using respondents
within their marital groups. In this case, note that never married males were more likely to commit
criminal activities when compared to other marital groups.

Page 119 of 136 CIRSM

 RRM 400

Triangulation

Triangulation involves checking the validity of findings and or/observational data themselves by cross-
checking (double or triple checking) with other sources of data (O‘Donoghue and Punch, 2003). The
technique is meant to: increase the credibility and validity of results; cross exam data from multiple
sources to search for regularities in research data (verification); give a more detailed and balanced picture
of the situation (Altrichter et al, 1996); avoid intrinsic biases and the problems that come from single
method, single observer, single-theory studies (Alexander, 2001); and, map out or explain more fully
the richness and complexity of human behaviour by studying it from one or more stand points (Cohen
and Manion, 1986). Triangulation is also used when building a case using different approaches. For
instance, detectives always use triangulation successful in establishing and corroborating findings with
little, yet elaborate instrumentation. They often use a modus operandi approach which consists largely
of triangulating independent indices. A detective would amass fingerprints, hair samples, alibis and
eyewitness accounts to identify a culprit. Note that such an approach produces a case which presumably
fits one suspect far better than others. Also note that all signs presumably point to one similar conclusion.
Further there are different kinds of measurements which provide repeated identification.

Goal of triangulation
Alexander (2001) asserts that the purpose of triangulation in specific contexts is to obtain confirmation
of findings through convergence of different perspectives seen to represent reality. However, ordinarily,
some researchers share a misconception that the goal of triangulation is to arrive at inconsistencies across
data sources or approaches. They argue that such inconsistencies may be likely given the relative
strengths and weaknesses of the different approaches. However, Patton (2002) argues that the
inconsistencies should not be seen as weakening the evidence, but should be viewed as an opportunity
to uncover the deeper meaning of the data.

Types of triangulation in research

1. Data Triangulation
This is the use of different sources of information to increase validity of a study. For instance,
in a study of armed robbery, a researcher can collect data from the robber‘s friends, parents,
neighbors, school, and other relevant stakeholders.
2. Investigator Triangulation
The researcher would be using different investigators in the analysis process. This approach is
typically common when an evaluation team consisting of colleagues within a field of study
where each investigator examines the programme within the same qualitative method. Findings
from each investigator would then be compared to develop broader and deeper understanding
Page 120 of 136 CIRSM
 RRM 400

of how the different investigators view the issue. If the findings from the different investigators
arrive at the same conclusion, then, the team‘s confidence in the findings would be heightened.

3. Theory Triangulation
This is a situation of pitting alternative theories against the same body of data. The main
idea will be to avoid the risk of researchers reaching theoretical conclusions, selecting only
those data which suit their pet views or developing small scale theory which has little
relevance beyond the immediate situation.
To triangulate theory the research should:
• Draw up a list of propositions which might explain or have relevance to the research
problem to be investigated;
• Identify the different ways each proposition might be interpreted;
• Carry out the research to ascertain the propositions which hold water;
• Discard those propositions which prove untenable;
• Carry out further research to identify the most likely interpretations still in the ring;
• Review the propositions which passed and failed the empirical test, along with their
parent theory; and,
• Arrive at the new theoretical understanding of the problem.

Depending on the outcome of the empirical work, this might include insights from a number of
apparently conflicting theories.

4. Environmental Triangulation
It involves the use of different locations, settings and other key factors related to the environment
in which the study took place. These factors could be time, day and season. The key to
environmental triangulation is to identify the environmental factors, if any, that might influence
the information that is received during the study. These environmental factors could be changed to
see if the findings are the same across settings. If the findings remain the same under varying
environmental conditions, then the validity has been established.
5. Methodological Triangulation
It involves the use of multiple qualitative and/or quantitative research methods during research or
to study the programme. Note that a study can be purely qualitative. However, within the
qualitative study, you can still triangulate your qualitative research methods to check for the
validity of the findings and also similarities and differences among the emerging issues from the
data. You can also triangulate quantitative and qualitative research methods. In this scenario, you
would use a survey to quantify the magnitude of your variables. Nonetheless, you would also use
the qualitative research methods such as the key informant interviews, FGDs, documentary review
and in-depth interviews to explain, describe and explore the factors underlying the quantified
variables. If the conclusions from each of the methods are the same, then validity is established.

Page 121 of 136 CIRSM

 RRM 400

When writing the methodology of Methodological Triangulation, the following sub-headings must be
explicitly stated:

Methodology for triangulated data collection methods a). Study area (Justify the
selection of the area)

b). Target population (Justify the selection of the target area)

c). Study Design

In your research design indicate that the study triangulated quantitative and qualitative data
collections methods. Briefly, state the methods and explain why each method was used.

d). Data collection methods (present each method write up to the selection of the respondent.
Note my examples in the preceding chapters). E.g
1. Survey
2. Key informant interviews
3. Focus group discussions
4. In-depth interviews
N.B. If you have used qualitative methods only, then present those up to the selection of the respondent.

e). Data Management & Analysis

f). Ethical Issues

Presenting findings from Methodologically Triangulated data collection methods

You should present the quantitative findings first and then substantiate with findings from the
qualitative research methods.

Below is an example of presenting findings of triangulated data on the consequences of gender-

based violence.

The study revealed that physically abused women are most commonly susceptible to experiencing

reproductive health problems (R.H). The respondents universally reported that they experienced at

least one R.H problem as a result of physical abuse. However, excessive bleeding was reported by

the majority, 43% of the respondents (Table 8.1). A significant proportion of the respondents, 20%,

reported that they had suffered from a miscarriage. Fourteen percent reported that they had suffered

Page 122 of 136 CIRSM

 RRM 400

a premature birth. A still birth and broken wounds and bones were reported by 13% and 10% of

the respondents, respectively.

Table 8.1: Percentage distribution of respondents who suffered RH problems as a result of

physical abuse.
Reproductive health problem Percentage

Miscarriages 20.0

Excessive bleeding 43.0

Still birth 13.0

Premature labour 14.0

Broken bones and other wounds 10.0

Total 100.0

n = 126
Qualitative data supported that physical abuse has adverse impact on women‘s reproductive health.

Miscarriages were commonly reported by the respondents during in-depth interviews and FGDs.

Note this remark from a woman aged 32 years:

Five years ago I experienced a miscarriage that almost took my life when my husband

attacked me with booted feet and clenched fists on my stomach whilst I was pregnant. I

ended up having a miscarriage.

This was a terrible incident in my life.

Data from FGDs also supported that miscarriages are typically common among the physically

abused women. Note this remark during a FGD of women aged 30-39 years, which received

support from fellow group discussants:

Page 123 of 136 CIRSM

 RRM 400

Pregnant women who are always suffer from severe physical abuse from their husband in

this community normally end up having miscarriages. These miscarriages are mainly a

result of excessive bleeding.

Deaths were also typical reported as consequences of physical abuse among women. During a

FGD of women between 40 and 49 years of age, one respondents, reported that some physically

abused women would die upon arrival to the hospital. Note her remark, which received some agony

from the fellow group discussants:

I have a neighbour who was pronounced dead on arrival to hospital due to excessive

bleeding. The bleeding was a result of physical abuse whilst she was pregnant.

Page 124 of 136 CIRSM

 RRM 400

Applying Research in the Risk and Security Environment

Final Research dissertation

Now that you are aware of what research entails, the qualitative and quantitative research
methodologies, presentation and analysis of both qualitative and quantitative data, you should be
able to carry out a study and write a dissertation. Below is an outline of the structure of your
dissertation. Note that each chapter starts with an Introduction. However, the introduction sections
of Chapters 2 to 5 will be basically summaries of issues to be covered by each of the respective
chapters.
Chapter I The introductory chapter

• Introduction (overview of the subject area)

• Background of the study (Backdrop to the problem under investigation)
• Statement of the Problem
• Justification of the Study
• Objectives
• Methodology (target population, study area and methods in brief) • Order of presentation

Chapter 2 – Literature Review

a) Literature Review
As you write your literature review, remember you are reviewing empirical and theoretical
literature up to the most recent and relevant. In addition, you should be responding to your
objectives. Literature review is not a mere presentation of data from various sources. However,
it is a synthesis of literature while at the same time explicitly indicating the gap(s) to be filled by
your current study.

b) Conceptual framework
- This should be presented at the end of the literature review chapter.
- Remember, you will be explaining the proposed relationship(s) between your dependent and
independent variables.

Page 125 of 136 CIRSM

 RRM 400

Chapter 3 – Methodology
a) Introduction

When presenting your methodology, make sure the following sunheadings are clearly
articulated:

b) Study area
c) Target population
d) Study Design
- Indicate whether or not you have triangulated qualitative and/or quantitative research methods.
Brief state the method highlighting why you have used each of the method.

e) Data collection Methods & Tools

• Remember to show the difference between a method and a tool.
• Justify the selection of each method (linking the method to your objectives)
• Describe each method right up to the selection of the respondent(s).
f) Data Management & Analysis
g) Ethical Issues

Chapter IV – Presentation of Findings

a) Introduction

b) Background characteristics
• Percentage distribution of the demographic and socio-economic characteristics of the
respondents
c) Respective sub-sections responding to the study objectives
• Remember when presenting your findings, you are responding to your objectives, while at the
same time being guided by your conceptual framework.

• If presenting both quantitative and qualitative data, they must be presented in a complementary
fashion not as separate entities.

Chapter V – Summary of findings, Discussion and Recommendations a) Introduction

b) Discussion
• You start by giving a brief summary of the key findings for each objective. By so doing, you
will be responding to each of your research questions.
• Compare your findings vis-a- visa your objectives, conceptual framework and literature review.

Page 126 of 136 CIRSM

 RRM 400

• State the implications of your key findings. In other words, you will be answering the question
‗So what?‖

c) Conclusions
• It is like the last chord of a song.
• Restate the objectives of your dissertation.
• Briefly summarise the main points of your findings.
• Remember, recommendations will be derived from the presented summary of findings.

d) Recommendations
• These are suggestions of the best course of action given the findings from your research.
• State what you propose to be done and the relevant implementer(s).

e) Further research
• Indicate the issues you have noted from your study which however you feel should be researched
further.
• You will be highlighting important knowledge gaps which you will have identified from your
study that call for further research.

REFERENCING
You acknowledge ALL the sources which you will have cited in dissertation. Remember, there are
different referencing styles. However, make sure you stick to one referencing style recommended
by your institution. As for CIRSM, use the Harvard Style of Referencing.

Page 127 of 136 CIRSM

 RRM 400

References

Blaxter L., Hughes C. & Tight M. (2001) How to research. Buckingham: Open University.

3. Merriam S. (1988) Case study research in education: a qualitative approach. San Francisco,
CA:Jossey-Bass.

5. Rudestam K. & Newton R. (1992) Surviving your dissertation. London:Sage.

6. Taylor D. & Procter M. (2008) The literature review: a few tips of conducting it. Health
Services Writing Centre:University of Toronto. http://www.utoronto.ca/writing/litrev.html

7. Wellington J., Bathmaker A., Hunt C., McCulloch G., & Sikes P. (2005) Succeeding with your
doctorate. London:Sage

Chinn P.L and Kramer M.K. (1999) Theory and Nursing: A Systematic Approach. 5th Edition. St Louis,
USA: Mosby.

Imenda S. (2014) Is There a Conceptual Difference between Theoretical and Conceptual Frameworks?
Journal of Social Science, 38(2): 185-195 Liehr P. and Smith M.J. (1999) Middle range theory:
Spinning research and practice to create knowledge for the new millennium. Advances in Nursing
Science, 21(4): 81-91.

Shi L. and D.A. Singh. (2011) The Nation's Health, Eighth Edition. Jones & Bartlett Learning: Sudbury,
MA

Sinclair M. (2007) Editorial: A guide to understanding theoretical and conceptual frameworks. Evidence
Based Midwifery 5(2): 39

Ajzen and Fishbein (1980) Understanding Attitudes and Predicting Social Behaviour, p8

Miles, MB. & Huberman, AM. (1994). Qualitative Data Analysis. (2nd edition). Thousand Oaks, CA:
Sage Publications.

Clandinin and Connelly,2000:18

Jean.J Schensul, 2005: Introduction to Ethnographic Research.
Kramp, 2004:104

Page 128 of 136 CIRSM

 RRM 400

Berg, B. L., & Lune, H. (2004). Qualitative research methods for the social sciences (Vol. 5). Boston:
Pearson.
Coyne, I. T. (1997). Sampling in qualitative research. Purposeful and theoretical sampling; merging or
clear boundaries?. Journal of advanced nursing, 26(3), 623-630.
Malterud, K. (2001). Qualitative research: standards, challenges, and guidelines. The lancet, 358(9280),
483-488.
Marshall, M. N. (1996). Sampling for qualitative research. Family practice, 13(6), 522-526.
Patton, M. Q. (2005). Qualitative research. John Wiley & Sons, Ltd.
Sandelowski, Margarete. "Sample size in qualitative research." Research in nursing & health2 (1995):
179-183.
Strauss, A., & Corbin, J. M. (1990). Basics of qualitative research: Grounded theory procedures and
techniques. Sage Publications, Inc.

Babbie, Earl R. The Practice of Social Research. 12th ed. Belmont, CA: Wadsworth Cengage, 2010;
Brians, Craig Leonard et al. Empirical Political Analysis: Quantitative and Qualitative Research
Methods. 8th ed. Boston, MA: Longman, 2011.
Given, Lisa M. (2008). The Sage encyclopedia of qualitative research methods. Los Angeles, Calif.:
Sage Publications

US Census Bureau. Technical Paper 63: Current Population Survey - Design and Methodology.
<http://www.bls.census.gov/cps/tp/tp63.htm> accessed 12 January, 2001. Page 14-8.
Frongillo, EA. Analysis of Complex Surveys. October 22, 1996. Cornell University, Office of Statistical
Consulting. College of Human Ecology and
Agricultural and Life Sciences, Division of Nutritional Sciences
<http://www.human.cornell.edu/Admin/StatCons/StatNews/stnews12.htm>

Henry, GT. Practical Sampling. Sage Publications, 1990. Applied social research methods series,
volume 21.
Turner, AG. Sampling Topics for Disability Surveys. United Nations Statistics
Division, Technical Notes, December 1996. Available at
<http://www.undp.org/popin/demotss/tcndec96/tony.htm>

10. Anderson, J. Web page for AM306 Market Research Methods

<http://www.alcd.soton.ac.uk/am306/> see section 4,

Page 129 of 136 CIRSM

 RRM 400

11. Turner
12. 13 also. Alecxih, L, Corea J and Marker D. Deriving state-level estimates from three national
surveys: A statistical assessment and state tabulations. Department of Health & Human
Services/ASPE. May 4, 1998. <http://aspe.hhs.gov/health/reports/st_est/> Accessed 11 January
11, 2001. Section II B, Precision of estimates.
16. Flores-Cervantes, I., Brick, J.M., and DiGaetano, R. Report No. 4: 1997
NSAF Variance Estimation, March 1999. One of the National Survey of America's Families
Methodology Series, at http://newfederalism.urban.org/nsaf/methodology1997.html
17 & 18. Carolina Population Center. Russia Longitudinal Monitoring Survey.
<http://www.cpc.unc.edu/projects/rlms/home.html>.
Coon, D. & Mitterer, J. O. (2010). Introduction to psychology: Gateways to mind and behavior with
concept maps. Belmont, CA: Wadsworth; 2010.
Jakob, Alexander 2001 (February), On the Triangulation of Quantitative and Qualitative Data in
Typological Social Research: Reflections on a Typology of Conceptualizing 'Uncertainty' in the
Context of Employment Biographies, Forum: Qualitative Social Research, Volume 2, Number 1.

Page 130 of 136 CIRSM