Social Research Methodology: An Introduction

Social Research Methodology

An Introduction

Denis Asiimwe Katebire

Printer
Contact address:

Human Rights and Peace Centre (HURIPEC)

Faculty of Law
Makerere University
P.O. Box 7062
Kampala

© Denis Asiimwe Katebire

All rights reserved,

including the right of reproduction
in whole or in part in any form.

Cover design:

Gerald Kasaija
Margaret Trowell School of Industrial and Fine Arts
Makerere University.

Printed by
Makerere University Printery
To the memory of my mother
Acknowledgments

I am greatly indebted to a number of people, singly and collectively, for

the various contributions towards the publication of this book. I would
like to specially acknowledge the contribution of my Masters students
for the years 2003/4, 2004/5, and 2005/6. Your inquisitive minds and
active participation during course seminars have greatly influenced the
content and quality of this book. I wish also to thank my research
associate, Nathan Twinomujuni, for the critiques on the drafts, Billy
Kalema for his help with the graphics, my research assistants, and
everyone else that contributed one way or another. I however take full
responsibility for the errors and omissions.

iv
Table of Contents

Part One: General Introduction to Social Research

Chapter One: The Nature of Social Research. . . . . . . . . . . . . . . . . 2
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Contemporary Social Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Controversy in Social Inquiry.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Social Research Paradigms and the Going Debate . . . . . . . . . . . . . . . 7
Qualitative versus Quantitative Research. . . . . . . . . . . . . . . . . . . . . . . 8
Mixed Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Social Research Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Other Classifications of Social Research. . . . . . . . . . . . . . . . . . . . . . 14

Chapter Two: Basic Concepts in Social Research. . . . . . . . . . . . . 17

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Methodology versus Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Variables in Social Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Independent Variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Dependent Variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Extraneous Variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Controlling Extraneous Variables.. . . . . . . . . . . . . . . . . . . . . . . 21
Operationalization of the Variables. . . . . . . . . . . . . . . . . . . . . . 25
Hypotheses and Research Questions. . . . . . . . . . . . . . . . . . . . . . . . . 26
Hypothesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Research Questions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Frame of Reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Features of a Framework. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Validity and Reliability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Validity.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Reliability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

v
Chapter Three: Population and Sampling.. . . . . . . . . . . . . . . . . . 36
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
From Population to Sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Population Size.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Sample Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Sample Size Determination. . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Sampling Techniques in Social Research.. . . . . . . . . . . . . . . . . . . . . 40
Qualitative Sampling Techniques . . . . . . . . . . . . . . . . . . . . . . . 41
Quantitative Sampling Techniques.. . . . . . . . . . . . . . . . . . . . . . 44

Chapter Four: Ethical Conduct of Social Research. . . . . . . . . . . 48

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Ethical Values and the Researchers’ Obligations. . . . . . . . . . . . . . . . 50
Obligations to Society.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Obligations to Sponsors and Institution of Affiliation. . . . . . . . 52
Obligations to Colleagues.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Obligations to the Subjects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Some Challenges to Consent and Confidentiality Principles. . . . . . . 55
Compliance and Regulation Mechanisms.. . . . . . . . . . . . . . . . . . . . . 57
Towards Ethically Responsible Research.. . . . . . . . . . . . . . . . . . . . . 57

Part Two: Qualitative Research Methods

Chapter Five: Qualitative Research Designs: An Overview.. . . . 62
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Action Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Case Study Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Historiography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Ethnography and Ethnomethodology. . . . . . . . . . . . . . . . . . . . . . . . . 72
Phenomenological Research.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Grounded Theory Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Chapter Six: Interviewing in Social Research. . . . . . . . . . . . . . . . 80

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Who Should be Interviewed?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Types of Interviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Structured Interviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

vi
Unstructured Interviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Semi-structured Interviews.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Successful Qualitative Interviewing. . . . . . . . . . . . . . . . . . . . . . . . . . 89
The Qualitative Interviewer’s Task. . . . . . . . . . . . . . . . . . . . . . . . . . 90
Probing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Evaluation of Qualitative Interviewing.. . . . . . . . . . . . . . . . . . . . . . . 91
Tips for Successful Interviewing. . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Chapter Seven: Focus Group Discussion. . . . . . . . . . . . . . . . . . . . 94

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
The Importance of FGDs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
The Focus Group Process.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
The Setting and Required Hardware. . . . . . . . . . . . . . . . . . . . . . . . . 98
The Facilitator and the Reporter.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Evaluation of the FGD Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Chapter Eight: Field Observation Methods.. . . . . . . . . . . . . . . . 104

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
The Nature of Field Observation. . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Common Concepts in Field Observation. . . . . . . . . . . . . . . . . . . . . 106
Observer Roles in Field Observation. . . . . . . . . . . . . . . . . . . . . . . . 108
Types of Observation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Recording and Reporting Observation Data.. . . . . . . . . . . . . . . . . . 114
General Observation of Observation. . . . . . . . . . . . . . . . . . . . . . . . 115

Chapter Nine: Participatory Rural Appraisal.. . . . . . . . . . . . . . 119

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
The Rapid Appraisal Methodology. . . . . . . . . . . . . . . . . . . . . . . . . 120
From Rapid to Participatory Rural Appraisal. . . . . . . . . . . . . . . . . . 123
The Principles of PRA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Process for PRA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Phase I: Before Fieldwork. . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Phase II: Fieldwork.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Phase III: After Fieldwork. . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
The Tools and Techniques of PRA. . . . . . . . . . . . . . . . . . . . . . . . . 133
Visualization Tools.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Sampling and Data Collection Tools. . . . . . . . . . . . . . . . . . . . 135

vii
Other PRA Tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

Chapter Ten: Document Study and Analysis Methods.. . . . . . . 138

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Content Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Conceptual Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Relational Analysis.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Evaluation of Content Analysis. . . . . . . . . . . . . . . . . . . . . . . . 144
Secondary Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Evaluation of Secondary Analysis. . . . . . . . . . . . . . . . . . . . . . 145
Meta Analysis.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Literature Review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Why Conduct A Literature Review?.. . . . . . . . . . . . . . . . . . . . 148
Organization and Structure of a Review.. . . . . . . . . . . . . . . . . 149
Which Literature is Reviewed?.. . . . . . . . . . . . . . . . . . . . . . . . 150
Evaluation of Literature Review.. . . . . . . . . . . . . . . . . . . . . . . 152

Part Three: Quantitative Research Methods

Chapter Eleven: Quantitative Research Designs: An Overview.154
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Experimental Research.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
True Experimental Designs. . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Quasi-Experimental Designs. . . . . . . . . . . . . . . . . . . . . . . . . . 159
Non-experimental Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Correlational Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Causal-Comparative (ex post facto) Design. . . . . . . . . . . . . . . 165
Survey Research Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

Chapter Twelve: Survey Design and Instrumentation. . . . . . . . 170

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Types of Surveys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
The Questionnaire Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
The Interview Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
The ABC of Survey Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Principles of Questionnaire Design. . . . . . . . . . . . . . . . . . . . . . . . . 182
Rules of Thumb in Questionnaire Design.. . . . . . . . . . . . . . . . 183

viii
 Format of Survey Questions. . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Questions to be Avoided . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

Chapter Thirteen: Measurements and Scaling Methods. . . . . 189

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
Measurement in Survey Research. . . . . . . . . . . . . . . . . . . . . . . . . . 189
Data Types and Levels of Measurement.. . . . . . . . . . . . . . . . . . . . . 190
Nominal Measurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Ordinal Measurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Interval Measurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Ration Measurement.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Beyond Measurement: Scaling in Social Research.. . . . . . . . . . . . . 193
The Concept of Dimensionality in Scaling.. . . . . . . . . . . . . . . . . . . 194
Dimensionality and Types of Scales.. . . . . . . . . . . . . . . . . . . . . . . . 195

Chapter Fourteen: Introduction to Statistical Methods. . . . . . . 209

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
The Role of Statistics in Social Inquiry. . . . . . . . . . . . . . . . . . . . . . 210
Methods of Statistical Data Presentation. . . . . . . . . . . . . . . . . . . . . 210
Text Presentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Tabular Presentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Diagrammatical Presentation. . . . . . . . . . . . . . . . . . . . . . . . . . 212
Graphical Presentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214

Chapter Fifteen: Statistics: From Description to Inferential. . . 219

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Measures in Descriptive Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . 220
Measures of Central Tendency. . . . . . . . . . . . . . . . . . . . . . . . . 220
Measures of Dispersion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Types of Distributions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
The Normal Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
Skewness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Kurtosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Beyond Description: Inferential Statistics. . . . . . . . . . . . . . . . . . . . 231
Estimation.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Prediction and Hypothesis Testing. . . . . . . . . . . . . . . . . . . . . . 233

ix
Inferential Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
The Chi-square. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Student t-Test.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Analysis of Variance.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Analysis of Covariance.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Regression Analysis... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

Part Four: Research Planning and Execution

Chapter Sixteen: The Research Proposal.. . . . . . . . . . . . . . . . . . 240
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Sources of Research Topics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Why Write a Research Proposal?. . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Elements of a Research Proposal. . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Chapter Seventeen: The Research Report. . . . . . . . . . . . . . . . . . 251

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
The Preliminaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
The Substantive Text of a Thesis/Dissertation. . . . . . . . . . . . . . . . . 253
The Structured Thesis/Dissertation.. . . . . . . . . . . . . . . . . . . . . 254
The Integrated Thesis/Dissertation. . . . . . . . . . . . . . . . . . . . . . 255
The Bibliography/Reference List. . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Appendices.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

Chapter Eighteen: Scholarly Writing and Citation Methods.. . 257

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Defining Scholarly Writing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Content of a Scholarly Paper. . . . . . . . . . . . . . . . . . . . . . . . . . 258
Format and Style.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Scholarly References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
The Iterative Process of Scholarly Writing. . . . . . . . . . . . . . . . 261
Scholarly versus Popular Writing: A Summary. . . . . . . . . . . . 262
Citation of Sources under the APA System. . . . . . . . . . . . . . . . . . . 263
The Reference List. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277

x
Preface

In this increasingly globalizing world where scholarship and practice are

pursued in a milieu of complex social problems, there is a growing need
for a holistic approach to the acquisition of knowledge and skills for one
to survive the challenges of academia and the work place. This calls for
concerted organization and regular reorganization and dissemination of
information and knowledge on social processes, social problems, and
social responses; hence the need for handy text materials.
This premier edition, which is a primer on social research methods,
was conceived out of the exigence of comprehensive textbooks
experienced during my several years of teaching, both at undergraduate
and postgraduate levels. Although there is a reasonable quantity of
literature on the subject, it is not without serious problems of
accessibility. First of all, most of the available works are either sector-
specific—focussing exclusively on areas like health, criminology,
education, agricultural extension, etc—or focussed on geographical case
studies. It takes a student considerable time to locate and access a
textbook that is relevant for multiple course modules. Secondly, and
more seriously, the cost of importation has made these texts
prohibitively expensive to an average student in a Ugandan—and indeed
African—university or polytechnic.
This book is arranged in four parts that are meant to systematically
take a student from one level of understanding of social research issues
to another, one higher. Part one begins with an exposition of
contemporary social problems and the historic milestones in the
development of scholarship for the study of such problems. This
provides a vital background to the understanding of the different
research paradigms and their typical traditions. One most important
feature of this part of the book is the introduction of the third research

xi
paradigm—mixed research. Almost all methods textbooks have in the
past tended to discuss research methods in the context of the traditional
qualitative-quantitative dichotomy. Yet, the mixed research paradigm
has over the last decades been steadily gaining ground.
Parts two to four of the book present a systematic journey through
the designs, methods and stages of research and reporting. Part two, for
instance, discusses the different qualitative research designs and
methods. Notable among the methods is the radical, project-oriented
Participatory Rural Appraisal, which is given ample coverage. Part three
discusses the different quantitative research designs and methods,
including especially the statistical methods. Part four deals with the
practical aspects of research—from the proposal to the report. Another
very interesting feature of this part is the in-depth discussion of
scholarly writing—the tradition of citation of sources and scholarly
references—which will prove invaluable both to students and academic
staff in universities and polytechnics.
Although manifestly comprehensive, this book does not claim to be
exhaustive, especially in the intricate statistical operations. It is an
introductory course meant to provide a firm foundation for more
rigorous statistical and other methodological operations, which I
promise in a future project. This book will prove a crucial and
indispensable tool to the student, lecturer and research practitioner.

Denis Asiimwe Katebire

Makerere University
February 2007.

xii
 Part One

General Introduction to Social Research

Chapter One

The Nature of Social Research

Introduction
Research, in whatever field and of whatever magnitude, may be defined
as a systematic process of investigation to discover exiting facts or
issues, or to develop a historical perspective on a subject. This may be
done through naturalistic observation, review of documents and
publications, collection and analysis of primary data, conducting
experiments, etc. Whichever methods and approaches adopted, and
irrespective of magnitude or scope, research must build on existing facts
as a point of departure, venture into the unknown and come out with
new findings about and interpretations of the phenomena investigated.
The question of whether social sciences are actually ‘sciences’ in
the same sense as the physical or natural sciences; and whether social
research can be viewed as ‘scientific research,’ has been one of
significant controversy. In spite of the fact that social science as a
discipline—and alongside it social research—has seen tremendous
development since the late nineteenth century, its pace of development
has not been matched with an inversely proportionate pace of waning
controversy. Quite to the contrary, questions in various forms—some
obvious and others more subtle—have persisted as to whether social
sciences are really sciences.
Compared to the physical and natural sciences on the one hand, and
the ‘non-scientific’ disciplines such as philosophy and the humanities on
the other, social science is quite nascent. Inasmuch as scholars have for
centuries taken interest in phenomena around their world, they either
focussed on distant non-social phenomena (as in astrology) or on social
phenomena in distant societies (as in anthropology). Hardly any
 The Nature of Social Research 3

attention was paid to social behaviour or developments in their own

society. But this was not to go on ad infinitum. Emerging sociological
processes—increase in population and its socioeconomic and other
implications; migration and social integration; urbanization and its
attendant challenges; natural disaster and human response; human
activity and its effect on the environment; conflict, poverty, perversion,
etc – sparked critical focus in sociology, psychology, and other ‘social’
sub-disciplines. Before delving into the origins, development, paradigms
and methods of social research, it is very important to understand the
spectrum of problems social researchers investigate.

Contemporary Social Problems

To appreciate nature and consequences of social problems, one must
understand the mesh of myriad social issues from which these problems
emerge and in which they manifest themselves. The Wikipedia defines
social issues as matters that can be explained only by factors outside an
individual’s control and immediate social environment. Social issues
usually concern more than a single individual and affect many
individuals in a society. They usually revolve around conflicting
viewpoints and tensions between people who take different stances.
There are lots of problems today that affect the lives of individuals,
groups and society as a whole, both in the public and private spheres.
These are the contemporary social problems.
Worsey (1971) defined a social problem as some piece of social
behaviour that causes public friction and/or private misery, and calls for
collective action to solve it. While social problems are often perceived
in terms of personal trouble, sociological studies tend to focus on
“public issues” rather than purely “personal troubles” (Mills, 1959:9).
This implies that what constitutes a social problem is itself quite
problematic. For instance, some actions and situations such as poverty,
child abuse, unemployment and blue-collar crime are clearly defined as
social problems, while others such as white-collar crime fail to attract
the same attention.* Similarly, some social problems, especially those of

* It is important to distinguish between white-collar and blue-collar crime.

Sutherland (1949) defined white-collar crime as a crime committed by a person of
respectability and high social status in the course of his occupation. White-collar crime
4 Social Research Methodology: An Introduction

a corporate and white-collar nature, are often defined to focus largely on

the individual culprit rather than the broader social, economic, political
and institutional implications.
In my view, the above contradictions are closely linked with the
defining process that is based on moral judgements about good and bad,
right and wrong, normal and deviant. This moralist stance presents a
shortcoming in the analysis of the nature and ramifications of social
problems. Take for example the problem of prostitution which is a social
problem. Its perception as deviant and immoral may inhibit a
comprehensive analysis and policy response to it as a social problem.
There are important issues that must be unearthed: to whom is
prostitution a problem? What problems prostitutes face that prostitution
solves for them? How do the problems prostitution solves (personal,
private) compare with those it causes (social, public)? Here below I
outline contemporary social problems to which similar questions and
concerns apply:
• family issues: marriage, parenting, divorce, orphanage, domestic
violence;
• drugs and substances: drug abuse, alcoholism, drug trafficking,
addiction, treatment;
• health issues: aging, mental health, and health care policies and
practices;
• welfare issues: poverty, gender inequality, disability, destitution,
affirmative action, welfare reform;

is thus associated with corporate environments in which large financial transactions

occur, presenting to employees the opportunity for fraud, bribery, insider trading,
embezzlement, money laundering, computer crime, and forgery—crime that is more
technical in nature.
Blue-collar crime is any crime committed by an individual from a lower social
class. Culprits are people employed in relatively unskilled environments, which present
fewer or no opportunities to exploit. They engage in crimes such as vandalism or
shoplifting—crimes that more often use physical force and attract more active police
attention. Thus, whereas blue-collar crime affects more obvious victims who report the
crime, in the corporate world, the identification of a victim is less obvious and reporting
is hampered by a culture of commercial confidentiality to protect shareholder value.
Note however, that white-collar crime is not the same as corporate crime: whereas
white-collar may be committed against the corporation, corporate crime is crime
committed by the corporation.
 The Nature of Social Research 5

• cultural pluralism: racial and ethnic differences, discrimination and

xenophobia;
• urbanization: urban growth, slums, housing and the quality of life;
• population issues: sexuality and adolescence, fertility and
contraception;
• moral issues: prostitution and procuration, rape and defilement,
abortion, paedophilia, pornography, homosexuality;
• migration issues: displacement, refugees and asylum, employment;
• global issues: war and conflict, terrorism, trade, the environment,
aid, technology transfer;
• crime: blue and white collar crime, criminal justice;
• problems on education, politics, the economy, human rights, social
justice;
• constitutional, legislative and judicial issues.
This list is by no means exhaustive, but it serves to give an overview of
the sea of problems social researchers grapple with.
A well considered analysis of these issues should consider two
important questions: the location of blame for social problems, and the
appropriateness of the resultant social policy. As Jamrozik and Nocella
(1998:39) point out, people in different social positions tend to perceive
different social problems; and they may also perceive the same social
problem in different ways. They suggest that since social problems are
defined by perceptions, an adequate understanding of a social problem
as a social phenomenon has to include the perception of its
genesis—that is, the causal links to social activities that ‘create’ the
problem, the social actors who identify the problem, the population
particularly affected by the problem, the nature of the perceived threat,
and applied methods of intervention intended to alleviate, control or
solve the problem. They suggest also that since perceptions may be
guided by theoretically guided research, the relationship between
disciplines of social science and those engaged in legislation,
administration, and formulation and delivery of social policy must be
analyzed and debated.

Controversy in Social Inquiry

The overarching controversy in social science does not so much revolve
around whether it is a science as to what constitutes the proper
6 Social Research Methodology: An Introduction

understanding of society and how that understanding can be achieved

(Bailey 1994:8). This controversy has shown significant resilience in the
social science discourse and indeed constitutes the philosophical
underpinnings that inform the dominant paradigms in social research.
The discourse has tended to position the discipline of social science in
relation to the physical or natural sciences. The emergency of three
schools of thought, as briefly discussed here below, represents major
milestones in the social science discourse.
The first school of thought, the positivist school, was proposed by
French sociologist Emile Durkheim. Durkheim believed that social
phenomena are orderly and thus generalizable. To him, social
phenomena are guided by social laws inasmuch as physical phenomena
follow physical laws. Therefore, the same research methods as those
used in the natural and physical sciences can be sufficiently employed
in the social sciences. The second school, the naturalist (also known as
phenemenological) school articulated by nineteenth century sociologist
Wilhelm Dilthey took an entirely opposite position. Dilthey posited that
human beings have free will and as such, their actions are spontaneous
and cannot be predictable or generalizable. That only unique social
events can be studied. This thinking implied that social research cannot
be conducted using systematic, scientific methods like those employed
in physical sciences. The third school of thought, proposed by German
sociologist Max Weber, adopted a middle ground and argued that while
social sciences can use all the methods used in the natural or physical
sciences, these methods are inadequate for the understanding of social
phenomena. Therefore, these methods should be used only whenever
they are sufficient, but should be supplanted with subjective intervention
of direct understanding.
The positivist position is rather extreme. Its unyielding faith in
orthodox scientific methods means that its social world is rather limited
to only what is physically observable, measurable, controllable,
predictable, reliable and generalizable. On the other hand, Dilthey’s
school fails to acknowledge the impact socialization, learning and
experience may have on human behaviour to make it predictable.
Patterns of behaviour under certain social circumstances can be fairly
accurately predictable (with a minimal margin of error). The Weberian
school acknowledges considerable legitimacy in the positivist criteria,
 The Nature of Social Research 7

but aptly points to their inadequacy in the study of qualitative aspects of

human subjects and human behaviour. Opinions and attitudes,
experiences, visions and insights, emotions and motivations, etc and the
extent to which they uniquely affect or are felt by individuals in
individual human circumstances need direct understanding.

Social Research Paradigms and the Going Debate

Developments in the social sciences and the debate on the nature of
social research have in the twentieth century taken leaps and bounds.
But these have remained located within the Durkheim-Dilthey-Weberian
frameworks. Research methods textbook writers in the 1960s and 70s,
for instance, almost took for granted the desirability of the positivist
approach. Bryman (1988:1-3) highlights some of these works to show
this tendency even among writers that recognized the limitations of
scientific methods. However, phenomenological ideas gave rise to a
growing awareness that the ‘scientific’ approach as applied in surveys
and experiments fails to take into account the differences between the
objects of the social sciences (people) and those of the physical and
natural sciences (things). There was thus a need to develop a strategy
that brought the researcher closer to the subjects of social inquiry – the
people. Hence, the emergency of ‘unscientific’ methods such as
participant observation and unstructured, in-depth interviewing.
These developments clearly brought out the dominant tendencies
in social research to paradigmatic levels, although the mid-road
tendency – the desirability of Durkheimian positivism and Diltheyian
naturalism to complement each other in Weberian fashion – was until
recently never forcefully articulated. A research paradigm as used here
follows Kuhn’s (1970) general definition to refer to a cluster of beliefs
and dictates which, for a scholar in a particular discipline, influence
what should be studied, how research should be conducted, and how
results should be interpreted. Thus, the three dominant traditions or
orientations in social research stemming from positivism, naturalism and
the mid-road school of Max Weber respectively are the quantitative,
qualitative, and mixed research paradigms.
By definition, quantitative research is a logical positivist study of
any aspects of the social world using quantitative (numerical/statistical)
and experimental methods to test the hypothetical-deductive
8 Social Research Methodology: An Introduction

generalizations of the research problem. Qualitative research, on the

other hand, may be defined as a phenomenological inquiry using
descriptive and naturalistic approaches to understand human experience.
Mixed research therefore is that one in which quantitative and
qualitative techniques are mixed in a single study.
The Weberian approach that emphasises complementarity of
methods, which in fact underlies mixed research, did not until recently
feature significantly in the social research paradigmatic debate. Yet, in
practice, social researchers have always—and ever more so of
late—combined qualitative and quantitative techniques and data in a
single study (Johnson & Onwuegbuzie, 2004; Tashakkori & Teddlie,
2006). Proponents of mixed research typically adhere to the
compatibility thesis as well as to the philosophy of pragmatism. They
attempt to use what is called the fundamental principle of mixed
research. According to this principle, the researcher should use a
mixture or combination of methods that has complementary strengths
and non-overlapping weaknesses. I will first deal with the qualitative
and quantitative paradigms before I turn to mixed research.

Qualitative versus Quantitative Research

There have been various positions or view points about these
dichotomous terms in the paradigmatic debate. One view point has been
that the terms ‘qualitative’ and ‘quantitative’ should not be taken merely
as approaches to data collection, devoid of philosophical underpinnings.
Rather, they should be viewed as divergent approaches driven by
assumptions on the nature of the social world and the purpose of
research in the social sciences. Quantification is quite important for
purposes of subjecting data to scientific manipulation and generalization
of findings and subsequent explanation. But that is only the far it goes.
Quite often, social issues transcend quantification and generalization,
and warrant direct understanding.
The qualitative-quantitative concepts have sometimes also been
viewed as competing view points about how social reality ought to be
studied. If this position is anything to go by, then the concepts may be
viewed as, to use Bryma’s (1988:5) words, divergent clusters of
epistemological assumptions; that is, what should pass as warrantable
knowledge about the social world. But the concepts have also been
 The Nature of Social Research 9

viewed as denotations of different ways of conducting social inquiry,

which may be simply varied to suit different research questions or even
integrated. This last position is very important, as social research
projects, big and small, are increasingly turning away from paradigmatic
characterization. Indeed, there is an increasing tendency today for
qualitative studies to incorporate quantitative data and vice-versa, for
complementation, triangulation and corroboration purposes.
Where then is the boundary between qualitative and quantitative
research? The major differences between the two lie in the design and
methodology of data collection and analysis. Accordingly, qualitative
research is characterized by data that can be expressed or described
verbally or non-numerically. This type of research explores
characteristics of individuals and settings that cannot easily be subjected
to quantification and manipulated through orthodox scientific methods.
Therefore, qualitative researchers attempt to describe events or
phenomena and to understand the nature, characteristics or behaviour of
an individual or group of individuals as they are. Quantitative research,
on the other hand, is characterized by data, which are expressed in
numerical terms. For example, a researcher may seek to compare the
average achievement scores of two or more groups of people.
Frequently, quantitative data are analyzed using various statistical
techniques ranging from descriptive to inferential.
Qualitative research can be more plausible as the researcher
personally gets to the heart of the subject, so to say. However,
qualitative data is generally more difficult to manage than quantitative
data since the unstructured methods the researcher uses make it possible
to gather lots of data. This is not to imply that either methodology is
superior; neither that qualitative and quantitative methods are
antagonistic or competing to catch the eye of the researcher. Rather, the
two should be viewed as alternative approaches to the same issues and
goals, which present the researcher with the opportunity to select the
alternative approach appropriate to the research question. As such, they
share various tools in their modus operandi: both employ observation
and interviewing; both conduct case studies (although with different
approaches and for different purposes) to gather data; both have their
sampling strategies; and both have techniques of quality assurance to
10 Social Research Methodology: An Introduction

ensure validity and reliability. However, the differences between them

are more apparent than their similarities as I will shortly summarise.

Mixed Research
The quantitative research paradigm was unmistakably dominant in social
research until the mid twentieth century. By the 1980s, however, the
qualitative paradigm had come of age as an alternative to the
quantitative paradigm, and it was often conceptualized as the polar
opposite of quantitative research (Bryman, 1988; Bailey, 1994). This is
notwithstanding the fact that mixed research has been conducted by
practicing researchers throughout the history of social research. By the
turn of the twenty first century, mixed research had become the
legitimate third paradigm (Tashakkori & Teddlie, 2003).
Mixed research is a general type of research in which quantitative
and qualitative methods, techniques, or other paradigm characteristics
are mixed in one overall study. It has got two major designs—mixed
method research and mixed model research. In mixed method research,
the researcher uses the qualitative research paradigm for one phase of a
research study and the quantitative research paradigm for another phase
of the study. For example, a researcher might conduct an experiment
(quantitative) and after the experiment conduct an interview study with
the participants (qualitative) to see how they viewed the experiment and
to see if they agreed with the results. Mixed method research is like
conducting two mini-studies within one overall research study. In mixed
model research, however, the researcher mixes both qualitative and
quantitative research approaches within a stage of the study or across
two of the stages of the research process. For example, a researcher
might conduct a survey and use a questionnaire that is composed of
multiple closed-ended or quantitative type items as well as several open-
ended or qualitative type items. For another example, a researcher might
collect qualitative data but then try to quantify the data.
Mixed researchers typically adhere to the compatibility thesis as
well as to the philosophy of pragmatism. The compatibility thesis is the
idea that quantitative and qualitative methods are compatible, that is,
they can both be used in a single research study. The philosophy of
pragmatism posits that researchers should use the approach or mixture
of approaches that works the best in a real world situation. A full insight
 The Nature of Social Research 11

into the philosophy of pragmatism can be found in James (1907), Peirce

(1955) and Sleeper (2001), among others. The fundamental principle of
mixed research, on the other hand, is that a researcher should use a
mixture or combination of methods that has complementary strengths
and non-overlapping weaknesses. In the table below I summarize the
differences between the three research paradigms.

Table 1: Differences between Research Paradigms

Difference Qualitative Quantitative Mixed research

Purpose of Description, Description, Multiple

research exploration, explanation, objectives
discovery prediction

View of Fluid, dynamic, Regular, Somewhat

human situational, social, predictable predictable
behaviour contextual,
personal

View of world Subjective Objective Commonsense

reality (personal and (different realism and
socially observers agree on pragmatic view of
constructed) what is observed) world reality (i.e.
what works is
what is ‘real’ or
true

Setting of the Natural Controlled Behaviour is

research environment. environment. studied in more
Behaviour is than one context
studied in or situation
situ,i.e.in the
context in which it
naturally occurs

Structure of Unstructured, Structured, rigid, Combination of

tools (methods flexible standardised structured, semi-
& instruments) structured and
unstructured

Number of Works with many Possible with one Uses various tools
tools tools tool
(triangulation)
12 Social Research Methodology: An Introduction

Form/type of Text, descriptive, Numeric, Multiple forms

data verbal statistical

Approach to Inductive analysis Deductive analysis Both inductive

data collection (bottom-up). The (top-down). The and deductive
and analysis researcher researcher
generates new formulates
hypotheses and hypotheses,
theories grounded collects and uses
in the data. data to test them

Process of data Search for Summarise data Qualitative and

analysis patterns, themes and identify quantitative
and holistic statistical
features relationships

Research Particularistic Generalizable Corroborated

findings findings. findings findings
Representation of
insider (emic)
viewpoint

Form of final Narrative, with Statistical, with Eclectic and

report contextual correlations, pragmatic
description and comparisons of
verbatim means, and
quotations from statistical
participants significance of
findings

Social Research Design

The foregoing paradigmatic classification of social research is by no
means adequate to provide a full understanding of every research
project. There are various other aspects in which social research can be
understood, planned, conducted or evaluated. Research design is perhaps
the most important of such aspects.
Research design can be thought of as the structure of a research
undertaking—the overall plan of an empirical study including topic
analysis, sampling strategy, measurement of key variables, and internal
checks for quality. It is a systematic plan to coordinate a study to ensure
the efficient use of resources and to guide the research according to
universally acceptable methods. Every research design must address the
following key issues: the entities (units of analysis) to be studied, the
 The Nature of Social Research 13

aspects or characteristics of these entities that are of interest, and the

anticipated kinds of relationships among such characteristics. The
common units of analysis in social science research are individuals,
social groups, formal organizations, social artifacts, concrete objects,
and social interactions. According to Trochim (2002), research design
provides the glue that holds the research project together. A design is
used to structure the research, to show how all of the major parts of a
research undertaking—the samples or subjects, measures or
observations, treatments, methods of assignment, etc—work together to
address the central research questions.
Social research designs have been classified differently by different
scholars. Mbaaga (2000) for example categorises all social research
simply into three designs: the census, the survey, and the case study.
There are many types of research this categorization cannot accomodate.
Ledbetter (2001) gives six designs: experimental, causal-comparative,
descriptive, correlational, historical and naturalistic. This may
accommodate the majority of projects. Trochim (2002) on his part
proposes a simple but well-thought out threefold classification premised
on random assignment of variables to groups. According to his scheme,
a design may be (1) a randomized or true experiment if random
assignment is used; it may be (2) a quasi-experimental design if random
assignment is not used but the design uses either multiple groups or
multiple waves of measurement; or it may be (3) a non-experimental
design if it meets none of the above criteria. Trochim thus seems to
think of research design only in the context of experimentation.
In this book, I categorise social research designs within the overall
qualitative-quantitative-mixed research framework. Thus, under the
qualitative paradigm are action research, historical research, case study,
ethnographic research/ethno methodology, phenomenological research,
and grounded theory research (See Chapter Five). Under quantitative
research, there are experimental designs (true and quasi experiments),
causal-comparative (ex post facto) design, correlational design, and
survey research (See Chapter Eleven). Under mixed research, there are
mixed method and mixed model designs (See page 10).
14 Social Research Methodology: An Introduction

Other Classifications of Social Research

Social research can also be classified on the basis of the original, overall
purpose for which it was conducted, the function the researcher
conducted it to serve, and the underlying philosophy and epistemology.

Classification by Purpose
Research can be undertaken out of curiosity or sheer interest to add to
the existing body of knowledge in the researcher’s discipline with no
immediate practical application of the findings in his/her mind. Such
research is called basic research (also known as pure or fundamental
research). The second type of research here is where research is done for
the expressed purpose of resolving or solving a problem. This type is
called applied research. Applied research can itself be divided into three:
action research, research and development (R&D), and evaluation
research.
Action Research is the process by which practitioners attempt to
study their problems in order to guide and evaluate their decisions and
actions, or find the solution to their day-to-day problems, e.g.,
improving welfare or security in a refugee settlement. It entails focussed
efforts to improve the quality of an organization and its performance.
Action research can be done by individuals or by teams of colleagues,
and is in the latter case called collaborative inquiry. Action research is
conducted for
R&D is the type of research intended to develop a new product
(e.g., a new model of a car, gun, farm power plant, etc) or a new
process/procedure of doing things (e.g., prototyping an information
system, improving housing through a low cost housing project, etc).
Once the products or processes are produced/designed, they are
pretested in the field to determine their effectiveness, and then improved
upon until a satisfactory level of effectiveness is achieved. Good
examples of R&D are in the production of drugs or software, and in
social pilot projects.
Evaluation research is a systematic process of collecting and
analyzing data in order to assess the merits or performance of a practice
or program after a given period of operation or application. For example,
after a period of protracted media campaign against HIV/AIDS through
the ‘ABC’ strategy in Uganda, an evaluation study may be conducted to
 The Nature of Social Research 15

assess people’s sexual behaviour and/or any reduction in HIV infection

or prevalence. Similarly, a study may seek to assess the level of
enrolment by children of school-going age after seven years of Universal
Primary Education (UPE). Such an evaluation is necessary to inform the
media campaigners whether the campaign strategy has been successful,
whether it should be modified, or whether it is a dismal failure to be
terminated immediately; and for the UPE programme that cannot be
terminated, to review the administrative, logistical, and other aspects of
the programme.

Classification by Function
Social research can be exploratory, descriptive or analytical; hence the
three functions: to explore, to describe or to analyse social phenomena
or events. Exploratory research is conducted to study a new
phenomenon or event, and is driven by the following objectives: (1) to
determine the feasibility of a more detailed study of the phenomenon or
event, and (2) to develop the methodology to be used in the detailed or
controlled study.
Descriptive research is conducted to collect detailed information
about people, situations or events to facilitate decision-making. For
example, a government campaign strategy against corruption might be
informed by details on the target population’s literacy and education
levels, income levels, religion, where they live, etc.
Analytical research involves the examination of different aspects
of a phenomenon or situation simultaneously. An analytical study may,
for example, seek to explain why Baganda of Uganda agitate for a
special constitutional status, or why the Batutsi of Rwanda are
ethnocentric.

Classification by Philosophy/Epistemology
Besides the foregoing classifications, social research may be variously
classified as objective or subjective; as focussing on the discovery of
general laws or the uniqueness of each particular situation; as aimed at
prediction and control or explanation and understanding; as taking an
outsider perspective or an insider one; etc. In short, social research
operates on some underlying assumptions about what constitutes “valid”
research and which research methods are appropriate.
16 Social Research Methodology: An Introduction

For our purposes, the most pertinent philosophical assumptions are

those that relate to the underlying epistemology which guides the
research. Epistemology refers to the assumptions about knowledge and
how it can be obtained. Basing on the underlying research epistemology
therefore, we can categorise social research into three: positivist,
interpretive, and critical research.
Positivist research assumes that reality is objectively given and can
be described by measurable properties which are independent of the
researcher and his or her instruments. It generally attempts to test theory,
in an attempt to increase the predictive understanding of phenomena.
Researchers are advised to follow the scientific method, which has as its
objective the elimination of subjectivity from the research process.
Positivist researchers are therefore assumed to be detached, external
observers and reporters of what they see.
Interpretive research adopts the epistemological assumption that
reality (whether given or socially constructed) consists of subjective
interpretations and cannot be studied objectively; and that it is subject
to multiple interpretations. The purpose of interpretive research is to
understand and analyze subjective interpretations and their
consequences. Interpretive studies generally attempt to understand
phenomena through the meanings that people assign to them.
Interpretive researchers do not predefine dependent and independent
variables, but focus on the full complexity of human sense making as the
situation emerges.
Critical research is based on the assumption that social reality is
historically constituted and that it is produced and reproduced by people.
Although people can consciously act to change their social and
economic circumstances, critical researchers recognize that their ability
to do so is constrained by various forms of social, cultural and political
domination. The main task of critical research is seen as being one of
social critique, whereby the restrictive and alienating conditions of the
status quo are brought to light. Critical research focuses on oppositions,
conflicts and contradictions in contemporary society, and seeks to be
emancipatory—that is, to help eliminate the causes of alienation and
domination.
Chapter Two

Basic Concepts in Social Research

Introduction
All academic disciplines—including social science research—evolve on
the basis of an internal process of consensus building on what should
constitute the discipline and its traditions. This process begins from
individual conceptions (ideas or mental images) that cannot be
communicated directly, to agreed-upon terms and procedures used in the
discipline. The terms enable scholars to progressively communicate and
eventually agree on what they (the terms) specifically mean. The
process of coming to an agreement about the meaning of the terms is
called conceptualization, and the resulting term is called a concept.
Concepts are thus constructs derived by mutual agreement from
conceptions. They summarize collections of seemingly related
observations and experiences, and only have the meanings that scholars
agree to give them.
There are three types of phenomena that social science researchers
measure. These are (i) direct observable phenomena—those physical
things we can observe rather simply and directly, such as settlements,
floods, etc; (ii) indirect observable phenomena—those that require
relatively more subtle, complex, or indirect observations, such as those
made by someone else in the researcher’s absence; and (iii)
constructs—theoretical creations that are based on observations but that
cannot themselves be observed directly or indirectly, such as
intelligence, bias, inequality, etc. Conceptualization is the refinement
and specification of these constructs (abstract concepts), while
operationalization is the development of specific research procedures
18 Social Research Methodology: An Introduction

(operations) that will result in empirical observations representing those

concepts in the real world.
As researchers discover the utility of a particular conceptualization
and operationalization of a concept, they often adopt it, leading to
standardised definitions of concepts. The conceptualization process
involves describing the indicators to be used to measure that concept,
and the different aspects of the concept, which are referred to as the
dimensions. An indicator is in fact a sign of the presence or absence of
the concept under study. For example, in a study of relative household
wealth, the indicators would include type/quality of housing, feeding,
clothing, etc. Likewise, a study on inequality would focus on
participation, gender balance, and similar indicators.
Whenever a researcher identifies a concept and sets about
specifying what he or she means by that concept, he or she often
discovers that the concept is multifaceted or multidimensional. This
discovery implies that concepts can be grouped and sub-grouped. The
term dimension, which is a technical term for the formed groups and
sub-groups, is actually a specifiable aspect of a concept. Specifying the
different dimensions of a concept makes it easy to identify the indicators
of interest, and thus facilitates a more understanding of what one is
studying. If several different indicators all represent, to some extent, the
same concept, then all of them will behave the same way that the
concept would behave if it were real and could be observed. However,
if there are variations in the dimensions, it should be seen if two sets of
indicators represent different dimensions of the concepts.
Conceptualization is a continuing process but it must be addressed
specifically at the beginning of any study design, especially rigorously
structured research designs such as those discussed in Part Three of this
book. Even in less structured research methods such as those discussed
in Part Two of this book, it is important to begin with an initial set of
anticipated meanings that can be refined during data collection and
interpretation. We begin observations with preconceptions; scientists
and observers must be conscious of and explicit about these conceptual
starting points. In this chapter I discuss the basic social research
concepts, starting with the differentiation of the often-mistaken-as-
synonymous, methodology and method, to the must-encounter ones and
their operationalization.
 Basic Concepts and Operations in Social Research 19

Methodology versus Method

The terms “methodology” and “method” are widely but erroneously
used synonymously, so it is very important here to draw a difference
between them. Research methodology should be understood as an
attempt by scholars and researchers to validate the rationale behind the
selected research design and provide justification of why it is
appropriate in solving the selected research problem. Generally, in a
methodology there is an overall philosophical approach to the problem.
A research method on the other hand, should be viewed as a recipe or a
codified series of steps taken to complete a certain task or to achieve a
certain objective. It can be viewed as just a part of a methodology. Most
disciplines have their own specific methods. The power of a method is
usually inversely proportional to the generality of that method. In other
words, the more specific the method, the more powerful it becomes.

Variables in Social Research

A variable is a characteristic of a person, object or phenomenon, which
can take on different values. The values of variables can be expressed in
numerical terms, and are in this case called numerical or quantitative
variables. Examples of numerical variables are age, distance, weight,
income, etc which are continuous and thus take on many different
values. Other values of variables can be expressed in terms of categories,
and these are called categorical or qualitative variables. Examples of
categorical variables are colour of eyes, blood group, sex, staple food of
a locality, etc, which are discrete.
Research involves dealing with variables and their relationships. A
researcher may, for example, be interested in the relationship between
motivation and job performance, legislation and its effect on civil order,
court decision and state-inspired violence, etc. A researcher can hardly
investigate all the variables on one given problem of study. He or she
must therefore select only those of particular interest to him or her.
Variables can be grouped into three broad categories: (i) independent
variables, (ii) dependent variables, and (iii) extraneous variables. It is
very important that a researcher identifies these as early in the process
of research as the time of problem selection and formulation of
hypotheses or research questions. We now discuss variables in detail.
20 Social Research Methodology: An Introduction

Independent Variables
An independent variable represents the conditions that a researcher
controls in an effort to understand their effect on a given outcome. It is,
in other words, what the researcher chooses to study. Independent
variables are therefore the factors that the researcher manipulates in
order to assess their possible effects on other variables. Independent
variables are themselves of two types: treatment variables and organic
or attribute variables.
Treatment variables are those that a researcher actually manipulates
and assigns to the participants or subjects of the study. They include
such things as actions, processes, etc., which have an influential effect
on other things. Take an example of a study of the effect of the size of
a defence team on a case decision. The size of the defence team is the
treatment (independent) variable because the researcher presumes it to
have an influential (potential or real) effect on the decision of the case,
and is studying it to confirm or disprove the presumption.
Organic variables, on the other hand, are the characteristics of the
participants, which are relatively natural or permanent and cannot be
physically altered or manipulated by the researcher. However, the
researcher can take these variables into account during the study or may
exclude them. Organic variables include such things as age, gender,
blood group, etc—variables that are determined by nature.

Dependent Variables
Dependent variables are those things, situations or outcomes, that are
presumed to be affected by independent variables or to result from them.
In research therefore, a dependent variable can be associated with the
outcome of the study and an independent variable with the input. Using
a similar example as the one above, a researcher may seek to answer the
following research question: “Is a case defended by five lawyers likely
to be better defended than that defended by two?” In this question, the
dependent variable is the quality of the defence because the researcher
is investigating whether it is many or few defence lawyers that would
cause better defence. The number of defence lawyers involved in the
case is thus the independent variable, because it is the one being
investigated how it affects the quality of the defence.
 Basic Concepts and Operations in Social Research 21

Extraneous Variables
There is a big problem about variables in social research: there are
always multiple independent variables that are likely to affect the
outcome of a study. Once a researcher has decided on the independent
variables to study, he or she must then direct attention to the potential
influence or effect of the other independent variables that are not being
studied. If those variables are not controlled, then the researcher may not
come out with plausible conclusions as to whether it is the independent
variables studied that actually account for the observed outcome or it is
due to the other variables. Such variables are called extraneous variables
or intervening variables. They are the independent variables that the
researcher has not selected for study.
When independent variables of interest in the study are mixed up
with extraneous variables that the researcher has not controlled, the
outcome of the study is mixed up. Let us consider the foregoing research
question again for example: “Is a case defended by five lawyers likely
to be better defended than that defended by two?” Whereas the
researcher is interested in the number or size of the defence team as the
independent variable influencing the quality of the defence, there are
other independent variables that may affect the quality of the defence.
These include quality of the defence lawyers involved (determined by
their education and experience), the merits of the case, political interests
in the case, independence of the courts, etc. If these are not being
studied, they become the extraneous variables, and they must be
controlled, lest they should lead to the researcher drawing flawed
conclusions.

Controlling Extraneous Variables

So extraneous variables are potential independent variables. Failure to
control their effects may result in a study that does not adequately test
the questions or hypotheses of the study. In other words, extraneous
variables interfere with obtaining a clear understanding of the
relationship between the outcome of the study and the independent
variable. Such variables make it possible to entertain an alternative
explanation of the outcome. Therefore, to be sure that the independent
variable we are interested in is the one causing the outcome of the study,
the extraneous variables must be controlled. There are at least seven
22 Social Research Methodology: An Introduction

basic techniques or procedures for controlling extraneous variables in a

study.

Randomizing extraneous variables

In randomization, the researcher first carries out random selection of
participants from the target population of interest and them, if necessary,
carries out random assignment or distribution of participants to different
treatment conditions in the study. Random selection means that every
member of the population has an equal chance of being selected to
participate in the study (as a member of the sample). Random
assignment means that every individual who has been randomly selected
to participate in the study has an equal chance of being assigned
(distributed or placed) to any condition included in the study. Thus,
randomization is the best single way to attempt to control the effects of
many extraneous variables in a study. Therefore, any quantitative
researcher should always randomize his/her extraneous variables and, if
necessary, randomly assign the participants to various conditions in the
study.

Eliminating extraneous variables

When possible extraneous influences can be entirely removed from the
experiences of participants in different conditions of the study, this
should be done. For example, if the researcher is convinced that
extraneous noise or interruptions or changes in the environment are
likely to affect the outcome of the study, then the researcher must try to
eliminate these conditions. This can be, say, by putting up a notice
“silence: exam in progress” so that everybody avoids making noise.

Holding extraneous variables constant

Sometimes the researcher may find it difficult to completely eliminate
an extraneous variable in the study. Under such circumstances, it might
be wise for the researcher to hold such an extraneous variable constant
across all the participants or all conditions in the study. For example, if
the researcher suspects that the instructions given to the participants on
how they are supposed to fill in the questionnaire may be differently
understood, then such instructions should be held constant across the
participants. This will simply be by giving the participants the same
 Basic Concepts and Operations in Social Research 23

written instructions so that they are all equally affected. Similarly, if the
researcher suspects that the time of the day affects the outcome of the
study, then data collection must be done every day at the same time so
that it is equally affected. More still, if intelligence is a variable in whose
effect the researcher is not interested, then he/she should ensure that the
participants in the study are on average of the same intelligence level. If
gender is an extraneous variable that may affect the outcome of the
study, then it should be made constant in the study by either taking only
males or females as participants.

Matching participants on an extraneous variable

If large groups of participants are used, then simple random assignment
is likely to yield groups that are comparable on all variables. However,
when working with such groups, researchers often use the process of
matched pairs to make sure that the two or more groups are initially
comparable on the identified extraneous variable. If for example I.Q. is
the extraneous variable, the first thing to do is to give an intelligence test
to the participants and then rank them on the basis of their intelligence.
If the researcher had two conditions in his/her study (condition A and
condition B), the following is done:
Participant I.Q. (ranked) Action

1 135 Randomly assign to conditions A

and B
2 130

3 130 Randomly assign to conditions A

and B
4 129

5 128 Randomly assign to conditions A

and B
6 128
Eventually, when the averages of the groups are taken, they will be
almost the same.

Including extraneous variables in the study

This is making extraneous variables independent variables. If an
extraneous variable cannot be eliminated or controlled using any other
24 Social Research Methodology: An Introduction

method, then the researcher can deliberately include it as an extra

independent variable. For example, supposing a researcher in the
Faculty of Arts is experimenting on the type of errors committed by the
students being trained as secretaries, but his interest is on the effect of
alcohol on typing errors, the researcher has two conditions of interest to
study:
A - no alcohol consumed
B - alcohol consumed
However, typewriters are the only machines available for the study and
there are not enough typewriters of the same make. Note that in this
study the effect of typewriters could be controlled by holding them
constant, that is, allowing the use of the same make of typewriters. Yet,
the researcher cannot do this due to lack of enough typewriters of the
same make. The choice therefore is to take as an extra independent
variable the make of the typewriter used.
Originally, the researcher was interested in alcohol as the only
independent variable. The researcher would then assign the participants
to the various conditions of the study , that is, “No Alcohol Consumed”
and “Alcohol Consumed.” After giving the participants an equal amount
e.g. a bottle of mineral water for the “No Alcohol” group and two bottles
of wine each to “Alcohol” group, then the participants are asked to type
a paragraph of the same content using the same type of machine. Check
for the mistakes and count them separately for each group. The average
of errors is computed for each group. If this average is higher for the
alcohol group, then the researcher will conclude that alcohol has an
effect on typing. On the other hand, when the researcher takes the
second variable into account, he/she will handle its effect when
analyzing the data, i.e. each independent variable separately from the
other. The researcher can then analyse the interaction effect when the
two variables (effect of alcohol and make of typewriter) are working
simultaneously.

Balancing the effects across different study conditions

The technique for controlling extraneous variables that cannot be
eliminated or held constant in a study is called balancing. This involves
distributing the effect of the extraneous variable across the different
treatment conditions of the study. For example, when the researcher
 Basic Concepts and Operations in Social Research 25

realizes that gender is an extraneous variable in the study, he or she has

two options to control it:
1) Simply use either males or females in the study. However if both
must be used in the study, he or she would need to balance the
effect of this extraneous variable.
2) If the researcher has two conditions in the study, then he/she would
balance by randomly assigning an equal number of males and
females to each of the two conditions in the study.
Note that balancing resembles holding the variable constant. The only
difference between the two is that in holding the extraneous variable
constant the variable is held constant at one level (use only females or
only males) whereas in balancing, the extraneous variable is distributed
equally to all the available conditions in the study.

Statistical adjustment for extraneous variables

Very often participants in a study may differ on an extraneous variable
even before the study is conducted. This initial difference that might
exist among the participants should be controlled through statistical
adjustment when the data is being analyzed. To achieve this, the
researcher obtains measurement for each participant on the identified
extraneous variable. Then he/she embarks on the study where he/she
now takes or obtains measurement (data) on the dependent variable of
interest. For example, if the identified extraneous variable is intelligence
but the researcher is not interested in intelligence, and if he chooses the
method of control, he would then administer an intelligence test to all
participants in the study. The initial differences in intelligence, which
can also affect the outcome of the study, can be statistically adjusted or
eliminated by use of statistical techniques such as analysis of covariance
(ANCOVA) or partial correlation.

Operationalization of the Variables

Once the researcher has identified his or her variables in the study, he or
she should come out clearly regarding how the variables will be
observed or measured. This clarification is called the operational
definition of the variables. It requires that the researcher specifies the
actions or operations that are necessary to measure or to identify those
variables. For example, suppose the researcher’s hypothesis is: “quality
26 Social Research Methodology: An Introduction

of school is correlated with students’ achievement.” The variables here

are quality which is an independent variable, and achievement which is
a dependent variable. The operationalization of these variables will be
to define what the researcher means by these variables. For example,
• Quality includes qualifications of teachers, availability of study
materials, and how students are fed.
• Achievement means academic performance at public examinations
such as Primary Leaving Examinations, and in continuous
assessment tests.

Hypotheses and Research Questions

Social researchers do not just collect any data relating to the problem
under study. They need only specific data relating to the specific aspects
of the problem being studied. This is done by the researcher making, as
a starting point, clear assumptions about the likely relationships between
the variables to be studied, or posing clear questions about such
relationships. It is these statements or questions that determine the kind
of data needed to verify or answer them. The statements are called
hypotheses and the questions are called research questions.

Hypotheses
A hypothesis is a tentative assumption made in order to draw out and
test its logical and empirical consequences. It is a proposition that is
stated in testable form and predicts a particular relationship or lack of it
between two or more variables. By test is meant to subject the
assumption to scientific manipulation in order to confirm it to our
satisfaction or to prove it wrong. Therefore, it must be a statement of
fact and susceptible to empirical investigation. Hypotheses translate the
objectives of the research into clear affirmative or negative statements
or predications of the expected results. There ate two types of
hypotheses, namely, statistical hypotheses and research hypotheses.
A statistical hypothesis, also known as the “null hypothesis” and
denoted as H0 is a statement claiming or predicting equality between the
variables under study. In other words, that there is no relationship or
difference between the variables; and it is this fact that constitutes the
equality. In short, a statistical or null hypothesis is a statement of
equality, and may be stated as follows: “There is no relationship
 Basic Concepts and Operations in Social Research 27

between judicial independence and the remuneration system of judicial

officers,” or “There is no difference in the average scores of LL.B
evening and LL.B day students on aptitude test.”
A research hypothesis, also known as “alternative hypothesis” and
denoted as H1 is a definite statement of inequality between the variables.
It assumes or predicts that a relationship or difference exists between the
variables; and it is such a difference or relationship that constitutes the
inequality. An alternative hypothesis can itself be divided into two
categories: (i) a directional research hypothesis, often called a one-tailed
hypothesis, and (ii) a non-directional research hypothesis, often called
a two-tailed hypothesis.
A directional research hypothesis indicates a specific direction that
the research expects to take: the researcher may be interested in a change
in one direction only. This is done through the use of adjectives such as
higher, worse, better, lower, more, less, etc. For example: “LL.B
evening students score higher average marks in aptitude tests compared
to BBA evening students,” or “There is a positive relationship between
the rewarding system for judicial officers and judicial independence.”
A non-directional research hypothesis states inequality between
variables or groups but it does not make specific prediction about the
direction of the outcome. For example: “There is a difference in average
scores in aptitude tests between LL.B evening students and BBA
evening students,” or “There is a relationship between remuneration and
judicial independence.”
Hypotheses are a feature of quantitative studies which require
statistical tests to prove or refute them, although research hypotheses are
also used in qualitative studies. Statistical/null hypotheses are always
formulated for purposes of statistical tests. In fact, all statistical tests use
null hypotheses with the view that the results would be the reverse.

Research Questions
A research question is a concise interrogative statement, which is
normally stated in the present simple tense and includes one or two
variables. Research questions normally focus on what variables are to be
described and what relationship might exist among them. By asking
specific questions, the researcher gives more details about the purpose
of the study. Research questions in one study may range from one to
28 Social Research Methodology: An Introduction

several; and in the latter case they should be listed in order of

importance to the research. A researcher must take care not to pose
questions for which he or she will not be able to collect suitable data and
arrive at plausible conclusions. Research questions may be stated as
follows: “Is there a relationship between remuneration and judicial
independence?” or “What is the relationship between the rewarding
system for judicial officers and judicial independence?” Research
questions are employed mostly in qualitative studies to guide the
researcher to the required data.

Frames of Reference
Once a researcher is familiar with the existing literature and the nature
of the problem, he or she must develop a frame of reference (commonly
known as a framework) within which to conduct the study. A frame of
reference is a theoretical analysis or abstract conceptualization that
places the study within a context of meaningful reference. The
development of this framework involves the identification of its basic
elements, namely, (i) the theory, (ii) the concepts or variables, and (iii)
the relationship among these variables. There are thus two types of
frameworks, of which only one may be used in a study. These are the
theoretical framework and the conceptual framework.
A theoretical framework is more appropriate for a research that is
based on one identified theory. That theory is discussed and analysed in
relation to the current problem under investigation. The rationale is to
ascertain, at the end of the study, whether that theory can satisfactorily
explain the occurrence and nature of the problem, or whether they is a
need for alternative theorising for the problem. A conceptual framework,
on the other hand, draws on concepts from various theories and previous
research findings. All the concepts and variables that are used in the
conceptual framework must be clearly linked or related together. The
possible relationships and their directions must be clearly identified.
Sometimes, this framework may be diagrammatic, and is in this case
called a conceptual model.

Features of a framework
There are four basic features of a framework. First, the variables or
concepts that are considered relevant to the study should be clearly
 Basic Concepts and Operations in Social Research 29

identified. Second, there should be a discussion stating how the variables

or concepts are related. This must be done for the important
relationships that are theorized to exist among the variables. Third, the
direction of the expected results, i.e., whether the relationships would be
positive or negative, should be indicated. This is especially so if the
nature and direction of the relationships can be theorized on the basis of
the findings from previous research. Fourth, there should be a clear
explanation as to why the relationship would be expected to exist in a
given manner. This should be drawn from the findings in previous
research. If a conceptual model is used, it should attempt to depict all
these features.

Validity and Reliability

Data collection instruments must posses certain qualities and standards
that will make them and the data they collect acceptable as appropriate
and authentic. Validity and reliability are used respectively to determine
this acceptability. Whenever a test or other measuring device is used as
part of the data collection process, the validity of that test or device and
the reliability of that data is important.

Validity
Validity refers to the extent to which the data collection instrument
collects data that have the characteristic or attribute the researcher wants
to measure. In other words, if the tools a researcher uses in the study
enable him or her to arrive at what was actually intended to be achieved,
then there is validity. Validity may thus be determined by the
appropriateness of a research tool (instrument), the research procedure
(technique) or the accuracy of the research findings (report). The
rationale behind validity is that an instrument must effectively measure
or test what it was designed to measure or test. After all, we are relying
on the results to show support or lack of support for our theory and if the
data collection methods are erroneous, the data we analyse will also be
erroneous. There are about six types of validity as I otline below.

Test or construct validity

This validity refers to the degree to which a test or other measuring
device is measuring what the researcher intended it to measure. A
30 Social Research Methodology: An Introduction

construct is basically the variable a researcher is testing or measuring.

Construct validity has to do with the logic of items which comprise
measures of social concepts. A good construct has a theoretical basis
which is translated through clear operational definitions involving
measurable indicators. A poor construct may be characterized by lack of
theoretical agreement on its content, or by a flawed operationalization
such that its indicators may be construed as measuring one thing by one
researcher and another thing by another researcher.
Examples of constructs include motivation, depression, anger, and
practically any human emotion or trait. For many constructs, or variables
that are artificial or difficult to measure, the concept of validity becomes
more complex. Construct validity is the term given to a test that
measures a construct accurately. If a researcher’s proposed construct is
at odds with the existing literature on related hypothesized relationships
using other measures, its construct validity is suspect. Thus, the more a
construct is used by researchers in more settings with outcomes
consistent with theory, the more its construct validity.
There are two aspects of construct validity: convergent validity and
discriminant validity. Convergent validity is based on two alternative
principles: (i) that the indicators for a given construct should be at least
moderately correlated among themselves; or (ii) that the researcher’s
proposed scale or measure of a given construct correlates with measures
of the same construct using instruments proposed by other researchers.
Discriminant validity, on the other hand, is based on the principle that
the indicators for different constructs should not be so highly correlated
as to lead to the conclusion that they measure the same thing. Some
researchers use the criterion that the correlations testing convergent
validity should be higher than those testing discriminant validity. This
is because items measuring the same thing should correlate more highly
with themselves than with other things. It is also notable that multi-
method, multi-trait methodologies (triangulation) lead to higher
construct validity.

Content or face validity

Content validity is concerned with two issues. The first issue is whether
a test measures what it claims to measure. It is possible for a study to be
internally and statistically valid, yet use measures that are lacking face
 Basic Concepts and Operations in Social Research 31

validity. Indicators may display construct validity, yet the label attached
to the concept may be inappropriate. The measures which operationalize
a concept should therefore be the ones which seem by common sense to
have to do with that concept.
The second issue is whether the items on the instrument measure
the full domain of a construct implied by their label. In other words,
whether the label is able to imply or represent all of the content of a
particular construct. It is important to ensure that the labels attached to
constructs are broad enough in domain. For example, if an instrument is
labelled “liberalism” but the items on it only limited to cultural issues
such as abortion, prostitution and gay rights, without any content on
economic or environmental issues, it lacks content validity. For such
contents, “cultural liberalism” would be a better label. The only sure
way to determine content validity is to seek expert opinion.

Concurrent or criterion validity

This type of validity is concerned with the correlation between
instrument measurement items, and the known and accepted standard
measures or criteria. It refers to the ability of a measurement device to
vary directly with a measure of the same construct or indirectly with a
measure of an opposite construct. It allows the researcher to show that
his or her test is valid by comparing it with an already confirmed valid
test.

Predictive validity
This is the ability of a test or measure or operationalization to predict
something it should theoretically be able to predict. Such a test should
thus be a valid screening device for some future behaviour. For example,
an interview (set of written questions) to a fresh graduate for an
administrative position in a company is valid if it predicts the ability of
that graduate to perform in that position. Predictive validity is
determined by computing a correlational coefficient comparing—in this
case—the graduate’s performance at an interview with the performance
appraisals on the job. If they are directly related, then we can make a
prediction regarding job performance based on the interview. In other
words, a high correlation would provide evidence for predictive
32 Social Research Methodology: An Introduction

validity—it would show that this test can correctly predict something
that was theoretically thought it should be able to predict.

Internal validity
Internal validity, mostly important in experimental and causal-
comparative research, refers to the ability of a study to determine if a
causal relationship exists between the variables under study. An
experimental researcher, for instance, may want to be reasonably sure
that the outcome of the experiment is a result of exposure or treatment.
Researchers must thus be aware of aspects that may reduce the internal
validity of a study and do whatever they can to control for these threats.
This may be done by introducing covert variables in order to control
sources of bias that would affect the cause-effect process. When there is
lack of internal validity, extraneous variables may be responsible for
part or all of the observed effect on the dependent variables. It should be
noted that internal validity is not at issue where a study is not concerned
with causality.

External validity
External validity relates to the generalizability of a study and possible
bias in the process of generalizing. Researchers want to be reasonably
sure that the results of a study based on a sample of the population truly
represent the entire population (including dimensions of generalisation
to other geographical and time settings). When there are threats to
external validity, the results within a sample group may by themselves
be normal, but their generalization to the wider population will be very
problematic. Naturally, when a sample of study is non-random in
unknown ways, such as in convenience sampling, the likelihood of
external validity is extremely low.

Reliability
The term reliability simply means the repeatability, consistency, or
dependability of a test, survey, observation, or other research method.
It refers to the degree to which a data collection instrument will at
different times give the same data or information about the attribute or
characteristic being measured or investigated (assuming that what we are
measuring is not changing). A reliability coefficient is often the statistic
 Basic Concepts and Operations in Social Research 33

of choice in determining the reliability of a method. This coefficient

merely represents a correlation, which estimates the intensity and
direction of a relationship between two or more variables. There are
about five types of reliability estimators as I outline below.

Test-retest reliability
This estimator is used to assess the consistency of an instrument or test
from one time to another. (It is sometimes referred to as “reliability
across time”). To determine the coefficient for this type of reliability, the
same test is given to the same or similar group of subjects on at least two
separate occasions. It is a three-step process: 1) you give a test during
one time period; 2) you give the exact same test to the same subjects
during a second time period; and 3) you correlate the two sets of scores
(same subjects, same test). If the two test scores have a “high and
positive” correlation, this is evidence of the consistency of the test score
across time: there is reliability.
This approach assumes that there is no substantial change in the
construct being measured between the two occasions. The amount of
time allowed between the measures or tests is thus very critical. One
major concern with test-retest reliability is what has been termed the
“memory effect” especially when the two tests are close together in time.
This can create an artificially high reliability coefficient as subjects
respond from their still fresh memory rather than the test itself. The
shorter the time gap, the higher the correlation, and vice versa. When a
pre-test and post-test for an experiment is the same, the memory effect
can play a role in the results.

Parallel forms reliability

One sure way to control or eliminate memory effects is to use a different
estimator called parallel forms reliability. This estimator, also called
equivalent forms reliability, is used to assess the consistency of the
results of two tests constructed in the same way from the same content
domain. A researcher first has to create two “parallel” or “equivalent”
forms. This may be by creating a large set of questions that address the
same construct and then randomly divide the questions into two sets.
The two sets are then administered to the same sample of subjects and
the average scores correlated to get the estimate of reliability. A high
34 Social Research Methodology: An Introduction

and positive correlation means that the two forms are parallel or
equivalent.
One major problem with this approach is that you have to be able
to generate lots of items that reflect the same construct, and this is not
always easy. Another problem is the assumption that the randomly
divided halves are “parallel” or “equivalent” since it is the halves and
not the items within the halves that are random.

Inter-rater reliability
This estimator, also called inter-observer reliability, is used to assess the
degree to which different raters or observers give consistent estimates of
the same phenomenon. This approach is commonly used in
observational studies where observations of behaviour are used as data
in research. Since the human observers are easily distractible, get tired
of monotony, and make misinterpretations, the issue for the researcher
is to ensure that these observations are reliable or consistent. One way
to determine this is to have two or more observers rate the same subjects
and then correlate their observations.
Inter-rater reliability can be estimated in two ways: using categories
and using continuous measurement. In the first instance, the researcher
constructs a “categories” observation or rating tool, and the observers’
or raters’ interest is in which category each observation falls. After
observation, the researcher then calculates the percent of agreement
between the different raters. In the second case, the measure must be a
continuous one (making use of a continuous scale, such as a Likert
scale). Here, all the researcher needs to do is calculate the correlation
between the ratings of the different observers.

Split-half reliability
Under this approach, a researcher randomly divides all items that purport
to measure the same construct into two sets, but these remain on the
same instrument. The entire instrument is then administered to a sample
of subjects and the total score for each randomly divided half
calculated. The two “half-test” scores are correlated. The split-half
reliability estimate is simply the correlation between these two total
scores. The split-half approach is much similar to parallel forms
 Basic Concepts and Operations in Social Research 35

reliability except that in the latter, the constructed parallel forms are
used independent of each other and considered equivalent measures.

Internal consistency reliability

This estimator is used to assess the consistency of results across items
within a test. Internal consistency is the extent to which tests or
procedures assess the same characteristic, skill or quality. It is a measure
of the precision between the observers or the measuring instruments
used in a study. A researcher administers a single measurement
instrument to a group of people on one occasion to estimate reliability.
The reliability of that instrument is judged by estimating how well the
items that reflect the same construct yield similar results. The
researcher’s interest is in how consistent the results are for different
items for the same construct within the measure. This type of reliability
often helps researchers to interpret data and predict the value of scores
and the limits of the relationship among variables.
Chapter Three

Population and Sampling

Introduction
When people talk of, say, the population of Uganda, they mean the total
estimate of all Ugandans as enumerated at a population census.
Scientists also normally talk of say a national cattle population, a
population of elephants in a national part, a population of a tree species
in a given forest, etc. The overall meaning one gets is that population is
the total collectivity of like biotic species in a given locale. In statistics,
the term ‘population’ refers to much more than a collectivity of people,
animals or plants. Statisticians may speak of a population of objects, or
events, or procedures, or observations; including such things as the
quantity of sugar in urine, the number of visits to the library, etc. To
them therefore, population refers to an aggregation of creatures, objects,
cases, or observations under study.
In social research, the term refers to people, either singly or
collectively as in a community, as the unit of analysis for the phenomena
being studied. The phenomena may be physical, such as natural disaster,
but the study of such phenomena will focus on such aspects as people’s
response, people’s attitude, and their other aspects of behaviour in
relation to those phenomena. It is to be noted however, that this term
does not include all people per se, even if they were staying in the same
house. It only includes those people who have something in common in
relation to the aspect of the phenomenon the researcher is studying.
Therefore, the term ‘population’ refers to all and only the people the
researcher can potentially obtain information from, as in interviews, or
about, as in observation.
Consider the problem of internal displacement, for example. If the
researcher is investigating, say, the effect of internal displacement on
 Population and Sampling 37

family health, the researcher’s population will consist of all people from
displaced families. However, the researcher will immediately discover
that not everybody in this population will be useful for this study. On the
contrary, it will be found that some members from these families cannot
provide useful information. It may also happen that some displaced
families moved far away and are not easily accessible. Oftentimes
therefore, one may not know the exact population size.
Where the population of the study is too big and widely dispersed
or even unknown, the question of population size may not be a problem.
The rules of probability will prove that the size is not that relevant. The
issue of population size may thus be safely ignored if it is too large
and/or unknown. The researcher therefore targets only some identifiable
category (or categories) from this population—those that are more ‘tied
together’ by very specific characteristics. These may be family heads,
members of families within a given area, members of families with
particular health problems, etc. These targeted people will constitute the
target population. Clearly then, researchers—especially in fairly large
scale studies such as surveys—have to move from the general relevant
population down to the target population to select the subjects (people)
to participate in the study.
It should be noted, of course, that research is generally always
hampered by problems of money, time, distance, weather, accessibility,
and other logistical, political or technical challenges. Even if the
researcher selected his or her target population, it might still be difficult
for him or her to reach every single subject in that population. The
subjects are still too many, and may due to various reasons be difficult
to access. Therefore, the researcher must aim at an accessible proportion
of the target population. This proportion is called a sample, and it is only
this that will provide the data needed in the study.

From Population to Sample

Since, as we have seen, it is quite difficult to reach everyone even
within the target population, the researcher has to plan and select a small
portion from that population to participate in the study. The process of
selecting the participants is called sampling and the selected participants
constitute the sample. The actual number of the subjects within the
sample is referred to as the sample size. As I will show in the this
38 Social Research Methodology: An Introduction

section, the sampling process is hardly ever a simple one due to the often
complex population dynamics.

Population Size
It is extremely important for the researcher to consider the number of
people there are in the population the sample is intended to represent.
This may be the number of people in a village or slum area being
studied, the number of people who have voting cards, the number of
people affected by a natural disaster, etc. Of course, for various reasons,
the researcher may oftentimes not be able to establish the exact size of
the population. There are instances where the population size matters
and others where it does not. If, for instance, the population is very
large, very scattered, and therefore unknown, its size becomes irrelevant
and may be ignored. However, if the population is relatively small and
known, such as a group of University students preparing for
supplementary exams, it must be taken seriously.

Sample Size
The larger the sample, the more sure the researcher can be that their
findings from that sample truly reflect the population. The most
important question concerning sampling therefore, is one of the sample
size required for a study. The question of sample size is influenced by
a number of factors, including most importantly the population size and
the purpose of the study. There are three criteria that a researcher should
follow to determine the appropriate sample size. These are (i) the level
of precision required to ensure representativeness of the sample to the
population, (ii) the level of confidence or risk that must be worked with
to ensure only the allowable sampling error, and (iii) the degree of
variability in the attributes being studied in order to guard against a
“bad” sample.

The level of precision

This is sometimes also called the sampling error, and is the range in
which the true value of the population is estimated to be. This range is
often expressed in percentage points (e.g., ±5 percent).
 Population and Sampling 39

The level of confidence

This is also called the level of risk, and is based on the major ideas of the
Central Limit Theorem. These are first, that when a population is
repeatedly sampled, the average value of the attribute obtained by those
samples is equal to the true population value; and secondly, that the
values obtained by these samples are distributed normally about the true
value, with some samples having a higher value and some obtaining a
lower score than the true population value. This theorem is discussed in
a good measure by Dudley (1999).

The degree of variability

The degree of variability in the attributes being measured refers to the
distribution of attributes in the population. The more heterogeneous a
population is, the larger the sample size required to obtain a given level
of precision. The less variable and more homogeneous a population, the
smaller the sample size required.

Sample Size Determination

Although, as I discuss in the following section, there are specific
sampling techniques both in qualitative and quantitative research, in
cases where the population size is difficult to determine, there are a
number of alternative measures a researcher may resort to. These are
outlined below.

Using a census
Where the study is dealing with a small population, the researcher may
choose to use the entire population as the sample. A census eliminates
sampling errors and provides data on all the individuals in the
population. Sampling the entire population also enables the researcher
to achieve a desirable level of precision.

Using a sample size of a similar study

A researcher may use the same sample size as was used in previous
similar studies. However, this requires first reviewing the procedures
that were employed in those studies, otherwise he or she risks a repeat
of possible errors that were made in determining the sample size for
40 Social Research Methodology: An Introduction

those studies. A review of the literature in one’s discipline can provide

guidance about “typical” sample sizes.

Using published tables and formulas

A third way to determine sample size is to rely on published tables
which provide the sample size for a given set of criteria; and the fourth
one is to calculate the necessary sample size for a different combination
of levels of precision, confidence, and variability through the application
of one of several available formulas. These tables and formulas are well
discussed in the University of Florida’s Institute of Food and
Agricultural Sciences document PEOD6, version of June 2003. This
document is available at the EDIS Web Site at http://edis.ifas.ufl.edu.

Sampling Techniques in Social Research

So a social researcher relies on data collected from a sample of subjects
drawn from a target population to study the problem that is affecting the
population. Depending on the nature of the problem and kind of data
required for his or her analysis, there are different sampling techniques
at his or her disposal to select from. Sampling techniques can generally
be grouped into two: (i) probability sampling that is adopted in
quantitative research, and (ii) non-probability sampling that is used in
qualitative research.
Probability sampling techniques are structured ‘scientific’
approaches, which take great care and spend more time calculating for
appropriate sample sizes and sampling errors. They are based on the
principle of randomness, whereby each item or element within the
population has an equal chance of being selected into the sample.
Probability sampling is highly sought after in large, structured,
quantitative studies for the single most important reason of reducing or
eliminating bias. Bias arises if the procedure used to create the sample
results in a sample that is systematically different from the population.
Take an example of a researcher conducting a study on people’s
attitudes towards the removal of Presidential term limits from the
constitution. Suppose the researcher publishes a questionnaire in the
local daily for people to tear off, fill and mail back, this would create a
highly biassed sample. The researcher has no knowledge and control
over who is actually filling the questionnaire. The sample is thus likely
 Population and Sampling 41

to be dominated by the agitators and sponsors of the removal of term

limits, who have the interest and therefore the time to fill out and mail
the questionnaire back. It should be noted that it is the procedure of
creating the sample rather than the sample itself that defines the bias.
Nonprobability sampling is generally less rigorous and relatively
less expensive than the probability, structured approaches. It is also
oftentimes so easy and convenient as to be done instantaneously.
However, the major weakness of non-probability techniques lies in the
fact that the researcher sets out to collect data without the real picture of
who and how many the respondents would be, and where they would be
found. For that matter, the research cannot generalise the findings based
the data collected from such samples to the population.

Qualitative Sampling Techniques

We noted that qualitative research is naturalistic, and interests itself in
the direct understanding of people, events or phenomena in their natural
settings and contexts. A qualitative researcher does not seek respondents
to ‘represent’ the entire population or to ‘generalize’ the findings
generated from such respondents to the entire population. Rather, he or
she directly interacts with respondents to understand their personal
experiences, fears, insights and other intrinsic qualities. The approach
taken to determine and recruit such respondents therefore, is likewise
less rigid, less structured and more reliant on the personal judgment of
the researcher. There are commonly four techniques in this category:
purposive sampling, quota sampling, snowball sampling, and accidental
sampling.

Purposive sampling
This technique, also sometimes referred to as judgmental sampling, is
where a researcher on his own judgment targets specific subjects to
participate in the study. The basis of his or her judgment is their
perceived knowledge or experience in relation to the problem under
investigation. This perception may be as a result of their level of
education and training, the nature of the office they hold, their position
in society, or the experience one has accumulated or gone through.
Suppose a researcher is investigating the problems associated with
indigenous capital accumulation and investment in an unstable economy
42 Social Research Methodology: An Introduction

such as Uganda. Rather than distributing a standardized questionnaire

to a randomly selected sample of industrialists and industrial workers,
the researcher would purposively seek and interview seasoned
industrialists such as the Mulwanas, Mukwanos or Wavamunnos who
have seen it all. Another good example could be of a study on the
traumatic experiences associated with HIV/AIDS where the researcher
would look out for counsellors in AIDS support organizations and
activists leaving with AIDS.

Quota sampling
This is a technique in which the population is divided into groups or
strata. The division is on the basis of the characteristics of that
population that are of importance to the study. The researcher’s foremost
task it to establish the proportions of these strata. The number of
subjects to interview in each stratum (called the quota) is set to be
proportionate to the representation of that stratum in the entire
population.
Take an example of a study of the attitudes of ‘veteran’ politicians
across the traditional political parties towards the current political trend
under the Movement regime in Uganda. The researcher would select the
area of the study, e.g. Kawempe South. He or she would then find out
from the registers at the party branch offices the membership of these
parties in order to establish their proportions to the entire population of
the area. Suppose the findings are that the DP constitutes 43%,UPC
41%, CP 12% and UPM 4%, when the researcher sets out to identify
these politicians in the field to interview, the total number of
interviewees, the quotas assigned to each party, should reflect this
proportionality.

Snowball sampling
In this technique, the researcher identifies a few individuals on the basis
of a rare characteristic they possess to participate in the study. They
could, for instance, be characterized by a shared secretive behaviour
(e.g. homosexuality) or secret identity (e.g. possessing a mark acquired
from devil worship). After successfully building rapport with them and
interviewing them, they help the researcher to identify other suitable
 Population and Sampling 43

subjects to participate in the study, and the process goes on until the
researcher is satisfied that he or she has got a sufficient sample.
Snowball sampling therefore is a phased sampling technique:
selection and interviewing of subjects is done in stages. The first stage
involves the identification and interviewing of the first few subjects
(they could even be as few as three!), who then become informants
about other potential subjects. The second stage involves tracing those
ones and interviewing them, who also help to identify yet others. The
third stage proceeds to identify and interview others until the researcher
deems the sample sufficient. As we have already indicated, this is the
most appropriate technique to select samples to participate in the study
of secretive behaviour.

Accidental sampling
This technique is also called convenience sampling. Under it, any
subjects that are part of the target population and are within easy reach
of the researcher are recruited into the sample. For example, in a study
of the problems faced by self-sponsored students in public universities,
the researcher may stand by any lecture hall and interview any private
student that comes by. If the researcher is aware of specific areas where
he or she is sure to particularly find such students, that would be the best
place to position himself or herself. If the researcher was at Makerere
University, for example, he or she would be sure to find such students
in a peculiarly winding queue at Stanbic bank going to pay their fees. He
or she would then interview any of them on a wide range of problem
issues until a “satisfactory” number is achieved.
This technique is the most appropriate to select participants for
studies where the phenomena under study are fast changing or passing
and there is no time to construct a sampling frame. It is commonly used
in journalism. Take an example of the funeral of late Palestinian leader
Yasser Arafat or Pope John Paul II: a radio or TV journalist would use
this sampling technique to recruit subjects from who to capture the
people’s emotions and the general mood at the funeral. If he or she
attempted to do this even one day after the funeral, he or she would hit
a dead end—the steam had evaporated. Effective reportage has to be at
the spur of the moment, and therefore people to form the sample for
interviewing have to be recruited there and then. Take another scenario
44 Social Research Methodology: An Introduction

of Makerere University students demonstrating after a speeding vehicle

knocked dead a fellow student, would a reporter have time to construct
a sampling frame and constitute a formal sample? Or would he or she
return two days later?

Quantitative Sampling Techniques

Among the major decisions that researchers have to make when they
design a quantitative study is how to select participants, and how many
of them to select. Quantitative researchers employ probability sampling
and ensure that each individual in the target population stands an equal
chance of being included recruited. In a target population of 100
subjects, each of the subjects has one out of 100 chances of being
selected. There are five commonly used probability sampling techniques
in social research: simple random sampling, systematic sampling,
stratified sampling, cluster sampling, and multi-stage sampling.

Simple random sampling

This technique aims to afford each individual in the target population an
equal chance of being selected. To obtain a simple random sample, a
researcher must first construct a sampling frame. This is a listing of all
the members of the target population from whom the sample is to be
drawn. After construction of the sampling frame, there are two methods
of determining who on that frame is drawn into the sample (i) the rotary
method, and (ii) the use of the table of random numbers.
The rotary method is preferred in a study where the target
population is small, e.g., an institution. The researcher writes the name
or any other identifier of each member of the target population on a slip
of paper (remember that not every member of the institution may be part
of the target population). The slips of paper are then folded and placed
in a container such as a box, shuffled and mixed thoroughly. They are
then randomly picked, one at a time. Each one picked identifies the
individual that is chosen to be part of the sample.
The use of the table of random numbers is common in studies that
are based on relatively large populations. This table is produced using
a computer, and the way the numbers are grouped is purely random. (A
good example of this table is reproduced by K.D. Bailey, 4th ed., 1994,
Appendix A, at pp 527-30. These numbers have multiple digits. To use
 Population and Sampling 45

these numbers, the researcher proceeds as follows: Each member of the

target population is given a number. The numbers assigned are selected
according to the sample size required. For example, if the researcher
wants to select 30 out of a population of 90, the numbers selected will
be the last two digits: 01,02,03 . . . 10, 11 . . . 90. For a population of
over 100, it will be the last three digits: 001, 002, etc. The table of
random numbers is consulted beginning randomly at any number in the
table.

Systematic sampling
This method is sometimes called interval sampling, and is often used
where random sampling is difficult to pursue. However, like random
sampling, it is used if a sampling frame is available and if it was
compiled in an unbiased manner. For example, if the researcher is
constructing a sampling frame of NGOs in Uganda that benefited from
the Global Fund, he or she does not list the alphabetically but just
anyhow.
There are two ways of employing this method of sample selection.
The first one is to blindly choose the starting point on the list. The
number should be between 1 and 9. If the number chosen is say 3, the
NGO identified by that number is selected, and thereafter every third
next NGO is selected until the sample is raised. The second one is to
compute the interval to be used. This is computed as follows: Suppose
the target population (sampling frame) is of 200 NGOs, and the sample
size required is 25. The researcher divides 200 by 25 to get an interval
of 8. This means that he or she will select every 8th NGO on the list
until the sample is raised. Where to start from should be a random
number between 1 and 10, which should be blindly chosen. If the
blindly selected number is 6, then select 14, 22, 30, 38, etc.

Stratified sampling
In cases where there is great variability within the population, the target
population can be divided into homogeneous groups on the basis of the
major attributes or traits necessary for the study. This is called
stratification, and the strata may be, e.g., male, female; school-going,
non-school going; alien, citizen, etc. The strata could even be three or
46 Social Research Methodology: An Introduction

more. Sampling is then done randomly group by group, and may either
be done proportionately or disproportionately.
In proportionate sampling, the researcher takes note of the total
number of individuals in the target population, and then the total number
in each stratum. Take for example a study where the strata are men and
women. If the target population is 365 people of whom 219 are women
and 146 men, the researcher will proceed as follows:

Women = 219 x 100 = approx. 60%

365

Men = 146 x 100 = approx. 40%

365

The final sample will be made up of 60% women and 40% men. This
means 60% of 219 women (approx. 131) and 40% of 146 men (approx.
58). Each of these participants is selected randomly, and the sample size
will be 131 + 58 = 189. In disproportionate sampling, the researcher
purposively takes more participants in some strata than in others.

Cluster sampling
This method is also called area sampling and it involves random
selection of areas (clusters) that are already well determined. This
method is particularly appropriate in the following situations: (i) where
the researcher cannot obtain a sampling frame, (ii) where the researcher
has little knowledge of the characteristics of the individuals in the
population, and (iii) where the target population is scattered over a wide
geographical area. To obtain a cluster sample, the researcher allocates
a number to each identified cluster and then using the rotary method, he
or she selects the required clusters. The people in each of the randomly
selected clusters constitute the sample.
This approach allows to overcome the constraints of costs and time
associated with a very dispersed population. If, for example, one wanted
to research on police response to reports of sexual offences in Kampala
district, he or she could decide that each police station represents one
cluster. If there are, say, 8 police stations, he or she may randomly select
3, and then contact the detectives in each cluster for interviews. When
 Population and Sampling 47

all the officers at the selected clusters are interviewed, this is referred to
as one-stage cluster sampling. If the subjects to be interviewed are
selected randomly within the selected clusters, it is call two-stage cluster
sampling. The two-stage cluster sampling technique is more appropriate
if the number of subjects within a cluster is very large.

Multistage sampling
Multistage sampling is any sampling where the subjects are chosen at
more than one stage. It may involve more than one method of sampling
used in combination, such as simple random and cluster sampling or
stratified and simple random sampling. Additionally, multistage
sampling uses a hierarchy of sampling units or clusters, so that at each
stage each subject belongs to only one sampling unit. The sampling unit
or cluster selected in the first stage is called the primary sampling unit,
and is very important, especially in computing variance.
Multistage sampling is a complex form of cluster sampling. Using
all the sample elements in all the selected clusters may be prohibitively
expensive or not necessary. Under these circumstances, multistage
cluster sampling becomes useful. Instead of using all the elements
contained in the selected clusters, the researcher randomly selects
elements from each cluster. Constructing the clusters is the first stage.
Deciding what elements within the cluster to use, is the second stage.
The technique is used frequently when a sampling frame is impossible
to construct.
Chapter Four

Ethical Conduct of Social Research

Introduction
Social research is an inter-sectoral undertaking, involving government,
academia, commercial, voluntary and non-profit organizations. It is also
interdisciplinary, cutting through sociology, psychology, management,
economics, political science, law, business, social work, and education.
It is therefore international and multi-problem based. This calls for
methodological innovation. New methods, however, pose new ethical
problems.
One single most important milestone in social research in recent
times has been the development of modern codes of ethics. The
importance of ethical codes of conduct in research involving human
participants was highlighted by remarkably ever increasing incidents of
unethical practices. Today, many countries have published codes of
ethical conduct in social and biomedical research, often observing a
succession of recent international documents. The international
documents demonstrate a trend towards making more explicit the ethical
standards, which must be met if research on or involving human beings
is to be ethically acceptable.
It is now widely accepted that all kinds of research involving or
impacting upon humans should conform to the highest standards of
academic integrity and ethical practice. The ‘humans’ in question
include all manner of people other than the community of professional
colleagues with special interests in the research. These are individuals,
groups or collectivities that may be affected in one way or another in the
course of conducting the research or communicating the results of
research.
 Ethical Conduct of Social Research 49

Normative Standards in Social Research

Ethical principles and normative standards extend to all spheres of
human activity. When people deal with each other, with animals, or with
the environment, they must be guided by – and are in fact expected to
exhibit – a certain minimum standard. Normative standards serve to
identify good, desirable or acceptable conduct and to justify it.
There are three universally acceptable norms that should govern all
social research. They are beneficence, respect, and justice (The Belmont
Report, 1979). These norms are intertwined in the entire research
process, and guide every step a researcher takes in the determination of
the research problem and objectives, selection of methods and subjects,
administration of research instruments, as well as the decisions
regarding interpretation of findings. They are at the heart of every
obligation a researcher must meet if the research has to pass the ethical
test.
Beneficence refers simply to the researcher’s obligation to
maximize good outcomes of the research while minimizing unnecessary
risk or harm. Harm, in this context, extends beyond physical harm to a
wide range of psychological or emotional distress, discomfort and
economic or social disadvantage. Researchers exercise beneficence in
assessing the risks of harm and potential benefits to participants, in
being sensitive to the rights and interests of people involved in the
research, and in reflecting on the social and cultural implications of their
work. Alongside this norm is the researcher’s obligation to do no harm,
and this is referred to as non-maleficence.
The norm of respect entails respect for the subjects of research, and
refers specifically to the protection of their autonomy. This norm
requires that individual research subjects be treated as autonomous
agents, and that persons with diminished autonomy be entitled to
protection. The overall import of this is that a researcher must have a
duty to a subject of research and a commitment not to use that subject
merely as a means to an end.
The norm of justice addresses the question of who ought to receive
the benefits of research, and who should bear its burdens. It means that
the researcher ensures reasonable, non-exploitative and carefully
considered research procedures with fair distribution of costs and
benefits. Questions of justice can also arise in relation to the use of
50 Social Research Methodology: An Introduction

public funds for research or balancing the interests of the external

funding agents.

Ethical Values and the Researchers’ Obligations

The pervading value in social research is that of integrity, which is
manifested in four principles. The first principle of integrity concerns
the integrity of the researcher. This includes: i) commitment to research
questions or hypotheses that are designed to contribute to knowledge, ii)
commitment to the pursuit and protection of truth, iii) commitment to
research methods and procedures that are appropriate to the discipline,
and iv) honesty. The second principle relates to the preservation of the
integrity of scholarship and research. The third focuses on the promotion
of the integrity of institutions and organizations responsible for research.
Finally, accountability both to the general public and to specific
individuals, groups, or collectivities is an important principle of integrity
a researcher must always uphold.
Researchers have an obligation to conform to the ethical values and
standards of the society in which they conduct their work. But they also
have a duty to ensure that they are informed about the legal or policy
framework in which their research has to be conducted and how that
framework might affect the conduct of their research. In Uganda,
research is cleared directly by the National Council of Science and
Technology (NCST) although “small” projects such as students’ work
may be cleared on its behalf by a designated research institution such as
a University, college or specialised research institute.
There are generally four categories of obligations all researchers
must meet. These are obligations to society as a whole, obligations to
faculty and colleagues, obligations to employers and sponsoring
agencies, and obligations to research subjects. Let us examine each of
these in turn.

Obligations to Society
If social research is to be of benefit to society as a whole and the groups
and individuals within that society in particular, the researcher must
conduct the study with utmost professional responsibility, taking
cognisance of the moral and legal order of a given society. Here below
I outline four specific obligations a researcher owes to society.
 Ethical Conduct of Social Research 51

Competence and the maintenance of high scientific standards

The researcher (or research team) must intellectually and professionally
measure up to the research task at hand. Similarly, the methods
employed in the collection and analysis of data must be appropriate and
valid. The researcher must not engage or collude in selecting methods
designed to produce misleading results, or in misrepresenting findings
by commission or omission.

Pursuit of objectivity
Social research is conducted within the value systems of given societies.
So what if the researcher comes from a society with different value
systems? Take an example of a ‘born-again’ novice researcher
conducting an interview in a polygamous household reeking of alcohol
and tobacco smell. Would she allow her radical religious values a place
in the evaluation of the interview? Researchers should at all times
uphold their professional integrity by remaining as impartial as possible,
especially when assessing and interpreting qualitative findings.

Widening the scope of research and reporting findings

Social researchers should use the possibilities open to them to extend the
scope of social enquiry and communicate their findings for the benefit
of the widest possible community. The reporting should be complete,
accurate, wide, and objective, and should contain full information on
any innovative methodologies employed. This serves to allow research
work to be assessed by colleagues and to increase public confidence in
its reliability.

Balancing of interests
Social enquiry is predicated on the belief that greater access to well
grounded information serves rather than threatens the interests of
society. However, in practice, societal interests are an aggregation of
interests of different sections of that society. And oftentimes these
interests are in conflict. Accessing information that may be deemed of
benefit to society may at the same time be offensive to some sections of
that society. Therefore, a researcher should throughout the research
project keep in mind and be able to balance the various conflicting
interests and gauge the likely consequences for society at large, groups
52 Social Research Methodology: An Introduction

and categories of persons within it, subjects, and possible future

research.

Obligations to Sponsors and Institution of Affiliation

A researcher’s relationship with and obligations to the sponsors or the
organization he or she is affiliated to should be clear and balanced. They
should not compromise his or her commitment to the standards
commensurate with professional integrity. Here below are three specific
duties a social researcher owes to these organizations.

Clarification of obligations and roles

A social researcher should clarify in advance his or her obligations to the
sponsor and the organization he or she is affiliated to, as well as the
obligations of these organizations themselves. He or she may, for
example, refer them to the relevant parts of a code of ethics to which
they are supposed to adhere. And due acknowledgement must be made
to these organizations for their respective roles in the research.

Impartial assessment of alternatives

A social researcher should consider the available methods and
procedures for addressing a proposed inquiry and should provide the
sponsoring and/or employing body with an impartial assessment of the
respective merits and demerits of alternatives.

Protection of privileged information

Sponsoring bodies and employers normally allow researchers access to
information that they may legitimately require to be kept confidential.
The researcher must oblige. However, this excludes information about
methods and procedures that have been utilised to produce published
data.

Obligations to Colleagues
Social research thrives on the maintenance of standards and appropriate
professional behaviour. This requires that methods, procedures, and
findings be open to collegial review without compromising obligations
to sponsors and employers, subjects or society at large. It also requires
 Ethical Conduct of Social Research 53

concern for the safety and security of colleagues and research assistants
when conducting field research. Below are four specific obligations.

Maintenance of confidence in the research

A social researcher should always endeavour to promote and preserve
public confidence in the study without exaggerating the accuracy or
direct benefits of their findings.

Exposure for review of methods and findings

A social researcher should provide adequate information about the
methods, procedures, techniques and findings to colleagues for
assessment. The assessment, however, should be directed at the methods
themselves rather than at the individuals who selected or used them.

Communication of ethical principles

It is sometimes necessary for a researcher to collaborate with colleagues
in other disciplines, as well as interviewers, clerical staff, and other
people in the conduct of certain studies. In these cases social researchers
should make their own ethical principles clear and take account of the
ethical principles of their collaborators.

Protection of field researchers

A social researcher must always endeavour to ensure the safety of the
field researchers. Simultaneously, he or she has a moral obligation to
attempt to minimise the risk of physical, mental or emotional harm to
him or herself, to their colleagues or research assistants from the conduct
of research.

Obligations to the Subjects

A social researcher must always strive to protect the subjects of research
from undue harm arising as a consequence of their participation in the
research. This requires that the subjects’ participation be voluntary and
fully informed. It also requires that their various rights and values, be it
individually or collectively, are fully respected by the researcher.
Attention must thus be paid to the following five specific obligations.
54 Social Research Methodology: An Introduction

Obtaining informed consent

Social research should as far as practicable be based on the freely given
informed consent of subjects. A researcher should not put the subjects
under the impression that their participation in the research is
mandatory. The subjects should be made aware of their right to
withdraw from participation at any stage for whatever reason and to
withdraw the information just supplied. Gaining informed consent is
necessary to ensure that the research subjects understand what is being
done to them, the limits to their participation and awareness of any
benefits and potential risks involved.

Avoidance of undue intrusion

A social researchers must keep in mind the intrusive potential of social
enquiry, but also the fact that he or she cannot as of right investigate
every social phenomenon. There are a number of political, social,
cultural and psychological contexts that pose ethical challenges to a
researcher’s licence. I will turn to these in some detail shortly.

Protection of the subjects from harm

A social researcher has a duty to protect the subject against potentially
harmful effects of participation. This duty is irrespective of the subject’s
free consent or any legal requirement for the subject to participate. A
researcher may do this by providing the subject with prior and adequate
information about the consequences of participating in the study.
Harm to the subjects may arise from undue stress through
participation, loss of self esteem, psychological injury or other side
effects. There are various factors a researcher should consider in
carrying out a risk-benefit assessment of a particular inquiry. These
include, among others, the probability of risk, the number of people at
risk, the severity of the potential harm, and the anticipated utility of the
findings. A number of such factors are not quantifiable.

Sensitivity to socio-cultural values

A social researcher has a duty to remain sensitive to, and cognisant of,
the social and cultural values that are different from his or her’s and all
other conflicting interests. He or she should therefore avoid any
questions and practices that may cause distress or offence.
 Ethical Conduct of Social Research 55

Confidentiality of information
Research data should be unconcerned with individual identities. The
identities and records of subjects must therefore be kept confidential,
whether or not confidentiality has been explicitly pledged. In social
research, there are genuine public concerns about privacy and the
storage of information of a personal nature. This information falls into
three categories: identified information, re-identifiable (coded or
potentially identifiable) information, and de-identified (not re-
identifiable, anonymous) information.
Identified information is information that allows the identification
of a specific individual. It comprises such data as the individual’s name,
date of birth, address, etc. Re-identifiable information is that one whose
identifiers may be removed and replaced by a code. In such cases it is
possible to use the code to re-identify the person to whom the data
relate. Personal information is de-identified if the identifiers have been
permanently removed or if the information has never been identified. In
that case, any attempt at identification or re-identification is glaringly
futile. Researchers have a responsibility to subjects in respect of each
category of personal information, and greater care is required where the
research involves identified and re-identifiable information.

Some Challenges to Consent and Confidentiality Principles

Developments in information technology in the recent past have
rendered the principle of anonymity hard to safeguard. The release of
non-anonymized data, such as in sharing data between governmental
agencies when the identities of individuals could be discovered, should
be agreed with participants in advance. Nevertheless, different forms of
social enquiry raise different ethical challenges. Technical or practical
considerations may oftentimes inhibit the achievement of prior informed
consent or adequate confidentiality. Here below I discuss some specific
examples.

Rights in observation studies

In observation studies, where behaviour patterns are observed without
the subjects’ knowledge, social researchers must take care not to infringe
the “private space” of an individual or group. Where practicable, social
researchers should attempt to obtain consent post hoc. In some cases, the
56 Social Research Methodology: An Introduction

subjects may be aware that they are being observed, and may feign
behaviour. In such a case, the researcher should interpret behaviour
patterns that appear deliberately to make observation difficult as a tacit
refusal of permission to be observed.

Dealing with proxies

In cases where a proxy is utilised to answer questions on behalf of a
subject, say because access to the subject is uneconomic or because the
subject is too ill or too young to participate directly, care should be taken
not to infringe the ‘private space’ of the subject or to disturb the
relationship between subject and proxy. Where indications exist or
emerge that the subject would object to certain information being
disclosed, such information must not be sought from the proxy. For
example, a researcher asking a wife to answer for the husband who is
not easy to locate or who would not consent should be careful about any
questions that the husband would not answer.

Secondary use of records

In cases where subjects are not approached for consent because a social
researcher has been granted access, say, to administrative or medical
records or other research material for a new or supplementary inquiry,
the permission to use the records should not substitute the researcher’s
responsibility to consider potential reactions, sensitivities and interests
of the subjects concerned. Where possible and appropriate, the subjects
could be approached afresh for consent to any new enquiry.

Duping potential subjects

In studies where the objectives preclude the prior disclosure of material
information to the subjects, social researchers must weigh up the likely
consequences of any proposed deception. To withhold material
information from, or to misinform the subjects involves a deceit,
whether by omission or commission, temporarily or permanently.
Instead, consideration should be given to informing the subjects in
advance that material information is being withheld, and when or if such
information will be disclosed. A serious problem arises for social
researchers when methodological requirements conflict with the
requirement of informed consent. Many cases exist in which the
 Ethical Conduct of Social Research 57

provision of background information to the subjects (say, about the

purpose or sponsorship of a study), or even the process of alerting them
to the fact that they are subjects (as in observation studies), would be
likely to produce a change or reaction that would defeat or interfere with
the objective of the study. These difficulties may lead social researchers
to waive informed consent and adopt either covert observation
techniques or deliberate deception in the interests of accuracy.

Compliance and Regulation Mechanisms

Freed-Taylor (1994) has broadly categorized into two the mechanisms
that have been put in place to ensure that these concerns are confronted.
The first category consists of externally imposed mechanisms such as
legislation, administrative and contractual arrangements, sanctions, and
the implementation of technical solutions. The second category consists
of mechanisms that assume the role of internal regulation, such as ethics
committees (also called institutional review boards), educational
programmes, and the development of codes of professional conduct.
These solutions are, of course, closely interrelated. The existence of one
solution may, in some cases, result from the lack of another, or may
inhibit the development of alternative solutions. The existence of
stringent legislation may, for example, make the development of more
self-regulatory trends difficult or impossible. Effective self-regulation
may, on the other hand, make legislative or punitive sanctions
unnecessary. While laws and regulations may avert the worst abuses of
ethical principles, they cannot produce positively ethical research.
Responsibility for this lies with the individual researcher, and it is
education and sensitization that will ensure that responsible research is
carried out.

Towards Ethically Responsible Research

I want to reiterate that self-regulation is the single most important
guarantee of responsible and ethically positive social research. It is
advisable therefore that researchers routinely check their projects, from
conceptualization to completion, to ensure ethical standards. I am
providing a summary of the major indicative questions and what would
be expected of the researcher to act as a checklist for ethical research.
This checklist is a mere guide and does not claim to be exhaustive.
58 Social Research Methodology: An Introduction

Indication Expectation

Title of research Should provide the broad theme and specific area of
the research, and so offer a quick reference for any
interested party.

Purpose and Clearly state the aims and objectives of the research,
significance of and indicate major questions, hypothesis testing,
research policy evaluation, and any potential “value” added to
the discipline or community.

Institutional affiliation Should be indicated for purposes of contact, assurance,

accountability and future collaboration.

Research team Give a list of names of all researchers, their

qualifications and functions in the proposed project.
This offers an estimate of competence together with a
chain of responsibility and accountability.

Funding agency Indicate organisation, individual or group providing

the finance for the study. This clears any suspicions as
to motives.

Design Indicate what will be done and the role of the subjects.
All procedural matters, especially research clearance,
should be clarified. Time commitments and data-
collection settings should be revealed. Data analysis
methods and procedures should also be clarified.

Risk-benefit equation Highlight the potential risks to the subjects—whether

physical, psychological or otherwise. Highlight the
possible benefit to the subjects or society as a whole.
What procedures have been established for the care
and protection of subjects (e.g. insurance, medical
cover) and the control of any information gained from
them or about them?

Subjects Indicate whether the subjects will be recruited with

their consent or if they might be “obliged” to
participate – as in the case of students, prisoners, or
patients. If participation is compulsory, the potential
consequences of non-compliance must be indicated to
the subjects.
 Ethical Conduct of Social Research 59

Consent Should indicate whether and how the subjects can

withdraw consent. Where appropriate, consent of
participants should be requested and put in terms
easily comprehensible to lay persons. This should
ideally be both orally and in writing.

Confidentiality Indicate the steps taken to safeguard the confidentiality

of records and any potential identifying information
about the subject must be revealed.

Monitoring Organisational procedures for monitoring the project

should be available for inspection.

Expected duration Give time scheduling of project. This indicates the

commitment required of the subjects and the
researchers.

Dissemination Indicate the anticipated use of the data, forms of

publication and dissemination of findings.
 Part Two

Qualitative Research Methods

Chapter Five

Qualitative Research Designs: An Overview

Introduction
We saw in chapter one that the various ways in which research can be
classified. In terms of paradigmatic classification, we saw that research
can be qualitative, quantitative or a mixture of the two. Qualitative
researchers can be found in many disciplines and fields, using a variety
of approaches, methods and techniques. Qualitative approaches for
collecting empirical data range from interviews (singly or collectively),
observational techniques, focus group discussions, participatory
appraisal, through to archival research, all aimed to enable the researcher
understand and explain social phenomena. These methods are the
subject of chapters six to ten.
Qualitative research is a generic term for investigative
methodologies described as ethnographic, naturalistic, anthropological,
field, or participant observer research. It emphasizes the importance of
looking at variables in the natural setting in which they are found.
Interaction between variables is important. Detailed data is gathered
through unstructured tools such as open-ended questions that provide
direct quotations. The interviewer is an integral part of the investigation.
This differs from quantitative research, which attempts to gather data by
objective methods to provide information about relations, comparisons,
and predictions and attempts to remove the investigator from the
investigation.
Qualitative research is characterized by a number of important,
distinct features. The first is holistic description. During qualitative
research, the investigator seeks to gain a complete picture of reality. A
 Qualitative Research Designs: An Overview 63

holistic description of events, procedures, and behaviours occurring in

a natural setting is often needed to make accurate situational decisions.
The second is corroboration, the purpose of which is not to confirm
whether people’s perceptions are accurate or true reflections of a
situation but rather to ensure that the research findings accurately reflect
people’s perceptions, whatever they may be. Corroboration helps the
researcher to enhance the probability that the findings will be seen as
credible or worthy of consideration by others. The third feature is
triangulation, which is one process involved in corroboration.
Triangulation involves the convergence of multiple data and sources, the
convergence of data collection techniques, or involving multiple
researchers in an investigation.
Just as there are various philosophical perspectives that inform
qualitative research, so there are various qualitative research designs. As
we learnt in chapter one, a research design is an overall strategy of
inquiry that moves from the underlying philosophical assumptions to a
specific research method and data collection techniques. The choice of
a research method influences the way in which the researcher collects
data. Specific research methods also imply different skills, assumptions
and research practices. The focus of this chapter is to provide an
overview of the qualitative research designs applicable in social
research. Six research designs are discussed here: action research, case
study research, historiographic research, phenomenological research,
ethnographic research, and grounded theory.

Action Research
Action research, also variously known as participatory research,
collaborative inquiry, emancipatory research, action learning, and
contextual action research, may simply be viewed as learning by doing.
There are various definitions of action research, but one of the most
widely cited is that of Rapoport, who defines it in the following way:

Action research aims to contribute both to the practical

concerns of people in an immediate problematic situation and
to the goals of social science by joint collaboration within a
mutually acceptable ethical framework (Rapoport, 1970:499).
64 Social Research Methodology: An Introduction

This definition draws attention to the collaborative aspect of action

research and to possible ethical dilemmas which arise from its use. It
also makes clear, as Clark (1972) emphasizes, that action research is
concerned to enlarge the stock of knowledge of the social science
community. A group of people, in which the researcher takes an active
role as a facilitator or process manager, identify a problem, do
something to resolve it, review how successful their efforts were, and try
again. While this is the essence of the approach, there are other key
attributes of action research that differentiate it from professional
practices, consulting, or routine common problem-solving activities.
The major attribute of action research is its emphasis on scientific
study—the researcher investigates the problem systematically and
ensures the intervention is informed by theoretical considerations. He
or she refines the methodological tools to match them to the problem,
collects data, analyzes and presents it on an ongoing, cyclical basis. The
other attribute is turning the people involved into researchers. The third
attribute is its social dimension—the research takes place in real-world
situations, and aims to solve real problems. Finally, the researcher
makes no attempt to remain objective, but openly acknowledges his or
her bias to the other participants.
Action research derives its unique flavour from a set of principles
that guide the research. Winter (1989) provides a comprehensive
overview of six key principles, which I summarise below.

1. Reflective critique. An account of a situation through notes,

transcripts or review of official documents, will make implicit claims to
be authoritative. This principle ensures that people reflect on issues and
processes and make explicit the interpretations, biases, assumptions and
concerns upon which judgments are made. In this way, practical
accounts can give rise to theoretical considerations.

2. Dialectical critique. Social reality is consensually validated and

shared through language. Social phenomena are thus conceptualized in
dialogue. Therefore, a dialectical critique is required to understand the
set of relationships both between the phenomenon being studied and its
context, and between the elements constituting that phenomenon.
 Qualitative Research Designs: An Overview 65

3. Collaborative resource. Participants in an action research project are

co-researchers. The principle of collaborative resource presupposes that
each person’s ideas are equally significant as potential resources for
creating interpretive categories of analysis.

4. Risk. The change process potentially threatens the status quo of doing
things in the organization and creates psychic fears among the members
of the organization. Some fears come from the open discussion of one’s
interpretations, ideas, and judgments. The researcher should use the risk
principle to allay others’ fears and invite participation by assuring that
everyone will be subject to the same process, and that whatever the
outcome, learning will take place.

5. Plural structure. The nature of the research embodies a multiplicity

of views, commentaries and critiques, leading to multiple possible
actions and interpretations. This plural structure of inquiry requires a
plural text for reporting. This means that there will be many accounts
made explicit, with commentaries on their contradictions, and a range
of options for action presented. A report of findings, therefore, acts as
a support for ongoing discussion among collaborators, rather than a final
conclusion of fact.

6. Theory, practice, transformation. For action researchers, theory

informs practice and practice refines theory in a continuous
transformation. The two are intertwined aspects of a single change
process. It is up to the researchers to make explicit the theoretical
justifications for the actions, and to question the bases of those
justifications. The ensuing practical applications that follow are
subjected to further analysis, in a transformative cycle that continuously
alternates emphasis between theory and practice.
The process of conducting a typical action research project begins
with the identification of the research problem. One should begin by
deciding on a question that is meaningful and important—to him or her,
to his or her organization, to his or her students, etc. The second step is
to plan the overall research strategy and data collection strategies. The
third step involves the collection of data. Good data are directly related
to the questions; one should use more than one strategy or source of
66 Social Research Methodology: An Introduction

information for each question. The fourth step is to analyze the data. It
is important to know that when one is no longer learning anything new
about the research questions, it is time to stop collecting data and start
analyzing. The second last stage of the research process is to share one’s
findings with others. After reaching conclusions about the research
question, it is establish the practical significance of the findings. The last
step is to take action based on the conclusions. At this point, participants
decide what to do and go on to carry it out. What is most important for
the study is what happens after this action. Whatever the action is and
whatever the outcome of the action, the group must reflect carefully on
the process and the outcome itself. This is a defining moment whether
the process is to start again or not.
Gerald Susman (1983) gives a somewhat more elaborate five-phase
cycle through which action research is conducted. Initially, a problem
is identified and data is collected for a more detailed diagnosis. This is
followed by a collective postulation of several possible solutions, from
which a single plan of action emerges and is implemented. Data on the
results of the intervention are collected and analyzed, and the findings
are interpreted in light of how successful the action has been. At this
point, the problem is re-assessed and the process begins another cycle.
This process continues until the problem is resolved.
Action research is used in real situations, rather than in contrived,
experimental studies, since its primary focus is on solving real problems.
But it can also be used for preliminary or pilot research, especially when
the situation is too ambiguous to frame a precise research question. It
is mostly preferred when the investigation requires flexibility, the
involvement of the people, or where there is a need for immediate
change Action research is more of a holistic approach to problem-
solving, rather than a single method for collecting and analyzing data.
Thus, it is conducted with several different research methods, which are
generally common to the qualitative research paradigm. These include
keeping a research diary, document collection and analysis, participant
observation, unstructured interviews, and case studies.

Case Study Research

A case study is a qualitative research design that involves an in-depth
examination of a single instance or event—a case—rather than using
 Qualitative Research Designs: An Overview 67

large samples and following a rigid protocol to examine a limited

number of variables. Case studies provide a systematic way of looking
at events, collecting data, analyzing information, and reporting the
results. Case study research is one of the most common qualitative social
research designs.
There are numerous definitions of case study research (Yin, 1981;
Ragin & Becker, 1992; Hamel, 1993; Yin, 1994; Stake, 1995; Yin,
2002). I will use the definition byYin (1994:13), who defines the scope
of a case study as “an empirical inquiry that investigates a contemporary
phenomenon within its real-life context, especially when the boundaries
between phenomenon and context are not clearly evident, and in which
multiple sources of evidence are used.” Case studies emphasize detailed
contextual analysis of a limited number of events or conditions and their
relationships. Case study research has long been used across a variety of
disciplines. Social scientists have particularly used it to examine
contemporary real-life situations and provide the basis for the
application of ideas and extension of methods.
Although case study research is largely qualitative, case studies should
not be confused with qualitative research because they can be based on
any mix of quantitative and qualitative evidence. A good example of this
is the mixed model case study design, which combines both qualitative
and quantitative tools and data in a single study.
There are different types of case study. An exploratory case study
is a descriptive study that is undertaken before implementing a large-
scale investigation. It helps the researcher to identify research questions
or hypotheses, select measurement constructs, and develop instruments.
An illustrative case study is also descriptive and is intended to add
realism and in-depth examples to other information about a program or
policy. A critical instance case study examines a single instance of
unique interest or serves as a critical test of an assertion about a
program, problem, or strategy. A programme implementation case study
is a case study that investigates operations, often at several sites, and
often normatively. A programme effects case study is a study that is
conducted to evaluate a programme . It examines causality and usually
involves multi-site, multi-method assessments. A cumulative case study
is one that brings together findings from many case studies to answer an
evaluation question, whether descriptive, normative, or cause-and-effect.
68 Social Research Methodology: An Introduction

A number of well-known case study researchers such as Robert E.

Stake and Robert K. Yin have suggested systematic procedures for
organizing and successfully conducting case study research. Generally,
case study research is conducted through a six-step process, as I
highlight below.

1. Determine and define the research questions. The researcher

establishes the focus of the study by forming questions about the
situation or problem to be studied and determining a purpose for the
study. The research object in a case study is often a program, an entity,
a person, or a group of people. The researcher investigates the object of
the case study in-depth using a variety of data gathering methods to
produce evidence that leads to understanding of the case and answers the
research questions. Case study research generally answers one or more
questions that seek to explore the “why” and “how” of the phenomenon.

2. Select cases and decide on research tools. The researcher determines

whether to select a single case or multiple real-life cases to examine in-
depth. This decision is normally dictated by the nature of the problem
or research questions. Note that in the event of multiple cases, each case
is treated as a single case. The major strength of case study research
involves using multiple sources of data and data collection techniques.
The researcher should at this stage determine what evidence to gather
and what techniques to use. The data collection tools to select from
include surveys, interviews, document review, observation, and the
collection of physical artifacts.

3. Prepare for field data collection. Advance preparations are necessary

to prevent the researcher from losing focus on the research purpose and
questions or becoming overwhelmed by the huge amount of data.
Preparations include training for field assistants, establishing clear
protocols and procedures of field work, designing the research
instruments, and pretesting or conducting a pilot study in order to
remove obvious barriers and problems. Training covers the basic
concepts of the study, terminology, processes, methods, and the proper
application of the techniques being used in the study.
 Qualitative Research Designs: An Overview 69

4. Field data collection. The researcher uses field notes and databases
to categorize and reference data so that it is readily available for
subsequent reinterpretation. Testimonies, stories, and illustrations from
the respondents must be systematically recorded for use in later reports.
The researcher must maintain the relationship between the issue and the
evidence gathered. He or she documents, classifies, and cross-references
all evidence gathered so that it can be easily and efficiently retrieved for
sorting and examination during the course of the study.

5. Evaluate and analyze the data. The use of multiple data collection
methods and analysis techniques enables case study research to provide
researchers with opportunities to triangulate data in order to strengthen
the research findings and conclusions. The researcher examines raw data
using many interpretations in order to find linkages between the research
object and the outcomes with reference to the original research
questions. Throughout the evaluation and analysis process, the
researcher remains open to new opportunities and insights. Data analysis
techniques include placing information into arrays, creating matrices of
categories, creating flow charts or other displays, and tabulating the
frequency of events. Quantitative data is used to corroborate and support
qualitative data, which is most useful for understanding the rationale or
theory underlying the relationships.

6. Report writing. Case studies present data in very publicly accessible

ways and may lead the reader to apply the experience in his or her own
real-life situation. Researchers pay particular attention to displaying
sufficient evidence to gain the reader’s confidence that all avenues have
been explored, clearly communicating the boundaries of the case, and
giving special attention to conflicting propositions. The report formats
vary. In some reports, each case may be handled as a separate chapter;
it may be treated as a chronological recounting; it may be reported as a
story.
Clearly, the case study design offers a method of learning about a
complex instance through extensive description and contextual analysis.
The findings articulate why the instance occurred as it did, and what one
might usefully explore in similar situations. Case studies can generate
a great deal of data that may defy straightforward analysis.
70 Social Research Methodology: An Introduction

Historiography
Historiography is the method of doing historical research or gathering
and analyzing historical evidence. Researchers adopt the historical
research design to systematically find evidence, evaluate it to establish
the facts, reflect on the facts to decide what they tell about the people
and events of the past, and publish their new knowledge. History has
always been a discipline of interpretation of past events. Documents,
records, reminiscences, artifacts, buildings, all of the materials produced
by human culture and activity serve as raw material from which
historical researchers fashion their understanding of the past.
There are four types of historical evidence: primary sources,
secondary sources, running records, and recollections. Primary sources
are also called archival data because they are kept in museums,
archives, libraries, or private collections. Historical researchers rely
mostly on this type of evidence, and emphasis is given to the written
word, although modern historiography involves any other media.
Secondary sources are the work of other historians writing history. This
includes all manner of information provided by a person who did not
directly witness the event, object or condition relating to the issue under
investigation. Examples are textbooks, periodicals, reference literature,
etc. Running records are documentaries maintained by private or
nonprofit organizations. Recollections are autobiographies, memoirs, or
oral histories.
There is a particularly interesting variety of historical research
called prosopography or prosopographic analysis. Prosopography is the
study of collective biography for individuals of a certain group. It has
been defined as an independent science of social history embracing
genealogy, onomastics and demography. Claude Nicolet defined its aim
as the history of groups as elements in political and social history,
achieved by isolating a series of persons having certain political or
social characteristics in common and then analyzing each series in terms
of multiple criteria, in order both to obtain information specific to
individuals and to identify the constants and the variables among the
data for whole groups (Keats-Rohan, 2003). By accumulating data on
individuals in a group, one learns more about the group.
There have been expressed doubts about the proper place of
prosopography in research methods—that it is more akin to “profiling”
 Qualitative Research Designs: An Overview 71

in political psychology than anything else (Stone, 1972). However,

prosopography involves the study of biographical details (family
background, childhood events, educational background, religion, etc.)
that are common or “in the aggregate” among a group of people. This
method sometimes yields significant insights by combining the common
background elements in individual profiles.
Once the decision is made to conduct historical research, there are
clear steps that should be followed to achieve a reliable result. The
process involves making a plan to find out systematically what primary
sources and secondary works are available. This could include finding
bibliographies by researchers who have already worked on the same
topic, talking to librarians and archivists about what secondary and
primary sources are available, or even interviewing experts who are
familiar with the field. The researcher should then make a plan for
studying each primary source as part of a larger systematic plan. New
interpretations of the past may be developed if evidence can be found in
primary sources to support such interpretations. Of course any document
can be read in many different ways. When reading any historic
documents, it is important to reflect on some basic issues such as i) the
historical context—who created the document, when, where, and why?
Who was the intended audience?; ii) conventionality—is it a typical
document or an unusual one? What problems, assumptions, ideas, and
values does the author share with others from this time period?; iii) the
author’s intentions—what was the author trying to accomplish?
Busha and Harter (1980:91) have proposed six steps for conducting
historical research. These are:
i) recognition of a historical problem or identification of a need for
certain historical knowledge;
ii) gathering of as much relevant information about the problem or
topic as possible;
iii) formulating of research questions or forming of hypotheses that
tentatively explain relationships between historical factors;
iv) rigorous collection and organization of evidence, and the
verification of the authenticity and veracity of information and its
sources;
v) selection, organization, and analysis of the most pertinent collected
evidence, and drawing conclusions; and
72 Social Research Methodology: An Introduction

vi) recording of conclusions in a meaningful narrative.

This process is by no means a simple one, especially to a novice
researcher or researchers using this design for the first time. Here below
is a checklist of dos and don’ts modified from Schlabach (1996)’s Ten
Commandments of Good Historical Writing.
1) Begin with an outline that builds your entire paper around your
central ideas.
2) Avoid self-conscious discussion of your intended purposes, your
strategy, your sources, and your research methodology.
3) You may covet other writers’ ideas but do not steal them.
4) Strive for clarity above cuteness; do not use jargon when common
language can serve, nor a large word when a small one can serve,
nor a foreign term when an English one can serve, nor an abstract
term where a vivid one is possible.
5) Keep your paragraph a significant unity; do not fragment your
discussion into one short paragraph after another, and do not write
a paragraph that fails to develop a topical idea.
6) Write as if your reader is intelligent—but totally uninformed on
this particular subject.
7) Use quotations sparingly and judiciously, only for colour and
clarity; if you must quote, quotations should not break the flow of
your own language and logic, and your text should make clear
whom you are quoting.
8) Never relegate essential information to your footnotes.
9) Write consistently in past tense, and in other ways keep your reader
firmly anchored in time.
10) Do not use passive voice.

Ethnography and Ethnomethodology

Ethnographic research comes from the discipline of social and cultural
anthropology where an ethnographer is required to spend a significant
amount of time in the field. Ethnographers immerse themselves in the
cultural life of the people they are studying and seek to place the
phenomena studied in their social and cultural contexts. Ethnography is
a form of research focussing on the sociology of meaning through close
field observation of socio-cultural phenomena. When used as a method,
ethnography typically refers to participant-observation conducted by a
 Qualitative Research Designs: An Overview 73

single investigator who “lives with and lives like” those who are studied,
usually for a year or more.
The ethnographic method involves observation and note taking,
which anthropologist Clifford Geertz called thick description (Geertz,
1973).* For about every half hour of observation, an ethnographic
researcher may write notes for about two hours. These notes contain
rich, detailed descriptions of everything that went on. There should be
no attempt at summarizing, generalizing, or hypothesizing. The notes
should capture as factual a description of the phenomenon as possible to
permit multiple interpretations, and most of all, to later infer cultural
meaning.
The ethnographic research design is generally associated with the
following features:
i) People’s behaviour is studied in everyday contexts, rather than
under experimental conditions created by the researcher;
ii) Data are gathered from a range of sources, but observation and/or
relatively informal conversations are usually the main ones;
iii) The approach to data collection is unstructured in the sense that it
does not involve following through a detailed plan set up at the
beginning; nor are the categories used for interpreting what people
say and do pre-given or fixed. This does not mean that the research
is unsystematic; simply that initially the data are collected in as raw
a form, and on as wide a front, as feasible;
iv) The focus is usually a single setting or group, of relatively small
scale. In life history research the focus may even be a single
individual; and
v) The analysis of the data involves interpretation of the meanings and
functions of human actions, and mainly takes the form of verbal
descriptions and explanations, with quantification and statistical
analysis playing a subordinate role if at all.
Ethnography is less specialized and less technically sophisticated than
approaches such as the experiment or the social survey; though all social

*. In anthropology and other fields, a thick description of a human behaviour

is one that explains not just the behaviour, but its context as well, such that the
behaviour becomes meaningful to an outsider.
74 Social Research Methodology: An Introduction

research methods have their historical origins in the ways in which

human beings gain information about their world in everyday life.
The ethnographic method starts with selection of a culture, review
of the literature pertaining to the culture, and identification of variables
of interest—typically variables perceived as significant by members of
the culture. The ethnographer then goes about gaining entrance, which
in turn sets the stage for cultural immersion of the ethnographer. The
middle stages of the ethnographic method involve gaining informants,
using them to gain yet more informants in a chain process, and gathering
of data in the form of observational transcripts and interview recordings.
Data analysis and theory development come at the end, although
theories may emerge from cultural immersion and theory-articulation by
members of the culture. However, the ethnographic researcher should
strives to avoid theoretical preconceptions and instead induce theory
from the perspectives of the members of the culture and from
observation. The researcher may seek validation of induced theories by
going back to members of the culture for their reaction.
There are three methodological principles that provide the rationale
for the specific features of the ethnographic method—naturalism,
understanding and discovery. These principles actually form the basis
for much of the criticism against quantitative research, notably, that it
fails to capture the true nature of human social behaviour; it relies on
artificial settings and on what people say rather than what they do; it
seeks to reduce meanings to what is observable; and it reifies social
phenomena by treating them as more clearly defined and static than they
are.
The principle of naturalism means that the researcher leaves the
natural phenomena alone and tries to be invisible—to watch and listen
carefully without being noticed. At some point, the researcher
disengages, retreats to a private place, and takes notes when the memory
is still fresh and laden. Qualitative social research normally aims to
capture the character of naturally occurring human behaviour, and this
can only be achieved by first-hand contact with that behaviour, not by
inferences from what people do in artificial settings like experiments or
from what they say in interviews. This is why ethnographers carry out
their research in situ—in settings that exist independently of the research
process, rather than in those set up specifically for the purposes of
 Qualitative Research Designs: An Overview 75

research. Another important notion of naturalism implies that social

events and processes must be explained in terms of their relationship to
the context in which they occur.
The principle of understanding is premised on the argument that
human actions differ from the behaviour of other animals—they involve
interpretation of stimuli and the construction of responses. To be able to
effectively explain human actions, one must gain an understanding of
the cultural perspectives on which these actions are based. Particular
groups, be they ethnic, occupational, or small informal groups (even
individual families or school classes) and individuals develop distinctive
ways of orienting to the world that may need to be understood if their
behaviour is to be explained. Ethnographers believe that it is necessary
to learn the culture of the group one is studying before one can produce
valid explanations for the behaviour of its members. That is why
participant observation and unstructured interviewing are the key
methods in ethnography.
The principle of discovery implies that ethnographic research
follows a process that is inductive or discovery-based rather than one
limited to the testing of explicit hypotheses. It is argued that if one
approaches a phenomenon with a set of hypotheses one may fail to
discover the true nature of that phenomenon, being blinded by the
assumptions built into the hypotheses. In ethnographic research, the
focus is narrowed and sharpened, and perhaps even changed
substantially, as the research proceeds. Similarly, theoretical ideas that
frame descriptions and explanations of what is observed are developed
over the course of the research.

ETHNOMETHODOLOGY is the study of the ways in which people make

sense of their social world. It is an ethnographic technique popularized
by American sociologist Harold Garfinkel in 1967. This technique
assumes a more active role for the researcher, one that involves
“breaking up” the standard routines of folk groups in order to see how
strongly and in what ways group members mobilize to restore the
cultural order.
Ethnomethodology differs from other sociological perspectives in
one very important aspect: it sets out with the assumption that social
order is illusory. Ethnomethodologists believe that social life merely
76 Social Research Methodology: An Introduction

appears to be orderly but in reality it is potentially chaotic. For them

social order is constructed in the minds of social actors as society
confronts the individual as a series of sense impressions and experiences
which he or she must somehow organize into a coherent pattern.
One common technique among ethnomethodologists is to disrupt
temporarily the world (social order) which people take for granted and
see how they react. This puts the researcher is then in a better position
to understand the fragile and fluid processes of social control, as well as
the rules that people use for maintaining cultural boundaries. The point
of this is to expose background assumptions that have been accepted as
reality for a long time. However, in spite of the great theoretical
potential of this research method, it is not commonly used. In fact, the
refined versions of this method are increasingly being referred to as
conversation analysis or socio-linguistics.

Phenomenological Research
Phenomenology may be defined as the examination of human
experiences through detailed descriptions of the people being studied.
The procedure involves studying a small number of subjects through
extensive and prolonged engagement to develop patterns and
relationships of meaning. As a sociological approach, phenomenology
seeks to reveal how human awareness is implicated in the production of
social action, social situations and social worlds.
The phenomenological research approach, which is traceable back
to the thinking of 19th century sociologist Wilheim Dilthey (see chapter
one), was developed by Edmund Husserl, who felt that the objectivism
of science precluded an adequate apprehension of the world (Husserl,
1931). Husserl presented various philosophical conceptualizations and
techniques designed to locate the sources or essences of reality in the
human consciousness. Subsequent scholars such as Alfred Schutz found
problems in popular theories and explanations of the social world, and
his critique saw phenomenology enter the domain of sociology and
social research (Schutz 1967). One such a theory was Max Weber’s
theory of action, in which Weber recognized the interpretation of human
actions as a subcategory of causal analysis, in which the agents’ visions
of desired outcomes, together with their beliefs about how to bring them
about, cause them to act as they do. Ever since Husserl and Schutz,
 Qualitative Research Designs: An Overview 77

interest in phenomenological research has been growing steadily as

researchers have sought to capture the richness of individual experience.
Phenomenology presents theoretical techniques and qualitative
methods that illuminate the human meanings of social life. It is based in
a paradigm of personal knowledge and subjectivity, and emphasizes the
importance of personal perspective and interpretation. As such, it is a
powerful approach for understanding subjective experience, gaining
insights into people’s motivations and actions, and cutting through the
clutter of taken-for-granted assumptions and conventional wisdom. Until
recently viewed as an adversary of the more conventional approaches to
social research, phenomenology is increasingly coming to be viewed as
an adjunctive or even integral part of the discipline of social research,
contributing useful analytic tools to balance the positivist/objectivist
approaches.
Phenomenological research has overlaps with other essentially
qualitative approaches, especially ethnography. Pure phenomenological
research seeks essentially to describe rather than explain, and to
integrate a grounded theory approach—starting from a perspective free
from hypotheses or preconceptions (Husserl, 1970). Phenomenological
methods are particularly effective at bringing to the fore the experiences
and perceptions of individuals from their own perspectives, and
therefore at challenging structural or normative assumptions. Adding an
interpretive dimension to phenomenological research, enabling it to be
used as the basis for practical theory, allows it to inform, support or
challenge policy and action.

Grounded Theory Research

Grounded theory is a qualitative research design that was developed by
Barney Glaser and Anselm L. Strauss in 1967. Grounded theory
researchers seek to systematically gather and analyze data in order to
develop theory that is “grounded” in that data. The researcher begins the
study without any assumed theoretical position or hypothesis about the
phenomenon under investigation. This contrasts with the hypothetico-
deductive methods (in quantitative research), where theories are
generated from cyclical testing and refining of a previously constructed
hypothesis. In grounded theory research, theory emerges from the
78 Social Research Methodology: An Introduction

systematic examination of the phenomenon through the interplay

between data collection and analysis phases.
Glaser (1978) coined the concept of theoretical sensitivity as an
important feature of grounded theory. According to him, theoretical
sensitivity is the process of developing the insight with which a
researcher comes to the research situation—an insight that should be
conceptual rather than concrete. This sensitivity is actually considered
a creative aspect of grounded theory and involves the researcher
working in the area to obtain experience and expertise. By gaining
theoretical sensitivity the researcher is able to recognise important data
and formulate conceptually “dense” theory.
Grounded theory presents a single, unified, systematic method of
analysis. Other qualitative methods often rely upon the application of
general principles rather than systematic method, making their
application and interpretation more difficult. Grounded theory also
allows for validation of findings and integration of quantitative data, as
both forms of data are necessary in many instances (Glaser & Strauss,
1967). A good source for a discussion of the various benefits of
grounded theory is Strauss and Corbin (1994), who argue that
quantitative researchers are becoming less satisfied with purely
quantified results and are turning increasingly to supplementary
qualitative analyses.
According to Strauss and Corbin (1990), in order for a theory to be
grounded it must meet four central criteria—fit, understanding,
generality, and control. Fit entails that the theory fits the substantive
data. Understanding entails that the theory be comprehensible to all
involved in the area of study. Generality entails that the theory is
applicable in a variety of contexts. Control implies that the theory should
provide a framework for action to achieve change in the issue being
investigated.
The grounded theory approach consists of a set of steps whose
careful execution “guarantees” a good theory as the outcome. Grounded
theorists believe that the quality of a theory can be evaluated by the
process by which the theory is constructed. This, of course, contrasts
with the orthodox, positivist perspective, that the quality of a theory is
determined by its ability to explain new data. Grounded theory follows
a systematic process that starts with the selection of a suitably complex
 Qualitative Research Designs: An Overview 79

research question that should be related to the aims of the study. The
second stage is data collection, the third data analysis, and the final stage
is report writing.
Data collection is done in the same ways, using the same techniques
as in other research methodologies, although interviewing normally
plays a leading role. Observations, document review, focus group
discussions, and any other method relevant to the research question may
be used. Clearly then, data may be qualitative or quantitative or a
combination of both. The analysis of data under this design is often
referred to as coding. There are three distinct but overlapping types of
coding in grounded theory research, and these are: 1) Open coding— the
process of selecting and naming categories from the analysis of the data.
It is the initial stage in data acquisition and relates to describing overall
features of the phenomenon under study. Variables involved in the
phenomenon are identified, labeled, categorized and related together in
an outline form. 2) Axial coding—putting data together in new ways and
trying to identify, understand and explain connections and causal
relationships between categories and sub-categories in order to
understand the phenomenon to which they relate. 3) Selective
coding—the process of selecting and identifying the core category and
systematically relating it to other categories. It involves validating those
relationships, filling in, and refining and developing those categories.
Categories are integrated together and a grounded theory is derived.
Positivist researchers generally tend to criticize qualitative methods
citing their inability to relate to aspects of the real world. However, it is
generally accepted that qualitative methods have their own internal logic
and validity. Yet, the relationship between theory and reality in
grounded theory research is no doubt a complex issue. All qualitative
studies aim to describe and understand the rich and complex phenomena
they engage in. The descriptions and understandings are usually placed
at a certain period in time and located in specific societies and social
contexts. But in a globalizing and a rapidly changing world, these
contexts are constantly changing, rendering the descriptions and
understandings of contemporary social processes and relationships only
temporary. Hence, the need to keep evolving theory.
Chapter Six

Interviewing in Social Research

Introduction
Nigel King (2004: 115) defines an interview as “a direct face-to-face
attempt to obtain reliable and valid measures in the form of verbal
responses from one or more respondents.” He also calls it “a
conversation in which the roles of the interviewer and respondent
change continually.” Mbaaga (2000: 38) simply defines it as a face-to-
face conversation between an interviewer and a respondent for purposes
of obtaining information. Clearly, by insisting on the element of “face-
to-face,” both definitions exclude any forms of mediated interviewing
that are increasingly gaining popularity through the use of telephone or
internet chat rooms. Similarly, these definitions are “general”—they do
not suggest that there are interviews whose purpose may be different
from obtaining information, as I shall shortly show.
For sure, one effective way to collect data is simply to ask
questions—as is the procedure of interviewing. Interviewing may thus
simply but comprehensively defined as the systematic collection of data,
through asking questions and carefully listening to people that have
experiences and knowledge being sought by the researcher,* and noting
or voice recording the answers. Interviews can take place in any
location, at any time, and with different individuals or groups of people.
They can be unstructured, structured, or semi-structured; and they can
also be held directly (face-to-face) or indirectly through a medium such

* These people being interviewed are variously called interviewees,

respondents, informants or subjects. In this chapter—and indeed elsewhere—these terms
are used interchangeably.
 Interviewing in Social Research 81

as a telephone or chat room. They may be formal and especially

arranged for the purpose, or informal, taking advantage of opportunities
to question people about a problem of interest.
Social research interviewing, as I hinted above, differs from other
forms of interviewing. While its purpose is generally to elicit
information from a respondent or group about a phenomenon of interest
to the researcher, the purpose of say a legal interview (Benchmark
Institute, 2005) or a medical interview (Coulehan & Block, 2005) is
widely viewed as to exchange information with the respondents, not just
to obtain it. Then there is the more common job interview or interview
for a place say in a school. The purpose of such interviews is to the
suitability of the candidates for the job or class, and then select in a
competitive context.
On their part, social research interviews provide the researcher with
access to a range of experiences, situations and knowledge that would
otherwise remain hidden within the interviewee. With these kinds of
interviews, the interviewee may describe private or sensitive behaviours,
events that happened in the past, or key locations that are unknown and
therefore inaccessible to the outsiders. Such interviews also provide
information on the meanings and definitions that people give to events
and behaviour. But for such interviews to be successful, at least three
simple conditions must be met. First, the information sought must be
available and within the reach of the respondent. Second, respondents
should be able to play the role of respondents. Third, they should be
motivated, either intrinsically or extrinsically,* to participate in the
interview.

Who Should be Interviewed?

Given the short time that is almost always available for the research
project, the research team should adopt a systematic and pragmatic
approach to selecting interviewees. Interviews should be conducted with

* Intrinsic motivation (motivation from within) means that respondents have

personal faith in the inherent goodness or necessity of the study being conducted, and
that the responses they give will help in achieving that good. Extrinsic motivation
(motivation from without), on the other hand, involves anything outside the respondent,
including pecuniary inducement, which makes them better disposed to participate in the
interview.
82 Social Research Methodology: An Introduction

a wide range of people, rather than a few individuals or groups of more

or less similar character or background.
The more specific the team can be about the information they want
to collect, the easier it will be for them to identify potential interviewees.
One way of doing this is for the team to reduce larger topic areas (such
as risk behaviour) into smaller, more manageable items. Discussion with
colleagues and key informants can be used to suggest who could be
contacted. Mapping the area of study is particularly useful, as it allows
the research team to identify potential interviewees in the local area. Key
informants can suggest and arrange access to individuals and groups that
the team may be able to interview. Interviews can occur spontaneously.
For example, when a researcher is conducting an observation and has a
chance or casual conversation with someone interesting or relevant to
the research, it may be recorded as an interview. Similarly, the team may
find that individuals, who previously refused interviews, change their
mind when they see other people talking to the research team members.

Types of Interviews
Social research interviewing may be classified from two perspectives:
first, on the basis of the structure of the interview; and second, on the
basis of the medium of interviewing. Going by structure, there are three
types of interviews: structured interviews, unstructured interviews, and
semi-structured interviews. Going by medium, on the other hand, there
are four types of interviews: face-to-face interviews, telephone
interviews, internet-mediated interviews, and group interviews. In this
chapter, I will deal with types of interviews by structure. Types of
interviews by medium of interviewing are discussed in chapter twelve.

Structured Interviews
The structured interview may be considered as an oral presentation of a
written questionnaire. According to Frey and Oishi (1995:1), it is “a
purposeful conversation in which one person asks prepared questions
(interviewer) and another answers them (respondent).” Nichols
(1991:131) who also refers to structured interviews as “closed”
interviews defines them as social surveys where “the range of possible
answers to each question is known in advance.” Under such interviews,
the interviewer reads out the questions and the interviewee gives his or
 Interviewing in Social Research 83

her response with a minimum of side interaction. Accordingly, the

questions used in structured interviews are closed questions.
Structured interviews tend to thrive on the assumptions that 1) there
is a common vocabulary for all potential respondents; 2) question
formats are equally meaningful to all; and 3) the context of each
question is obvious. Accordingly, the following characterise structured
interviews:
i) Questions are set in advance;
ii) Each interview is conducted in exactly the same way;
iii) The questions and their order are the same for all respondents;
iv) The range of possible responses is determined in advance by the
researcher;
v) In case of probes, they are usually scripted beforehand; and
vi) Only closed questions are used.
Structured interviews are quantitative, and are normally used when
the researcher wants more control over the topics and the format of the
interview. A detailed response schedule or questionnaire known as an
interview schedule is used by the researcher to record the responses.
Structured interviews are often undertaken after exploratory research
that yields qualitative data. This helps the researcher to use findings
from other methods to identify topics that require further investigation.

The interview schedule

Before undertaking a structured interview, the researcher must prepare
an interview instrument—the interview schedule. Since, as I have
indicated, a structured interview requires a more detailed instrument
than an unstructured interview, the interview schedule must be clearly
detailed. It lists all the questions, topics and issues that the research team
wants to cover during the interview. It may also include instructions on
how to respond to certain answers, the order and wording that questions
should be asked in, and the probes and prompts to encourage the
respondent to produce more information or talk about certain topics.
An interview schedule is so much like a structured questionnaire
discussed in chapter twelve. Interview schedules should avoid
ambiguity, double barrelled questions, leading questions, complex
questions, and in some cases, technical questions. The construction of
an interview schedule follows the following process:
84 Social Research Methodology: An Introduction

i) Identify appropriate topics and questions. The researchers should

be guided by the specific objectives of the study and the hypotheses
or research questions. All the topics and areas to cover in the
instrument should be discussed with colleagues, and the questions
should ensure that the data gathered will help fill any gaps in the
existing knowledge.
ii) Decide on the level of detail. The schedule may include broad
topics that act as reminders, but should have specific questions
which the researcher must ask in a precise order.
iii) Draft the questions. This involves careful attention to the language
used in phrasing the question, formulating the questions and listing
the alternative answers the respondents will select from, ordering
the questions, and listing any probes or prompts.
iv) Pre-test the questions. This involves conducting a pilot test of the
questions on a few potential respondents to gauge how suitable or
otherwise the instrument is.

Evaluation of structured interviewing

Structured interviewing has got a number of strengths. First of all, since
each question has a list of alternative answers, structured interviews are
quick and easy to answer. Secondly, a standard format for each
interview makes it easier to code the answers, analyse them and compare
data. Another strength is that the interview schedule helps the researcher
not to spend unnecessarily long hours discussing each question or topic
since the direction and length of the enquiry is clear. This has got the
effect of ensuring that interviews are not used to prioritise questions
when only a short amount of time is available and the question order
given. Again, the straightforward ‘factual’ information elicited ensures
a high degree of reliability. Finally, detailed interview schedules allow
inexperienced researchers to conduct interviews while at the same time
they reduce the possibility of interviewer bias.
However, this method is not without significant weaknesses. One
of these is inflexibility. The rigid adherence to the schedule may prevent
the collection of unexpected but relevant information. Secondly,
although a standard format is used, interviewees may nevertheless hear
and understand the questions in different ways. This can affect
comparison between respondents. Thirdly, because the number of
 Interviewing in Social Research 85

possible responses is often limited, participants may be forced into

giving responses which do not reflect their true feelings about an issue,
and this raises a validity problem. Finally, the data gathered by a
structured approach always lacks the richness obtained by in-depth,
open-ended interviews.

Unstructured Interviews
Unstructured interviews, also sometimes variously called “in-depth,”
“one-on-one” or “intensive” interviews are usually designed to collect
qualitative information from a small-sized sample. Oftentimes, they are
used to explore conceptual issues at an early stage in the development
of a questionnaire for a wider survey. In-depth interviews involve open-
ended questions asked by a researcher to an individual respondent. They
enable the interviewees to offer their opinions, knowledge and
experience freely. Probing techniques are used to encourage respondents
to give the fullest answer possible. Instead of basing on tight, rigid
schedules of questions, unstructured interviews use interview guides that
have a list of themes or issues that the researcher aims to explore.
Nichols (1991:131) defines an unstructured interview as :

an informal interview, not structured by a standard list of

questions. Field workers are free to deal with the topics of
interest in any order and to phrase their questions as they think
best.

This flexible approach means that the order and wording of questions in
each interview will vary from respondent to respondent. The researcher
encourages the respondent to provide as much detail and be as frank as
possible.
This type of interview is based on the knowledge of, or assumption
that, the respondents have had a particular experience, on which they
can elaborate. The interview thus focuses on the respondents’ subjective
experiences. It allows the respondents to describe in detail the situation
as it is meaningful to them, and gives the interviewer the latitude to
determine which questions to ask, how they are phrased, and when to
use probes and prompts. Unstructured interviews require good
communication and facilitation skills. An interviewer must listen
86 Social Research Methodology: An Introduction

carefully to the respondents and be aware of any new or interesting

information.

Approaches to unstructured interviewing

There are two approaches to unstructured interviewing: the informal,
conversational approach; and the in-depth, interview guide approach.
The informal conversational interview may occur spontaneously in the
course of field work and the respondent may not know that the interview
is taking place. Questions emerge from the immediate context, so the
wording of questions and even topics are not predetermined. The major
advantage here is that the interview is highly individualized and relevant
to the individual. Thus, it is likely to produce information or insights
that the interviewer could not have anticipated. This kind of interview
requires an interviewer who is very knowledgeable and experienced in
the content area and strong in interpersonal skills, since he or she will
have considerable discretion in directing the interview. However, since
different information is collected from different people, this kind if
interview is not systematic or comprehensive and it can be very difficult
and time consuming to analyse the data.
The in-depth, interview guide approach is the most widely used
format of qualitative interviewing. Here, the interviewer has an outline
of topics or issues to be covered but is free to vary the wording and
order of the question to some extent. The major advantage is that the
data is rather more systematic and comprehensive than in the informal
conversational interview, while the tone of the interview still remains
conversational and informal. Like in formal conversational interview,
this type of interview requires an interviewer who is relatively skilled
and experienced, since he or she needs to know when to probe for more
in-depth responses or guide the conversation to make sure that all the
topic on the outline are covered. A possible draw back is that sticking to
the outlined topics will prevent other important topics from being raised
by the respondent. Also, while this format is more systematic than the
conversational interview, it is difficult to compare or analyze data
because different respondents are responding to somewhat different
questions.
 Interviewing in Social Research 87

When to opt for unstructured interviews

In-depths are used when a particular topic 1) needs to be explored in
great depth or detail, or 2) in situations when focus groups are neither
appropriate nor practical for the audience of interest. In general, they
may be used for the same purposes as focus groups. Most often, they are
used to develop a detailed inquiry into attitudes, motivations, emotions,
and other intrinsic feelings and behaviours of the respondents. For
example, they are often used in areas of personal experience such as
among victims of an embarrassing condition, who may withdraw in a
focus group setting. They are commonly used with physicians,
pharmacists, attorneys, or direct competitors, since focus groups among
these types of professionals often create a self-conscious, posturing
environment. The in-depth format eliminates these distractions, letting
respondents focus on the issues at-hand.

Evaluation of unstructured interviewing

One of the major strengths of this approach is that it allows flexibility:
there are no restrictions on what can be discussed. In-depth interviews
are flexible enough to allow the interviewers to modify their line of
enquiry, follow up interesting responses, and investigate underlying
motives. Secondly, this approach respects the respondent’s viewpoint:
respondents can answer in their own words, and the nature of the
responses in not limited. The third strength, and directly following from
the above, is that the open-ended method yields much richer data in
terms of validity and depth. Meanings and attitudes are sufficiently
explored. Another good point is that the respondents may have more
control over the conduct of the interview in that they are often allowed
to discuss issues as they arise and not necessarily in an order
predetermined by the interviewer. In-depth interviews are useful for
collecting background data in the early stages of a survey, when a
researcher (or team of researchers) has little knowledge of the topic or
problem. An in-depth is particularly useful as a pilot study, to test what
people’s responses would be to a particular issue.
The weaknesses of unstructured interviews include the largest
potential for interviewer bias. This may be made worse if inexperienced
interviewers are involved in the research, who may pose poorly worded
questions. Therefore, unlike structured interviews, in-depth interviews
88 Social Research Methodology: An Introduction

require interviewing skills, and this has cost implications concerning

training. Another weakness is that lack of standardization may
encourage the respondent to talk about irrelevant and unimportant
issues. Again, since each interview tends to be unique, it may be
difficult for the researcher to code, analyse and compare the data from
different respondents. This is not only time consuming but also has
serious implications for the reliability of the data.

Semi-structured Interviews
A semi-structured interview is a standardised open-ended interview that
uses a schedule of questions very much like a questionnaire. The
questions are usually open-ended, and the responses are voice-recorded
for later transcription. In a semi structured interview, it is permissible for
the interviewer to stray from the subject area and ask supplementary
questions. However, at least three principles should be adhered to: 1)
never change the order of the questions; 2) never allow your opinions to
show through; and 3) probe only by asking such questions as “could you
give me more detail on that point?” or by just repeating the question.
Semi-structured interviews are still considered qualitative
interviews rather than quantitative largely because their responses are
open-ended. They are useful for reducing bias when several interviewers
are involved, when interviewers are less experienced and
knowledgeable, or when it is necessary for the researcher to compare the
responses of different respondents. This may be the best method if the
researcher is to rely on volunteers or inexperienced interviewers, or if
one has limited time and money available for analyzing data. The major
drawback is that the interviewer has a little flexibility to respond to the
particular concerns of the individual, and there is no guarantee that the
questions asked tap into the issues that are most relevant to this
particular respondent.

Group Interviews
Group interviews involve asking several interviewees a question at the
same time, with participants providing answers individually. Unlike in
a focus group discussion discussed in chapter seven, the researcher
usually does not encourage the interviewees to discuss the question
amongst themselves. Group interviews can use unstructured and
 Interviewing in Social Research 89

structured interviewing techniques. Information from group interviews

cannot be treated like data from individual interviews. Answers may be
influenced by group dynamics: prominent individuals can dominate,
sensitive issues may be suppressed, or group pressure to express a
‘common’ view can stop other views from being expressed.
Group interviews are easy to organise when interviewees gather in
naturally occurring groups such as friends, colleagues, students or clinic
patients. However, the researcher often has less control over who takes
part. This can lead to conflict between interviewees with directly
opposing views. Such interviews are not normally useful in tackling
delicate or personal issues.

Successful Qualitative Interviewing

As already indicated, qualitative interviews use unstructured interview
technique with no posited question written in advance. The questions are
spontaneously composed as the interview goes along providing minimal
guidance to the respondent. An interviewer assigns only a topic. Under
unstructured interviews, open ended questions are asked. This requires
the respondent to dictate an extemporaneous statement (statement
spoken or made without previous thought) on s topic about which he or
she may not have thought before. In the process, the respondent will
ramble and move away from the designated area in the researcher’s
mind. At this moment the researcher introduces the probing method so
as to lead the respondent.
Contact should be made with the potential interviewees prior to the
interview, and an appointment made about the venue and time of the
interview. The interviewer must at this stage secure reasonable
confidence of and rapport with the interviewees. Explain the purpose of
the interview, allay any fears and suspicions, correct any
misconceptions, and give all assurances about the exercise (interviewees
may distrust strangers investigating them). Mention any incentives
(gifts, money, or refreshments) offered to participants to take part.
Interviews should be conducted in a location that facilitates
discussion. Ideally, this should be neutral, free from interruptions, and
as comfortable as possible. If a number of interviews are being
conducted over the course of a few days, the researcher(s) could for
example request for or hire a room in a local school or community centre
90 Social Research Methodology: An Introduction

that is accessible. The common practice however, is for interviewers to

visit interviewees in their own homes.

The Qualitative Interviewer’s Task

As the interview kicks off, the interviewer should ensure a harmonious
relationship with the interviewee. A good rapport helps the respondent
as much as the interviewer to feel at home and express himself or herself
freely and willingly. The interviewer should therefore endeavour to
motivate participants to share their thoughts and feelings on the topic
under study. He or she needs to be able to listen carefully, to know when
to seek clarifying information, and to pay attention to non-verbal as well
as verbal responses. The interviewer should perform the following tasks:

1. Set a conducive environment. One must ensure that the interview

is fixed at a time and place that is most convenient for the interviewee.
This means that appointment is made prior to the interview.

2. Be polite. Do not forget to greet the participants before you start,

or to thank them for their time and help after the interview.

3. Be interested. It is crucial that the interviewer is genuinely

interested in both the subject matter and in the participants. One should
maintain a natural attitude with respect to conversational issues.

4. Explain what you are doing. Tell the participants what will
happen during the interview and how the information will be used. Make
sure the participants understand that any information collected will be
kept confidential.

5. Ask questions. Ask one question at a time and allow the

respondent to exhaust his or her answer without interruption. Repeat the
question if necessary to ensure that the respondent has understood it
well. Listen attentively and carefully to the respondent’s answers and
read his or her body language or expressions and tones.

6. Take notes. It can be very helpful for the interviewer to rate the
participant and the responses in some way that is meaningful. This may
 Interviewing in Social Research 91

be done, for instance, by putting a star in the margin next to the

questions about which the participant seems to feel particularly strongly.
It may also include developing some kind of rating system to be
completed at the end of the interview. Items may include issues such as
whether the participant appeared comfortable or anxious, truthful or
dishonest, sober or under the influence of alcohol, etc. This will help
you make decisions later during the analysis phase about the validity of
a given interview, and can be invaluable especially if the data do not
“make sense” upon careful reading.

Probing
An interviewer can use probes or follow up questions whose main
function is to lead the respondent to answer more fully and accurately
or at least provide a minimally acceptable answer. The other function of
probes is to structure the respondents’ answers and make sure that all
topics of interest to the interviewer are covered and the amount of
irrelevant information reduced. In structured interviews, probe questions
may be written in advance for each question. Probes are thus essentially
contingency questions to be used only if the respondent’s answers to
earlier questions in a certain way are unclear or incomplete. Examples
of neutral probes include repeating the question, repeating the answer,
indicating understanding and interest, pause, and neutral question or
comments like ‘tell me more.’ Generally, qualitative interviews are
intended to probe the respondents’ deepest and most subjective feelings.

Evaluation of Qualitative Interviewing

Qualitative interviewing is associated with the following advantages:
i) It provides a detailed description of the social setting of the
investigation.
ii) Qualitative interviewing exhibits a preference for ‘contextualism’
in its commitment to understanding phenomena (events, behaviour,
etc) in their context. The researcher can repeat and explain the
nature of the questions being asked.
iii) Unstructured interviews are flexible and enhance the opportunity
of coming across entirely unexpected issues which may be of
interest to the researcher.
92 Social Research Methodology: An Introduction

iv) Interviewers can probe for more specific answers and can repeat the
questions when the respondent indicates he or she must have
misunderstood it.
v) The interview tends to have a better response rate than most other
methods. The persons who are unable to read and write can still
answer questions in an interview and others who are unwilling to
spend energy to write may be glad to talk.
vi) The interviewer is able to observe non-verbal behaviour and to
assess the validity of the respondent’s answers.
vii) The interviewer can record the exact time, date and place of the
interview.
viii) Complicated questionnaires can be simplified by explanation to the
respondent.

However, qualitative interviewing has got a number of shortfalls. These

include the following:
i) Different people view similar questions differently. This makes it
rather hard for the researcher to come to a correct interpretation of
the responses being given.
ii) Qualitative studies are commonly seen as case studies, and this
creates a problem of generalizing such information got from
different persons or institutions.
iii) Interviewing can be extremely costly.
iv) Interviews are often lengthy and may require the interviewer to
travel long distances hence taking a lot of time.
v) Te respondent’s answers can also be affected by his or her reaction
to the interviewer’s sex, race, social class, age, dress and physical
appearance or accent.
vi) The interview does not allow the respondent to check records or
consult. It relies of memory which is fallible and can suffer loss.
vii)The respondent may give answers that are less efficient merely
because the interviewer arrived at the wrong time.

Tips for Successful Interviewing

The following summary may provide a useful check-list for successful
interviewing:
i) Arrive early at the location where the interview is to take place.
 Interviewing in Social Research 93

ii) Try and ensure that the location is as quiet and as free of
interruptions as possible.
iii) Ensure that you have all the materials and equipment (note books,
pens, recorders, etc).
iv) Introduce anyone present to the participant.
v) Assure participants that everything discussed will be confidential
and ask for consent.
vi) Use clear and simple language. Allow participants time to think and
speak.
vii)Reflecting peoples’ answers back helps check that you understand
what they are saying.
viii)Be a good listener and ask why and how.
ix) Check with the respondent that it is acceptable to continue an
interview if it looks as though it may last longer than expected.
x) Always collect demographic information such as age, gender, and
ethnicity.
xi) Do not give advice or answers that you are not in a position to
offer.
xii)Tag and date any materials, including tapes, notes, maps or other
items.
xiii)Immediately after the interview reflect on what happened. Review
any materials that have been produced. Are there any weaknesses
in the way the interview was conducted? Were any topics missed?
What new issues arose?
Chapter Seven

Focus Group Discussion

Introduction
Very many times, popular methods such as the personal interview or
direct observation have enabled researchers to collect rich, in-depth
qualitative data. Very many other times, however, such methods have
been deemed highly inappropriate. There are a number of reasons for
such inappropriateness, mostly logistical, but also mostly occasioned by
the nature of the problem and the time it would take to gather data
through such methods. Where the social researcher deems it more
convenient and more expeditious to engage small groups of respondents
in joint structured discussions, he or she will employ the collective
interview method known as the focus group discussion (FGD).
A focus group discussion is a structured group discussion of
approximately six to nine participants guided by a facilitator. An FGD
is a qualitative method geared toward obtaining in-depth information on
concepts, perceptions and ideas of a group that are of interest to the
study. Although an FGD is actually a structured focus group interview,
in effect it is more than a question-answer interaction. The participants
talk freely and spontaneously about a certain topic. This method is
particularly useful for exploring attitudes and feelings and to draw out
precise issues that may be unknown to the researcher.
Typically, focus groups are composed of homogeneous people, all
representing a particular segment of the population. People from similar
backgrounds or experiences (e.g., prison warders, ex-convicts, medical
student interns, computer programmers, or legislators) are brought
together to discuss a specific topic from within their world of
 Focus Group Discussion 95

knowledge, practice, or experience that is of interest to the researcher.

Homogeneous samples are preferred because heterogeneity has
oftentimes proved counterproductive, since it tends to inhibit free
expression of views from members that may feel somewhat inferior.
A focus group session should normally last about 60-90 minutes,
but it can under certain circumstances stretch to two hours as the
absolute maximum time. The researcher, who is the bona fide group
facilitator, keeps the discussion on track by asking a series of open-
ended questions meant to stimulate discussion and generate the requisite
data.
The FGD method derives from market research strategies in which
theories of social psychology and communication were applied and
incorporated into social science research methods. Today, FGDs have
increasingly gained prominence in applied social science research.
Social science researchers prefer FGDs to (but often in combination
with) other qualitative methods for a number of projects.

The Importance of FGDs

While the data collected through the use of focus group interviews
would no doubt possibly be collected by alternative means, there are a
number of situations where FGDs present a unique choice to researchers.
Focus group research techniques are, for instance, the unrivalled
alternative where the researcher seeks to focus research and develop
relevant research hypotheses by exploring in greater depth the problem
to be investigated and its possible causes. Consider, for example, a
scenario where the Ministry of Agriculture receives reports of an
unusually large scale starvation and famine in one region of the country.
If the problem cannot outrightly be attributed to obvious factors such as
draught, floods, large scale pest invasion or a reported disease outbreak,
the Ministry may decide to organise about three focus group discussions.
It might be helpful to engage district agricultural officers/extension
workers in the region in one FGD, farmers in another focus group, and
local leaders in yet another group discussion. This will help the Ministry
identify both the potential and actual causes of the problem and then
develop appropriate interventions.
Quite often, social researchers want to formulate appropriate
questions for more structured, larger scale surveys. A discussion with
96 Social Research Methodology: An Introduction

mothers of children under two in Western Uganda (this author

conducted a study there two years ago) would, for instance, reveal that
they associate childhood diarrhoea with challenges of growing up such
as sitting up, crawling, and developing milk teeth. Because this is a
semi-literate rural population that does not appreciate the link between
feeding habits and hygiene to childhood diarrhoea, such a discussion
comes in handy if the researcher is developing a questionnaire for a
large scale survey. Using the community’s notions and terms, the
researcher would more carefully formulate questions and appropriately
use concepts to yield the most authentic data.
Focus group discussions have become a time-tested means for
policy planners and community workers to develop appropriate
messages for extension work and community intervention programmes
(especially in agriculture, health and education) and later evaluate the
messages for clarity and effectiveness. The district health educator in
Rukungiri district wanted to develop a health education programme on
malaria among children under five. The programme was to focus on the
identification and immediate response to malaria cases, as well as the
problems most often encountered by child care providers and what to do
about them. Focus group discussions were used to explore relevant local
concepts and perceptions of childhood malaria.* They were also used to
develop messages and test draft instruments on child care providers of
different socioeconomic and other backgrounds in the district.
Community interventions are definitely not always without
unexpected problems that require identifying and fixing. Focus group
discussions may, whenever such scenarios arise, be called into action.
Suppose an analysis of an intervention programme such as Universal
Primary Education (UPE) in, say, Mayuge district reveals that enrolment
has on average improved from 43% over the last seven years to 89%.
Suppose also that there is a markedly sharp variation in enrolment, say
50-93%, in the different sub-counties of the district. A number of FGDs
may be conducted for different stake holders in these sub-counties to

* For example, convulsions, which are a condition of acute malaria among

infants, were not associated with malaria per se, but were considered a disease on its
own. It was also variously known as ebuyaga, ebihungu, engundu, and ebyankore. They
also had local responses which unfortunately did not work and resulted in death if no
immediate medical attention was sought.
 Focus Group Discussion 97

find out why enrolment is not hitting 100% and why in some sub-
counties it has remained almost non-responsive to the intervention.
Other kinds of problems that may warrant the use of FGDs include
where a researcher seeks to collect a wide variety of local terms and
expressions used to describe an act, event or phenomenon; to explore the
meanings of survey findings that cannot be explained statistically; or to
explore controversial topics that different segments of society may view
from quite different perspectives.

The Focus Group Process

A focus group session can be regarded as a mini-study in itself.
Therefore, a successful FGD must follow a well thought out process:
careful planning of the study, clarity of the major concepts and data
required, the sources of data (who participates in the sessions), requisite
expertise to steer and record the discussions, up to the final report.
First and foremost, the researcher must identify the issues that are
the subject of the study, and then formulate clear and very specific
objectives relating to these issues. This is a very crucial stage as a
flawed starting point will surely lead to a flawed outcome. These
objectives are the single most important guide to the research team in the
formulation of the discussion questions.
It is imperative at this stage that the researcher performs a thorough
scansion of the environment ahead of preparation of interview questions
and selection of participants. As a matter of fact, any FGD requires good
knowledge of local conditions and the nature of the population.
Communities are almost always heterogeneous, and their differences
will be reflected in their perceptions of the problems they suffer from
and their motivations and aspirations. A researcher must therefore be
aware of these differences, lest he or she should fail to constitute
appropriate groups with the right participants to deliver the required
information. The researcher must thus explore the area and identify
possible target groups. This analysis of the environment should
preferably be carried out in a participatory manner, with persons from
that very population.
These two stages thus accomplished, the researcher should now be
in a position to prepare a set of questions that will provide the overall
direction for the discussion. The questions should flow from the general
98 Social Research Methodology: An Introduction

to the specific in order to motivate the participants and bring them

systematically along. Questions should be open-ended, simple, unbiased,
and focused to the issue at hand. The questions are merely a guide
whose ultimate purpose is to stimulate discussion. Undoubtedly the
discussion will illicit more questions and bring up issues that the
facilitator may want to follow-up on.
The researcher is at this stage ready to identify and recruit the
participants. Participants should have a similar background in relation
to the issue under investigation. The researcher should take ample care
to identify the types of people that are able to provide the answers
needed. He or she must think about the key population groups that may
have an interest in the issues being researched. One may want to form
several different, separate groups that represent different viewpoints.
The groups can be formed based on several different characteristics: age,
income, gender, race, place of work, place of abode, unemployed, single
mothers, students, retired, education, etc. Whatever the composition of
the group, what is important is that it should not inhibit free discussion.
Quite often, a researcher may need to obtain information on a topic
from several different categories of informants who are likely to discuss
it from different perspectives in separate FGDs, although these groups
may at a later stage be joined. If the researcher is an outsider in the area
of the research, he or she may have to rely on the key informants for the
first selection of participants in FGDs. These key informants to whom
the purpose and process of the study have been thoroughly explained
might each suggest some individuals who could be invited to a focus
group discussion.

The Setting and Required Hardware

For an effective FGD, communication and interaction during the session
should be encouraged in every way possible. Attention should be paid
to such details as the sitting arrangement. Increasingly, a circular sitting
arrangement tends to be very appropriate as it promotes the sense of
involvement and increases mutual participation. All manner of
distraction should be controlled, and the facilitator should ensure
sufficient quietness, adequate lighting, optimal temperature, etc. He or
she should ensure a neutral setting that will encourage participants to
freely express their views.
 Focus Group Discussion 99

A range of materials are used to support FGDs. These normally

include notebooks, pens and voice recorders to record the proceedings;
and video, sound recordings, and pictures to introduce topics for
discussion. Recording the discussion on tape has the advantage of being
able to play it back and pick up salient points after the discussion is
over. The disadvantage is that transcribing from tape takes a long
time—it can take up to five hours to transcribe an hour’s tape recording
of a focus group discussion plus another couple of hours to listen to the
tape again and check the transcription for accuracy. Even then important
points may be missed if the tape recording is not accompanied by
detailed notes on who the participants were, the order in which they
spoke, and the non-verbal language which accompanied what was said.
If the discussion took place in a language not understood by the
researcher, translation can mean added time and financial costs.
Nevertheless, the importance of the FGD as a powerful qualitative data
collection method is not diluted by these limitations.

The Facilitator and the Reporter

It is not possible that the researcher will always be in a position to
facilitate the discussions himself or herself. There are a number of
obvious reasons: he or she may not share the language with the
participants, he or she may lack the necessary expertise at facilitation,
he or she may have communication difficulties, etc. One of the members
of the research team—or even a person hired specifically for the
purpose—should act as a ‘facilitator’ or ‘moderator’ for the focus group
discussion, and another as a ‘recorder’ or ‘reporter’. The facilitator’s
role, however, is not one of an expert on the topic: he or she must be
able to stimulate and support discussion as the recorder takes care of the
proceedings.

The functions of the facilitator

Stimulation and support of the discussion mean much more than what
a chairperson does during a school debate. The fact that one is not
‘playing the expert’ does not make him or her less one. Therefore, a
number of functions are to be expected of a facilitator.
100 Social Research Methodology: An Introduction

1. Introducing the session. It is the facilitator to call the session to

order. This entails introducing himself or herself, the recorder, and
inviting the participants to one-by-one introduce themselves. It also
entails explanation of the purpose of the FGD (although they will have
known by now) and the kind of information required.

2. Building rapport. The success of the discussion will ultimately

depend on the relationship between the facilitator and the participants.
The participants must be made to feel at home and comfortably express
themselves. A soft drink at the start of the session may warm their
hearts. A free, informal interaction with individuals especially before the
formal session kicks off may bring you closer to them. As the session
begins, assure them of confidentiality of their opinions, but you may
also request their co-operation to keep what has been discussed
confidential.

3. Encouraging involvement. Every participant will feel good to

know that he or she is ‘educating’ you. The facilitator should thus pose
questions, encourage as many participants as possible to express their
views, and impress it upon them that he or she is very interested
in—even enlightened by—their ideas. As such, the facilitator should
express utmost enthusiasm in all the answers (there are not supposed to
be right or wrong answers). It should be borne in mind that an FGD
ought not be a question-and-answer session. The facilitator should
therefore always endeavour to ask for clarification; to reorient the
discussion when it goes ‘off the track’; to use one participant’s remark
to direct a question to another; to thank the participant and change the
subject; etc. Even a reluctant participant can be brought on board by , for
example, frequently seeking his or her opinion.

4. Directing the flow of the discussion. The facilitator can also play
the traditional role of the chairperson by movin the discussion from
topic to topic. As long as he or she ‘injects life’ in the discussion,
participants will spontaneously jump from one topic to another. This is
not altogether bad since useful additional information may surface.
However, the facilitator should, after a while, summarise the points
brought up and reorient the discussion. Careful adherence to the time
 Focus Group Discussion 101

allocated to various topics is extremely necessary, not only for purposes

of maintaining interest, but also of completing the exercise.

5. Handling sensitive questions. A discussion may lose tempo or

stop altogether as a result of the sensitivity of the topic under discussion.
To keep the session going without losing out on the information at stake,
the facilitator may seek written anonymous responses or opinions. If
such a topic has been discussed elsewhere in the past, the facilitator may
summarise for the group some of the opinions from such discussions,
focussing on one or two major contrasting opinions. However, in case
these strategies do not apply, the facilitator may encourage the session
to split into working groups, summarise the opinions, and have one
member of each group present the summary at the plenary so that these
opinions can still be discussed.

6. Summarising and cross-checking issues. It is extremely

important to time the discussion so that at the end there is still some time
to summarise the issues and cross-check with the participants for
agreement or disagreement. Additional comments should also be
incorporated. The facilitator should at the end thank the participants and
assure them that their ideas have been a valuable contribution and will
be used for planning the proposed larger survey, intervention, or
development of materials.

Functions of the recorder

The recorder should be no less a knowledgeable person than the
facilitator. He or she should keep a record of the content of the
discussion as well as emotional reactions and important aspects of group
interaction. The tone of the meeting and the mood created by the group
process are important indicators of the validity of the information
collected during the FGD.
Ideally, the recorder should aim to record the date, time, place, and
the names and characteristics of the participants. He or she should then
capture the general description of the group dynamics such as the level
of participation, the presence of a dominant participant, the level of
interest, etc. More importantly, however, the recorder should get a full
verbatim record of the opinions of the participants, especially for key
102 Social Research Methodology: An Introduction

statements. The emotional aspects of the discussion such as

demonstrated reluctance of some participants, strong feelings attached
to certain opinions, etc; and the terminology used, particularly in FGDs
that are conducted for purposes of developing questionnaires or
intervention materials must also be recorded.

Evaluation of the FGD Method

Implementation of FGDs is an iterative process; each focus group
discussion builds on the previous one, with a slightly elaborated or
better-focussed set of themes for discussion. Provided the groups have
been well chosen, in terms of composition and number, FGDs can be a
powerful research tool, which is associated with the following
advantages:
i) they provide valuable spontaneous information in a short period of
time and at relatively low cost;
ii) they enable the researcher get rich data in the participants’ own
words and develop deeper insights on the problem under study;
iii) respondents are able to build on one another’s responses and come
up with ideas they might not have thought of in a one-on-one
interview;
iv) their homogeneity helps to keep participants interested and
involved, which guarantees a high response rate
v) if properly planned and facilitated, they are suitable for obtaining
data from children or people with low levels of literacy;
vi) they provide the opportunity to involve participants in data
analysis. The researcher may ask such questions as: “out of the
issues we have talked about, which ones are most important to
you?”
vii) participants can act as checks and balances on one another, by
identifying factual errors or extreme views;
viii)when used in combination with other data-gathering tools, they
generate good quality data.
However, the FGD method is riddled with some shortcomings,
which include the following:
i) there is a big problem of group influence on individual responses.
Because participants build their views on those of others during the
discussion, their responses are not independent;
 Focus Group Discussion 103

ii) a few dominant focus group members can skew the session: they
may monopolize or hijack the discussion for their own purposes;
iii) FGDs require a skilled and experienced facilitator, who is not
always easy to find. Poor facilitation can stifle the discussion and
render the investigation futile;
iv) a facilitator has less control over proceedings than with individual
interviews;
v) this method might not be appropriate for discussing very sensitive
issues: group members may hesitate to air their feelings and
experiences freely;
vi) data that results from a focus group discussions requires skill and
experience to analyze, and this makes the method less attractive to
novice researchers.

In a Nutshell
FGDs should not be used for quantitative research purposes, such as the
testing of hypotheses or the generalisation of findings to larger areas,
which normally require more elaborate surveys. However, FGDs can
profitably complement such surveys or other qualitative techniques.
Depending on the topic, it may also be risky to use FGDs as a single
tool. In group discussions, people tend to centre their opinions on the
most common ones, on “social norms.” Yet in reality, opinions and
behaviour are more diverse. Therefore, it is advisable to combine FGDs
with at least some key informant and in-depth interviews. Explicitly
soliciting other views during FGDs should also be encouraged.
Chapter Eight

Field Observation Methods

Introduction
In quantitative research, the researcher is basically not a participant
observer—he or she neither participates in nor influences what is being
studied. In qualitative research, however, the researcher can learn the
most by participating and/or being immersed in a research situation. The
researcher’s choice to use this approach is determined, not only by sheer
researcher interests and the nature of audience the researcher seeks
findings from, but also by consideration of factors such as the time
frame, money and other resources available.
Until the 19th century, naturalistic or direct field observation was
the chief—if not sole—method of data collection in the natural sciences
(Mbaaga, 2000), and the object of observation was limited to non-human
organisms and their behaviour. Field research based on direct
observation of social and institutional order of communities was born
out of the two related disciplines of anthropology and sociology.
Ethnographic fieldwork—the study of native cultures by learning the
native language, observing and taking part in native life—can be said to
have originated with the founders of modern anthropology. However,
modern field-based social research is more associated with the social
reform tradition of sociology than anthropology. In the late nineteenth
century, social reformers and sociologists turned first to social surveys
as a departure from the more orthodox experimental research and often
case studies. Increasingly, however, lots of questions about social
behaviour, especially in organic social settings, remained unanswered.
 Field Observation Methods 105

There was an obvious need to move to a more flexible methodology of

social investigation based primarily on field observation.
Today, field observation is an indispensable method in a number of
situations. First of all, it is an undisputable choice for a researcher with
a keen interest in the study of values and behaviour of communities such
as ethnic and other social groups. This provides the researcher with an
inside view of reality. Secondly, a researcher seeking to establish the
immediate impact of an event on a given community or social group
finds field observation handy. This is not limited to where the event may
be rarely occurring and ad hoc (such as natural calamity or declaration
of curfew), but also where it is quite dynamic (such as continual
admission of inmates in a prison community. Thirdly, the method may
be used in the study of a phenomenon on which there is not much
knowledge. This is particularly so in cases where the researcher adopts
a grounded theory approach to research. Finally, field observation is
resorted to in situations where methodological problems, resources, or
ethical challenges preclude the adoption of other research strategies.

The Nature of Field Observation

Ethnographic research offers an orientation to understanding the process
and structure of a social setting, and employs data collection methods
that are consistent with this orientation. Field observation, which enables
the researcher direct access to information on social phenomena, is a
very important such a method. Through this method, the researcher
becomes a first hand eye-witness of everyday social actions. Researchers
observe different forms of overt phenomena which manifest themselves
generally as behaviour, actions, and/or interactions.
In social surveys and other quantitative studies, the importance of
the researcher not becoming “personally involved” with the respondent
is greatly emphasize. The researcher maintains a significant personal and
social distance from the people on whom the research is being
conducted. In field observation, however, the researcher aims to
understand the social world and social reality from the subjects’
perspectives.* This method involves the researcher “getting to know” the

* Max Weber refers to this verstehen, meaning direct understanding, as the

motive or rationale of phenomenological research (Bryman, 1988: 56-7).
106 Social Research Methodology: An Introduction

people, groups or communities he or she is studying; and this he or she

does by entering their world and participating either openly or secretly
in that world. The researcher attempts to experience events in the way
his or her subjects experience them. Thus, as a matter of fact, the results
of observational research are inevitably influenced by the researchers’
perceptions and expectations.
Field observation is, in a nutshell, a naturalistic method because it
involves direct understanding, through personal experience of the
researcher, of phenomena in their natural setting. It is ethnographic since
it involves studying cultural behaviour by getting as close to that
behaviour as possible in order to achieve an authentic understanding of
it. The researcher attempts to understand the motives and meanings of
people’s behaviour by “seeing through” those people.

Common Concepts in Field Observation

The field observation method, as we have seen, is employed in situ,
where the events or behaviour to be observed are occurring naturally. As
such, the researcher adopts the Mahomet-going-to-the-mountain
approach, which brings into play all manner of problems associated with
direct personal interaction. This method is thus characterised by a
number of concepts, which I now briefly turn to.

Social distance
This concept connotes researcher objectivity, and refers to the ability of
the social researcher to remain detached, aloof or personally separated
from the people he or she is investigating. Objectivity is concerned with
the researcher’s ability 1) to avoid letting his or her personal beliefs,
attitudes and values creep into the research process; and 2) to avoid
influencing the way respondents behave or respond to the questions. In
observational studies, especially those involving direct participation
(complete participant and participant-as-observer), social distance is
largely elusive.

Subjectivity
This is the researcher’s deliberate attempt to “see for himself or herself”
the behaviour that people describe in an interview or questionnaire. The
researcher attempts to express the nature or quality of people’s
 Field Observation Methods 107

behaviour from his or her own interpretation. The biggest danger is

obviously the researcher’s attitudes, beliefs and other inclinations
creeping in and influencing the interpretation.

Empathy
This is the ability, as human beings, to see ourselves as we think other
people see us. It is a valuable human ability that the researcher should
exploit in order to understand how people experience the social world.
Researchers who use participant observation aim to discover the nature
of social reality by understanding the subject’s perception,
understanding, or interpretation of the social world. The participant
observer exploits the human ability to empathise—the main objective
being to participate in a social group while, at the same time, employing
the insights and understanding of a scientific observer. The researcher
strives to observe and experience the world as a participant, while
retaining an observer’s stance for understanding, analysis and
explanation.

Integration
The concept of “integration” involves the idea of having a sense of
belonging to a group. If, for example, a researcher sees himself or
herself as being a part of the group being studied (in terms of such things
as norms, values, roles and so forth), he or she is well-integrated into
that group. Over-integration is a real problem in participant observation
for it erodes the much needed social distance from the group being
observed and leads to the problem of “going native.”

Getting access
Access to a group can be understood in two main perspectives. The first
relates to the question of “gaining entry” into a group one seeks to
observe; and the second to the information one is able to collect once
one is part of the group. In stratified, hierarchical or exclusive groups,
it is not easy for a researcher to conduct covert observation because the
he or she does not have the same characteristics as the group. The
108 Social Research Methodology: An Introduction

alternative is overt observation, where access to such groups may be

secured through a sponsor.*

Going native
In participant observation, the researcher must strive to separate the roles
of “participant” and “observer.” Otherwise, one might become so
immersed in the behaviour of the group, cease to be an observer
(researcher) and simply become a participant (just another member of
the group). This is called “going native,” where the researcher can no
longer balance the roles of participant and observer and, instead, simply
participates like any other group member. A researcher who “goes
native” effectively stops being a researcher.

Observer Roles in Field Observation

A researcher conducting field observation proceeds from a well
considered set of observation dynamics—the context of the research,
type of group to observe, time and resources available and so forth. A
researcher becomes a player of sorts; and taking into consideration the
above dynamics, the role the researcher plays is very crucial for the kind
of observation he or she makes and type of data they gather.
Gold (1969) and Mbaaga (2000) point out four alternative roles or
levels of participation a field observer can assume in a group. These, one
may say, are the physical positions a researcher takes in order to
observe. In other words, it is where in relation to the group/community
being observed the researcher”stands” to observe. These levels are 1)
complete participant, 2) participant-as-observer, 3) observer-as-
participant; and 4) complete observer.
Under the role of complete participant, the researcher disguises and
immerses himself or herself in the group and takes a covert position so
the group he or she is studying is not aware they are being studied. The
researcher thus makes covert observation from within the group.
Whereas the researcher gets 100 percent access to the events and

* In field observations, a “sponsor” is someone within the group you want

to study who is willing to both vouch for you and explain your presence to the other
group members. Sponsors are often a good way for researchers to gain entrance into a
group, especially if the sponsor has a relatively high status within the group. That way,
the other group members are less likely to show hostility to the researcher.
 Field Observation Methods 109

behaviour of the group he or she is studying, this role is very risky in

that one tends to eventually lose the stranger value (which is an
important researcher-identity) and go native. Once gone native, the
researcher may lose the ability to evaluate the observed behaviour and
appreciate its meaning.
The second role is that of participant-as-observer, where the
researcher actively participates in the life and activities of the group
being studied, but makes the group members aware that they are being
studied. This researcher is therefore overt, and is not compelled to do
everything the group members do. For example, if one is studying the
Imbalu rights of passage among the Bamasaba of Eastern Uganda, one
may participate in the preparations and songs and dances, but one may
not offer to be a candidate for circumcision. However, this researcher
also risks going native. But even more seriously, since the group
members are aware that they are being studied, they may feign
behaviour.
The third observer role is that of observer-as-participant, where the
researcher makes interactions with the group members but does not
participate in their activities. The group members are aware he or she is
studying them, and his or her role indeed remains that one of a
researcher. It is therefore overt observation, and has the disadvantage of
the researcher failing to access some of the important aspects of the
event or behaviour being observed.
The fourth role is that of complete observer, where the researcher
keeps a reasonable social distance and observes the group without its
members being aware that they are being observed. The researcher thus
makes covert observation from outside the group. It is to be noted that
whereas covert observation (as in this and the first role) is the best way
of accessing behaviour in its most naturally occurring mode, it
nevertheless has serious ethical implications. One such is the question
of observing people without their knowledge or consent.
From all the above four observer roles or levels of participation,
two most popular classifications of observation emerge: those indicating
the role of the researcher (participant and non-participant observation)
and those indicating the position taken by the researcher (overt and
covert observation). What determines the role the researcher chooses to
110 Social Research Methodology: An Introduction

play, the position taken, and the limitations faced are the subject of the
succeeding sections.

Types of Observation
As I have indicated, the different classifications of field observation are
couched on the foregoing roles. Chief of these are participant and non-
participant observations, and then overt and covert observations.

Participant Observation
In participant observation, the researcher is part of the situation or social
phenomena being observed. Sometimes participant observations are
disguised, so that other members in the group do not know they are
being observed. At other times, however, they are undisguised. The
researcher participates actively for an extended period of time, and
sometimes it may require him or her to live or work in that area. This
requirement assumes that the researcher/observer will become an
accepted member of the group or community. Historically, field research
has been associated most strongly with participant observation.
There are both advantages and disadvantages associated with
participant observation. The advantages are mainly to do with giving the
observer access to situations in which a formal, detached observation
would not be welcomed. Likewise, participation gives the researcher a
first hand sense of the quality of the experience, and what it feels like to
be in that situation. Therefore, by combining the role of participant and
that of observer the researcher may be able to understand and explain
situations and events an ordinary member of the group is incapable of.
Secondly, by looking at behaviour “from inside” the researcher is able
to explain why people would want to behave in certain ways. Moreover,
by becoming a part of a group the researcher is able to see things from
its members’ perspective, and get a clearer understanding about their
behaviour.
However, there are important potential disadvantages as well. First
of all, the observer may, through their actions, have an impact on the
situation being studied, and through active participation, lose the
capacity to be objective. Secondly, given that one of the purposes of
participant observation is to experience the world from the viewpoint of
the people being studied, this purpose may fail if the researcher does not
 Field Observation Methods 111

become sufficiently involved with the group. More still, the data
collected through observation is based on the subjective impressions of
the observer. Readers of the research must thus dwell on trust that the
behaviour, events and experiences were exactly as described by the
researcher.

Non-participant Observation
In non-participant observation, the researcher observes participants and
their behaviour at a distance. He or she attempts to observe people
without interacting with them, and without their knowledge that they are
being observed. He or she is more or less an eavesdropper. According
to our discussion of observer roles in the preceding section, a researcher
in non-participant observation is actually a complete observer. This form
of observation is most routinely used by psychologists studying children
and animal behaviour.
The major strength of this type of observation is that the researcher
generally does not influence the behaviour of the observed, so what is
observed is likely to be a genuine picture of the reality. However, by not
interacting with the observed people, there is a likelihood that the
observer and the observed remain in different worlds. The researcher’s
perspective on the observed behaviour could remain limited. There is
also the ethical problem of observing people without their knowledge
and consent.

Overt Observation
In this form of observation, the researcher becomes open to the group he
or she is going to study. He or she informs the members of the group
(either personally or through a sponsor) about such things as the purpose
of the research, its scope, how long the research will last and so forth.
Therefore, the research is done with the permission and co-operation of
the group and the fact of being open with the group being researched
upon carries with it certain advantages and disadvantages as far as the
overall conduct of the research is concerned.
The technique on overt observation is that of “hanging-around” the
group, observing behaviour, asking questions and recording what is
happening. The researcher is involved, to some extent, with the group
itself (but not as a full participating member) and experiences things as
112 Social Research Methodology: An Introduction

group members experience them. The researcher endeavours to separate

the roles of participant and observer, thereby reducing the chance of
becoming so immersed in the behaviour of the group and becoming just
another member of the group (going native). A researcher who “goes
native” is in effect no longer a researcher.
Overt observation is associated with a number of advantages. First
of all, because the researcher is able to join a group openly, the problem
of access is fairly easily overcome. Secondly, data can be recorded while
observing people’s behaviour since the people are aware of what is
going on. Also, it is less likely that the researcher will go native.
However, the knowledge of being openly observed may affect the way
participants behave, and this has a significant effect on the validity of
the data collected. This is called the “observer effect” (or Hawthorne
effect). Secondly, this form of observation puts demands on the
researcher, not only in terms of observing and recording behaviour
accurately, but also in terms of interpreting data.

Covert Observation
This form of observation may be participant or non-participant: it
involves the researcher studying a group—whether from within or
outside—without informing the members of that group. In this respect,
the research is carried-out secretly. A complete participant and a
complete observer are both covert observers since the groups they are
observing are not aware of their being observed.
This method, too, has certain advantages and disadvantages for the
researcher, who has to balance the twin roles of researcher and
participant while keeping the former role secret from other group
members. Some of the advantages of covert observation are, first, it
makes it possible for the researcher to gain access to groups that would
not normally allow themselves to be studied. A researcher can
investigate behaviour that is normally hidden from the public. This may
be illegal behaviour (e.g., human sacrifice), deviant behaviour (e.g.,
homosexuality), and various forms of secret behaviour (e.g., hacking or
devil worship). Secondly, it avoids the problem of “observer effect”
since the group members are unaware that they are being studied. The
researcher can thus be sure that what he or she is observing is the
people’s “normal behaviour.”
 Field Observation Methods 113

However, it may not be possible for a researcher to covertly join a

group with exclusive characteristics, e.g., a man in a group of nuns, a
civilian in a military barracks, etc. Secondly, it is not easy to generate
and record data because the fact of being in the group secretly would
make the group suspicious if one openly recorded conversations and
other behaviour. Also, there is the problem of the researcher “going
native” and thereby ceasing to observe the group objectively. Another
problem is that even where the researcher quietly writes down
observations in a field diary at the end of the day, there are clear
problems of accuracy, memory and interpretation. This also raises
questions about the reliability and validity of covert research. Finally,
there are general ethical questions of observing people without their
consent.

Other Classifications
We can talk of naturalistic versus controlled observation or observation
with intervention. Naturalistic observation occurs when a researcher
conducts observations in a naturally occurring situation, without
becoming actively involved. The researcher makes no attempt to control
or change what happens. The research task is to make a detailed record
of the events that occur and of apparent relationships between these
events. Controlled observation on the other hand involves conducting
observations in a situation that has been staged or arranged in some way
by the researcher.
Another common classification is structured versus unstructured
observation. Structured observation is used when the researcher is
looking for specific behaviours, actions, and/or interactions. When using
structured observation techniques, the researcher identifies beforehand
which behaviours are to be observed and recorded. This method enables
the researcher to have greater control over the sampling process and
size, measurement of error, and permits stronger generalizations and
checks on reliability and validity. Unstructured observation on the other
hand, is the unplanned, informal, watching and recording of behaviours
as they occur in a natural environment. It places no restriction on what
the observer may note: all behaviour in the episode under study is
monitored.
114 Social Research Methodology: An Introduction

Recording and Reporting Observational Data

Observational data is normally recorded in a field log—a diary or not
book. The recording function is quite important and determines the
quality and authenticity of what is reported. There are three types of
notes an observer may make. The first type is called the mental notes.
These are not physical notes per se—the researcher stores the observed
events or behaviour in his or her memory, and later relies on this
memory to make a full transcript of the observation. The second type is
called the jotted notes. These are shorthand notations (phrases,
statements, key words, etc) that are intelligible to the observer, which he
or she will later rely upon to write a full report of the observation. The
third type is called field notes. These are detailed descriptions of people,
events, settings or patterns of behaviour that were observed. Field notes
are normally written on a very regular basis, commonly at every end of
day.
It should be remembered that observation is a protracted method,
lasting several weeks onwards. In cases where it stretches into several
months, the researcher faces a risk of going native. It has been
suggested, however, that in order to avoid this risk, the researcher should
occasionally leave the field for short lapses (Mbaaga, 2000). The
researcher should also have his or her field notes peer-reviewed by
colleagues that are not in the field.
It should be remembered also that qualitative data analysis
normally goes hand-in-hand with data collection. This is the case with
observational data. There are generally three stages of observational data
analysis. The first involves identification and definition of a range of
problems, questions and objectives behind the observation. The second
stage, as the observation sets in motion, is concerned with the selection
of a few of the preliminary questions and objectives to pursue. The third
stage is to synthesize the observational findings into a model that
explains the observed phenomena, or to use those findings to build a
theory. Such models and theories are usually tested in subsequent
studies.
After data analysis, the researcher embarks on reporting the
findings. Observational reports are narrative, and may either be
anecdotal or detailed narrative accounts of behaviour recorded in a
sequential manner. An anecdotal report is a brief narrative describing a
 Field Observation Methods 115

specific incident or aspect of behavior that was of interest to the

observer. It should be factual, objective, and graphical. A detailed
sequential narrative, on the other hand, reports on all behaviour as it
happened.
In order to ensure “objective” observation, the researcher should
always endeavour to record only the facts, and record every detail
without omitting anything. One should not interpret as he or she
observes and records, and should not record anything they do not see.

General Evaluation of Observation

The observation method is associated with a number of relative strengths
as well as weaknesses. A researcher’s choice of this method is usually
influenced by the nature of the problem to be investigated (i.e. if it
cannot be studied by any other method), or by its relative advantage over
other methods.

Advantages of Observation

1. Flexibility. As the study gets under way, the researcher can still
adjust the goals and objectives of the study. Similarly, because a
researcher cannot decide in advance what is important when studying
social behaviour, he or she can react to events or ideas, follow leads,
pursue avenues of research that had not occurred to him or her before
their involvement with a group.

2. Spontaneity. Observation enables the researcher to study behavior as

it occurs. If, for example, one is studying reactive behaviour to
accidents, one need not bother with issues of sampling and instrument
design in readiness for a “specific” accident.

3. Alternative for unwilling or unable respondents. There are some

instances when informants cannot express themselves meaningfully. For
example, in research dealing with children or other persons who cannot
speak or understand the issues at hand, observation the method. In some
other instances, individuals may not be willing to be interviewed or to
fill in a questionnaire. These would be studied through observation.
116 Social Research Methodology: An Introduction

4. Quality and depth of information. Field observation generates a

highly detailed, high-quality, and first hand account of the phenomena
being studied. This is primarily because the researcher does not impose
structures on the observed phenomena, but instead attempts to analyze
the complexities found in those phenomena.

5. The opportunity for understanding (empathy). It is easier for the

researcher to understand the group dynamics—social pressures,
influences, norms, etc—that create particular forms of behaviour. This
gives the researcher sustained insights into individual and group
behaviour, and authority to formulate theories that explain such
behaviour.

6. No responses errors. The observation method is free of the response

errors that are largely associated with interviews and questionnaire
situations. The researcher does not have to ask questions to selected
respondents or follow-up respondents to collect back research
instruments—which often are either not fully answered or not recovered
in surveys.

7. External validity. It has been shown that the observation method has
better external validity. In some studies, such as those on natural
disaster, observational results may be satisfactorily generalized—a
feature of quantitative research. Note that even in quantitative research,
the researcher merely relies on a sample to generalise to the population.
The problem is that in a number of cases, the samples are not accurately
representative, a situation that may lead to wrong generalization.

Limitations of Observation

1. Scope and scale of observational studies. Most field observation is

restricted to fairly small-scale studies carried out over a long period.
Moreover, the group being studied is unlikely to be representative of any
other social group. It may also be unlikely that a researcher will
adequately be able to generalise their findings from one study to the
next.
 Field Observation Methods 117

2. Researcher’s level of participation. A researcher has to learn the

culture of a group if he or she is to participate fully in their behaviour
and this may not always be easy or possible. If a researcher is too young,
too old, too male or too female for the group they want to research
about, this will cause problems of participation. A researcher’s ability
to blend seamlessly into a group—which is rare—is absolutely crucial
to the success or failure of the research project.

3. The skills required of a researcher

Field observation requires a great deal of skill and commitment from the
researcher. The crucial qualities are the ability to fit-in with the people
being studied, ability to communicate with groups members, ability to
separate the role of participant from that of observer, tact, etc, which are
a rare endowment to a single individual researcher.

4. Time frame. Ethnographic research is generally very time-consuming

– it may take months, even years for a researcher to become fully-
involved with a group, writing and publishing the findings.

5. Project timing and convenience. With this method of data collection,

the researcher may have to wait for a long time before the phenomenon
or event to be observed occurs. This implies that many phenomena often
pass un observed since researchers might not be aware when they would
occur.

6. Attitudes and opinions. Observation cannot be adequately employed

to study attitudes and opinions. This is because the readiness and
predispositions of people to behave or react in certain ways to stimuli
cannot readily be investigated with observation.

7. Replication. Participant observation studies, by their very nature, are

not replicable—a fact that gives them limited reliability. Also, there are
many simultaneous actions or events in any complex social group. This
poses difficulties for the researcher about who or what to observe, and
when to observe them. Significant evidence may be missed, leading to
invalid interpretations of the group’s behaviour.
118 Social Research Methodology: An Introduction

8. Naturalism. Whether people are being studied in situ, and whether

the research is carried out with the permission and co-operation of the
participants, there is no way of measuring how the presence of the
observer influences the behaviour of group members.

9. Influence of environmental factors. Observation research relies

heavily on the sense organs, which are quite inadequate. What one
observes or perceives is determined by his or her current physiological
and psychological state, which is itself a function of the state of the
external environment. The observer may thus be tired, moody, or
indisposed, and this may hamper his or her alertness or proper judgment.
Chapter Nine

Participatory Rural Appraisal

Introduction
Formal social research methods such as structured surveys, experiments,
case studies, etc., are powerful tools for use in gathering and analysing
data needed for socioeconomic policy and planning. Indeed, in many
developing countries, the traditional approach to producing a
socioeconomic policy or development plan has been to contract a team
of experts to carry out detailed formal surveys to gather data, analyse it,
and formulate policies and plans. However, combined with digital
processing, these surveys often produced spurious quantification,
usually in a long period of time and at a great cost. With time, many
policies and plans that were formulated on the basis of such surveys
failed, and this led to two developments. First was the realization that
decision-making in the real world is very much a time-bound activity.
Although management may commission research, real-time decision-
making processes are based on the rapid collection and analysis of data.
Successful enterprises, projects and programmes thus depend on the
immediate reaction to situations. Second was the realisation that a great
deal of the experience and expertise needed for local development could
be found from the local inhabitants. It became obvious that without their
involvement in the analysis and decision processes, their sense of
ownership of the programmes was diminished as were the chances of
successful implementation of those programmes.
By the 1980s, it was so obvious that a development practitioner
such as a project or programme manager in need of information for
urgent decision-making would not be adequately served by the formal
methods. He or she sought to rely on methods that were not only quicker
120 Social Research Methodology: An Introduction

and more cost-effective to gather and analyse information, but also those
that involved the beneficiaries of the programmes in the planning
process. An approach called “rapid appraisal” was developed to fill this
gap. Over time, this approach has evolved into two related methods: the
Rapid Rural Appraisal (RRA) and the Participatory Rural Appraisal
(PRA). Consequently, development workers in rural settings have taken
a significant shift in emphasis from socio-statistical surveys and related
methods based on sampling theory to the more flexible, rapid and
participatory methods. While the hard statistics gathered from large
samples and generalised to entire populations undisputedly have a
legitimate place in economic and social analysis, informal and more
rapid surveys are increasingly gaining both academic and practical
legitimacy as a formidable complement, if not an alternative, to the
formal, structured methods.

The Rapid Appraisal Methodology

It is quite important that we discuss the nature, application and core
methods of the rapid appraisal regime generally before we can highlight
the nature of RRA and delve into the details of the PRA as research
methods. It is only then that a clear understanding of the underlying
philosophical presumptions, rationale and working environment of the
PRA can be appreciated.

The nature of rapid appraisal

The rapid appraisal methodology is generally a hybrid of two entirely
different intellectual traditions of social science research–logical
positivism and phenomenology.* This hybrid in a given development
situation will reflect the peculiar circumstances within that situation and
how the elements from each of the said intellectual traditions have been
balanced. In fact, Kumar (1993:11) has defined rapid appraisal as falling
within a continuum of various formal and informal modes of data
collection used to provide decision-related information in development
settings. On one extreme end of this continuum are highly informal
modes of data collection that have no clear, established procedures.
These modes rely on the researcher’s intuition, experience, and common

* These are discussed in chapter one of this book.

 Participatory Rural Appraisal 121

sense; and the information they generate is neither systematic nor

verifiable. Examples would include conversations with affected people,
short and impromptu visits to the sites of existing or planned projects,
general perusal of official records, etc. The strength of such methods is
that they are quick, inexpensive and generally easy. However, there is
considerable uncertainty about the quality of information they generate.
There is also a problem of biases and prejudices that accompany the
researcher, which are inimical to the overall credibility of the
information and the decisions based on it.
The opposite extreme of the continuum constitutes highly formal
modes of data collection, which are based on sampling theory and large
samples, and include surveys, censuses, experiments, and comparisons.
They have formal procedures and generate statistical data leading to
generalised conclusions. The strength of these methods lies in the
accuracy of their data, the minimization of bias, and the possibility of
verification of information, which all lead to strong credibility of
decisions based on their data. However, these methods require longer
time and considerable resources, and it is always questionable whether
the development workers always need such extensive data to take
decisions. It is in-between these extremes thus that lie rapid appraisal
methods, which are neither highly informal nor fully formalised.

Application of rapid appraisal

The rapid appraisal methodology should be discerned in the context of
applied research. Applied research does not seek to solve theoretical
problems within the related disciplines of social science but to facilitate
a more rational decision-making process in real-life situations. Rapid
appraisal is thus best applicable in the context of this goal of applied
research – to provide timely and relevant information to decision-makers
in a situation of limited resources. Kumar (1993:9) has emphasised this
point in the following words:

the success of applied research is to be judged, not only by the

scientific criteria of validity and reliability, but also with
reference to the relevance of the research to the problem and
its timely delivery in a cost-effective fashion.
122 Social Research Methodology: An Introduction

The rapid appraisal methodology may not score highly in the validity
and reliability tests, but its ability to procure an effective balance
between time, cost and effectiveness makes it an undisputable
alternative for rural settings.

The core methods of rapid appraisal

Picciotto (1993) has outlined five rapid data collection methods as the
core methods of rapid appraisal. These are key informant interviews,
focus group discussions, community interviews, direct observation, and
informal surveys. Here below I only summarise these methods since
they are already discussed in detail in separate chapters of this book.

• Key Informant Interviews (KIs). These are the most popular

among development practitioners (project workers, extension
workers, Community Based Organizations, etc), and have two
special features. First, they are essentially qualitative and use
interview guides that list topics and issues to be covered in the
interview. Second, they require only a small number of KIs who are
selected on the basis of their knowledge, experience, and position.
The interview method is discussed in Chapter six.

• Focus Group Discussions (FGDs). This method was until recently

a popular preserve of market researchers who employed it to gauge
the reactions and attitudes of the populace towards certain brands
of products and services. Today, development experts use FGDs to
design, implement and evaluate health, agricultural and poverty
eradication projects in mostly rural areas. This method is discussed
in Chapter seven.

• Community/Group Interviews. These are somewhat similar to

FGDs in that investigators ask questions, raise issues and ask for
responses from the participants. Unlike FGDs which emphasize
interaction amongst the participants however, the interaction in
community interviews is between the investigators and the
participants.
 Participatory Rural Appraisal 123

• Direct Observation. This involves careful gathering of data, based

on well designed observation forms. Form designing takes into
consideration the nature of the object to be observed. Sometimes,
it also involves individual or group interviews, or both. It should be
noted here, that direct observation is not the same as the
ethnographic method of participant observation. There are at least
three differences. First, direct observation can be completed in days
or weeks whereas participant observation is a long term process,
taking several months or years. Second, direct observation can deal
with physical objects such as a road, housing, or produce whereas
participant observation focusses on social and cultural phenomena.
Third, researchers in participant observation empathise with the
people to gain insider perspective, but not so in direct observation.
Observation is discussed in Chapter eight.

• Informal Surveys. These are most popular in rural development

interventions. They are based on open-ended questionnaires,
administered to 25 to 50 respondents, whose selection is not based
on the random, probability techniques as used in formal surveys.
One popular technique of sample selection in this method is
convenience sampling, where respondents can be found in a bazaar,
church, rally, etc. The interview process is also marked by a
significant level of flexibility. It would appear that KI interviews
are similar to informal surveys, but there are three clear differences.
First, in KI interviews the respondents provide information about
others whereas in informal surveys, they provide information about
themselves. Second, in KI interviews different sets of questions
may be used whereas in informal surveys, questions are similar or
the same. Third, in KI interviews, the number of respondents is
smaller.

From Rapid to Participatory Rural Appraisal

The RRA and PRA methods are approaches to development planning
where local communities guide outside experts in data collection,
analysis and development of plans, and outside experts empower local
people to do their own data collection and plan development and
implementation.
124 Social Research Methodology: An Introduction

RRA is a process of learning about rural conditions in an iterative

and expeditious manner. More often than not, it is multi-disciplinary in
nature and has an in-built flexibility in the process of collecting
information. Kashyap (1992) has defined it as “any systematic activity
designed to draw inferences, conclusions, hypotheses or assessments,
including acquisition of new information in a limited period of time;”
and Dunn (1994) as “a qualitative survey methodology using a multi-
disciplinary team to formulate problems for agriculture and research
development.” Both these sources emphasise that RRA reflects the new
thinking about development, needs, and people-oriented responsibilities.
It is a process that is highly systematic, relying on interdisciplinary
teamwork and special strategies for data collection and analysis such as
triangulation, probing, and iteration.
PRA is an outgrowth of RRA and often the difference between the
two tends to be blurred. PRA has been defined as an “approach and
method for learning about rural life and conditions from, with, and by
rural people” (Chambers, 1992; Dunn, 1994). Chambers further stated
that PRA extends into analysis, planning and action. PRA closely
involves villagers and local officials in the process, and is considered
one of the popular and effective approaches to gather information in
rural areas—and now increasingly in urban settings as well, especially
slum areas. This approach was developed in early 1990s with
considerable shift in paradigm from top-down to bottom-up approach,
and from blueprint to the learning process. It is indeed a shift from
extractive survey questionnaires to experience sharing by local people.
The acronym PRA may at a glance look misleading because of the
elements ‘Rural’ and ‘Appraisal,’ yet it is applicable in urban and peri-
urban areas as well; and it can be conducted beyond the appraisal stage
to action planning, project design, and implementation. What is
important is that the process is characterized by ‘Participation,’ which
may be the participation of community members or members of an
organization. A process is considered fully participatory only when
participants are involved in needs assessment, goal-setting, planning,
policy-making, implementation, and evaluation. A PRA process can
focus on an entire community, on particular sectors or stakeholder
groups in one community or across several communities, or on an
organization. It can be applied to “communities of place” or
 Participatory Rural Appraisal 125

“communities of interest.” Data collection and analysis are undertaken

by participants, with the investigators acting to facilitate rather than to
control these processes.
The relationship between RRA and PRA illustrates the values that
led to the emergence of rapid appraisal and that still guide it. RRA is a
collection of techniques developed by rural development practitioners
for the collection of useful agricultural, social, and cultural data on their
target populations without the unwieldy investment of time usually
required for formal, scientific studies, traditional anthropological
participant observation, and tedious and often inaccurate questionnaire
surveys. In the RRA approach, researchers work more closely with
subject populations than has been typical of rural development research.
Together with the local beneficiaries, they undertake such activities as
diagramming aspects of the rural activity such as the farming system and
drawing maps of a village’s resources. The PRA repertoire includes
several of the RRA techniques.
The major difference between the two approaches lies not in any of
the techniques, but in the emphasis that PRA places on participation.
RRA generates information for planners while PRA shifts the
presentation and analysis of information, planning and implementation
to community members. This is the distinction between extracting
information from the participants versus assisting participants to
assemble, analyze, and use information for their own purposes; the
distinction between the agenda of the outside agents versus the agenda
of the people they are ostensibly attempting to help. Another key
difference between the two is that in PRA rushing is replaced by
relaxation and there is a stronger rapport with community members.

The Principles of PRA

PRA methods are based on principles that are aimed at offsetting
deficiencies of the formal investigative approaches. The PRA principles
are thus the special features that identify it and give it a distinction from
other development approaches. They are as follows:

1. Participation. The input of the local people into PRA activities

is very essential to its value as a research and planning method and as a
means for diffusing the participatory approach to development.
126 Social Research Methodology: An Introduction

2. Multi disciplinary teamwork. To the extent that the validity of

PRA data relies on informal interaction and brainstorming among those
involved, it is best done by a team that includes local people with
perspective and knowledge of the area’s conditions, traditions, and
social structure and development experts with multi disciplinary
backgrounds and experience. Communities, rural leaders and specialists
work together closely to better understand community problems,
constraints, needs and opportunities. A well-balanced team should
represent the diversity of socioeconomic, cultural, gender, and
generational perspectives.

3. Flexibility and informality . The relationship between the PRA

team and the community is such that each is free to discuss. PRA does
not provide a blueprint to its practitioners. As such, the methods and
techniques used to generate information are varied in combinations that
are determined by the size and skills of the team, time and resources
available, and the topic and location of the project.

4. Relaxed learning. The premise of this principle is that PRA

should not be conducted in a hurried and rigid manner. Learning takes
place in the field and analysis of information gathered is an integral part
of the fieldwork.

5. Triangulation. This principle is premised on the fact that PRA

works with qualitative data. Triangulation thus involves crosschecking
information from different sources using different approaches. This may
be done by using different techniques, groups, locations, etc.
Triangulating ensures validity and reliability of the findings.

6. Optimal ignorance. The PRA team should only seek information

that is relevant to their objectives, and must avoid unnecessary detail,
accuracy and over-collection of data as the case in formal surveys. The
team should asks, “what kind of information is required, for what
purpose and how accurate does it have to be?” This is especially
important if one has to ensure efficiency both in terms of time and
money.
 Participatory Rural Appraisal 127

The Mt. Elgon Integrated Conservation and Development Project

(1999) has proposed the following additional principles:

7. Offsetting bias. The different viewpoints of team members

complement each other and provide a more comprehensive picture.
Because the exercise is relaxed and not rushed, the researcher listens and
does not lecture; is unimposing instead of acting important; probes
keenly; and seeks out poorer people and women, then learns their
concerns and priorities.

8. Seeking diversity. Variability is sought in PRA work, rather than

averages. It involves sampling in a non-statistical manner. It means
deliberately looking for, noticing and investigating contradictions,
anomalies and differences.

9. Facilitating. This principle is premised on the “they can do it”

conviction. As long as they are properly guided, rural people themselves
can conduct investigation, analysis, presentation and learning by
themselves so that they own the outcomes and also learn from the
process. This entails an outsider starting the process then sitting back or
leaving the group to work without interrupting.

10. Self-critical awareness and responsibility. Facilitators must

continuously examine their behaviour and try to improve their skills.
This includes embracing error as an opportunity to do better; using one’s
judgement at all times; and accepting personal responsibility rather than
vesting it in a manual.

The PRA Process

The application of different tools and techniques of PRA may be done
in any combination, and this may vary from case to case. However, the
process for conducting PRA remains essentially the same. A group of
people as an interdisciplinary team is required to execute the various
activities throughout the PRA cycle. Generally, the PRA process goes
through three phases—the preparatory phase just before fieldwork, the
fieldwork phase, and the phase after fieldwork.
128 Social Research Methodology: An Introduction

Phase I: Before fieldwork

This phase comprises of the preparatory tasks that must be performed
before field data collection can be undertaken. There are generally six
steps to be followed by the researcher or development worker in order
to be able to collect reliable and precise information from the field.
These are discussed below.

Site selection and clearance

Today, many people, organisations, institutions, and interested groups
are increasingly becoming aware of PRA as a tool for sustainable rural
development. Depending on who is initiating a PRA, the site for the
PRA may be selected in any of the following ways: First, leaders
(religious, community or grassroots local government) in a given area
who have become aware of the potential of PRA in mobilising rural
people for sustainable development, may approach government
institutions, donor agencies or NGOs to conduct PRA in their area. In
collaboration with such organisations, a PRA site can be selected.
Second, field workers of different backgrounds and training may request
for PRA to be carried out in areas where they work in order to
strengthen results. That way, they will propose the site. Third, many
projects today, especially in the developing world, are using the PRA
approach to produce sustainable results in all areas of intervention.
Many projects are site specific, and therefore the sites to carry out the
PRA may be pre-determined on a wider scale but the specific area will
be chosen on the basis of other criteria. Whichever way, the selected site
should be endorsed by the local authorities.

Setting PRA objectives and developing topics

Depending on the results desired, there are different types of PRA in
terms of size, team composition and tool combination, that can be
carried out. The objectives of the PRA will determine the type of PRA
and the information to be generated. Generally, a brainstorming session
of experienced people is organized for developing objectives, topics and
sub-topics. The session should also discuss how to carry out field work,
especially the tools to be applied for collecting the required information.
Of course, the choice of tools depends on the objectives, topics and
expected output.
 Participatory Rural Appraisal 129

Review of secondary data

Information from secondary sources is quite helpful in the development
of topics, sub-topics or checklists to be used during the PRA, and forms
a basis upon which primary data from the community will be
crosschecked and cross analysed. Having an overview of the area that
you intend to work in is important because it sheds some light on the
issues likely to be encountered during the PRA. It also influences
planning for the PRA as to what organisations, institutions and
departments should be invited to the PRA. There could also be various
projects already going on, or others that were started, and for various
reasons, stalled. An understanding of the underlying factors will help in
pursuing other alternatives during the PRA. Therefore, a careful review
and assessment of the secondary data is necessary before fieldwork.
Sources of secondary data include reports, books, management plans,
project plans, maps, etc.

Formation of an interdisciplinary team

The PRA methodology is extremely effective if carried out by a team
consisting of especially trained interdisciplinary persons. The number
and composition of the PRA team is decided on the basis of the
expertise required and an official invitation is made informing them of
the site, objectives, and dates of the PRA. Generally, a PRA team need
not be so big, as a small team facilitates close cooperation and sets team
members for role sharing. For a project on wetland conservation, for
example, it may consist of an environmentalist, an anthropologist, an
agricultural extension expert, and a lawyer. On wood fuel flows, it may
consist of a community forester, a forest hydrologist, a statistician, and
a sociologist. On poverty redaction, it may consist of a development
planner/economist, sociologist, agriculturalist, and an animal scientist.
If the constituted team has no prior PRA experience, then it becomes
necessary to train them. Otherwise, a review of PRA objectives and
secondary data is done.

PRA training for the team

Training is an expensive exercise but has long-term benefits. It is
dependent on the availability of resources, but is actually motivated by
the long-term benefits accruing from it.
130 Social Research Methodology: An Introduction

Preliminary visits
It is preferable to have short field visits to the PRA site to identify key
informants and try out the data collection tools. This task may be
assigned to a single team member, preferably someone who is familiar
with the locality. As earlier indicated, it is essential to obtain the prior
informed consent of local leaders and communities before embarking on
a PRA exercise in a village. It is important at this stage to make as many
contacts as possible. The number of contacts made will depend on a
number of factors and will also vary from one area to another. In some
areas, you may meet your contacts all in one day, whereas in others, it
may take several days to meet all of them. On the basis of these field
visits, the pre-designed tools may have to be changed or modified.

Phase II: Fieldwork

The field work is people-oriented. The team seeks information ranging
from indigenous knowledge to local customs and daily life practices. A
variety of different data collection methods are combined in a flexible
sequence at each stage of field work. The actual sequence used in each
village varies with the context of each case and may generally not be
fixed before the team goes out to the village. However, a checklist is
prepared with the objectives and protocols of the PRA. A project
protocol provides guidelines for all the team members to follow. The
field dynamics that affect the success or failure of the PRA are in fact
subject to two extremely important initiatives—rapport building and
maintaining protocols.

Rapport building
Rapport building is an important factor if the team is to collect reliable
information. It is usually done to develop communication and to
establish working relationships with the local people. Generally, rapport
building is initiated immediately as the team enters a village, and helps
bring the team closer to the village people. Some tips:
• Start greeting and talking to the rural people whenever you meet
them, whether at the village spring, agricultural fields, market
place, or in homes.
• Treat and respect rural people as per their local custom. For
example, in areas such as Buganda, you should concentrate on the
 Participatory Rural Appraisal 131

greetings first. Then, depending on the person, you may ask a few
informal questions such as “what are you doing? Do you live in this
village? How far is your house from here? etc.
• Ask the knowledgeable people about a subject or area in a village.
• Try to meet with local leaders and officials before starting work in
a village.
• Clearly explain the reasons for coming to the area.
• Show genuine interest in the local issues.
• Choose time and venue that are convenient for the local people.

Maintaining protocols
Maintaining protocols in the team is the basic rule in rapid appraisal
methods. The protocols should provide a way of introducing members
of the team to the public, stating the purpose of the PRA, how it will be
conducted and what the outcomes will be. RRA protocols are actually
rules of interpersonal behaviour by which a team agrees to operate.
Some of the protocols earlier used have included the following:
• A facilitator (from the team) controls the interview process by
singling start, filling gaps, etc.;
• Each team member pursues a sub-topic, following one’s own line
of questioning and reasoning;
• The interviewing sequence (who starts, who follows, who finishes)
is predetermined, and members take notes during each other’s turn;
• Unanticipated questions that arise are handled later;
• After each day’s sessions, the members debrief as a group, discuss
to verify the findings, and incomplete information is noted down
for recollection.

One the rapport is developed in the village, the researchers should

start collecting information by applying appropriate tools and
techniques. In order to collect more in-depth and reliable information,
researchers should also apply probing and iterating techniques. The
range of data collection methods and techniques normally employed in
PRAs are discussed under the section “Tools and Techniques of PRA.”
132 Social Research Methodology: An Introduction

Phase III: After fieldwork

After the fieldwork, the PRA team should meet and discuss the collected
information and if it is necessary arrange to go back for more
information. Each team member is assigned tasks and the results are
upheld by the members’ consensus. The findings are also presented in
workshops or seminars for comments, and the comments are taken care
of and incorporated in the final reports. After the field data collection,
the PRA team’s engagement with the community proceeds towards the
formulation and implementation of a Community Action Plan (CAP).

Report Writing
Compiling the information into a clear, usable document is a very
important part of the PRA process. It is quite important that by the time
the CAP is completed, a summary of the field data has been completed.
Upon adopting the CAP, the first draft of the report, which is a
compilation of different pieces of information generated by different
groups from the community, is produced. At this stage, a smaller group
of 3-4 people can be assigned the task of completing the final draft of
the report for publication.

Implementation of the CAP

Details of workable programmes to implement the action plan are done
by the individual organisations and development workers in
collaboration with the community, along guidelines found within the
different implementing bodies. In some cases, further studies may be
carried out before the identified opportunities are put into workable
programmes. It is necessary for the PRA team to prepare local groups
for the task of implementation of the CAP. The local leaders should be
helped and encouraged to organise work groups, design follow-up
schedules, coordinate activities with extension workers, ensure that
materials are gathered, and maintain contacts with administrative
officials and any other relevant institutions.

Follow-up
Theoretically, PRA concludes with the preparation of the CAP, but in a
practical sense this is just a transitional stage from research to action.
After laying strategies for implementation of the CAP and kick-starting
 Participatory Rural Appraisal 133

the physical work, the different stakeholders should continue providing

technical support in order to get through with the CAP. It is only at this
point, with the requisite technical advice, that the PRA process is
considered complete.

The Tools and Techniques of PRA

There are a variety of approaches, tools and techniques to PRA that are
used during the PRA exercise in order to elicit community participation
in the generation of information. These tools and techniques may always
be modified to fit the size of the client group, location, time, type of
information required, and other characteristics of the participants and
their situation. I will discuss these methods under two broad categories:
1) visualisation tools; and 2) sampling and data collection techniques.
More often for the PRA team to get full participation of the community,
all of the methods are used for best results.

Visualisation tools

Participatory mapping and modelling

Making a community map is normally the best approach to get both the
PRA team and the community started. It involves the rural people
marking and drawing sketches on the ground, using local materials such
as sticks, stones, grasses, coloured sand, etc. and often outside materials
such as chalk, pens, and paper. The sketches provide spatial data, i.e.,
reference points within which data collection and analysis will take
place. The data may include communal facilities, personal and family
buildings, assets and liabilities, etc. The maps and models show the
perspective of the drawer and reveal much about local knowledge of
resources, land use and settlement patterns, or household characteristics.
Maps and models can later lead to transact walks, in which greater detail
is recorded.

Participatory diagramming
As in the foregoing case, diagramming is also done on the ground, using
the same materials. The diagrams include the following:
• Trend lines. These show the people’s perceptions of changes over
time in key sectors. They can be used for many variables such as
134 Social Research Methodology: An Introduction

crop yields, livestock population, price changes, rainfall,

environmental degradation, etc. These tools will focus community
attention on positive and negative changes over time in terms of
what people consider important.
• Seasonal calendar. This is a snapshot of the village activities,
problems, opportunities, and possible interventions represented
month-by-month or season- by-season throughout the year.
• Venn diagrams. These are diagrams showing institutional
importance and interaction. A good example of these is the
institutional analysis chart, which helps the community and PRA
team to learn about the activities of various groups and
organisations within community, understand how the community
views these institutions and how they rank them in terms of their
contribution to development, and to assess the relationships among
institutions.
• Pie charts. These are used to show different resource contributions
to livelihood, constraints, needs and opportunities for the whole
community or individual households.

Matrix scoring and ranking

Matrices are used to study, collect, analyse and compare information on
diverse subjects such as resource use, conflicts, constraints,
opportunities, resources availability trends, etc. They are a flexible tool
and adaptable to a wide range of topics. They are also used to rank
problems, opportunities, preferences and wealth. Before ranking, there
must be a community inventory in place.
• Creating a community inventory. The inventory, and especially the
process of making it, is the most important and central element of
PRA. It is crucial to aim for an inventory that assesses the strengths
and weaknesses of the community. The community members
should be guided to construct the inventory as a group.
• Preference ranking. This is a ranking list of preferences of the
different interest groups within the community, drawn by the
groups themselves. Preference ranking is a good icebreaker at the
beginning of a group interview, and helps focus the discussion.
• Wealth Ranking. This is a particularly useful method of (1)
discovering how the community members define poverty, (2)
 Participatory Rural Appraisal 135

finding who the really poor people are, and (3) stratifying samples
of wealth. This is best done once you have built up some rapport
with the community members. How the groups categorize members
of the community, and the reasons they give for making those
categories and for putting different households into each category,
are very revealing about the socioeconomic composition of the
community.

Transect walks
A transect is a diagram of land use zones. It compares main features,
resources use, problems and opportunities of different zones.

Sampling and data collection tools

Time line
This is a historical profile tool used to provide an aggregation of
important events in the history of the community over the years. The
PRA team learns what local, national, and international events the
community considers to be important in its history. Discussions provide
a good opportunity to ask elders about previous trends and traditional
community responses, as well as about the possible opportunities to
resolve current problems.

Direct observation
Direct observation is used to generate on-the-spot questions to ask the
community without formal questionnaires. Methods of direct
observation include measurement of field sizes, distances, and weight of
produce in the field; noting directly observable objects, events, process
or relationships; recording in notebooks, record sheets, and photographs;
sight-seeing places such as markets, places of worship, homes, schools,
places of entertainment, roads, water points, etc.; and the use of other
senses.

Farm sketches
These enable the PRA team to compare resources use, opportunities,
problems and strategies in different zones. They compare, on a relative
basis, socioeconomic status within the community in terms of standards
136 Social Research Methodology: An Introduction

of resources management, productivity, and income. Sketches reveal

differences in farm size, cultivated crops, livestock management, and
other variables of household and resources use. Whatever families
decide on resources management affects collective decisions,
opportunities and environmental impact of the whole community.

Interviews and discussions

• Group interviews. These interviews are used to obtain community

level information. They have several advantages which include
providing 1) access to a large body of knowledge, 2) immediate
cross-checking on information, and 3) opportunity to easily single
out members with divergent views for further discussion.
• Focussed group discussions. Discussion is held on predetermined
topics, with individuals who share characteristics such as gender,
profession, age, challenges, etc.
• Semi-structured interviews. Semi-structured interviewing is a
guided discussion where only some of the questions are pre-
determined and new questions come up during the interview. The
interviewers prepare a list of topics and questions rather than
utilizing a fixed questionnaire. Semi-structured interviews are held
with individuals, Key informants, and groups.

Other PRA tools

• Probing. The main function of a probe is to encourage the
respondent to answer more fully and accurately. Furthermore, it
also helps to structure the respondent’s answer and make sure that
all topics of interest are covered. Always start questions with who,
what, why, when , who and how (the ‘six helpers’) for helping to
establish the basic situation.
• Iteration. Iteration is one of the techniques for collecting
information in RRA/PRA methods. In this technique, the same
question is repeatedly asked in different situation for conforming
the given information. A very high pay-off from flexibility of the
methodology through iteration occurs in the ability to reform
questions and formulate new questions, especially within the
interview itself.
 Participatory Rural Appraisal 137

Conclusion
PRA has become very popular in gathering reliable and precise
information in rural areas. It is considered good for learning about rural
life and conditions from the rural people themselves, and involving them
in planning and implementing the plans as a way of intervening in their
situation. Reliable and precise information can be collected with the
available PRA tools and techniques, depending on how these tools are
combined and how the process is followed by the interdisciplinary team.
Some cautionary procedures such as rapport building and keeping
protocols are also important in the PRA process. These help researchers
and practitioners to be more familiar with the area and thus to collect
reliable and precise information. Equally crucial is the understanding of
the three distinct phases in the PRA process, i.e., before fieldwork,
during fieldwork and after fieldwork. This enables the researchers or
practitioners to conduct the PRA without the fear of a stalling project.
By looking in a new way at village data, rural priorities, aspirations
and potential interventions, PRA enables restructuring the manner and
process of rural planning and resources management without interfering
with national or local government activities and programmes. PRA is
thus a complementary methodology that gathers site-specific data,
integrates sectors, involves beneficiaries in the planning process, links
extension services directly with rural communities, and sets up an
implementable plan that local institutions can handle.
Chapter Ten

Document Study and Analysis Methods

Introduction
Documents are an essential element of day-to-day work in organizations,
be they public or private organizations, and whether they are service or
purely business-oriented. The documents may be generated within an
organization, be about the organization, or just be acquired by the
organization. They may be formal (official), informal (personal), or
public. Official documents include policies and regulations, almanacs,
mission statements, reports, strategic plans, organizational charts,
circulars, memos, correspondence, standing orders, contracts, minutes,
etc. Personal documents include personal correspondence, diaries,
confidential health reports, autobiographies, wills, etc. Public documents
are all manner of literature that is released for general public use, and
include books, magazines, newspapers, press releases, etc.
Documents are a very important source of information for
researchers, social or otherwise. Researchers require documented
information at the planning stage of a research project, which they use
as evidence of existence and severity or magnitude of a problem as well
as the gaps in knowledge and interventions. They also require such
information during the study to establish the cause, nature, or any other
aspects of the problem. At the end of the study, documents are still
needed to enable researchers corroborate their findings and come to
appropriate conclusions.
Document analysis refers to the systematic examination of written
documents, including artefacts such as films and videos. How deeply a
researcher analyses the documents depends on his or her research
 Document Study and Analysis Methods 139

questions and objectives. However, while examining documents,

researchers aim to identify repeating ideas and larger themes. The
following guidelines adopted from McNamara (1998) should be found
helpful to researchers:
i) Read all the relevant documents.
ii) Label ideas as you go along.
iii) Organize ideas with similar content and identify repeating ideas.
iv) Identify themes (patterns or associations and causal relationships)
among repeating ideas.
These guidelines are, of course, general. It is important to consider that
there are different analysis methods, and different methods have
different objectives and different procedures. The different methods of
document analysis are content analysis, secondary analysis, meta
analysis, and literature review.

Content Analysis
Content analysis is a method researchers use to quantify and analyze the
presence, meanings and relationships of words and concepts, and then
make inferences about the messages within the texts. By texts is meant
generally books, book chapters, newspaper articles, essays, speeches,
diaries, advertisements, etc. Krippendorf (1980) defines content analysis
as a research technique for making replicable and valid references from
data to their contexts. To conduct a content analysis, the text is broken
down into manageable categories on a variety of levels such as word,
phrase, sentence, or theme, and then examined. This process is called
coding.
Content analysis can be categorized into two: conceptual analysis
and relational analysis. Conceptual analysis is the strategy where the
researcher establishes the existence and frequency of concepts – most
often represented by words or phrases – in a text. Relational analysis, on
the other hand, is where the researcher goes a step further by examining
the relationships among concepts in a text. Let us examine these
categories in some detail.

Conceptual analysis
Under this type of analysis, a concept is chosen for examination, and the
analysis involves quantifying and tallying its presence. The researcher
140 Social Research Methodology: An Introduction

looks at the occurrence of selected terms within a text or texts, whether

the terms are implicit or explicit. Explicit terms are obviously quite easy
to identify as compared to implicit terms whose level of implication is
somewhat complicated by the need to make subjective judgments. In
order to limit subjectivity, the coding of implicit terms usually involves
the use of either a specialized dictionary or contextual translation rules,
or both.
Conceptual analysis follows after the determination of research
questions and selection of a sample or samples of texts. The texts are
then coded into manageable content categories that are indicative of the
research question. The process of coding is basically one of “selective
reduction,” whereby the researcher reduces the text to categories
consisting of sets of words or phrases (concepts). It should be
emphasized that in conceptual analysis, the researcher simply wants to
examine presence of the concepts with respect to his or her research
question.
Carley (1992) has proposed an eight-step process of coding as I
summarize below.

1. Decide the level of analysis. The researcher must decide

whether. to code for a single word or for sets of words or phrases.

2. Decide how many concepts to code for. This involves

developing a pre-defined or interactive set of concepts and categories.
The researcher must decide on a number of issues: For example, code
for every single positive or negative word that appears, or only certain
ones that are most relevant to the research question? Code only from the
pre-defined set, or allow addition of relevant categories as found in the
text? Note that while pre-determining a certain number and set of
concepts allows a researcher to examine a text for very specific things,
coding flexibility allows the incorporation of new, important material
that could have significant bearings on the results.

Step 3: Decide whether to code for existence or frequency of a

concept. When coding for existence, a concept (word or phrase) is
counted only once, no matter how many times it appears. This is a very
basic coding process and may give the researcher a limited perspective
 Document Study and Analysis Methods 141

of the text, whereas the number of times the concept appears might be
more indicative of its importance. These are important considerations for
the interpretation the researcher makes.

Step 4: Decide on how to distinguish among concepts. Should the

concepts be coded exactly as they appear, or should they be generalized
(be recorded as the same even when they appear in different forms)? For
example, for concepts such as “enquiry” and “investigation” the
researcher needs to determine if they mean radically different things or
if they are similar enough that they can be coded as being the same
thing.

Step 5: Develop rules for coding your texts. After taking the
generalization of concepts into consideration, a researcher has to create
translation rules that will help him or her to streamline the coding
process and ensure maximum consistency and coherence.

Step 6: Decide what to do with “irrelevant” information. Weber

(1990) suggests that “irrelevant” information (words that add nothing to
the quantification of the concepts of interest and can therefore be
disregarded without impacting the outcome of the coding) should be
ignored. Carley (1992), however, thinks that they can be used to
reexamine and/or alter the coding scheme.

Step 7: Code the texts. This may be done manually (reading

through the text and recording concept occurrences), or by the use of a
computer. Coding with a computer is a very popular and efficient
approach to conceptual analysis, provided the researcher has done
thorough preparation and reliable category construction. Then computer
programs can easily automate the coding process and examine huge
amounts of data, and a wider range of texts, quickly and efficiently.

Step 8: Analyze the results. On completion of coding, the researcher

examines the data for possible conclusions and generalizations. It should
be noted, however, that while conceptual analysis is extremely useful
and effective for providing information on existence and frequency of
142 Social Research Methodology: An Introduction

concepts when done right, it cannot explore the relationships that exist
between these concepts. One must thus turn to relational analysis.

Relational analysis
Like conceptual analysis, relational analysis begins with the act of
identifying the presence of concepts in a given text or set of texts.
However, relational analysis goes beyond presence by exploring the
relationships between the concepts identified. Palmquist et al. (1997)
have used the term “semantic analysis,” to reflect the thrust or focus of
the method to look for semantic, or meaningful, relationships. Individual
concepts, in and of themselves, are viewed as having no inherent
meaning. Rather, meaning is a product of the relationships among
concepts in a text.
The process of relational analysis has achieved a high degree of
computer automation but still is, like most forms of research, time
consuming. Perhaps the strongest claim that can be made is that it
maintains a high degree of statistical rigor without losing the richness of
detail apparent in even more qualitative methods. Following are the
steps or strategies that can be followed to code a text or set of texts
during relational analysis.

Step 1: Identify the research question. The question is important

because it indicates where you are headed and why. Without a focused
question, the concept types and options open to interpretation are
limitless, and this makes the analysis difficult to complete.

Step 2: Choose a sample or samples for analysis. Once the question

has been identified, the researcher must select works written on the
subject. For relational content analysis, the primary consideration is how
much information to preserve for analysis. The researcher must take
special care not to take on so much that the coding process becomes too
heavy and extensive to supply worthwhile results.

Step 3: Determine the type of analysis. After determination and

selection of the sample for analysis, it is necessary for the researcher to
determine what type(s) of relationships to examine. There are different
subcategories of relational analysis that can be used to examine the
 Document Study and Analysis Methods 143

relationships in texts. Once the subcategory of analysis is chosen, the

selected text must be reviewed to determine the level of analysis. The
researcher must decide whether to code for a single word or for sets of
words or phrases.

Step 4: Reduce the text to categories and code for words or

patterns. At the simplest level, a researcher can code merely for
existence. At a higher level, however, the requirement of one’s specific
research question may necessitate deeper levels of coding to preserve
greater detail for analysis. The researcher might also choose to code for
what words he used that have such an ambiguous nature in relation to
the importance of the information directly related to those words.

Step 5: Explore the relationships between concepts. There are three

aspects of relationships – strength, sign, and direction of relationship.
Strength of relationship refers to the degree or extent to which two or
more concepts are related. Strength is the easiest relationship to analyze,
compare, and graph when all relationships between concepts are
considered to be equal. Sign of a relationship refers to whether or not the
concepts are positively or negatively related. Direction of the
relationship refers to the type of relationship categories exhibit. Various
types of directional relationships include “X implies Y,” “X occurs
before Y,” “if X then Y,” etc.

Step 6: Code the relationships. One of the main differences

between conceptual analysis and relational analysis is that in the latter,
the statements or relationships between concepts are coded.

Step 7: Perform statistical analyses. This is enabled by various

computer programs for relational analysis.

Step 8: Map out the representations. In addition to statistical

analysis, relational analysis often leads to viewing the representations of
the concepts and their associations in a text (or across texts) in a
graphical form. Relational analysis is also informed by a variety of
different theoretical approaches: linguistic content analysis, decision
mapping, and mental models.
144 Social Research Methodology: An Introduction

Evaluation of Content Analysis

Advantages
Content analysis offers several advantages to researchers who consider
using it. In particular, content analysis:
• looks directly at communication via texts or transcripts, and hence
gets at the central aspect of social interaction;
• can allow for both quantitative and qualitative operations;
• can provides valuable historical/cultural insights over time through
analysis of texts;
• allows a closeness to text which can alternate between specific
categories and relationships and also statistically analyzes the
coded form of the text;
• can be used to interpret texts for purposes such as the development
of expert systems (since knowledge and rules can both be coded in
terms of explicit statements about the relationships among
concepts);
• is an unobtrusive means of analyzing interactions; and
• provides insight into complex models of human thought and
language use.

Disadvantages
Content analysis suffers from several disadvantages, both theoretical and
procedural. In particular, content analysis:
• can be extremely time consuming;
• is subject to increased error, particularly when relational analysis
is used to attain a higher level of interpretation;
• is often devoid of theoretical base, or attempts too liberally to draw
meaningful inferences about the relationships and impacts implied
in a study;
• is inherently reductive, particularly when dealing with complex
texts;
• tends too often to simply consist of word counts;
• often disregards the context that produced the text, as well as the
state of things after the text is produced; and
• can be difficult to automate or computerize.
 Document Study and Analysis Methods 145

Secondary Analysis
Secondary analysis involves the utilization of existing data, collected for
the purposes of a prior study, in order to pursue a research interest which
is distinct from that of the original work (Heaton, 1998). This research
interest may be a new research question or an alternative perspective on
the original question. In this respect, secondary analysis differs from
systematic reviews and meta-analyses of qualitative studies which aim
instead to compile and assess the evidence relating to a common concern
or area of practice. Secondary analysis can involve the use of single or
multiple quantitative or qualitative data sets, as well as a mixture of
qualitative and quantitative data sets. Secondary analysis of quantitative
data is a common and generally accepted mode of inquiry, but secondary
analysis of qualitative data is yet to gain such acceptance.
This is not to say, however, that there is no growing interest in re-
using qualitative data. On the contrary, there are at least two factors that
are fronting for secondary analysis of qualitative data. First of all,
limited opportunities and funding for conducting primary research and
the enormous task of qualitative data analysis have prompted researchers
to consider maximizing use of the data already available to them.
Secondly, the advent of software to aid the coding, retrieval and analysis
of qualitative data means that both the archiving and availability of
qualitative data for secondary analysis will be more and more easily
done. These factors are similar to those that informed the method of
secondary analysis of quantitative data.

Advantages of secondary analysis

• the approach can be used to generate new knowledge, new
hypotheses, or support for existing theories;
• it reduces the burden placed on respondents by negating the need
to recruit further subjects;
• it allows wider use of data from rare or inaccessible respondents.
• it is a more convenient approach for particular researchers, notably
students.

Disadvantages
• where the researcher was not part of the original research team the
approach is best only employed by experienced researchers
146 Social Research Methodology: An Introduction

• the use of the approach does not necessarily preclude the possibility
of collecting primary data. This may, for example, be required to
obtain additional data or to pursue in a more controlled way the
findings emerging from the initial analysis.
• there may also be a need to consult the primary researcher(s)
(assuming that they are available) in order to investigate the
circumstances of the original data generation and processing.

I need to stress, all in all, that despite the interest in and arguments
for developing secondary analysis of qualitative data, the approach has
not been widely adopted to date—it is taking a very slow pace.
Moreover, most of the studies employing this method have mainly been
conducted by researchers re-using their own data rather than by
independent analysts using data collected by others.

Meta Analysis
This is a technique, where a researcher combines the results of several
different studies dealing with the same research question. According to
Glass (1976:3), meta-analysis is the analysis of analyses...the statistical
analysis of a large collection of analysis results from individual studies
for the purpose of integrating the findings.
Meta-analysis is basically a statistical technique for amalgamating,
summarizing, and reviewing previous quantitative research. By using
meta-analysis, a wide variety of questions can be investigated, as long
as the researcher has a reasonable body of primary research
studies. Selected parts of the reported results of primary studies are
entered into a database, and this “meta-data” is “meta-analyzed,” in
similar ways to working with other quantitative data. In essence
therefore, meta analysis is decidedly quantitative, but it involves some
of the same sorting and coding techniques found in qualitative research.
The appeal of meta analysis is that it in effect combines all the
research on one topic into one large study with many participants. Its
danger, however, is that in amalgamating a large set of different studies
the construct definitions can become imprecise and the results difficult
to meaningfully interpret. And although this method is prized for its
objectivity, its value nevertheless depends on making some qualitative-
type contextualization and understanding of the objective data. Meta-
 Document Study and Analysis Methods 147

analysis has in fact been successfully used to give helpful insight into 1)
the overall effectiveness of social interventions, 2) the relative impact
of independent variables, and 3) the strength of relationships between
variables.

Literature Review
Quite often, research will have been done on a topic similar or related
to the one under investigation. In some cases, there will be abundant
such concluded research and in some others, there will hardly be any. In
the former case, the researcher runs the risk of duplication if he or she
does not carefully select and review what is of utmost relevance to the
problem being investigated. In the latter case, the researcher can at best
make use of the literature on related research (research on other aspects
of the topic than the one under study, or similar aspects but in different
settings or contexts).
A literature review is a critical or analytical survey and synthesis of
carefully selected scholarly works, and aims to offer an overview of
significant literature published on the topic under investigation
(Bourner, 1996; Cooper, 1998; Galvan, 1999; Muskal, 2000; Macauley;
2001; Taylor & Procter, 2005)). It examines an area of research or a
theory, provides a description or summary, and makes a critical
evaluation of each work. It usually has an organizational pattern and
combines both summary and synthesis. A summary is a recap of the
important information of the source, but a synthesis is a re-organization,
or a reshuffling, of that information. It might give a new interpretation
of old material or combine new with old interpretations. Or it might
trace the intellectual progression of the field, including major debates.
And depending on the situation, the literature review may evaluate the
sources and advise the reader on the most pertinent or relevant.
In short, therefore, although you need to summarize relevant
research, it is also vital that you evaluate this work, show the
relationships between different works, and show how it relates to your
present study. You should thus strive to compare and relate different
theories rather than just summarize them individually. Your review
should also be done along a particular focus or theme. It does not have
to be an exhaustive account of everything published on the topic, but it
should discuss all the significant academic literature important for the
148 Social Research Methodology: An Introduction

study. It should provide an overview and analysis of the current state of

research on a topic or question. The aim should be to evaluate and
compare previous research on a topic in order to provide an analysis of
what is currently known and also to reveal controversies, weaknesses
and gaps in that work.

Why Conduct a Literature Review?

A literature review is based on the assumption that knowledge increases
day-by-day and we learn from and build on what others have done.
Researchers read studies to compare, replicate or critique them. A
literature review may constitute an essential chapter of a thesis or
dissertation, or may be a self-contained review of writings on a subject.
Reviewing the literature is so crucial that its omission represents a void
or absence of a major element in research.
According to the foregoing sources, the reasons for reviewing the
literature before embarking on a research project include the following:
i) to place each work in the context of its contribution to the
understanding of the subject (problem or research question) under
investigation;
ii) to describe the relationship of each work to the others under
consideration;
iii) to identify new ways to interpret, and shed light on any gaps in
previous research;
iv) to resolve conflicts amongst seemingly contradictory previous
studies;
v) to identify areas of prior scholarship in order to avoid “reinventing
the wheel,” which is an unnecessary duplication of effort. This
saves time and it can stop one from making the same mistake as
others;
vi) to carry on from where others have already reached. Reviewing the
research already done in the field allows one to build on the
platform of existing knowledge and ideas
vii) to increase breadth of knowledge on one’s subject area; and
viii)to provide an intellectual context for one’s work enabling one to
position one’s project relative to other work.
 Document Study and Analysis Methods 149

In sum, literature review provides the researcher with a handy guide

to a particular topic. If one has limited time to conduct research,
literature review can provide an overview or act as a stepping stone. For
professionals, literature reviews are useful reports that keep them up to
date with what is current in the field. And for scholars, the depth and
breadth of the literature review emphasizes the credibility of the writer
in his or her field. The literature review itself, however, does not present
new primary scholarship.
After the review of literature, the researcher decides whether to go
ahead and carry out the proposed study, to modify it, or abandon it if
similar studies have already been conducted.

Organization and Structure of a Review

The organization of a literature review depends on the type and purpose
of the review, as well as the topic being reviewed upon. Literature
reviews are organized thematically, which makes it easier to examine
contrasting perspectives, theoretical approaches, methodologies, and
findings; and to analyse the strengths and weaknesses of and point out
any gaps in previous research. In other words, literature review can be
used where there are conflicting ideas to achieve two main goals: 1)
establishing any facts that can be agreed upon, and 2) suggesting new
explanations to reconcile the conflict.
The literature review should be sectioned according to the sub
problems of the study. It should have an introductory section in which
the researcher gives his or her impression about the literature available
and introduces the rest of the sub topic of the review. Each section and
sub section should be briefly introduced with a sentence or two before
citing the literature.
A good literature review should comprise the following elements:
First, an overview of the subject, issue or theory under consideration,
along with the objectives and methodology of the literature. Second, a
division of works under review into categories (e.g. those in support of
a particular position, those against, and those offering entirely
alternative theses). Third, an explanation of how each work (or group of
works) is similar to and how it varies from the others in order to show
its relevance and point out gaps. Finally, conclusions as to which works
150 Social Research Methodology: An Introduction

make the greatest contribution to the understanding and development of

their area of research.
The literature being reviewed must be linked to the proposed
research so that it can be established that not only is the proposed study
distinctive and different from previous research, but that it will fill the
gaps left by previous research. The following aspects are indicative of
the research gaps in question:
i) Aspects of the research topic which are not yet investigated;
ii) Priority areas which are not yet studied;
iii) Long lapse of time since the reviewed studies were done;
iv) Reviewed studies were conducted in different geographical areas;
v) Different methods and instruments were used from those proposed;
vi) The reviewed studies have methodology weaknesses.

After identifying the gaps, the researcher explains how the

proposed study will fill them. This identification and explanation can be
done in three ways. First, it can be done in a study-by-study analysis, i.e.
linking each single work reviewed with the proposed study. This should
be adopted if the literature reviewed consists of only a few studies,
otherwise the links will become monotonous and have less impact.
Second, it can be done when the review is clearly done under themes so
that linking is done at the end of each section. Third, it can be done in
a separate section often entitled “The Research Gaps” or “The Present
Study.” Linking under the second and third methods is adopted when
many studies are reviewed.

Which Literature is Reviewed?

The type of scholarship and research for which the literature is being
reviewed may be empirical, theoretical, critical/analytic, or
methodological in nature, but it oftentimes combines several of these
aspects. The literature to be reviewed therefore, should be that with
scholarly content (see chapter eighteen). According to Cooper (1998),
Muskal (2000), and Taylor & Procter (2005), a literature review uses as
its database reports of primary or original scholarship, and does not
report new primary scholarship itself. These reports are published in
different media. Here below are the recommended media as sources of
reports/literature to be reviewed for intellectual scholarship.
 Document Study and Analysis Methods 151

1. Journal articles. These are a highly recommended type of

literature for review, and are frequently used in literature reviews
because they offer a relatively concise, up-to-date format for research.
They are also a popular and reliable source because all reputable
journals that publish them are refereed.

2. Conference proceedings. This source of literature is also highly

recommended because it provides the latest research, as well as research
that has not been published. It also provides information on which
people are currently involved in which research areas, and so can be
helpful in tracking down other work by the same researchers.

3. Theses and dissertations. These are very useful sources of

information, especially those written for advanced degrees and
supervised by expert research scholars. The findings in dissertations for
Bachelor’s degrees should be treated with more caution, though, because
undergraduate researchers are generally novices.

4. The Internet. This is the fastest-growing source of information

for research. However, great care must be taken in determining the
quality and reliability of that information—who is the author of that
work? Which audience is targeted? In case of e-journals, is the journal
refereed? It is advisable to seek information in electronic journal
databases, and pay scant attention to the general information accessed
through browsers such as Google.

5. CD-ROMS. Today, there are a number of specialized CD-ROMS

that publish bibliographies for use in academic libraries. These can be
a very valuable tool in searching for the literature required for review.

6. Textbooks. Text books are not recommended for inclusion in a

literature review since they are intended for teaching, and not for
research. They tend to be less up-to-date as it takes longer for a book to
be published than for a journal article. However, they are an
indispensable source of theories, concepts, and models.

7. Government and corporate reports. Many government

departments and corporate bodies commission or carry out research
152 Social Research Methodology: An Introduction

whose findings can provide a very useful source of information for

review.

8. Newspapers. Newspapers are not a source of literature for review

because they are intended for a general audience, and are lacking in
scholarly content. However, they are helpful as providers of information
about recent trends, changes, or developments, and may therefore be
used as a source of evidence while writing the background to the
research problem.

9. Magazines. Popular magazines, which target a general audience,

are not useful for literature review. Specialized magazines useful only
as a starting point by providing news or general information about new
discoveries, policies, etc. that you can utilise while writing the
background section.

Evaluation of Literature Review

As with other methods of data collection, literature review has a number
of advantages. First of all, literature review is versatile—it can be
conducted for almost any topic and can provide information, either at the
overview level or in-depth. Secondly, literature review is relatively
inexpensive and efficient. A large amount of data can be collected
quickly at minimal cost. Thirdly, no scheduling or coordination is
involved. The cooperation of others is not required since you are just
dealing with documents. Lastly, literature review can be an excellent
first step in a project or study because it provides a conceptual
framework for further planning and study.
However, this method is not without its shortfalls. An effective
literature review requires a high level of skill in identifying resources,
analyzing the sources to identify relevant information, and writing a
meaningful summary and synthesis. Moreover, literature review is
limited to collecting information about what has happened in the past,
and cannot provide data about current actual behaviour.
 Part Three

Quantitative Research Methods

Chapter Eleven

Quantitative Research Designs: An Overview

Introduction
Research design is the investigator’s overall strategy for answering the
research questions. It involves the setting of the study, the assigning of
the participants to the conditions of the study, and the order of
presentation of the independent variables. The choice of the design is the
first decision the investigator has to make about how the study will be
performed. The various types of designs have different strengths and
weakness and some are better for answering some types of questions
than others. Feasibility and cost of the research are also important
determinants of the type of design to adopt. Quantitative designs can be
classified in two broad categories: experimental designs and non-
experimental designs.

Experimental Research
An experiment is a research design where the investigator manipulates
one or more variables and then assesses the impact of the manipulation.
An experiment isolates the effect of a variable by holding all other
variables constant while manipulating this variable. The purpose of
experimental research is thus to examine the cause-effect relationships
(causality) between the variables in a highly controlled environment. It
involves selecting the dependent variable to be observed or measured as
well as the independent variables that are assumed to be the cause.
The independent variable in experimental research is referred to as
the experimental variable (or treatment variable). It is the variable
manipulated by the researcher under controlled conditions to determine
 Quantitative Research Designs: An Overview 155

any changes in the dependent variable. The dependent variable is

sometimes referred to as the outcome variable or criterion variable.
Experimental research can itself be subdivided into two: True
experimental designs and quasi-experimental designs. The distinction
between the two is determined by the amount of control, which the
researcher is able to exercise over the research conditions.

True Experimental Designs

Under this category, there is a specific research (experimental) design.
Examples of true experimental designs include:
i) Pretest-posttest control group design;
ii) Posttest only control group design;
iii) Solomon four group design; and
iv) Factorial design.

In the discussion of these designs, the following symbols are commonly

used:
R = Random assignment of participants to the study
E = Experimental group
C = Control group
X = Exposure of a group to an independent variable
— = No exposure of a group to an independent variable.
We now turn to the true experimental designs.

Pretest-posttest control group design

This design can be diagrammed as follows:

Group Pretest IV (treatment) Posttest

E (experimental group) O1 X O2

C (control group) O1 — O2

Group members are randomly selected and randomly assigned to the E

and C groups. The groups are given a pretest (O1). For example, if you
are going to teach a new concept, first give a pretest to gauge what they
know about the concept before you intervene. Then intervene in the E
group by using a new method while the C group goes on being taught
156 Social Research Methodology: An Introduction

normally. Finally, give both groups another test, the posttest. The C
group is the baseline group on which the intervention is assessed.
This research design is used (1) when the researcher has some
concern that initial differences between the groups still exist even after
randomly selecting the participants from the target population and
assigning them to the groups, and (2) when the researcher has a limited
number of participants for the study (less than forty in each group).
When a pretest-posttest control group design is used, the results are
statistically analyzed as follows:
i) If the two groups do not differ in the O1 results, then the researcher
can assess the effect of the intervention by comparing the averages
of the two groups using only the O2 results. The statistical test
applicable here is the t-test for the two entirely independent groups.
Statistical tests are discussed in Chapter Fifteen.
ii) If initially the two groups were clearly different, the researcher can
handle as follows: subtract the pretest results for each participant
from the posttest results. Then find the average of these differences
for each group. It is the averages of the differences that are
compared using the t-test for the independent groups. These
differences in scores (results) are called “gain scores.”
iii) If the researcher does not does not want to analyze the gain scores,
then he can remove the initial differences reflected in the two
groups by using a statistical method called Analysis of Covariance
(ANCOVA).

Posttest only control group design

This design is often used when a researcher is unable to locate or
construct an appropriate pretest. Also this design is appropriate when
a researcher suspects that exposing the participants to a pretest might
sensitize them. This design is diagrammatically represented as follows:

Group (randomly IV (treatment) Posttest

assigned)

E X O1

C — O2
 Quantitative Research Designs: An Overview 157

When data is collected using this design, it can be statistically analyzed

by comparing the posttest averages for the E and the C groups using the
t-test for the separate, independent groups. For example:

E C
A 18 20
B 16 15
C 12 12
D 14 13
Ó 60 60
[The Greek letter sigma (Ó) is a symbol for summation]

The average is Ó/4 = 60/4 = 15 both for the E and C groups. These
averages are then compared using the t-test (see chapter fifteen).

Solomon four group design

This design combines both the pretest-posttest control group design and
the posttest-only control group design. This is so when the researcher
seeks to simultaneously test the effects due to the treatment/intervention
variables, as well as the effect of a pretest on a posttest performance.
The design diagrammatically depicted as follows:

Group (R) Pretest IV Posttest

E1 O1 X O2

C1 O1 — O2

E2 — X O1

C2 — — O1

The most interesting and useful aspect of this design is the many
comparisons that can be made to determine the effect of the pretest and
the intervention. The t-test for the independent group is used in these
various comparisons. It should be noted that only three of these
comparisons are very important. Therefore, the researcher must report
these three analyses:
158 Social Research Methodology: An Introduction

i) For the effect of the treatment/intervention, the researcher

compares the mean posttest scores of the experimental group 1 (E1)
with the mean posttest scores of control group 1 (C1). Simply
ignore the pretest.
ii) The effect of the treatment can again be compared for those groups
that did not have a pretest (i.e. E2 versus C2).
iii) For a test of the effect of the pretest sensitization on the later
posttest outcomes, a comparison of mean posttest scores of control
groups (C1 versus C2) is made.

Factorial design
The word “factor” is synonymous with the word “variable”.
Accordingly, factorial means more than one independent variables. The
factorial design is used when the researcher wishes to study two or more
independent variables simultaneously. This is in order if one wants to
study the independent effect of each variable on the dependent variable
as well as their interaction effect (their effect which shows
simultaneously). For example, a researcher may be interested in
investigating the relative effectiveness of two teaching methods, X1 and
X2. The researcher is also interested in gender as an independent
variable. That is, he wants to find out whether male and female students
achieve differently under the teaching methods X1 and X2. The
researcher will randomly select the participants and then randomly
assign a group of males and females to each of the teaching methods.
After exposing the groups to the various teaching methods, they are
given a test to check on their achievement. The scores obtained will give
an indication of their achievement (dependent variable) under the two
methods. This situation can be diagrammatically depicted as on the
following page.
If in the analysis you are interested in the method and not gender,
you combine both genders. If you are interested in both method and
gender, observe separately males and females for each method.
 Quantitative Research Designs: An Overview 159

First I.V.
X1 X2
(teaching method)

Second I.V.
(gender) Male Femal Male Female

All 4 groups involved in the study

Quasi-Experimental Designs
The word “quasi” means as if or almost, so a quasi-experiment means
almost a true experiment. Quasi-experimental designs have some, but
not all, of the properties of true experiments. They are however
relatively strong in terms of internal validity. Some common
characteristics of quasi-experiments include the following:

1. Matching instead of randomization is used. For example,

someone studying the effects of a new police strategy in town would try
to find a similar town somewhere in the same geographic region,
perhaps in a 5-state area. That other town would have citizen
demographics that are very similar to the experimental town. That other
town is not technically a control group, but a comparison group, and this
matching strategy is sometimes called nonequivalent group design.

2. Time series analysis is involved. Time series is a type of

longitudinal research (discussed under survey designs), which is a study
of changes occurring in a given society (or among different sections of
society) over an extended period of time. A time series is employed to
examine changes in the dependent variable over time, in which time
functions as the independent variable. Examples of variables that change
over time are unemployment rates, household incomes, voting
preferences, public opinion on an issue, etc. We can talk of interrupted
and non-interrupted time series. Interrupted time series involve before
160 Social Research Methodology: An Introduction

and after measurement. For example, someone might use a time series
to look at crime rates as a new law is taking effect. This kind of research
is sometimes called impact analysis or policy analysis. Non-interrupted
time series involve continuous collection and analysis of data.

3. The unit of analysis is often something other than people. Quasi-

experiments are best suited for the study of “fuzzy” or contextual
concepts such as sociological quality of life, pervasion, disorganization,
morale, etc. This kind of research is sometimes called contextual
analysis. Quasi-experimental research is often conducted to capture
longer time periods and a sufficient number of different events to control
for various threats to validity and reliability. The hope is that the design
will generate stable, reliable findings and tell something about the
effects of time on a given dependent variable. In fact, for a non-
interrupted time series, time itself is the independent variable. Almost
all quasi-experiments are somewhat creative or unusual in what they
attribute the cause of something to, and this is the case because we are
not using a true experiment where we manipulate some independent
variable in order to assess causality. Instead, at best, we have a statistical
baseline and some interventions that have occurred naturally (like the
passage of a law) or were created by the researcher (such as some public
relations campaign).

Non-experimental Research
Non-experimental research designs are generally referred to as
descriptive designs. They include correlational designs, comparative
designs, and survey designs; and they are actually the major quantitative
research designs employed in the social sciences. Unlike experimental
designs that focus only on causality, these designs in different fashions
look beyond the cause-effect relationship between variables to consider
how different variables co-vary under different circumstances and over
time. Let us consider each of them in turn.

Correlational Designs
One of the easiest and most widely used relations in quantitative social
research is correlation. Correlation is used when investigating the
relationship between two or more quantitative, continuous variables at
a single point in time. If for example a researcher wanted to find out
 Quantitative Research Designs: An Overview 161

whether residential settings determine people’s opinions on whether

prostitution should be legalized or not, one could use a correlational
study of two or more different residential settings. The relationship
between these variables (settings and opinions) is represented by a
correlational coefficient, which is denoted as ‘r.’ A correlational
coefficient typically ranges between –1.0 and +1.0 and provides two
important pieces of information regarding the relationship: the intensity
of the relationship and the direction of the relationship.
Intensity refers to the strength of the relationship and is expressed
as a number between zero (meaning no correlation) and one (meaning
a perfect correlation). These two extremes are rare as most correlations
fall somewhere in between. A correlation of 0.30 may be considered
significant and any correlation above 0.70 is almost always significant.
The absolute value of ‘r’ represents the intensity of any correlation.
Direction refers to how one variable moves in relation to the other
and is described as positive or negative. A positive correlation (or direct
relationship) means that two variables move in the same direction, either
both moving up or both moving down. For instance, it is often observed
that people who are very active in sports also have a strong appetite. As
such, it may be accurate to state that being active in sports and having
a sharp appetite are positively correlated. A negative correlation (or
inverse relationship) means that the two variables move in opposite
directions: as one goes up, the other tends to go down. For instance,
depression and self-esteem tend to be inversely related because the more
depressed an individual is the lower his or her self-esteem. As
depression increases, self-esteem tends to decrease. There are a number
of measures of correlations in social research as I highlight below.

The product-moment correlation

This correlation coefficient, commonly known as the Pearson's
correlation coefficient or Pearson’s r, is a measure of the intensity or
strength and direction of the relationship between two variables. It also
measures the extent to which both the independent and dependent
variables co-vary or change together. However, the correlation
coefficient does not give a satisfactory explanation for the change in the
dependent variable. Thus, the coefficient of determination is usually
preferable since it enables a researcher to measure the precise amount of
change in the dependent variable that arises from a change in the
162 Social Research Methodology: An Introduction

independent variable. The coefficient of determination, which is the

square of the correlation coefficient (r2 ), is equal to the percentage of the
variation in one variable that is related to the variation in the other
variable. For example, if in a study of the relationship between
liberalisation of education and literacy levels in Uganda we obtain an r
value of .4, the coefficient of determination would be .42, which would
be .16 (16%). This means that 16% of the change in literacy levels is
explained by the liberalisation of education. The other 84% change is
explained by other unknown factors.
To study the relationship between two continuous variables, one
first draws a scatter plot of the variables to check for linearity. The
correlation coefficient cannot be calculated if the relationship is not
linear. The independent variable is normally plotted on the x-axis
(horizontally) and the dependent variable on the y-axis (vertically). The
nearer the scatter of points is to a straight line, the higher the strength of
relationship between the variables. The relations are interpreted as
follows:

Value of 'r' Indication Interpretation

r = -1 data lie on a perfect straight as one variable increases, so the

line with a negative slope other decreases, and vice versa

r=0 no linear relationship no correlation

between the variables

r = +1 data lie on a perfect straight both variables increase or

line with a positive slope decrease together

r2 proportion of variation of one

variable ‘explained’ by the other

Spearman's coefficient of rank correlation

When data are measured on at least an ordinal scale, the ordered
categories can be replaced by their ranks. (Scales of measurement are
discussed in chapter thirteen). A Pearson’s correlation coefficient can
then be calculated on these ranks. This is called Spearman’s rank
correlation coefficient, which is denoted as rs. The rs thus provides a
measure of how closely two sets of rankings agree with each other. It is
 Quantitative Research Designs: An Overview 163

not a measure of linear association. Take an example of two teachers of

fine art assessing the perfomance of eight students. The students are
ranked according to performance from 1 (best) to 8 (worst) as follows:

Student Teacher No.1 Teacher No. 2

1 4 5

2 1 3

3 3 1

4 2 2

5 6 6

6 5 4

7 8 7

8 7 8

A significant association between the sets of ranks by calculating

Spearman’s rank correlation coefficient (rs) is indicated by p = 0.05, as
usual. The value of rs ranges from -1 to +1. Results indicate that there is
evidence to suggest good agreement (rs = 0.86) between the teachers’
assessments (p = 0.007).
The following formula is used to calculate this coefficient:

rs = 1 - Ód2 (n2 - 1)
n

where Ód2 is the sum of the squared differences between the pairs of
ranks, and n is the number of pairs. Thiscoefficient can be used for any
data that can be ranked.

Other correlations
When there are more than two variables, when the variables are
dichotomous (true/false or yes/no) or rank ordered, or when the
variables have a nonlinear or curved direction, different types of
correlations are used. The biserial and point biserial correlations are
164 Social Research Methodology: An Introduction

used when one variable is dichotomous and the other is continuous, e.g.
gender and income. The phi correlations are used when both variables
are dichotomous, e.g. gender and marital status. The Spearman’s rho
correlation is used with two rank ordered variables and eta is used when
the variables are nonlinear. For a detailed study of these correlations, see
Campbell & Machin (1999).

Correlation and causality

One common mistake often made with regard to the interpretation of a
correlational coefficient concerns causality. Let us take our earlier
example of depression and self-esteem. Upon observing that depression
and low self-esteem are negatively correlated, there may be a tendency
to conclude that depression causes the decrease in self-esteem. This is
highly misleading. Two variables may often be related just because of
a third variable that is not accounted for. Thus, it could be that
depression causes self-esteem to go down, or that low self-esteem results
in depression, or that a third variable causes the change in both. When
looking at a correlational coefficient therefore, one should recognize that
the variables may be correlated but that this in no way implies that the
change in one causes the change in the other.
Correlational studies are useful in a number of scenarios. First of
all, they are used if a researcher is seeking to establish relationships
between certain individual characteristics and some observed behavior,
e.g., education level and political participation, income and consumption
behaviour, gender and opinion on social issues, etc. Secondly, studies
based on individual observations enable researchers to generalize to
group behaviour and to generate hypotheses. For example, a
correlational study of the relationship between age and hypertension
might allow a researcher to generalize that the older the people become,
the more likely they are to become hypertensive. Thirdly, correlational
designs are relatively easy and inexpensive to carry out.
However, correlational studies are not without limitations. To begin
with, they are very weak in terms of establishing causality. The
researcher can only establish whether or not there is a relationship, and
whether the relationship is weak, moderate or strong; positive or
negative. Secondly, there is a problem of ecological falacy (Freedman,
 Quantitative Research Designs: An Overview 165
*
1999). In a study of household incomes and tuberculosis (TB)
prevalence, a researcher may not know whether those individuals within
a population who have lower household incomes are actually those who
contract and often die of TB even if on average populations with low
incomes also have higher rates of TB. Thirdly, there may be other
factors about the populations that are associated with the problem under
study. In our above example, if those with a lower income are more
likely to eat and drink unhygeinecally and/or communally or smoke
crude drugs, these would be more likely causes for increased TB
prevalence. In this case income would merely be a marker for feeding
and smoking habits.

Causal-Comparative (ex post facto) Design

Causal-comparative research is a useful tool that can be employed in
situations where experimental designs are not possible. The researcher
always keeps in mind that demonstrating a relationship between two
variables—even a very strong relationship—does not prove that one
variable actually causes the other to change. If for example one wanted
to find out whether unhygeinic consumption behaviour is linked to TB
prevalence, a correlational study would be appropriate. But if one
wanted to see whether health extention work could reduce the
prevalence of TB, a correlational design would not apply because the

* The ecological fallacy occurs when a researcher bases on aggregate data for
a group to make an inference about an individual. Suppose a researcher examines the
aggregate data on household incomes for a given area and discovers that the average
household income for the residents of that area is UGX 12,000,000 per annum. To state
that the average income for residents of that area is UGX 12,000,000 is true and
accurate. But to state, based on this data, that people living in the area earn about UGX
12,000,000 may not be true at all—it may be an ecological fallacy. On close
examination of the area, one might discover that there are two housing estates: one of
low-to-modest income earners, earning on average UGX 6,000,000 per annum, and the
other of high-income earners garnering UGX 25,000,000 per annum. Stating that
individuals who live in this area earn UGX 12,000,000 (the mean rate) without
considering that this average is computed from two disparate groups is to miss the point.
In fact, there may be extremely few individuals—if any at all—earning UGX
12,000,000. Assumptions or inferences made about individuals based on aggregate data
are thus vulnerable to the ecological fallacy. The process of aggregating or
disaggregating data may conceal the variations that are not visible at the larger aggregate
level, and researchers should be careful of this fact.
166 Social Research Methodology: An Introduction

researcher is now looking at prevalence in a time frame. To see whether

the extension work intervention has an effect, the researcher has to
compare the TB prevalence before and after the intervention was
effected. Therefore, the researcher would use an ex post facto design .
There are two basic approaches to ex post facto research: the first
one is to begin with subjects that differ on an independent variable (such
as their principal instrument/voice) and then study how they differ on
dependent variables (such as levels of performance, anxiety, or music
theory test scores). The second one is to begin with subjects that differ
on a dependent variable (such as attrition from music—comparing those
students who drop out of music with those who persist) and then study
how they differ on various independent variables (such as how much
they practice, how they feel about their relationship with their teacher,
how they feel about themselves as musicians, etc.).
There are often likely to be alternative explanations as to what
causes what. The first one is the situation where there is a common
cause of the two correlated factors. Both the independent and dependent
variables being studied may be influenced by the same thing. For
example, one may observe a strong correlation between the number of
students who wear sleeveless shirts and the rate of consumption of
mineral water at the school canteen. However, it would obviously be
faulty reasoning to claim that one thing causes the other bacause there
could be a third factor such as hot weather which dictates what to wear
and how much to drink, thereby influencing both. The second
explanation is reverse causality—it may appear that one variable is
causing the other when actually it is the other way around. For example,
it may appear that heavy smoking causes lung cancer when actually
cases of lung cancer progressively develop an urge to smoke. The third
explanation is the possibility that the observed effects are being caused
by one or more unknown independent variables.
To establish that a causal relationship exists between the variables,
the following kinds of evidence must be adduced: 1)one must establish
a significant statistical relationship between the independent and
dependent variable; 2) one must show that the independent variable
occurred before the dependent variable; and 3) one must show that no
extraneous variables are influencing the dependent variable.
 Quantitative Research Designs: An Overview 167

Survey Research Designs

Surveys represent one of the most common types of quantitative, social
science research. In survey research, the researcher selects a sample of
respondents from a population and administers a standardized
instrument (structured questionnaire or interview schedule) to them
(Survey research is discussed in Chapter Thirteen). Surveys are not a
preserve of social research—they can be employed by almost any
discipline. It is perhaps their capacity for wide application and broad
coverage that makes surveys an indispensable method of social
investigation. There are generally three survey designs: longitudinal
surveys, cross-sectional surveys and cross-sequential surveys.

Longitudinal surveys
A researcher may seek to assess and explain the changes occurring in a
given society (or among different sections of society) over an extended
period of time. He or she will design a longitudinal survey in which he
or she will investigate the same groups of subjects for a given period.
The results of longitudinal studies can provide valuable data (including
qualitative data) regarding the differences in development between
various groups. For example, to assess the differences in the
performance of Universal Primary Education (UPE) in rural and urban
settings, a researcher may choose a sample of subjects from each of
these settings and assess them on various measures every six months for
a period of six years. The major problem with longitudinal research, in
fact, is the length of time it takes to complete a single study.
Longitudinal designs are themselves of three types.

1. Trend studies. These studies focus on a given population, which

is scrutinized repeatedly over a period of time. Each time, different
samples are drawn from that same population. But since trend studies
are conducted over a long period of time, they do not have to be
conducted by just one researcher or in one research project. At any one
time, data from several observations of the population may be analysed
(compared) in order to establish a trend.

2. Cohort studies. Cohort studies also focus on a particular

population, but with a focus on a specific cohort within that population.
A cohort is a well-defined group of people who have had a common
168 Social Research Methodology: An Introduction

experience or exposure, who are then followed up for any developments

or changes. A group of law or education students that was admitted at
University X in 2001 or graduated at college Y in 2003 is a cohort. A
study could be conducted on samples drawn from such a group every
after two years in order to establish any changes. If the researcher
studied law student admitted at University X in 2002 some years later,
it would not be a cohort study.

3. Panel studies. In panel studies, the researcher studies the same

sample of subjects every time to find out why changes in the population
are occurring. That sample is called a panel. If you selected and
interviewed 25 people just recruited into one department of Rwanda
Revenue Authority (RRA), you would interview the same people on the
same issues every after a period, and ask them the reasons for any
changes in their responses.

Cross-sectional surveys
Cross-sectional surveys are used to gather information on a population
at a single point in time. If for example a researcher wanted to study the
reactions of private students about the fees increment at Makerere
University, he or she would supply a questionnaire or conduct
interviews across a range of students just once. In other cases, a cross
sectional study might seek to examine the soundness of a theory linking
academic and social development, and decide to conduct it on small
groups of say three year olds, six year olds, nine year olds and 12 year
olds at the same time. The assumption is that the differences between the
age ranges represent natural development and that if a longitudinal study
had been used, similar results would be found. The obvious benefit is in
the length of time it takes to complete the study, but the assumptions that
the six year old group will achieve the same academic and social
development as the nine year old group can be invalid.
In cross-sectional studies both the independent and the dependent
variables are determined and investigated at the same time. For example,
a cross-sectional survey might determine whether the subjects were
current smokers (smoking being the IV) and whether they had evidence
of TB infection (the disease being the DV). Because both exposure and
disease are determined at the same time, it is not possible to determine,
 Quantitative Research Designs: An Overview 169

in most cases, which came first. Therefore, cross-sectional studies can

suggest associations between IVs and DVs, but cannot prove causality.

Cross-sequential studies
Cross-sequential studies combine both longitudinal and cross sectional
methods in an attempt to both shorten the length of the research and
minimize developmental assumptions. For example, pupils enrolling in
Primary One under UPE may be studied for a period of three years to
assess both their progress and whether an average P.1 pupil is similar to
a P.3 pupil after three years under a UPE programme.
Chapter Twelve

Survey Design and Instrumentation

Introduction
Occasionally, authorities—whether public or private, national or
local—may require facts on some socioeconomic or other problem
within their jurisdictions for purposes of making informed policy
planning, intervention or other decision-making. They may seek to
establish the cause-effect and other relationships about social
phenomena, or just to understand why certain important social,
economic, political and other forces are the way they are. With
increasing globalization and reliance on Information and
Communication Technologies (ICTs) in every aspect of human
endeavour, organizations require a prompt and accurate flow of
information on preferences, needs, problems and behaviour of the
people. It is in response to this critical need for information on the part
of the government, business, and social institutions that so much reliance
is placed on surveys. Surveys provide a speedy and economical means
of determining facts about society and about people’s knowledge,
attitudes, beliefs, expectations, and behaviours.
A social survey is a structured, large-scale inquiry into any aspects
of the above or related concerns, basing not on all of the individuals
within the population for the requisite information, but on a
representative sample drawn from this population. The
representativeness of the sample must reflect a number of important
issues: the demographic characteristics of the populations in question;
the socio-political environment; economic activities of the populations;
their cultures, mores and beliefs; their attitudes, opinions, and
aspirations; the effect of non-social or non-human forces such as natural
 Survey Methods and Instrumentation 171

calamities; etc. The information from the sample is then generalized to

the entire population.
This mesh of concerns and processes, and the individual and social
attributes of the population readily suggest the diversity of contexts
within which social surveys are conducted. They imply also that social
surveys differ from each other in respects ranging from the purpose for
which they are conducted, their scope and magnitude, the sources of
information, to the methods of getting and analyzing this information.

Types of Surveys
Social survey research is a broad research design, encompassing any
measurement procedures that involve selecting representative samples
of respondents and asking the respondents questions. Such asking may
be through a paper-and-pencil exercise executed by the respondent
himself or herself. The researcher designs and pretests the data
collection instrument called the questionnaire before hand, and hands it
to every respondent to fill in by themselves. Alternatively, the asking
may be interactive and conducted through a structured one-on-one
interview between the researcher (interviewer) and the respondent
(interviewee). In this case, the data collection instrument known as the
interview schedule is retained by the interviewer. The interviewer then
asks questions to the respondent and records the responses, either in a
notebook or using a voice recorder, or both. The different types of
surveys may thus be roughly divided into two broad areas:
Questionnaire surveys and interview surveys.
Quite often, it is hard to tell the difference between a questionnaire
and an interview schedule.* Both are written and pretested instruments;
questions thereon are asked in a systematic sequence; and both are used
to gather quantitative data. It can be said therefore, that an interview
schedule is a questionnaire designed for oral administration. The major
difference between the two may be discerned on the choice the
researcher has to make on the basis of the literacy level of the
population. A questionnaire is necessarily used on a population that can
read it and fill it independent of the researcher. It may be structured (all

* Moser and Kalton (1971:308) in fact use what they call the “more
convenient” term questionnaire to refer to both documents for purposes of instrument
design. They are of course well aware of the differences between the two.
172 Social Research Methodology: An Introduction

questions pre-coded and closed-ended), semi-structured (some questions

closed and others open-ended), or unstructured (all questions open-
ended). An interview schedule on the other hand is used by a researcher
to orally administer an interview. It is fully structured and suitable to use
when the population is illiterate, or where it is distant so that contact has
to be through telephone.

The Questionnaire Survey

One of the most important considerations, right from the time of
planning the objectives of the survey, is the population of the study.
Where do they live? What are their education and literacy levels? Any
unique characteristics? How will information be gathered from them?
These and other questions help the researcher to decide whether a
questionnaire will be circulated amongst the sampled respondents, or
whether the respondents will be interviewed. If the population is easily
accessible, able to read and write, and generally inclined to respond to
the questions without prodding, a questionnaire survey is preferable.
There are generally four types of questionnaires for this type of survey:
mail questionnaire, self administered questionnaire, group administered
questionnaire, and electronic questionnaires.

Mail questionnaire
To employ a mail questionnaire, the researcher identifies the postal or
physical addresses of the respondents and mails to them the
questionnaire. The questionnaire must be accompanied by a detailed
covering letter explaining the nature and purpose of the survey,
requesting consent and cooperation, and pledging confidentiality and
other obligations. The researcher must also enclose a self-addressed and
stamped envelope which the respondent will use to mail the completed
questionnaire back.
There are a number of advantages associated with mail surveys or
mail questionnaires as I outline below:
• Although the postal system and courier services in Uganda (and
East Africa as a whole) are generally expensive, mail surveys are
relatively inexpensive to administer. This is especially so where the
population is dispersed and the respondents much far apart that
other forms of instrument administration would be more expensive.
 Survey Methods and Instrumentation 173

• Mail surveys are time-saving. A researcher can send the exact same
instrument to a wide number of respondents at the same time, and
they are sent back within the same short period.
• A mail questionnaire enables the respondent to fill it out at their
own convenience. This may allow for more accurate information
because it allows for consultation.
• Because of the impersonal nature of the relationship created by the
physical distance between the researcher and the respondent, the
possibility of bias is highly minimised if not eliminated altogether.
• There is greater assurance of anonymity, and this favours a higher
response rate in terms of completion of instrument.

However, this form of survey is not without disadvantages. These

include the following:
• Response rates from mail questionnaires are often very low. This
is because some people simply do not attend to the questionnaire
since they neither see nor know the researcher. They may thus not
feel any obligation, and therefore not send it back.
• Mail questionnaires are not the best medium for seeking detailed
written responses since they lack flexibility.
• It is not possible to cross-check the accuracy or authenticity of the
responses, so they have to be taken as final.
• Mail questionnaires are inappropriate where spontaneous responses
are sought. For example, where the researcher is testing knowledge
and is interested in the actual views of an individual at the very
instance of interrogation, this method cannot work.
• One cannot ascertain who actually filled the questionnaire.
• In populations of lower educational and literacy levels, response
rates are often too low to be useful. Experience shows that the best
response levels are achieved from highly-educated people and
people with a particular interest in the subject (which, depending
on your target population, could lead to a biased sample).

Self administered questionnaire

In a self-administered questionnaire (also called a household drop-off
survey), a researcher goes to the respondent’s home or business and
hands the respondent the instrument. In some cases, the respondent is
asked to mail the instrument back, and this requires that the researcher
174 Social Research Methodology: An Introduction

leaves a self-addressed and stamped envelope behind. Otherwise, the

researcher returns to pick it up. This approach has a number of
advantages.
• The respondent can attend to the questionnaire in private, when it
is convenient. This way, he or she face less or no pressure and is
therefore likely to give it full attention.
• The researcher makes some personal contact with the respondent,
and not just send an impersonal survey instrument. The respondent
can therefore ask questions about the study and get clarification on
what is to be done before the researcher goes away. Generally, this
would be expected to increase the number of people willing to
participate and therefore the response rate.
• Self-administered questionnaires preserve confidentiality.
Anonymity and privacy of the respondent encourage more candid
and honest responses, which can lead to more valid and authentic
responses.
• Self-administered questionnaires are less expensive than interviews
since they do not require a large staff of skilled interviewers.
• They are a relatively faster method of data collection, since they
can be administered in large numbers all at one place and time.
They thus enable speed of administration and analysis.
• Since there is no interviewer, the problems associated with
interviewer error or bias are eliminated.
• Self-administered questionnaires are suitable for computer based
research methods.

This method, however, is also faced with a number of disadvantages.

These include the following:
• There is often a low response rate, since the absence of the
researcher leaves the respondent with no urgency or obligation to
fill the questionnaire.
• Questions can be misunderstood, and the researcher is not at hand
to explain them.
• In some forms of self-administered surveys, there is no control by
the researcher over who actually fills out the questionnaire.
• The respondent may very well read part or all of the questionnaire
before filling it out, and this may potentially bias his or her
responses.
 Survey Methods and Instrumentation 175

• There is considerable time loss as the researcher traverses the area

of the survey locating the households and respondents, going back
to collect the filled out questionnaires, as well as chasing after the
respondents that did not fill out the questionnaires.

Group-administered questionnaire
In a group-administered questionnaire, a sample of respondents is
brought together and asked to respond to a structured sequence of
questions. Each respondent is handed an instrument which he or she
completes while in the room. The administration of questionnaires in
group settings is basically for convenience. The researcher gives the
questionnaire to those who are present and is fairly sure that there is a
high response rate. If the respondents are unclear about the meaning of
a question they ask for clarification.
The group-administered questionnaire presents to the researcher a
number of advantages:
• They are generally administered to a sample of respondents in a
group setting, and this guarantees a high response rate.
• A fairly large group of respondents can be surveyed at one time,
and this makes it very tim-saving.
• Because each individual attends to his or her instrument alone, this
anonymity may provide more honest answers especially to
sensitive questions.
• The costs are kept significantly low since less staff is needed than
for personal interviews.
• Group administered questionnaires can be very versatile, allowing
for a spectrum of open and closed ended types of questions, and
can serve a variety of specific purposes, particularly if one is
surveying a very specific group of people.
But this method has a number of weaknesses as well, and these include
the following:
• It requires a relatively small sample, and is therefore not the best
method for large scale surveys. It is only useful in cases that call
for very specific information from specific groups.
• Since this method requires a group of respondents to answer the
survey together, there is a need to set aside time that is convenient
for all respondents. This is not always easy to arrange.
176 Social Research Methodology: An Introduction

• This method is rarely suitable for most populations being studied,

especially since members of the group are not independently
chosen.

Electronic surveys
This category combines all manner of surveys conducted through the use
of electronic technology. They include questionnaires administered via
e-mail and those published on the intranet or Internet. Electronic surveys
are associated with a number of advantages, and these include the
following:
• They are both very economical and very fast. An e-mail or web
page questionnaire can gather several thousand responses within a
day or two.
• There is practically no cost involved once the set up has been
completed. And large samples incur the same cost as smaller ones.
• It is possible and easy to attach pictures and sound files onto the
questionnaire.
• The novelty element of an electronic survey often stimulates higher
response rates than ordinary mail surveys.
• Web page questionnaires can use complex question skipping logic,
randomizations and other features that are not possible with paper
questionnaires. These features enhance better data.
• Respondents are inclined to give more honest answers to questions
about sensitive topics, such as drug use or homosexuality, when
giving their answers to a computer, instead of to a person or on
paper.
The disadvantages of electronic questionnaires include the following:
• Electronic media depend solely on electricity, which is very erratic,
even non-existent in many parts of Uganda and other developing
countries.
• Some people are likely to respond several times or pass the
questionnaire along to friends to answer.
• A number of programs have no mechanism to check and eliminate
people responding multiple times to bias the results.
• Some people do not respond to unsolicited e-mail, and this is likely
to affect the response rate.
 Survey Methods and Instrumentation 177

• Respondents can easily quit in the middle of a questionnaire. They

are not as likely to complete a long questionnaire on the Web as
they are if talking with a good interviewer.
• Electronic surveys do not reflect the population as a whole. This is
true even if a sample of Internet users is selected to match the
general population in terms of age, gender and other demographic
characteristics. Therefore, one cannot generalize the findings to the
whole population.

The Interview Survey

Interviews are a far more personal form of research than questionnaires.
However, not every interview can possibly be direct to make it personal,
and not every one can be a one-on-one. There are thus three types of
interview surveys: the personal interview, the telephone interview, and
the group interview.

Personal interview
An interview is called personal or direct when the interviewer asks the
questions face-to-face with the interviewee. Personal interviews can take
place in the home, on the street, in office or polling place, etc. The
interviewer works directly with the respondent; and unlike in
questionnaire surveys, the interviewer has the opportunity to probe or
ask follow-up questions. Interviews are generally easier for the
respondent, especially if what is sought are opinions or impressions. The
interviewer is considered a part of the measurement instrument,
therefore he or she has to be well trained in how to respond to any
contingency.
Like other survey methods, personal interviews have got their
relative advantages and shortcomings. The advantages include the
following:
• Personal interviews are conducted at the place and time convenient
to the respondent. This is because appointment is normally secured
prior to the time of the interview. As a result, the respondent is
always ready to give full attention to and complete the interview.
• Longer interviews are tolerated. People may be willing to talk
longer face-to-face than to someone on the phone.
178 Social Research Methodology: An Introduction

• The researcher is able to cros-check the accuracy or genuineness of

the response through the respondents body language. This includes
facial expressions to express dislike, uncertainty, contempt, etc.
• Probing is used to get a full response.
The disadvantages include the following:
• Personal interviews are very time consuming as it may take up to
one hour or more to complete a single interview.
• They are resource-intensive, especially in terms of the cost
involved in the training of interviewers.
• Personal interviews usually cost more per interview than other
interview methods. This is particularly true of interviews conducted
in homes where travel time is a major factor.
• There is a higher level of interviewer bias. The interviewer may be
from a different background from that of the respondent, and so the
personal impression he or she gets of the respondent may influence
the way he or she interpretes the findings.

Telephone interview
Quite often it is very difficult to find every respondent at their home or
office. This is especially so where the population of the study is
scattered across a wide geographical area and involves great distances.
In such cases, the researcher finds contact via telephone handy if the
telephone addresses can be accessed. Telephone interviews are
particularly preferred in the conduct of public opinion polls, and they are
particularly popular with journalists.
There are a number of advantages associated with the use of this
method. These include the following:
• Telephone interviews enable a researcher to gather information
rapidly from a wide area, especially remote and difficult-to-access
areas. If the interviewer is taking advantage of the ICTs and using
computer-assisted telephone interviewing (CATI), the results can
be available shortly after completion of the last interview.
• They reduce travel costs while permitting interaction with remote
participants.
• Like personal interviews, telephone interviews allow the
interviewer to ask probing and follow-up questions, and
clarification of unclear responses. Skilled interviewers can often
 Survey Methods and Instrumentation 179

elicit longer or more complete answers than respondents would

give on their own to other forms of surveys.
• Telephone interviews can supplement site visits, face-to-face
interviews, and other methods. For example, after a site visit or a
questionnaire survey one may gather additional data by conducting
a phone interview.
• Telephone interviews have a better response rate than mail surveys.

The disadvantages associated with telephone interviews include the

following:
• Telephone interviews are quite expensive, especially in Uganda
where the call rate per minute is prohibitively high.
• Telephone interviews are not as good as face-to-face interviews
when the researcher is dealing with complex issues.
• If the telephone interview is conducted when the interviewee is at
home, it faces many potential distractions such as visiting
colleagues, kids, calls on other lines, background noise, etc.
• If the survey involves showing or sampling products, it cannot be
done by phone.

Group interview
In a group interview, the respondents are not given the research
instruments. Instead, they work in groups to answer the questions
together while one person takes notes for the whole group. The
interviewer facilitates the session, while the group, listening to each
others’ views and comments, answer the questions. A general evauation
of interviewing as a data collection method is made in Chapter Six.

The ABC of Survey Design

Admittedely, social surveys differ in many respects—size, method of
data collection, nature of problem being investigated, character and
location of population, etc. Therefore, while one survey may be
conducted by a handful of people and last only a few days, another may
employ several researchers and take several months to complete. While
one may focus on opinions and attitudes, another may seek factual
information such as housing status, income levels, etc. Nevertheless, all
surveys are defined by a common planning and execution process.
180 Social Research Methodology: An Introduction

The survey design process goes through five stages before a

researcher goes into the field to administer the instruments for purposes
of data collection. The process begins with the determination of the
purpose of the study, and proceeds to determine the population and
sample to be used. Then the researcher decides the type of survey type
that is suitable for the study, whether questionnaire or interview, designs
the data collection instruments, and pretests them before proceeding to
do research. I highlight these stages below.

The purpose of the survey

The first stage in any survey is deciding what you want to learn. The
aims and objectives of the survey determine the respondents to recruit,
the duration it will take, and the information to gather from them. If not
properly set, the objectives will yield poor results. Social surveys
normally pursue such general objectives as the potential market for a
new product or service;
ratings of current products or services; employee attitudes; customer
satisfaction levels; reader/viewer/listener opinions; association member
opinions; opinions about political candidates or issues; corporate
images; etc. These examples represent general areas. The more specific
a researcher can make them, the easier it will be to get useful
information.

Population and sample selection

There are two main components in determining whom you will
interview or supply a questionnaire to. The first is deciding the target
population. If the researcher seeks to conduct an employee attitude
survey in a given sector or organization, or an association membership
survey, the population is obvious. If tye aim is to predict the outcome of
a presidential election, the target population may be less obvious.
Correctly determining the target population is therefore critical,
otherwise the objectives will not be achieved. The second issue to decide
is how many people one needs to interview or supply the questionnaire
to. Surveys are conducted on representative samples from the
population, and as a general principle, the larger the sample, the more
precisely it represents the target group. Population and sampling
processes are discussed in Chapter Three.
 Survey Methods and Instrumentation 181

Selection of survey method

Once a researcher has decided on his or her target population and sample
size, he or she must decide on the method of data collection. There are
many factors to guide the choice between questionnaire survey and
interview survey, and which types of questionnaire and interviews. Each
method of course has its strengths and weaknesses. In general, however,
the researcher should consider the nature of the problem under
investigation; the nature, size and location of the population; the time
frame, personnel and other resources available, among others.

Instrument design
This is a critical stage and it requires very careful and concerted
attention to (1) the objectives of the survey—what do we want to
achieve? (2) the hypotheses—what assumptions are to be tested? (3)
research questions—what questions is this study meant to answer? (4)
required data—what specific questions should be asked to gather
specific data to address all the foregoing issues? Every question posed
in the instrument should have particular relevance, otherwise it will be
redundant and useless. The researcher should ask himself or herself:
“what information will this question bring?” “what will the respondent
understand this question to mean?” “what purpose will this information
serve?” “will it satisfactorily serve that purpose?” etc. The instrument
must pass the validity test (validity and reliability tests are discussed in
Chapter Two).

Pre-testing the instrument

After the instrument has been completed, it is pretested on a small
number of people before conducting the main survey. Pretesting is for
purposes of adjustment of the instrument in order to ensure maximum
validity. It can reveal unanticipated problems with question wording or
instructions to skip questions, and it helps to show if the respondents
understand the questions and can give useful answers. Ideally, the
instrument should be pretested on the same kind of people as will be
involved in the main study.
There are about eight techniques of pretesting a survey instrument.
These fall into two categories: those used before fieldwork, and those
used during fieldwork. The pre-field techniques include the following:
182 Social Research Methodology: An Introduction

1. Respondent focus groups. These are focus group discussions that are
conducted early in the instrument development process. They are used
to assess the question-answering process.

2. ?Think-aloud” interviews. These are one-on-one interviews using a

structured interview schedule or questionnaire, in which the respondents
describe their thoughts while answering the questions. The other
techniques that test the survey instruments under operational conditions
include the following:

3. Behaviour coding. This is the systematic coding of the interactions

between the interviewer and the respondent, either from live or voice
recorded interviews.

4. Respondent debriefings. These involve incorporating structured

follow-up questions on the respondents' interpretations of the survey
questions. The main purpose is to determine whether concepts and
questions are understood in the same way and as originally intended.

5. Interviewer debriefings. The interviewer utilises the opportunity of his

or her direct contact with the respondents to enrich the instrument
designer's understanding of the answering problems.

6. Split-panel tests. These are controlled experimental tests on

questionnaire variants or modes of interviewing in order to determine
the better one or the differences between them.

7. Analysis of nonresponse rate. This is done by examining how often

items on the questionnaire are not responded to.

Principles of Questionnaire Design

Constructing a survey instrument is an art in itself. There are a number
of decisions that must be made—about content, wording, format,
placement, etc—that can have important consequences for the entire
study. There are some guidelines one may follow that might increase
one’s chances of developing a valid and reliable instrument. I will
highligh these in three perspectives: what must be done, the formats of
survey questions, and the kind of questions to avoid.
 Survey Methods and Instrumentation 183

Rules of Thumb in Questionnaire Design

1. Design to fit the medium. Keep in mind the medium through which
the questionnaire will be administered. If you intend to use the Web, you
may design a questionnaire with animated pictures and catoons for
illustration. You may even incorporate voice to guide the respondents
if they do not understand a question or guide on the way questions are
to be answered. Intimate or sensitive questions are sometimes best
handled this way, because anonymity is most assured.

2. Be courteous. Start with an introduction of yourself, your institutional

affiliation, and your mission. Seek the respondent’s consent, and make
all the assurances the respondent requires. A good introduction or
welcome message encourages people to complete the questionnaire.

3. KISS—keep it short and simple. If you present a big questionnaire, say

one with 50 questions taking about 12 pages, most potential respondents
will be demotivated as soon as they receive it. Long questionnaires, or
those that require the respondent to think hard, are associated with poor
response rates or nonresponses.

4. Phraseology. When designing a questionnaire, it is important to

ensure the questions elicit responses that accurately tell you what you
want to learn. The way you phrase a question can change the answers
you get. Try to make sure the wording does not favor one answer choice
over another—the questions should accept all the possible answers.

5. Know the expected information. A researcher should put himself or

herself in the position of a respondent and gauge what response he or
she could give to the question. If you cannot give yourself a satisfactory
answer, leave the question out. Consider the following three categories
to see where a given question belongs: “must know,” “useful to know,”
and “nice to know.” The last category should be discarded, unless the
previous two groups are very short.

6. Question sequencing. Ask yourself how the order of the questions and
answer choices is likely to encourage or discourage your respondents to
complete the questionnaire. It is advisable to begin with questions that
184 Social Research Methodology: An Introduction

are easy and pleasant to answer, and move progressively with harder or
more sensitive ones. Grouping together questions on the same topic also
makes the questionnaire easier to answer.

7. Question content. This is about the kind of information sought by the

qquestion—whether factual information or opinion information. A
factual question is one whose answerr can be confirmed or validated
either through observation or by consulting documents. Opinion
questions, on the other hand, are not confimable since a respondent’s
opinion can change any time. If a factual question is to be asked, one
should ensure that the respondent will understand the question, will
know the answer, and will answer correctly. If opinion questions are to
be asked, gauge whether the question is likely to elicit a genuine
opinion. Opinion questions are often asked in a form of statements and
the respondent is asked to agree or disagree with the statements.

8. Flexibility and exhaustiveness. The questionnaire designer should not

assume that the alternative answers to a given question are exhaustive.
Similarly, a rigid set of alternative answers may lead some respondents
to feel they are being coerced into giving an answer they do not want to
give; and they will most likely not complete the questionnaire. In such
cases, one should include such alternatives as “Don’t know,” “Not
applicable,”“Other,” “None,” or “Decline to state.”

Formats of Survey Questions

Survey research questios fall under two broad formats—open-ended
questions and close-ended questions. Any of these question formats may
require text or numeric responses. Hence, there are open-end text, open-
end numeric, closed-end text, and closed-end numeric questions.
Open-ended questions are unstructured questions which allow the
respondents the latitude to present their responses. An example of such
questions:

What do you think should be done to improve household

incomes?
__________________________________________
__________________________________________
__________________________________________
 Survey Methods and Instrumentation 185

__________________________________________

There is no restriction as to the amount of information the respondent

may give or the order in which it should be given. This is an open-end
text question, which is sometimes referred to as a “verbatim.” Below is
an example of an open-end numeric question:

What is your average monthly income? ________

It can clearly be seen that the space provided for the response is small,
because what is expexted is just a figure.
Close-ended questions are structured, multiple choice questions in
which the researcher controls the way the question should be answered
by providing alternative answers to select from. The researcher must be
versed with the problem at hand in order to be able to structure the
question. Quite often, in fact, a baseline study (normally through FGDs)
is conducted to generate factual information, hypotheses and
terminology to be used while designing a larger survey. Closed-ended
questions may be numeric or text, and include categories that are
subjected to measurement scales. Here below are examples of various
types of closed end questions.

An example of a closed-end numeric question:

What is your average household income?

Q Below US $200
Q US$ 201 - 400
Q US$ 401 - 600
Q US$ 601 - 800
Q Over US$ 800

Two examples of a closed-end text question:

What is the single most important threat to child health in this

area?
Q Measles
Q Malaria
Q Diarrhoearal diseases
186 Social Research Methodology: An Introduction

Q Acute Respiratory Inflation

Q Other (specify)
_______________________________

Which region of the country do you come from?

Q Northern
Q Central
Q Eastern
Q Western
Q Not applicale

Examples of questions that are subject to measurement scales:

What is your rating of the class average performance?

Q Excellent
Q Good
Q Fair
Q Poor

On an interest-measuring scale where 1 means least

interested and 7 most interested, how would you rate
yourself on the following?
Jurisprudence _______
Economics _______
Sports _______
Debating _______
Gardening _______
Politics _______
Relationships _______

How would you agree with each of the following statements?

Strongl Agre Disagr Strong
y agree e ee ly
disagr
ee
 Survey Methods and Instrumentation 187

Uganda’s NRM is a
political paragon
East African Federation
will be 100% success
Politics is a dirty game
Only FDC and NRM
are serious parties, the
rest are jokers
Corruption in Uganda is
institutionalised
Wnen FDC takes
power corruption will be
history

The process of structuring questions for scaling and the different types
of scales are discussed in Chapter Thirteen.

Questions to be Avoided
There are a number of bad questions in a survey that should be avoided.
These are questions that affect not only the response rates but also the
overall validity and reliability of the survey instrument. The following
types of questions should be avoided:

Leading questions. These are questions that bias the respondents to

answer in a given direction. The rsepondent is inclined towards a
specific answer—rather like question tags and answer tags in English
lessons. A leading question may read as follows: ?Since graduated tax
affects the poor most, shouldn’t it be abolished?”

Ambiguous questions. These are questions that are more or less two
in one, so they pose to the respondents a difficulty of answering them.
They are also sometimes called double barrelled questions. It is not easy
to tell whether the answer is to one aspect of the question or the other,
or indeed both. Take an example of a YES/NO question that reads as
188 Social Research Methodology: An Introduction

follows: ?Are you satisfied with the performance of the Minister and his
deputies?” Suppose the deputies are performers and the Minister a non-
performer, what would the respondent’s answer be?

Vague questions. These are questions that are amenable to different

interpretations, so different respondents tend to understand and respond
to them differently. If for instance you asked a question: ?Are you a
serious cinema goer?” what would the term <serious’ mean?

Presumed questions. These are questions which assume that the

respondents engage in given activities, possess certain qualities, hold
certain opinions, or are inclined in certain directions. The designer thus
formulates a question on the basis of such preconceived positions. For
instance, you cannot ask someone the type of movies they watch unless
you have already asked whether they watch movies at all. Presumed
questions are not necessarily bad. When the same questions are
predicated on the responses of previous questions, they are quite
acceptable. In that case, the previous questions are called filter
questions, and the subsequent ones contingency questions. Here is an
example: ?Have you ever used contraception?” [filter question]. ?If your
answer to the above question is YES, what contraception methods have
you used before?” [contingency question].

Embarassing questions. There are some questions that may only be

answered in a situation of assured total anonymity. These are questions
of a personal nature, which touch on the negative aspects (history,
personality, habit, etc) of the respondent that he or she feels extremely
ashamed and uncomfortable about.

Hypothetical questions. These are questions that require the

respondents to assume roles or imagine a situation and answer the
questions from that perspective. For example, asking inmates what they
would do if they were made prisons authorities will not bring out
rational data for penal reform. Hypothetical questions are therefore bad
questions because they have poor predictive value.
Chapter Thirteen

Measurement and Scaling Methods

Introduction
The first issue to determine in any survey design is always: “What is to
be measured?” Much as the problem statement or research question of
the study may clearly indicate the concept that is to be investigated, it
often does not say anything about the measurement of that concept.
Suppose the researcher wants to evaluate the performance of a
programme such as Universal Primary Education (UPE). This may be
defined in numerical terms such as total enrolment, average performance
of schools, percentage increases, etc. The researcher could also measure
more subjective factors such as satisfaction of teachers, morale of pupils,
attitudes of parents, etc.
In any research that relies on quantitative techniques, the objective
is to express in numeric terms the differences in responses. Hence, a
scale is used to represent the item being measured in the spectrum of
possibilities. The values assigned in the measuring process can then be
manipulated according to certain mathematical rules. This chapter
presents a discussion of the concepts of measurement and scaling, and
shows that scaling may be considered an extension of measurement.

Measurement in Survey Research

Measurement is the procedure of assigning numbers, letters or other
symbols to variables according to pre-specified rules (Miller, 1991;
Folz, 1996). The variables range from quantifiable measures such as
distance, amount, age, etc or objects such as persons, instruments,
computers, etc to more subjective attributes such as satisfaction,
perception, attitude, emotion, opinion, preference, etc. In survey
190 Social Research Methodology: An Introduction

research, such as market research or opinion poll, numbers are usually

assigned for one or two reasons. First, numbers permit statistical
analysis of the resulting data. Second, numbers facilitate the
communication of measurement rules and results. The most important
aspect of measuring is the specification of rules for assigning numbers
to characteristics or attributes. The rules for assigning numbers should
be standardized and applied uniformly. They must not change over
objects or time.
Let us take an example of a researcher who is interested to measure
the preferences of political parties among his or her population. He or
she may arbitrarily assign the values 1, 2 and 3 to the parties in question,
say, 1 = FDC, 2 = NRM, 3 = UPC. The level of measurement describes
the relationship among these three values. The researcher is simply using
the numbers as shorter placeholders for the lengthier text terms. It
cannot be taken that higher values mean “more” or “better” of something
and lower numbers signify “less” or “worse.” The value of 2 for instance
does not mean that the NRM is twice preferred to FDC. The value of 3
assigned to UPC cannot mean that UPC is the least preferred party. The
values are used only as identifiers or codes for the attributes being
measured. Here, the researcher would describe this level of measurement
as “nominal,” – derived from “name.” There are of course other levels,
as I shall soon show.
There are at least two reasons why one ought to know the level of
measurement. First, it helps one to decide how to interpret data from a
given variable. When you know that a measure is nominal (like the one
just described), then you know that the numerical values are just
placeholders for the longer names. Second, it helps the researcher to
decide what statistical analysis is appropriate on the values that are
assigned. Statistical tests and analyses are discussed in Chapter 14.

Data Types and Levels of Measurement

There are typically four types of data that give rise to four corresponding
levels of measurement. These are nominal, ordinal, interval, and ratio.
Let us examine these in detail.
 Measurement and Scaling Methods 191

Nominal Measurement
At this level of measurement, data is partitioned into discrete categories
where the categories have no relationship to each other. The numerical
values assigned to the data simply “name” the attribute uniquely. No
ordering of the cases is implied. For example, the numbers assigned to
vehicles in a motor rally are measures at the nominal level. They are
simply identifiers or labels for the contesting motorists and their
navigators. If Katebire’s car is No. 8 and Kamuhanda’s is No. 4, it does
not imply that Katebire is twice better than Kamuhanda. Nor does it
imply that Kamuhanda is four something ahead of Katebire.
It has become a highly contentious question whether nominal
measurement is actually measurement since it only assigns numbers for
the purposes of catogorizing events, attributes or characteristics (Miller,
1991; Babbie, 1995; Sarantakos, 1998). A nominal measure does not
express any values or relationships between variables. The only
mathematical or statistical operation that can be performed on nominal
data is a frequency run or count. One cannot determine an average,
except for the mode – the number that holds the most responses. Not
even addition and subtraction operations can be done. On a
questionnaire with nominal scale questions, it is important that the
response categories must include all possible responses. The categories
provided must also be mutually exclusive, that is to say, thay should not
overlap or duplicate in any way.

Ordinal Measurement
Data at this level of measurement has an implied order of magnitude but
the distances between each category are not assumed to be equal. When
items are classified according to whether they have more or less of a
characteristic, the measurement used is ordinal. In ordinal measurement,
the attributes can be rank-ordered, but the distances between these
attributes do not have any meaning. The main characteristic of the
ordinal scale is that the categories have a logical or ordered relationship
to each other, so measurement of degrees of difference is possible, but
not the specific amount of difference.
For example, a researcher may rate the respondents’ level of
education as 1 = no education, 2 = primary education, 3 = secondary
education, 4 = tertiary education.
192 Social Research Methodology: An Introduction

The researcher may know that respondent X (secondary education) has

lower education attainment than respondent Y (tertiary), but he or she
has no idea how much lower X is from Y. Respondent X may have
completed Senior 1 (close to primary education) or Senior 6 (close to
tertiary education). One cannot tell that the distances between answer
alternatives are equal. Hence only the mode and median can be
calculated, but not the mean. The range and percentile ranking can also
be calculated (See Chapter 15). This level of measurement is very
common in opinion, satisfaction and attitudinal research.

Interval Measurement
Data at this level of measurement, as with ordinal level data above, has
an implied order of magnitude, and the distances between the
magnitudes are equal. Interval measurement takes the notion of ranking
items in order one step further from ordinal, since the distance between
adjacent points on the scale are equal. Take an example of a weighing
scale with values spanning to 60 kg. The distance from 10-20 is the
same as that from 40-50. The interval between these values is therefore
interpretable. As another example, think of a clock. The hourly intervals
are equal. Thus, unlike in ordinal measurement, one can compute
averages (modes, means, medians) of an interval variable.
There is no true or absolute zero point in interval measurements,
which means that although one can add and subtract, one cannot
multiply values or create ratios (you cannot say 12.00 is twice as much
as 6.00 am or pm; nor that 80° is twice as warm as 40°). What is
important in determining whether a scale is considered interval or not is
the underlying intent regarding the equal intervals.

Ratio Measurement
A ratio measurement is largely like an interval one with the difference
that the former has an absolute zero that is meaningful. Data at this level
of measurement has an implied magnitude with equidistant spacing
between values, and the zero point on a ratio scale represents a true zero
rather than an arbitrary one. If we ask respondents their ages, the
difference between any two years would always be the same, and ‘zero’
signifies the absence of age or birth. Hence, a 60-year old person is
indeed twice as old as a 30-year old one.
 Measurement and Scaling Methods 193

In applied social research, most “count” variables are ratio. These

include variables such as age, height, distance, income, etc. Ratio scales
are therefore used to gather quantitative information on such variables.
In order to respect the notion of equal distance between adjacent points
on the scale, the researcher has to make each category the same size. Let
us use an example. If the researcher puts the first category of age as 0-
20 years, the second should be 21-40, the third 41-60, etc. Obviously,
categories should never overlap and should follow a logical order, most
often increasing in size.
From the preceding discussion, it can clearly bee seen that there is
a hierarchy implied in the level of measurement. At lower levels of
measurement, assumptions tend to be less restrictive and data analyses
tend to be less sensitive. At each level up the hierarchy, the current level
includes all of the qualities of the one below it and adds something new.
Thus, ratio measurement is the most sophisticated, since it incorporates
all the characteristics of nominal, ordinal and interval.

Beyond Measurement: Scaling in Social Research

Scaling is the branch of measurement that involves the construction of
an instrument that associates qualitative constructs with quantitative
metric units. It was born out of efforts in education and psychology to
measure abstract or subjective constructs such as motivation, self
esteem, depression, xenophobia, etc that were traditionally not
“measurable.” Today, scaling may be considered an extension of
measurement that involves creating a continuum upon which measured
objects are located. Such a continuum is actually a scale whose range
corresponds with the range of values to be measured. Suppose a
researcher wants to measure the level of a respondent’s agreement with
a statement. The levels on the scale may range from 1 to 5 indicating the
following values: 1 = Strongly agree, 2 = Agree, 3 = Undecided, 4 =
Disagree, 5 = Strongly disagree.
In most scaling, the objects are text statements, usually statements
of attitude, opinion or belief. Scaling involves procedures that a
researcher does independent of the respondent so that he or she can
come up with a numerical value for the object. In true scaling research,
one uses a scaling procedure to develop his or her instrument (response
scale) and uses such an instrument to collect responses from participants.
194 Social Research Methodology: An Introduction

One may ask: Why do researchers do scaling at all, and not just
create text statements or questions and use response formats to collect
the answers? First, sometimes scaling is done to test a hypothesis. We
might want to know whether the construct or concept is a single
dimensional or multidimensional one (more about dimensionality later).
Sometimes, we do scaling as part of exploratory research. We want to
know what dimensions underlie a set of ratings. For instance, if you
create a set of questions, you can use scaling to determine how well they
“hang together” and whether they measure one concept or multiple
concepts. But probably the most common reason for doing scaling is for
scoring purposes. When a participant gives their responses to a set of
items, we often would like to assign a single number that represents
that's person's overall attitude or belief. For the figure above, we would
like to be able to give a single number that describes a person's attitudes
towards immigration, for example.

The Concept of Dimensionality in Scaling

According to the Weekipaedia, a dimension is a parameter or
measurement required to define the characteristics of an object—i.e.
length, width, and height or size and shape. In mathematics, dimensions
are the parameters required to describe the position and relevant
characteristics of any object within a conceptual space—where the
dimensions of a space are the total number of different parameters used
for all possible objects considered in the model. Generalizations of this
concept are possible, and different fields of study will define their spaces
by their own relevant dimensions, and use these spaces as frameworks
upon which any other study in that area is based.
In social research, a scale can have any number of dimensions in it,
although most scales have only a few dimensions. Think of a dimension
as a number line. If we want to measure a construct, we have to decide
whether the construct can be measured well with one number line or
whether it may need more. For instance, height is a concept that is
unidimensional or one-dimensional. We can measure the concept of
height very well with only a single number line (e.g., a ruler). Weight is
also unidimensional—we can measure it with a scale. Thirst might also
be considered a unidimensional concept—you are either more or less
thirsty at any given time. It is easy to see that height and weight are
unidimensional. But what about a concept like self esteem? If you think
 Measurement and Scaling Methods 195

you can measure a person's self esteem well with a single ruler that goes
from low to high, then you probably have a unidimensional construct.
What would a two-dimensional concept be? Many models of
intelligence or achievement postulate two major dimensions—
mathematical and verbal ability. In this type of two-dimensional model,
a person can be said to possess two types of achievement. Some people
will be high in verbal skills and lower in math. For others, it will be the
reverse. But, if a concept is truly two-dimensional, it is not possible to
depict a person's level on it using only a single number line. In other
words, in order to describe achievement you would need to locate a
person as a point in two dimensional (x,y) space.
And a three-dimensional concept? According to Hinton (1993), any
objects can be distinguished or differentiated from each other along
three dimensions—activity, evaluation, and potency. This general theory
of meaning is called the semantic differential, and it essentially states
that you can rate any object along those three dimensions. In other
words, in order to describe the meaning of an object you have to locate
it as a dot somewhere within the cube (three-dimensional space).

Dimensionality and Types of Scales

There are four types of unidimensional scales used in social research.
These are the Thurstone scale propounded by R.L. Thurstone in 1928,
the Likert scale by R. Likert in 1932, the Guttman scale by L. Guttman
in 1944, and the Social Distance Scale by E.S. Bogardus in 1925. The
multidimensional scales include Semantic Differentials articulated by
Charles Egerton Osgood and others in 1957, and the Diamond of
Opposites scale (Carlson-Sabelli, L. et al., 1997). Let us discuss each of
these scales in more detail.

The Thurstone Scale

Named after its inventor, the Thurstone scale consists of three related
techniques of developing a unidimensional scale. These are the
technique of equal-appearing intervals (most popular one), the technique
of successive intervals, and the techniqque of paired comparisons. These
techniques lead to more or less the same scale.
The scaling process starts with defining the focus for the scale one
is trying to develop. The researcher must ensure that the concept he or
she is arttempting to scale is clearly unidimensional. The description of
196 Social Research Methodology: An Introduction

this concept should be as clear as possible so that the respondents who

are going to create the statements have a clear idea of what is to be
measured. After the defnition of the focus, the researcher generates
potential scale items—a large pool of statements from which to select
your final scale items. The statements must take a similar grammatical
or structural pattern. For instance, one should not have a mixture of
questions and statements, or of statements of taste and those of opinion.
The third step is to rate the scale items—rating each statement on a scale
of 1 to 11 according to the degree of favourableness towards the concept
being measured (i.e. 1 = extremely unfavourable, 2 = extremely
favourable). The fourth step is to analyze the rating data by computing
the median and the interquartile range.* Finally, the researcher selects
statements that are at equal intervals across the range of medians,
preferring statements that have the smallest Interquartile Range. These
will constitute the scale for measuring the concept in question.
For example: To develop a Thurstone scale to measure people’s
attitudes towards AIDS, Trochim (2000) generated the following
statements from a class exercise he did in 1997.
1. people get AIDS by engaging in immoral behavior
2. you can get AIDS from toilet seats
3. AIDS is the wrath of God
4. anybody with AIDS is either gay or a junkie
5. AIDS is an epidemic that affects us all
6. people with AIDS are bad
7. people with AIDS are real people
8. AIDS is a cure, not a disease
9. you can get AIDS from heterosexual sex
10. people with AIDS are like my parents
11. you can get AIDS from public toilets
12. women don’t get AIDS
13. I treat everyone the same, regardless of whether or not they have AIDS
14. AIDS costs the public too much
15. AIDS is something the other guy gets
16. living with AIDS is impossible
17. children cannot catch AIDS
18. AIDS is a death sentence
19. we should focus our resources on prevention instead of curing
20. people who contract AIDS deserve it

* These and other statistical measures are discussed in Chapter Fifteen.

 Measurement and Scaling Methods 197

21. AIDS doesn't have a preference, anyone can get it

22. AIDS is the worst thing that could happen to you.
23. AIDS is good because it will help control the population.
24. if you have AIDS, you can still live a normal life.
25. people with AIDS do not need or deserve our help
26. by the time I would get sick from AIDS, there will be a cure
27. AIDS will never happen to me
28. you can't get AIDS from oral sex
29. AIDS is spread the same way colds are
30. AIDS does not discriminate
31. you can get AIDS from kissing
32. AIDS is spread through the air
33. condoms will always prevent the spread of AIDS
34. people with AIDS deserve what they got
35. if you get AIDS you will die within a year
36. bad people get AIDS and since I am a good person I will never get AIDS
37. I don't care if I get AIDS because researchers will soon find a cure for it.
38. AIDS distracts from other diseases that deserve our attention more
39. bringing AIDS into my family would be the worst thing I could do
40. very few people have AIDS, so it's unlikely that I'll ever come into contact with
a sufferer
41. if my brother caught AIDS I'd never talk to him again
42. people with AIDS deserve our understanding, but not necessarily special treatment
43. AIDS is a omnipresent, ruthless killer that lurks around dark alleys, silently
waiting for naive victims to wander passed so that it might pounce.
44. I can't get AIDS if I'm in a monogamous relationship
45. the nation's blood supply is safe
46. universal precautions are infallible
47. people with AIDS should be quarantined to protect the rest of society
48. because I don't live in a big city, the threat of AIDS is very small
49. I know enough about the spread of the disease that I would have no problem
working in a health care setting with patients with AIDS
50. the AIDS virus will not ever affect me
51. everyone affected with AIDS deserves it due to their lifestyle
52. someone with AIDS could be just like me
53. people infected with AIDS did not have safe sex
54. AIDS affects us all.
55. people with AIDS should be treated just like everybody else.
56. AIDS is a disease that anyone can get if there are not careful.
57. it is easy to get AIDS.
58. the likelihood of contracting AIDS is very low.
59. the AIDS quilt is an emotional reminder to remember those who did not deserve
to die painfully or in vain
60. the number of individuals with AIDS in Hollywood is higher than the general
public thinks
61. it is not the AIDS virus that kills people, it is complications from other illnesses
(because the immune system isn't functioning) that cause death
198 Social Research Methodology: An Introduction

62. AIDS is becoming more a problem for heterosexual women and their offsprings
than IV drug users or homosexuals
63. a cure for AIDS is on the horizon
64. mandatory HIV testing should be established for all pregnant women.

He then had ‘judges’ rate each statement on a 1-to-11 scale in terms of

how much each statement indicated a favorable attitude towards people
with AIDS. On a continuum, 1 represented an extremely unfavorable
attitude towards people with AIDS and 11 an extremely favorable
attitude towards them. For each statement, he computed the median and
the interquartile range, as summarized in the table below.

Statement No. Median Q1 Q3 Interquartile Range

23 1 1 2.5 1.5
8 1 1 2 1
12 1 1 2 1
34 1 1 2 1
39 1 1 2 1
54 1 1 2 1
56 1 1 2 1
57 1 1 2 1
18 1 1 1 0
25 1 1 1 0
51 1 1 1 0
27 2 1 5 4
45 2 1 4 3
16 2 1 3.5 2.5
42 2 1 3.5 2.5
24 2 1 3 2
44 2 2 4 2
36 2 1 2.5 1.5
43 2 1 2.5 1.5
33 3 1 5 4
48 3 1 5 4
20 3 1.5 5 3.5
28 3 1.5 5 3.5
31 3 1.5 5 3.5
19 3 1 4 3
22 3 1 4 3
37 3 1 4 3
41 3 2 5 3
6 3 1.5 4 2.5
21 3 1.5 4 2.5
 Measurement and Scaling Methods 199

32 3 2 4.5 2.5
9 3 2 3.5 1.5
1 4 3 7 4
26 4 1 5 4
47 4 1 5 4
30 4 1.5 5 3.5
13 4 2 5 3
11 4 2 4.5 2.5
15 4 3 5 2
40 5 4.5 8 3.5
2 5 4 6.5 2.5
14 5 4 6 2
17 5.5 4 8 4
49 6 5 9.75 4.75
50 8 5.5 11 5.5
35 8 6.25 10 3.75
29 9 5.5 11 5.5
38 9 5.5 10.5 5
3 9 6 10 4
55 9 7 11 4
10 10 6 10.5 4.5
7 10 7.5 11 3.5
46 10 8 11 3
5 10 8.5 11 2.5
53 11 9.5 11 1.5
4 11 10 11 1

From this summary, he selected the final statements for his scale. These
were statements at equal intervals across the range of medians, and
which had the least amount of variability across ‘judges.’ The following
were selected:

• People with AIDS are like my parents (6)

• we should focus our resources on prevention instead of curing (5)
• people with AIDS deserve what they got (1)
• AIDS affects us all (10)
• people with AIDS should be treated just like everybody else (11)
• AIDS will never happen to me (3)
• it is easy to get AIDS (5)
• AIDS doesn't have a preference, anyone can get it (9)
• AIDS is a disease that anyone can get if they are not careful (9)
• if you have AIDS, you can still lead a normal life (8)
200 Social Research Methodology: An Introduction

• AIDS is good because it helps control the population (2)

• I can't get AIDS if I'm in a monogamous relationship. (4)

To administer this scale to the respondents, he subjected the statements

to an agree-disagree measure. Here is how to interprete the results of the
attitudes of each respondent.

Statement Agree Disagree

people with AIDS are like my parents T

we should focus our resources on prevention instead T

of curing

people with AIDS deserve what they got T

AIDS affects us all T

people with AIDS should be treated just like T

everybody else

AIDS will never happen to me T

it is easy to get AIDS T

AIDS doesn't have a preference, anyone can get it T

AIDS is a disease that anyone can get if they are not T

careful

if you have AIDS, you can still lead a normal life T

AIDS is good because it helps control the population T

I can't get AIDS if I'm in a monogamous relationship T

This respondent checked eight items as Agree and four as Disagree. The
average scale values for these eight items are 6, 5, 10, 11, 9, 9, 8, and 4.
The Agree final value for this respondent is the average of these values,
which is 62/8=7.75. This means that on the 1 to 11 continuum that
measures attitudes towards persons with AIDS, this particular
respondent would fall at 7.75. Look at another respondent.
 Measurement and Scaling Methods 201

Statement Agree Disagree

people with AIDS are like my parents T

we should focus our resources on prevention instead T

of curing

people with AIDS deserve what they got T

AIDS affects us all T

people with AIDS should be treated just like T

everybody else

AIDS will never happen to me T

it is easy to get AIDS T

AIDS doesn't have a preference, anyone can get it T

AIDS is a disease that anyone can get if they are not T

careful

if you have AIDS, you can still lead a normal life T

AIDS is good because it helps control the population T

I can't get AIDS if I'm in a monogamous relationship T

For this second respondent, there are only four Agree scores: 1, 3, 2, and
4, all of which are on the negative end of the continuum. The Agree final
value for this respondent is the average of these values, which is
10/4=2.5. This shows that this respondent has a very negative attitude
towards people with AIDS.

The Likert Scale

Also named after its inventor, a Likert scale is a unidimensional scale
alternatively called a “summated instrument” scale or just “summative”
scaling. This is because the items making it up are summed to produce
a total score. A Likert scale is a widely used rating scale where
respondents are required to indicate a degree of agreement or
disagreement with each of a series of statements about the concept being
measured. Each scale item has five categories, with scale values ranging
202 Social Research Methodology: An Introduction

from -2 (strongly disagree) to +2 (strongly agree), 0 being a neutral

response [-2 -1 0 +1 +2].
The process of developing a Likert scale is much similar to that in
other scales. The first step is to define what the researcher is attempting
to measure. This is followed by generating a set of potential scale items.
The researcher may upon his or her intimate understanding of the
subject matter create the items by himself or herself, or may—as is quite
often the case—engage a number of people in the item-creation activity.
The items must be those whose “favourableness” can be rated on a 1-to-
5 response scale. The third step is to engage a group of ‘judges’ to rate
the items on a 1-to-5 rating scale where: 1 = strongly unfavorable to the
concept; 2 = somewhat unfavorable to the concept; 3 = undecided; 4 =
somewhat favorable to the concept; and 5 = strongly favorable to the
concept. (Note that these levels correspond with the -2 to +2 scale
above). The fourth step is to compute the ratings of the judges in order
to determine which items to retain for the final scale. The highly rated
statements find their way onto the final scale to administer to the
respondents in the field. Thus, the final step is to administer the scale.
A respondent's final score on the scale is the sum of their ratings for all
of the items—the reason it is sometimes called a "summated" scale.
Here is an exmple of a 10-point Likert scale for measuring people's
attitudes towards Uganda's multiparty politics. SD = Strongly Disagree;
D = Disagree; U = Undecided; A = Agree; SA = Strongly Agree.
Respondents are supposed to check the appropriate column for each
statement.

Statement S.D D U A S.A

The NRM is the most popular party

Unless opposition parties unite, none of

them will ever take power

The NRM will lose power at the next polls

Uganda is still under a one-party system

All state institutions have been politicised

If another party takes power, they will not

control the army
 Measurement and Scaling Methods 203

The military may overthrow the NRM

government

Uganda is a defacto two-party state

The NRM's cost of public administration is

abnormally big

Uganda's economy is Africa's best, thanks

to the NRM government.

The Guttman Scale

This is the third type of a unidimensional scale, variously known as
cumulative scaling or scalogram analysis. Named after its inventor like
the former two, the Guttman scale is used establish a one-dimensional
continuum for a concept to be measured. The researcher endeavours as
much as possible to come up with a list of statements, where a
respondent who agrees with any specific statement also agree with all
previous statements. If, for example, a respondent scores a five on an
eight-item cumulative scale, it should mean that he or she agreed with
the first five statements. The researcher would simply need to find a set
of items that perfectly matches this pattern, which is highly
impracticable. He or she therefore use scalogram analysis to examine
how closely a set of items corresponds with this idea of cumulativeness.
Like in the foregoing scaling methods, the process of developing
a Guttman scale begins with the definition of what the researcher is
attempting to measure. With the focus on the concept now clear, he or
she then develops a large set of statements that reflect on the concept.
Whether the researcher develops the statements himself or herself, or
whether they engage other people for the purpose, the third step is to rate
these statements on a YES or NO basis in terms of their
“favourableness” to the concept using judges. These judges are, of
course, not respondents to give their own agreement to a statement, but
are simply helping the researcher to determine how related the statement
is to the concept. The fourth step is to analyze the rated statements and
develop a cumulative scale. This is done by constructing a matrix or
table that shows the responses of all the judges on all of the items, and
then sorting and grouping the judges/statements according to the
204 Social Research Methodology: An Introduction

answering pattern. This results into the final scale that is then
administered to the actual respondents by simply asking them to check
the statements they agree with. Note that each scale item has a scale
value associated with it from the analysis in step four. A respondent’s
total score, which indicates his or her attitude towards the concept, is
obtained by summing up the scale values of every AGREE item.

The Bogardus Social Distance Scale

Created by Emory S. Bogardus in 1933, the Bogardus Social Distance
Scale is a psychometric* scale devised to empirically measure social
intercourse. Bogardus sought to measure people’s willingness to
participate in social contacts of varying degrees of closeness with
members of diverse social groups. With the help of one Robert Park, he
produced the scale that came to be widely used in social research on
prejudice, including religious prejudice. The term social distance refers
to the measure of degree of tolerance or prejudice between social
groups, including religious groups. It is measured by presenting a series
of statements that show the attitudes towards a given social group. The
scale is a cumulative one, more or less a Guttman scale, because
agreement with any item implies agreement with all preceding items. It
asks people the extent to which they would be accepting of each group.
If, for instance, a researcher wanted to find out the attitudes of
Baganda toward Banyankore. He or she will generate and ask the
Baganda respondents a series of questions about how closely they are
willing to associate with Banyankore. The questions should evoke
attitudes from the least possible contact to more intimate interaction. For
example:

Question Yes No

Are you willing to let Banyankore live in our country?

Are you willing to let Banyankore live in your village or

town?

* Psychometrics is the field of study concerned with the theory and technique
of educational and psychological measurement, which includes the measurement of
knowledge (achievement), abilities, attitudes, and personality traits.
 Measurement and Scaling Methods 205

Are you willing to let Banyankore live in your

neighborhood?

Are you willing to let Banyankore live next to you?

Are you willing to let your children play with

Banyankore children?

Would be willing to let a child of yours marry a

Munyankore?

Logically, respondents that are not willing to interact with Banyankore

at the first levels of social distance will also refuse closer contact as in
the later levels.
The process of constructing a social distance scale is similar to that
for a Likert scale, but with several modifications. One must have a
proper conceptual analysis, identify what is to measure, and write a
series of culturally appropriate and clear questions–normally 5 to 10
questions. As the preceding table indicates, the questions are arranged
in the order of narrowing social distance and increasing contact.
Social distance scales are generally scored as 1 for each YES
response and 0 for each NO. Thus, total scale scores range from 0 for
any respondent that is NOT AT ALL willing to associate with members
of a “foreign” group to any number equal to the number of items
answered as YES. The higher the scores therefore, the greater the
willingness to intercourse with–and ipso facto, the more favorable the
attitude toward--the group in question.

The Semantic Differential

Developed by Charles Egerton Osgood and others in 1957, the Semantic
Differential (SD) measures people’s reactions to stimulus words and
concepts in terms of ratings on bipolar scales that are defined with
contrasting adjectives at each end (Heise, 1970: 235). An SD is a seven-
point rating scale, whose end-points are associated with bipolar labels
that have semantic meaning. All SD scales are tri-dimentional, their
three basic dimensions of response being Evaluation, Potency, and
Activity (EPA). It is these that account for most of the co-variation in
the ratings. A typical SD survey question reads like:
206 Social Research Methodology: An Introduction

How would you rate the effectiveness of the following

contraception methods on these scales?

1. Natural cycle: Effective 3 2 1 0 1 2 3 Ineffective

2. Withdrawal: Effective 3 2 1 0 1 2 3 Ineffective
3. The coil: Effective 3 2 1 0 1 2 3 Ineffective
4. The condom: Effective 3 2 1 0 1 2 3 Ineffective
5. The femidom: Effective 3 2 1 0 1 2 3 Ineffective
6. The pill: Effective 3 2 1 0 1 2 3 Ineffective
7. The injection: Effective 3 2 1 0 1 2 3 Ineffective
8. Spermicide: Effective 3 2 1 0 1 2 3 Ineffective
9. Vasectomy: Effective 3 2 1 0 1 2 3 Ineffective
10. Abstinence: Effective 3 2 1 0 1 2 3 Ineffective

The alternative format of the SD scale is:

Effective 1 2 3 4 5 6 7 Ineffective

The format in the first example is the more common, where the 0
position is normally the equivalent of “neutral” or “not sure,” the 1
positions “slightly,” the 2 positions “quite,” and the 3 positions
“extremely.” Such scales measure the E (evaluation) dimension of the
SD, and two perspectives of the dimension are particularly measured:
the directionality of a reaction (effective versus ineffective) and its
intensity (from slight to extreme). Note that the principle is actually the
same in the alternative format, where the 4 position is taken as the
equivalent of 0 position in the first format.
The process of constructing an SD scale is guided by two basic
criteria of scale selection–relevance and factorial composition. In the
first instance, a scale is considered relevant if it meaningfully relates to
the concepts being judged, and if it can make distinctions that are
familiar. If, for example, one used a scale like sweet-sour to rate people,
the scale would be irrelevant to what is being judged or measured. If one
used friendly-unfriendly or helpful-unhelpful, it would be more relevant
and easy to use.
The second criterion of factorial composition entails a fairly
rigorous factor analysis. Note that the basic goal in an SD study is to get
independent measurements on the EPA dimensions, which factor
 Measurement and Scaling Methods 207

analyses show. Therefore, appropriate scales should have high factor

loadings on the EPA dimensions and thus be able to measure these
dimensions. Ideally, factorially pure scales are selected on the basis of
actual factor analyses--i.e., each scale having a high loading on just one
dimension, thereby giving a relatively pure measure of that dimension.
Another very important issue is one of how many scales should be
included in the final instrument. Suppose a researcher has obtained a set
of relevant scales, each of which loads on one and only one of the EPA
factors. The ideal is to have more than one scale for each dimension
since this improves the reliability of factor scores. However, research
has shown that including more than ten scales to measure a single
dimension impacts negatively on the SD’s reliability (Heise, 1970). Four
scales per dimension can normally give adequate sensitivity for most
purposes.
The SD methodology presents a number of important endearments
to researchers. First of all, its bipolar adjective scales are a simple,
economical means of obtaining data on any people's reactions to stimuli,
provided these scales are tailored to such people. Secondly, the ratings
on bipolar adjective scales tend to be correlated. Thirdly, some adjective
scales are almost pure measures of the EPA dimensions. For example,
good-bad, effective-ineffective, etc measure Evaluation; strong-weak,
powerful-powerless, tough-simple, etc measure Potency, and active-
passive, fast-slow, etc measure Activity. Properly balanced to match the
EPA dimensions, a few pure scales of this sort can help a researcher to
conveniently obtain a respondent’s overall response to something.
Fourthly, EPA measurements are appropriate when one is interested in
affective responses – the EPA formulation is a multi-variate approach to
affect measurement. It is also a generalized approach, applicable to any
concept or stimulus, and thus it permits comparisons of affective
reactions on widely disparate things. Finally, the SD has been used as a
measure of attitude in a wide variety of projects. Osgood, et al., (1957)
used the SD to assess attitude changes resulting from mass media
programs, as well as to measure attitude. The SD has since been
successfully used by other researchers to study attitude formation,
208 Social Research Methodology: An Introduction

attitudes toward organizations, attitudes toward jobs and occupations,

and attitudes toward minorities.*

The Diamond of Opposites

The Diamond of Opposites, a type of two-dimensional plot, is a
sociometric scaling method that simultaneously measures positive and
negative responses to a statement. A sociometric scale is a quantitative
scale used to measure social relationships. The Diamond of Opposites
can illuminate the presence of contradictions in processes that cannot be
detected by any single questionnaire item using a traditional format (e.g.
a Likert scale). It is commonly used in psychodrama groups.**
Unlike traditional question formats, especially the semantic
differential format where the respondent must choose a point on a one-
dimensional scale anchored by two semantically opposite terms, the
Diamond of Opposites allows the respondent to express attraction and
repulsion independently. Suppose a researcher is attempting to measure
the employees’ relationships with their employers. He or she may ask a
pair of questions as follows:
• Indicate your level of attraction on a scale of 1 (low) to 10 (high)
• Indicate your level of repulsion on a scale of 1 (low) to 10 (high).
The two variables are plotted on two orthogonal axes. Suppose a
respondent indicates his attraction on the x axix at 8, and his repulsion
on the y axis at 3. The point (x',y') [8,3] represents the current attitude
of the respondent toward his employer.

* Heise (1970) gives a very detailed account of a variety of studies that

employed the SD methodology since Osgood et al (1957).

** Psychodrama is a form of drama therapy that explores, through action, the

problems of people. It is a group working method, in which each person becomes a
therapeutic agent for others in the psychodrama group. It has strong elements of theatre,
often conducted on a stage with props.
Chapter Fourteen

Introduction to Statistical Methods

Introduction
Why statistics in social research? Whenever we apply numbers to data
in order to organize, summarize, or interpret it, we are using statistical
methods. These methods can range from somewhat simple computations
such as calculating sums and averages of distributions, to very complex
computations such as analysing interactions of variables and the effect
of such interactions within a complex data set. As a matter of fact,
quantitative research generates masses of data. A relatively small study
based on say 180 subjects will distribute 180 questionnaires. If there are
15 items on each questionnaire, there will be 2,700 items of raw data. To
make sense of this data it needs to be summarised in some way, so that
the reader has an idea of the typical values in the data, and how these
vary. To do this researchers use descriptive or summary statistics: they
describe or summarise the data, so that the reader can construct a mental
picture of the data and the people, events or objects they relate to.
Statistics may be defined as a body of concepts and methods that
deal with the collection, organization, summarization, presentation,
analysis and interpretation of data in order to draw valid conclusions that
will enable reasonable decision-making. The term may be used to denote
the data itself or the numbers derived from the data, that is, numerical
information such as averages. There are two major branches of statistics,
each with specific goals and specific formulas. These are descriptive
statistics and inferential (also known as inductive) statistics. Descriptive
statistics refers to the use and analysis of numbers to describe an entire
population, which is a known data set. Inferential statistics, on the other
hand, refers to the methods concerned with analysis, predictions or
210 Social Research Methodology: An Introduction

inferences about a set of data that is merely a sample of the population.

Both descriptive and inferential statistics are the subject of Chapter
Fifteen.

The Role of Statistics in Social Inquiry

On the whole, it can be said that statistics gives a general idea of what
happened in the past, what is happening now, and what will happen in
the future. The impact of statistics has been felt in almost all areas of
human endeavour, where planning action or evaluation of performance
or progress have more and more relied on collection and analysis of
data.
In the social world, statistics is associated with a number of
benefits. First of all, it facilitates social policy planning by enabling the
collection and analysis of data about the social problems of a given
population, the opportunities available and challenges to be overcome
before such policies can have a headway. Secondly, statistics ensures
that only the relevant information is paid attention. It is possible quite
often that governments, non-governmental and community-based
organizations or any other players in the social sector have too much
information on which to base a plan or intervention. If this is so, social
sector planners may wish to reduce the data to a single average figure
that is representative of the whole field of data, or reduce the data to a
diagrammatic form in order to ascertain the best trend.
The third benefit is that statistical methods help planners and social
service providers, through evaluation research, to check the quality of
the services. Fourthly, through time series analysis, past performance
figures are used to produce a forecast for the future. This is particularly
important in planning for ever going services such as education,
immunization, and maternal health programmes, or for disaster
preparedness. Finally, the techniques of correlation and regression help
the planners and social service providers to determine the relationship
between important variables such as cost and method of service delivery,
media campaign and behaviour change, etc.

Methods of Statistical Data Presentation

There are basically four ways of presenting statistical data. These are
text presentation, tabular presentation, diagrammatical presentation, and
graphical presentation. Let us examine each of them.
 Introduction to Statistical Methods 211

Text Presentation
This is a narrative presentation of the statistics. A researcher may report
as follows: “In an interview with sixty six clients about how satisfied
they were with the product, seventy percent expressed satisfaction,
twenty percent dissatisfaction, while ten percent said the product could
be slightly modified.” As can clearly be seen, this mode of presentation
is quite verbose and time consuming, and is not suitable for a manager
or planner who would be interested in visibly summary information.

Tabular Presentation
A table is a systematic organization of data in rows and columns. A
good table should be concise and easy to read and understand. Tables
generally have the following features:
(i) Title: Every table should have a title just above it, which should be
brief but complete enough to identify all the data in the table. The
wording of the title should include the most important
considerations in the what, where, how and when order.
(ii) Prefatory note: This is a phrase that is sometimes placed below the
title and in lower case of less prominent letters. It provides an
explanation concerning the entire table or a substantial part of it.
(iii) Source note: This is placed just below the table (as a caption)and
gives the source of the data contained in the table. If the data is
from a secondary source, the source should quote the author, title,
volume (if applicable), publisher and date of publication, and page
number. This is important for purposes of verification of figures or
obtaining of additional information by the reader.
(iv) Footnote: This is an explanation below the table and the source
note concerning individual figures of a column or row. Footnotes
referring to a row or column of figures are usually identified by
superscripted numbers such as 1,2,3, etc while those referring to
specific figures are identified by symbols such as *,$,#, etc.
(v) Stub and caption: Stubs describe data row-wise and captions
describe data column-wise.
(vi) Units of measurement: If these are not self explanatory, they should
be clearly indicated in the stub or caption. If the units of
measurement are uniform for all figures in the table, their
description could be done using a prefatory note.
212 Social Research Methodology: An Introduction

(vii) Totals: The subtotals for each separate class of data and the
grand total for all combined classes should be indicated where
appropriate.

Diagrammatical Presentation
There are two ways statistical data may be diagrammatically presented.
One is through the use of pictographs and the other by pie charts. These
are simply alternative modes of presentation, and whichever is opted for
is purely out of the researcher's preference. Let us look at each of them.

Pictograph
Pictographs are icons (small pictures) used to represent the quantity or
variable under study. Each icon is assigned a value. A researcher may,
for instance, want to present housing statistics from a given study. He or
she will present them as follows:

Year Population

2000 ----

2001 --------

2002 ----------

Key: 1:5,000

This pictograph means that each small house represents 5,000 permanent
housing units, implying that in 2000, there were 20,000 units, which
doubled in 2001 and increased to 50,000 in 2002. This method is
commonly used in geographic atlases, where icons are used on maps to
show population densities and distributions in different areas.

Pie chart
A pie chart is a circle that is divided by radial lines into sections or
subsections of different angles. The area of each section of the chart,
which is a fraction of the total sample of the study and corresponds to
equivalent degrees out of a total of 3600 of the circle, represents the
percentage of the frequencies of the responses on a given item on a
 Introduction to Statistical Methods 213

research instrument. A pie chart is thus constructed keeping in mind the

fact that the 100% of the frequencies corresponds to the 3600 of the chart
and the total sample of the study. Take our earlier example of the 66
interviewed clients. They could be presented on the pie chart as follows:

Levels of Satisfaction with the Product

According to this chart, the actual numbers and degrees corresponding

with the indicated percentages of clients at the different levels of
satisfaction with the product are as follows:

70 x 66 = 46 equivalent to 70 x 360 = 2520

100 100

20 x 66 = 13 equivalent to 20 x 360 = 720

100 100

10 x 66 = 7 equivalent to 10 x 360 = 360

100 100
214 Social Research Methodology: An Introduction

Graphical Presentation
Graphical presentation of data may be in form of line graphs (different
formats), bar graphs (different formats) or histograms.

Line graphs
A line graph is used to show continuous data; how one thing is affected
by another. In other words, it summarizes how two pieces of information
are related and how they vary depending on one another. It is clear to see
how things are going by the rises and falls a line graph shows. This kind
of graph is needed to show the effect of an independent variable on a
dependent variable. The numbers along a side of the line graph are called
the scale. Here below is an example of a line graph showing one Denis'
average anual weight in kilograms for a period of five years.

This graph shows how Denis' average weight varied from 2002 to 2006.
The weight scale runs vertically, while the time scale is on the horizontal
axis. We can see that Denis gained weight from 2002 to 2004. He was 65
kg in 2002, 70 kg in 2003, and 74 kg in 2004. He then began losing it,
and was 72 kg in 2005 and 71 kg in 2006.
 Introduction to Statistical Methods 215

Bar graph
A bar graph consists of an axis and a series of labeled horizontal or
vertical bars that show different values for each bar. The numbers along
a side of the bar graph are called the scale. Bar graphs are called simple
bar graphs if each strands singly, they are compound bar graphs if in a
group, and composite bar graphs if each bar is divided. Any of these bar
graphs can be vertical or horizontal. Below are examples of a vertical and
a horizontal simple bar graph.

Daily Fruit Sales in Mityana Central Market

This bar graph shows in kilograms the different fruits sold in an urban
market in a day. The sales for each fruit isclearly indicated by the vertical
axis. It is a simple bar graph by the fact that only two pieces of
information, the type and the amount of fruit sold, are indicated. The
same information is given in the horizontal bar graph below.

Daily Fruit Sales in Mityana Central Market

216 Social Research Methodology: An Introduction

A compound bar graph is similar to a simple bar graph, except that it

gives two or more pieces of information for each item on the vertical or
horizontal axis. The bar chart below shows in kilograms the amounts of
different fruits sold on two different days by a local market. This lets us
compare the sales of each fruit over a two-day period, not just the sales
of one fruit compared to another. We can see that the sales of all the
fruits were higher on Saturday and reduced on Sanday. The sales of

passion fruit and oranges remained the highest on both days, showing
that they are the most consumed. The reverse was true for jack fruit.

Weekend Fruit Sales in Mitayana Central Market

 Introduction to Statistical Methods 217

Histograms
Histograms are similar to simple bar graphs except that each bar
represents a range of independent variable values rather than just a single
value. What makes this different from a regular bar graph is that each bar
represents a summary of data rather than an independent value. For this
type of graph, the dependent variable is almost always a scalar scale
representing the count, or number, of how many of a sample fall within
each range of the independent variable.
The histogram below gives a summary of the population of all the
females in geograhical area X. The independent variable is age, which
has been grouped into ranges of 5 years each. The ranges for each bar
should be kept uniform (5 years in this case), with the exception possibly
being the first and/or last range.
218 Social Research Methodology: An Introduction

Age Categories of the Female Population of X

Students should be encouraged to do more reading and acquaint

themselves with what mode of data presentation is most suitable for what
kind of data.
Chapter Fifteen

Statistics: From Description to Inference

Introduction
We saw in the foregoing chapter that there are two branches of
statistics—descriptive and inferential. Descriptive statistics is the basic
branch of statistics, which is used to describe the basic features of the
data in a study. These statistics provide simple summaries about the
sample and the measures. With the use of simple graphics analysis,
descriptive statistics form the basis of virtually every quantitative
analysis of data. Descriptive statistics are thus used to present
quantitative descriptions in a manageable form.
Inferential statistics are typically distinguished from descriptive
statistics. Whereas with descriptive statistics one is simply describing
what is or what the data shows, with inferential statistics, one is trying to
reach conclusions that extend beyond the immediate data alone. Thus, we
use inferential statistics to make inferences from our data to more general
conditions. For instance, one may use inferential statistics to conclude
from the sample data how the population might vote.

Descriptive Statistics
As already hinted upon, all quantitative studies have some level of
descriptive statistics. For example, sample size, maximum and minimum
values, averages and measures of variation of the data about the average
are all quantitative aspects of the study, and they are amenable to
statistical manipulation. In many quantitative studies this is a first step,
prior to more complex inferential analysis. Social research studies
normally employ descriptive statistics to describe the various
measurements performed on the data as well as the distribution character
220 Social Research Methodology: An Introduction

of data sets. My discussion of descriptive statistics in this chapter is thus

centred on the measures and distributions of data.

Measures in Descriptive Statistics

There are two main types of measures social researchers employ in
descriptive statistics—measures of central tendency (the convergence of
scores) and measures of dispersion or variability (the spread of scores).

Measures of Central Tendency

These measures are often crudely viewed as “averages,” but they are
much more than that. A measure of central tendency refers to the sample
value around which the distribution of data is centred. A single value is
calculated and used to describe the distribution of scores. These values
are called measures of central tendency because they estimate the
tendency of scores (whatever characteristics these scores represent) to
converge or cluster. There are three measures of central tendency and
each one plays a different role in determining where the centre of the
distribution or the average score lies. These are the mean, the median,
and the mode.

The mean
The mean is actually the arithmetic mean or statistical average. To
determine the mean of a distribution, all of the scores are added together
and the sum is then divided by the number of scores. If the scores were
10, 12, 13, and 15, the mean would be (10 + 12 + 13 + 15)/4 = 50/4 =
12.5. The following symbols are used:

Ó = summation
0 = the mean
n = the number of scores
x2, x2, x3, ... xn = the individual scores.

Thus, the formula for the mean is

0 = x2, x2, x3, ... xn
n

= Gx
n
 Statistics: From Description to Inference 221

The mean is the most widely used of all measures of central tendency,
and it is used more frequently in advanced statistical procedures.
However, it is also the most susceptible to extreme scores. If for example
you had a class of 10 students and they scored as follows: 17, 28, 38, 46,
69, 74, 86, 88, 97, 98, the mean is the sum of these scores (623) divided
by the number of scores (10) = 623/10 = 62.3% A look at this mean mark
would give an impression that the students did well, but this will be not
until we notice that half the class actually failed (that is, scored below
50%).

The median
The median is another method for determining central tendency and is
the preferred method for highly skewed distributions. The median is
simply the middle score. To get the median, the scores are first arranged
in an ascending or descending order and the middle score identified. For
an even number of scores there will be two middle numbers and these are
simply added together and divided by two in order to determine the
median. For the following scores: 6, 8, 7, 12, 9, 8, 5, 13, 4, the array is:
4, 5, 6, 7, 8, 8, 9, 12, 13 and the median is 8. For the following: 6, 7, 12,
9, 8, 5, 13, the array is: 4, 5, 6, 7, 8, 9, 12, 13, the median is the average
of 7 and 8, which is 7.5.

The mode
The mode is the least used measure of central tendency. It is simply the
most frequently occurring score. For distributions that have several
peaks, the mode may be the preferred measure. There is no limit to the
number of modes in a distribution. If two scores tie as the most
frequently occurring score, the distribution would be considered bimodal.
Three would be trimodal, and all distributions with two or more modes
would be considered multimodal distributions.
It is quite interesting to note that in a perfectly normal distribution,
the mean, median, and mode are exactly the same. However, as the skew
of the distribution increases, the mean and median begin to get pulled
toward the extreme scores. The mean gets pulled the most which is why
it becomes less valid the more skewed the distribution. The median gets
pulled a little and the mode typically remains the same. You can often
tell how skewed a distribution is by the distance between these three
222 Social Research Methodology: An Introduction

measures of central tendency. We shall see these details as we discuss the

types of distributions shortly.

Measures of Dispersion
These are also called measures of variability. Dispersion or variability
refers to how spread apart the scores of the distribution are or how much
they vary from each other. These measures/statistics thus describe how
the data varies or is dispersed (spread out). Rather than showing how
data are similar, they show how data differ (their variation, spread, or
dispersion). These measures can be divided into two categories: measures
of absolute dispersion and measures of relative dispersion. Measures of
absolute dispersion are expressed in the same units as the observations
or scores under study. These measures include the range, inter-quartile
range, the variance, and the standard deviation. Measures of relative
dispersion, on the other hand, are expressed as percentages. They show
some measure of dispersion as a percentage of the average about which
it is computed. The most common of these measures is the coefficient of
variation. We shall turn to this later.

The range
The range is the distance between the highest value and the lowest in a
given data set. For ungrouped data, the range represents the difference
between the highest and lowest score in a distribution. In the following
distribution: 4,5,8,11,14,15, the range is 15 – 4 = 11. The range is rarely
used because it considers only the two extreme scores. But it may be
more useful (and is sometimes quoted this way in research reports) to
state the range as 4 to 15 (or to state these as minimum and maximum
values), so the reader is aware of the values themselves, as well as the
difference between them.
For grouped data, there are two approaches of determining the
range: the use of class limits and the use of class marks. Under the class
limits, the range is the difference between either the upper class limit of
the last class and the upper limit of the first class interval, or the lower
class limit of the last class and the lower limit of the first class interval.
Under the class marks, the range is the difference between the class mark
of the last class interval and the class mark of the first class interval.
Take an example of the following class:
 Statistics: From Description to Inference 223

20 - 25
26 - 31
32 - 37
38 - 43
44 - 49
The range = 49 - 25 = 24 (upper classes)
= 44 - 20 = 24 (lower classes)
= 46.5 - 22.5 = 24 (mid-classes or class marks).
The interpretation of the range is simply that the wider the range, the
greater the dispersion and vice versa.
One of the problems of the range is that it is susceptible to extreme
outliers. Outliers are extreme values that may be very different from the
majority of the data, and therefore quoting the range alone may be
misleading. Similarly, the range does not tell the arrangement of the
observations that fall in between the extreme values.

The interquatile range

The interquartile range measures the difference between the outermost
scores in only the middle fifty percent of the scores. In other words, to
determine the interquartile range, the score at the 25th percentile (1st
quartile) is subtracted from the score at the 75th percentile (3rd quartile),
representing the range of the middle 50 percent of scores.
To compute the quartiles, the observations are arranged in an
ascending order and the position of the quartile (Q) is located using the
formula (n+1) i
n
where n is the number of observations, and i is the quartile number. Thus
in the following observations: 441, 510, 420, 490, 438, 435, 430, 447,
445, 500, 515, the quartiles are located as follows:
Ascending arrangement: 420, 430, 435, 438, 441, 445, 447, 490, 500,
510, 515.

Q1 = (11+1) i = 1(11+1)/4 = 12/4 = 3

4

Q2 = (11+1) i = 2(11+1)/4 = 24/4 = 6

4
224 Social Research Methodology: An Introduction

Q3 = (11+1) i = 3(11+1)/4 = 36/4 = 9.

4

Therefore, Q1 = 435, Q2 = 445, Q3 = 500.

The interquartile range = Q3 - Q1 = 500 - 435 = 65. The interpretation
of the interquartile range is the same as that of the range.

The variance
The variance is the mean of the squared differences of each score from
the mean. It is denoted by ó2 for population and S2 for sample variance.
To calculate the variance, the following formulae are applicable:
For the population variance:

ó2 = Ó(xi – ì)2
N

where ì = Gxi is the population mean

N
N = number of observations in the population
xi = ith observation.

For the sample variance:

S2 = G(xi – 0)2
n–1

where n = sample size and 0 = sample mean.

The standard deviation

When the square root is taken of the variance we call this new statistic
the standard deviation (SD). Since the variance represents the squared
differences, the standard deviation represents the true differences and is
therefore easier to interpret and much more commonly used. The
standard deviation is usually quoted along with the mean. It is a measure
of the variability of the sample data around the mean. For example, in a
study comparing two groups of subjects in a trial, details are given of the
age of subjects in each group so the extent to which the groups are
similar can be determined, and are therefore comparable.
 Statistics: From Description to Inference 225

Experimental group n Control group n = 30

= 30
Mean age 70 (5) yrs 71 (10) yrs
(SD)

In this table (which is very typical of what might be seen in a research

report) it can be observed that the mean age for each group is very
similar, but the standard deviation is quite different. This might suggest
that the groups are not as similar as the mean suggests, and the greater
variability in the control group might have an influence on the findings.
Since the standard deviation relies on the mean of the distribution,
however, it is also affected by extreme scores in a skewed distribution.
It should be noted that quoting both a measure of central tendency
and the relevant measure of dispersion for one set of data always gives
a much better picture of the data being described than quoting one
measure alone.

The coefficient of variation

As I hinted earlier, this statistic shows the standard deviation as a
percentage of the mean. It is denoted as V and expressed as follows:
V = ó/ì * 100 for a population
V = S/0 * 100 for a sample.
Like the standard deviation, we consider the magnitude of the coefficient
of variation when making assertions about the dispersion of a data set.
The bigger the value of V, the greater will be the dispersion. Therefore,
given two populations, the one with the large value of V is said to exhibit
greater dispersion.

Types of Distributions
When data sets are graphed they form a picture that can aid in the
interpretation of the distribution of the observations or scores.
Distributions of scores can assume any shape from symmetrical to
asymmetrical to tell whether the distribution is normal or skewed.
Similarly, even if the shape is symmetrical, it may be sharp or have a
sharp peak as I show below.
226 Social Research Methodology: An Introduction

The Normal Distribution

The most commonly referred to type of distribution is called a normal
distribution or normal curve, and is depicted as a bell shaped curve. A
normal distribution is symmetrical, meaning that the distribution and
frequency of scores on the left side matches the distribution and
frequency of scores on the right side. The mean, median, and mode of a
normal distribution are identical and fall exactly in the centre of the
curve. Here below is a typical normal curve.

The actual class scores may appear as in the following diagram:

In the above, the three averages are the same: the mean, mode and
median as the first diagram indicates. An exception to this is the case of
a bi-modal symmetrical distribution. In this case the mean and the
median fall at the same point, while the two modes correspond to the two
highest points of the distribution. An example follows:
 Statistics: From Description to Inference 227

Usually, the normal distribution is transformed into a standard normal

distribution so that it has a mean of 0 and a standard deviation of 1.
Normal distributions can be transformed to standard normal distributions
by the formula:
z=x-ì
ó
where x is a score from the original normal distribution, ì is the mean of
the original normal distribution, and ó is the standard deviation of
original normal distribution. The standard normal distribution is
sometimes called the z distribution. A z score always reflects what
number of standard deviations above or below the mean a particular
score is. If a student scored a 70 on a test with a mean of 50 and a
standard deviation of 10, then he or she scored 2 standard deviations
above the mean. Converting the test scores to z scores, an x of 70 would
be:
z = 70 - 50 = 2
10
So, a z score of 2 means that the original score was 2 standard deviations
above the mean. Note that the z distribution will only be a normal
distribution if the original distribution (x) is normal.
The application of the above formula will always produce a
transformed variable with a mean of zero and a standard deviation of
one. However, the shape of the distribution will not be affected by the
transformation. If the score (x) is not normal then the transformed
distribution will not be normal either. One important use of the standard
normal distribution is for converting between scores from a normal
distribution and percentile ranks. Also, the shape of the curve allows for
a simple breakdown of sections. For instance, we know that 68% of the
228 Social Research Methodology: An Introduction

population fall between one and two standard deviations from the mean
and that 95% of the population fall between two standard deviations from
the mean. Study the following diagram:

Areas under portions of the standard normal distribution are shown to the
right. About .68 (.34 + .34) of the distribution is between -1 and 1 while
about .96 of the distribution is between -2 and 2.

Skewness
The skewness of a distribution is the degree of its asymmetry or
departure from the symmetry of a distribution. For such a distribution,
the mean, median and mode assume different values. Thus , such
distributions will show a pronounced tail on one side indicating that a big
fraction of the observations falls on one side of the mean. One way to
remember the order of the mean, median, and mode in a skewed
distribution is to remember that the mean is pulled in the direction of the
extreme scores. The more skewed a distribution is, the more difficult it
is to interpret.

Positive skewness
A positively skewed distribution is asymmetrical and points in the
positive direction. When a curve has extreme scores on the right hand
side of the mean, it is said to be positively skewed. In other words, most
of the observations or scores cluster around a relatively low value,
resulting into a long tail to the right of the mean (the positive direction).
In such a distribution, the mean is higher than the median and the median
higher than the mode. In a practical world, if a test was very difficult and
 Statistics: From Description to Inference 229

almost everyone in the class did very poorly on it, the resulting
distribution would most likely be positively skewed.
The mean is the balance point of the distribution. Because points
further away from the balance point change the centre of balance, the
mean is pulled in the direction the distribution is skewed. In a positively
skewed distribution, the extreme scores are larger, thus the mean is larger
than the median.

Negative skewness
A negatively skewed distribution is asymmetrical and points in the
negative direction. When a curve has extreme scores on the left hand side
of the mean, it is said to be negatively skewed. In other words, most of
the observations or scores cluster around a relatively high value, resulting
into a long tail to the left of the mean (the negative direction). In such a
distribution, the mean is lower than the median and the median also
lower than the mode. The diagram below shows the different types of
skewness.

Note that the normal curve represents no skew, meaning a normal

distribution.

Measures of skewness
There are two common measures of skewness: the Pearson coefficient of
skewness and the Moment coefficient of skewness. The Pearson
coefficient of skewness is computed as follows:

S.K = mean - mode

standard deviation
230 Social Research Methodology: An Introduction

where S.K = 0 implies symmetry; S.K > 0 implies positive skewness;

S.K < 0 implies negative skewness. A value of -0.5 # s.k # 0.5 indicates
moderate skewness. The Moment coefficient of skewness is denoted by

á3 = m3
%m23

where m3 is the third sample moment about the mean; m2 is the second
sample moment about the mean; á3 = 0 is symmetry; á3 > 0 is positive
skewness; and á3 < 0 is negative skewness.

Kurtosis
Kurtosis refers to the peakedness or flatness of a distribution. A normal
distribution or normal curve is referred to as a perfect mesokurtic
distribution. Curves that contain more scores in the centre than a normal
curve tend to have higher peaks and are referred to as leptokurtic.
Curves that have fewer scores in the centre than the normal curve and/or
more scores on the outer slopes of the curve are said to be platykurtic.

Statistical procedures are designed specifically to be used with certain

types of data, namely parametric and non-parametric. Parametric data
consists of any data set that is of the ratio or interval type and which falls
on a normally distributed curve. Non-parametric data consists of ordinal
or ratio data that may or may not fall on a normal curve. When
evaluating which statistic to use, it is important to keep this in mind.
Using a parametric test on non-parametric data can result in inaccurate
results because of the difference in the quality of this data. Remember,
 Statistics: From Description to Inference 231

in the ideal world, ratio, or at least interval data, is preferred and the tests
designed for parametric data such as this tend to be the most powerful.

Beyond Description: Inferential Statistics

Social research studies employ inferential statistics for two purposes—to
estimate what might be happening in a population, basing on a sample of
that population; and to predict what might happen in the future. To use
inferential statistics, only a sample of the population is required, unlike
in descriptive statistics where the entire population is used. The
descriptive statistical methods are of course used a great deal in
inferential statistics, but the data is taken a step further in order to enable
generalization or prediction.
As we saw in descriptive statistics, it is quite easy to determine the
mean of a known sample of subjects or scores. One simply adds up all of
the scores and divides by the number of subjects. Clearly then, the mean
of a sample is a known variable. However, it is not easy to estimate the
mean of a population that is not known or to predict the mean of a test
that has not yet been taken. Because these variables are not known, the
researcher has to rely on what is known and make assumptions.
Researchers employ inferential statistics to rely on what is known and
make estimations and predictions about what is not known.

Estimation
In estimation, a researchers uses a sample to estimate a population
parameter. A parameter is a numerical quantity or a measure of some
aspect of a population of scores. Such a measure may be the mean which
measures central tendency, standard deviation which measures
dispersion, correlation which is a measure of co-variation, or proportion
which is a measure of relative quantity/composition. Parameters are
rarely known and are usually estimated by statistics computed in
samples.
For a quicker understanding of estimation issues, it is important to
distinguish between a population parameter and a sample statistic. A
parameter is a numerical summary of a population. Note that because
populations have so many subjects, their actual numbers can never be
known unless a census is taken. However, it is feasible to select a sample
of subjects and then compute some measures (e.g. means, standard
deviations, etc) from that sample. A statistic is any numerical measure
232 Social Research Methodology: An Introduction

computed from a subset of the population—typically a sample. Thus, a

parameter describes a full population and a statistic describes a sample
of that population. Whenever a researcher uses a sample to estimate a
population parameter, he or she must construct a confidence interval
about the estimate. I shall turn to this shortly.
The following table gives examples of some common measures, the
Greek letters used to represent the corresponding parameters and the
Arabic letters used to represent the statistics used to estimate the
parameters using samples.

Measure Parameter Statistic

Mean ì M

Standard deviation ó s

Correlation ñ r

Proportion ð p

Confidence interval
Suppose the Uganda Media Centre conducted an opinion poll on 2400
registered voters during the 2006 Presidential campaigns, and the results
showed that 60% of the voters would vote for Museveni, 36% for
Besigye, and the remaining 4% for the rest of the candidates. This result
would allow a plus-minus value, say of 3% (± 3%), meaning that the
stated results could go up or down by this value. This value is called the
margin of error. Thus, Museveni could poll anywhere between 57 and
63%, and Besigye between 33 and 39%. This margin of error actually
gives a six-percentage point spread between the estimated final results
(57-62 = 6; 33-39 = 6). This spread is the confidence interval. By
definition therefore, a confidence interval is the range within which the
researcher is fairly sure that the true value of the parameter being
estimated lies.

Confidence level
If independent samples are taken repeatedly from the same population,
and a confidence interval calculated for each sample, then a certain value
of the intervals will include the unknown population parameter. This
 Statistics: From Description to Inference 233

value is called the confidence level and is calculated in percentages.

Confidence intervals are commonly calculated at a confidence level of
95%, although levels of 90% or even 99% are also used. The confidence
level is actually the probability value (1 - á) associated with a
confidence interval. For example, the pollster in the above example put
his or her á at 0.05, which is 5%. This means that the confidence level 1 -
á = 1 - 0.05 = 0.95, which is 95%.

Prediction and Hypothesis Testing

Prediction is done through the process of hypothesis testing. Statistical
tests help researchers to separate significant effects from mere luck or
random chance—in other words, can a given situation or effect be
scientifically attributable to a specific phenomenon or sheer chance? All
hypothesis tests have unavoidable, but quantifiable, risks of making the
wrong conclusion. There are in fact two types of errors involved in
predicting the cause-effect relationships of phenomena. I will turn to
these shortly. The process of hypothesis testing involves stating two
hypotheses—the null hypothesis (H0) and the alternative hypothesis
(H1)—and applying an appropriate statistical test to reject one of the
hypotheses and either accept or fail to reject the other.

Statistical significance
Literary, this concept is like practical insignificance. The concept of
statistical significance is based on the assumption that events which
occur very infrequently by chance are significant. In fact, events which
occur less than 5 out of 100 times are said to be statistically significant.
If a result of a test is statistically significant, it implies that the
probability of obtaining the observed effect or value of a parameter is
small—typically 5% or less.

Significance test
Quantitative researchers always want to determine whether an observed
value of a statistic differs enough from a hypothesized value of a
parameter in order to draw the inference that the hypothesized value is
not the true value. They thus conduct significance tests to decide whether
to reject or not to reject the hypotheses. This is done by calculating the
probability of obtaining a statistic as different or more different from the
null hypothesis (given that the null hypothesis is predicted as correct)
234 Social Research Methodology: An Introduction

than the statistic obtained in the sample. If this probability is sufficiently

low, then the difference between the parameter and the statistic is said to
be statistically significant. If the effect is statistically significant, the null
hypothesis is rejected and vice-versa.

Level of significance
According to the significance test, one may ask how low is sufficiently
low to make the effect or value statistically significant? In statistics, the
level of significance—commonly referred to as the p-value— is used.
The p-value is equal to the probability of obtaining the observed
difference, or one more extreme, if the null hypothesis is true. If the null
hypothesis is true, the p-value is the probability of obtaining a difference
at least as large as that observed due to sampling variation.
Consequently, if the p-value is large the data support the null hypothesis
and if it is small the data support the alternative hypothesis.
Conventionally, the following p-values lead to the following
corresponding decisions in accepting or rejecting the null hypothesis:
(i) for a p-value greater than 0.05 (5%), there is no evidence to reject
H0;
(ii) for a p-value between 0.01 and 0.05, there is some evidence to
reject H0 (therefore accept H1);
(iii) for p-value between 0.001 and 0.01, there is strong evidence to
reject H0 (therefore accept H1); and
(iv) for p-value less than 0.001, there is very strong evidence to reject H0
(therefore accept H1).

Note, of course, that the p-value has nothing to do with the importance
of a finding. As already seen, the symbol á (Greek letter alpha) is
oftentimes used to indicate the significance level.

Type I and II errors

There are two kinds of errors that can be made in significance testing: (1)
a true null hypothesis can be incorrectly rejected and (2) a false null
hypothesis can fail to be rejected. These are the Type I error and Type II
error respectively. The probability of a Type I error is designated by the
Greek letter á and is called the Type I error rate; while the probability of
a Type II error is designated by the Greek letter beta (ß) and is called the
Type II error rate.
 Statistics: From Description to Inference 235

A Type II error is only an error in the sense that an opportunity to

reject the null hypothesis correctly was lost. It is not an error in the sense
that an incorrect conclusion was drawn since no conclusion is drawn
when the null hypothesis is not rejected. A Type I error, on the other
hand, is an error in every sense of the word. A conclusion is drawn that
the null hypothesis is false when, in fact, it is true. Therefore, Type I
errors are generally considered more serious than Type II errors. The
probability of a Type I error (á) is called the significance level and is set
by the researcher. Its rate is almost always set at .05 or at .01, the latter
being more conservative since it requires stronger evidence to reject the
null hypothesis at the .01 level then at the .05 level.

Inferential Procedures
There are specific procedures—tests and analyses—that are used to test
hypotheses and make inferences about an unknown population parameter
or unknown score. These vary, mainly depending on the type of data
used and the purpose of making the inference. Here below I briefly
describe some of these procedures.

The Chi-square
The chi-square (denoted as ÷2) is a non-parametric test of statistical
significance for bivariate tabular analysis. It is used to determine whether
an association or relationship between two categorical variables in a
sample is likely to reflect a real association between these two variables
in the population. The sample data is used to calculate a test statistic, the
size of which reflects the probability (p-value) that the observed
association between the two variables has occurred by chance, ie due to
sampling error. The formula for the chi-square is:

÷2 = 3(O - E)2
E
where O is the observed frequency, and E is the expected frequency. The
degrees of freedom for a one-dimensional chi-square statistic is [df = C -
1], where C is the number of categories or levels of the independent
variable.
Any appropriately performed test of statistical significance enables
the researcher to know the degree of confidence he or she can have in
accepting or rejecting an hypothesis. By employing a chi-square, the
236 Social Research Methodology: An Introduction

researcher seeks to establish whether two different samples are different

enough in some characteristic or behavior aspect so that one can
generalize from those samples that the populations from which they are
drawn are also different in those behaviours or characteristics.

Student t-Test
The t-test (developed by W.S. Gossett under the guise of
“Student”–hence the name Student t-test) is a parametric test used to
assess whether the means of two groups are statistically different from
each other. It is arguably the simplest of the inferential statistics. To
compare these groups, the t-test statistical formula includes the means,
standard deviations, and other data sets for each group. Each of these sets
of data can be derived by using descriptive statistics. This test is
preferred when the samples in question are small, preferably 30 or less.
And because of this, the t-test can be computed by hand in a relatively
short amount of time.

Analysis of Variance
Analysis of Variance (ANOVA) is also a parametric test normally used
when there are one or more independent variables and two or more
dependent variables. Whereas a simple t-test would suffice when the data
have one independent and one dependant variables, in the case of data
with several variables it would mean performing one t-test for every pair
of data. If there were, say, eight pairs of data, it would take eight t-
tests–which is time consuming. The ANOVA can analyzes all eight sets
of data at once. There are two reasons that make ANOVA a superior
technique for analysis of complex relationships within the data. The first
is its ability to combine complex data into one statistical procedure. The
second is the ability to determine the interaction effects.

Analysis of Covariance
Analysis of covariance (ANCOVA) is a statistical test used for reducing
experimental error or for removing the effect of an extraneous variable.
Its main purpose is statistical control of variability when experimental
control cannot be used. Statistical control is obtained by using a covariate
along with the dependent variable. ANCOVA employs linear prediction
equations or regression to predict the values (scores and means) of the
dependent variable on the basis of the values of the covariate variable,
 Statistics: From Description to Inference 237

and then subtract these scores and means from the corresponding values
of the dependent variable. In so far as the measures of the covariate are
taken in advance of the experiment and they correlate with the measures
of the dependent variable they can be used to reduce experimental error
or control for an extraneous variable by removing the effects of the
covariate from the dependent variable.

Regression Analysis
We saw in chapter eleven how correlation is used to determine the
strength and direction of the relationship between two or more variables.
Regression analysis goes beyond correlation. A researcher creates a
regression formula based on the known data, and uses it to predict a
student’s score say on the final examination merely by knowing his or
her score on the course work. Suppose these two variables (course work
and final exam) were correlated at +1.0 or –1.0 (perfect correlations), the
prediction would be extremely accurate, and much less so if the
correlation coefficient was less perfect. In fact, the farther from a perfect
correlation, the less accurate the results of the prediction. When the data
set is much larger and the correlation less than perfect, prediction
requires the use of the statistical regression, which is basically a
geometric formula used to determine where a score falls on a straight
line.
 Part Four

Research Planning and Execution

Chapter Sixteen

The Research Proposal

Introduction
Before starting the research one should have at least some idea of what
he or she wants to do. This is the most difficult yet the most vital part of
the process of proposal writing. A researcher should identify the broad
problem area that is very closely related to his or her professional goals
(area of in-depth specialization e.g., constitutionalism, management,
human rights, curriculum, rural development, etc). This broad problem
area is called the research topic. Broad research topics may read as
follows: “Constitutionalism in post 1995 Uganda;” “Advances in
Information Technology;” “Pedagogical Issues in the Information Age;”
“Corporate Responsibility in a Globalized Economy;” etc. To arrive at
a suitable research problem, it is important that a person proposing the
research systematically moves from the broad to the specific aspect of the
problem. On narrowing down, the more specific and more manageable
research problem may read as follows: “Constitutional review in Uganda:
the legal implication of the Ssempebwa Commission;”
“Constitutionalism at crossroads: an analysis of the origin and rationale
of the “Third Term” questions in Uganda;” “People’s attitudes towards
a referendum on political systems in Uganda: a case study of Kampala
district, etc.”
In conceptualising a topic, one should consider the following
important aspects: the subject or problem area to be studied, the issues
or arguments involved in the study, the procedures or methodology
issues, and the sources or the literature available. Whatever topic selected
and the title of the project finally arrived at, its suitability is generally
determined by a number of even more important considerations. These
 The Research Proposal 241

include the relevance of the research to one’s future career, whether the
research topic will sustain the researcher’s interest, the existing literature
on the problem, past studies, the availability of jobs, the feasibility of the
methodology, the ease of data gathering, and access to data sources.

Sources of Research Topics

The are a number of sources where a researcher can turn for a
researchable topic. The most common of these, which I highlight below,
are personal experiences, deduction from existing theories, literature
sources, practical issues in everyday work, deductive reasoning, and
inductive reasoning.

Personal experiences
Among the most fruitful sources for novice researchers are their own
experiences and observations as professional practitioners. Sometimes,
the researcher may get challenges and insights through seminars or
lectures, especially regarding problem areas where knowledge is not yet
adequate. Also, discussion with peers may reveal areas where there is a
gap in knowledge that could be investigated.

Deduction from exiting theories

In one’s reading of textbooks, research reports and other literature, and
attending lectures and workshops, one is likely to grasp theories that are
relevant to specific topics. These topics could have been formulated in
a different cultural setting and so the researcher may question the extent
to which they can apply in another setting, and may thus be interested to
replicate the study.

Literature sources
The existing literature is a very important source where the researcher
may be intrigued by the authenticity of research procedures or findings,
provoked by the novelty of the subject, or simply encouraged by a
recommendation for further research on a given aspect of the problem.

Deductive reasoning
Deductive reasoning refers to the process of concluding that something
must be true because it is a special case of a general principle that is
known to be true. Deductive reasoning works from the more general to
242 Social Research Methodology: An Introduction

the more specific. Sometimes this is informally called a “top-down”

approach. One might begin with thinking up a theory about one’s topic
of interest, and may then narrow that theory down into more specific
hypotheses that can be tested. These are narrowed down even further
when one collects data to address the hypotheses. This ultimately leads
that researcher to be able to test the hypotheses with specific data in
order to confirm or refute the original theories.

Inductive reasoning.
Inductive reasoning works the other way, moving from specific
observations to broader generalizations and theories. Inductive reasoning
is the process of reasoning that a general principle is true because the
special cases under it that have been seen are true. Informally, this is
sometimes called a “bottom up” approach. In inductive reasoning, one
begins with specific observations and measures, then proceeds to detect
patterns and regularities, formulates some tentative hypotheses that can
be explored, and finally ends up developing some general conclusions or
theories.
Inductive reasoning, by its very nature, is more open-ended and
exploratory, especially at the beginning. Deductive reasoning is more
narrow in nature and is concerned with testing or confirming hypotheses.
Even though a particular study may look like it is purely deductive (e.g.,
an experiment that is designed to test the hypothesized effects of some
treatment on a given outcome), most social research involves both
inductive and deductive reasoning processes at some time in the project.

Why Write a Research Proposal?

A research proposal is a document giving account of what is going to be
investigated and a detailed plan of how the investigation is going to be
conducted. Even where a researcher has clearly determined the problem
he or she intends to investigate, and regardless of the research experience
one may have in that subject area, it is still absolutely necessary to first
write and present a research proposal. There are several functions of a
proposal, as I outline below.
The first and foremost function of a research proposal is to persuade
a panel or committee of scholars or sponsors that the project meets the
three kinds of merit all disciplines value, namely, conceptual innovation,
methodological rigour, and a rich, substantive content. The panel needs
 The Research Proposal 243

to satisfy itself that the research project is worth the time and resources,
and not merely reinventing the wheel.
Secondly, the proposal serves the researcher as a guiding framework
or checklist of the activities and tasks to be accomplished and the
procedures through which to accomplish them. In this way, the
researcher is sure to remain on course, and not stray into unknown
terrain.
Thirdly, because the research proposal lays down in a systematic
manner the problem, the required data, and the procedures of
investigation, it helps the supervisors or promoters to ascertain whether
the procedures, techniques and tests fit the problem. A mismatch of
methods to the problem will certainly lead to a flawed project.
Fourthly, the proposal must indicate the research team leader, the
team members and all their profiles and competencies. This will help the
research committee, hosting institution or the funding agency whether
the research team has the competencies that it takes to accomplish the
project.
Fifthly, the scope and magnitude of the project, the time frame and
the possibility that the budget will be supported are important elements
of the proposal that assure the supervisors of the likelihood that the
project will be accomplished or abandoned along the way.
Sixthly, the proposal helps the funders and the authorising agency
(in the case of Uganda, NCST) to gauge where in the wider research
programme the proposed research fits, whether it addresses priority
areas, and therefore whether it is worth channelling resources to.
Finally, the proposal helps the Ethics Review Board or its equivalent
organ to gauge whether the proposed research meets the standards for
ethics approval.

Elements of a Research Proposal

It should be noted that while the form and the organization of a proposal
may be matters of taste, largely varying from individual to individual and
institution to institution, it is important to consider a typical research
proposal. Remember that most proposals are reviewed by panels or
committees of multidisciplinary experts. A reviewer of a proposal from
another field expects to meet the proposer halfway. As much as possible,
jargon should be avoided; and when technical language is really needed,
it should be restricted to those terms that truly lack equivalents in
244 Social Research Methodology: An Introduction

common language. When additional technical material is needed, or

when the argument refers to complex ancillary material, putting it into
appendices decongests the main text. A typical proposal thus has the
elements I outline below.

The title of the project

This is the distinctive name given to the research proposal, which
describes the work scope in a specific, clear and concise manner. The
title should indicate the major variables of the study and the population
on which the study will be conducted. It should accurately reflect the
scope and content of the study in about 16 words. It should be descriptive
yet discrete.

The abstract
Every proposal, even brief ones, should have an abstract. An abstract,
normally 200 to 300 words, provides a quick concise overview of the
proposal, and should appear on a page by itself. The abstract speaks for
the proposal when it becomes separated from it, provides the reader with
the first impression of the intention, and, by acting as its summary,
frequently provides the reader also with the last impression. The major
objectives and questions of the project and the procedures to be followed
in meeting these objectives and answering these questions should be
mentioned. The abstract is thus a very important element in the proposal.

The table of contents

Very brief proposals with few sections ordinarily do not need a table of
contents, bus a proposal of 10 pages and above should have it. The
guiding consideration in this is the reader’s convenience. The table of
contents should list all the major parts and divisions (including the
abstract, even though it precedes it). Very detailed subdivisions usually
need not be listed.

Background to the study

This is the rationale of the study providing evidence and conditions of the
existing situations to make the reader feel the urgency of the problem and
the need to study it in order to solve it or contribute to its solution. The
background is presented logically and systematically from the broad
context to the specific, hinging on the available literature and statistics
 The Research Proposal 245

as sources of evidence of the prevalence of the problem. By the end of

the background, the problem should have naturally become apparent and
the researcher needs only to derive it and state it succinctly in his or her
own words. Please note that without providing reasonable evidence of
the problem in this section, when it is finally stated, no one will believe
it. One should therefore strive to reference a number of sources in this
section. This section may be three to four pages long.

Statement of the problem

This refers to the problem, whether general or specific, that has been
detected and needs a solution in the practical or theoretical world. A
research problem is a kind of discrepancy between what ought to be and
what is or seems to be. This discrepancy constitutes a research gap. The
researcher should clearly state the nature of the problem and its known
or estimated magnitude or extent. Sometimes, the problem may be
expressed in a form of a broad question. The statement of the problem
should be concise and brief, in the researcher’s own words, normally one
to two paragraphs.

Purpose and objectives of the study

The purpose is the overall aim or general objective of the study. It always
reads like a restatement of the title of the research, except that it gives
more detail than the title. Because of its general nature, the purpose
normally has to be broken down into specific, more manageable and
achievable objectives. The objectives are normally enumerated, each
objective addressing a specific relationship between given variables. The
statement of objectives is very crucial—they determine the actual data
required from the field, and therefore the methods and instruments to be
used in the research.

Hypotheses/research questions
Hypotheses are investigative assumptions or propositions that guide the
study. They should be stated in testable form, so that the data collected
are used to test them. Research questions, on the other hand, are specific
and pertinent questions that may be considered in determining the
variables and delimiting the scope of the study. They are normally
specific objectives restated in interrogative form. Note that hypotheses
are normally used in quantitative research and research questions in
246 Social Research Methodology: An Introduction

qualitative research (and often in descriptive quantitative studies). They

may be brought immediately after the objectives or after the literature
review.

Theoretical/conceptual framework
A theoretical framework is an examination of existing or self-formulated
theories in relation to the researcher’s objectives. After the analysis, the
researcher chooses one theory he or she thinks most favourably explains
the problem under study. This will be revisited after data analysis to
ascertain whether it sufficiently explains the problem, or whether there
is an alternative explanation. A conceptual framework
is a scheme of concepts (variables) and their relationships which a
researcher will operationalise in the study in order to achieve the set
objectives. A conceptual framework can be presented diagrammatically.
Only one of the two may be used, and it is mandatory in proposals for
higher degrees.

Scope of the study

This refers to the inclusive frame of reference as well as the limits or
boundaries of the study. The scope may be conceptual (subject
conceptualization), geographical (area coverage), temporal (period
covered), as well as theoretical (in case of many theories, only one is
selected to guide the study).

Justification/significance of the study

Justification for the research is a kind of extension of the discussion
presented in the background. It focuses on further evidence of the
problem and the need for such a study in order to solve such a problem.
It is thus presented from the point of view of the problem, and may be
brought immediately after the statement of the problem. Significance of
the study, on the other hand, refers to the relevance of the study in terms
of academic contributions and practical use that might be made of the
findings. It focuses on the specific benefits of the findings to specific
publics, i.e., which beneficiary for what form of benefit. The importance
is projected in terms of knowledge creation, policies (community,
national, etc), emerging realities, and further research.
 The Research Proposal 247

Literature review
A literature review is an analysis of the existing body of literature related
to the study being proposed. It is a discussion on how the research
proposal is related with the current researches in the field. Literature
reviews should be selective and critical. Especially pertinent works and
their analysis are key to an effective review. Discussion of research work
accomplished by others should therefore lead the reader to a clear
impression of how the proposal is building upon what has already been
done, and how the proposed study differs from them. Literature reviews
should focus on problem areas previously researched, and the focus
should be in terms of the purpose of those studies, the methods used, the
theories, and the findings. The thrust is to identify the relationship with
the proposed research and the gaps to be filled.

Research methodology
Methodology is one of the most important parts of the proposal. It refers
to the detailed technical or scientific activities, tools and procedures
taken to plan, gather, and analyse data. The proposal must specify the
research operations one intends to undertake and the way one intends to
interpret the results of these operations in terms of the central problem.
A methodology is not just a list of research tasks but a justified narrative
as to why these tasks and procedures add up to the best attack on the
problem.
The process of gathering data and moving from data to
interpretation tends to follow disciplinary customs, which are more
standard in some fields than in others. It is thus important to show which
parts of one’s methodology are standard, and which are innovative. One
should be as specific as possible about the activities one plans to
undertake to collect information, about the techniques to be used to
analyze it, and about the measures to take for data quality control. Most
proposals fail because they leave reviewers wondering what the applicant
will actually do. Specify the archives, the sources, the respondents, and
the proposed techniques of analysis.
The statement of methodology should, to a reasonable degree of
detail, address the following issues:
i) Research design. This describes the nature and pattern the research
intends to follow e.g whether it is qualitative or quantitative;
whether it is a historical study, case study, descriptive survey,
248 Social Research Methodology: An Introduction

experimental or quasi experimental, etc.; and whether it is

comparative, analytical, etc.
ii) Area of the study. This is a description of the geographical area
where population of the study exists.
iii) Population and sampling. This describes the population of the study
from which samples will be selected. It also describes the sampling
strategies, by which the researcher will select representative
elements or subjects from the population.
iv) Data collection. This is the description of the methods, instruments,
tools and procedures to be used in the research.
v) Research procedure. This is a narrative of the entire process of
research from clearance and authorization to self presentation to the
various local authorities and research subjects.
vi) Data analysis. This is the description of the methods, procedures,
techniques, formulas and software to be used during data
organization, analysis and interpretation.
vii) Data quality control. These are the measures the researcher takes to
ensure that the quality of the research is not compromised. They
include triangulation of methods, pretesting the instruments, and
validity and reliability tests.
viii) Ethical considerations. These consist in the moral justification of
the investigation or intervention. They include issues such as abuse
of the process, consent, confidentiality of information, the safety,
social and psychological well-being of the subjects and/or
community, etc.
ix) Limitations (and delimitations). This refers to anticipated
constraints imposed by the methods, location, or other
circumstances of research, and how the researcher envisages to
contain them. They are thus important considerations before
embarking on a study to ascertain the feasibility of accomplishment
of the study.

Definitions
Researchers do not always adopt uniform definitions, so key and
controversial terms should be defined in order to indicate the positions
taken in a specific research. The term being defined should be
highlighted (in italics or in bold), and the format for presenting each of
the definitions should be standard. Definitions should match the
 The Research Proposal 249

underlying assumptions of the research and researchers may need to

justify some of their definitions. Therefore, researchers should strive to
use definitions of authorities wherever possible, so that the results of the
research can be fitted into the body of literature and so that the report
(thesis or dissertation) can survive attacks by examiners with trivial
personal preferences.

Summary of chapters
Each chapter is briefly described in terms of its general content. Some
chapters follow the standard structure of a thesis/dissertation, while
others are crafted on the specific objectives (See Chapter Seventeen).

Working bibliography
This is list of all works cited in the proposal and it should be written
according to the approved format (See Chapter Seventeen). It is placed
at the end of the text proper and just before the appendices. The style of
bibliographical item itself depends upon the disciplinary field. The main
consideration is consistency: whatever style is chosen should be followed
scrupulously throughout.

Appendices
These include all items that are not considered the substantive part of the
proposal, but which are very important reference items. They include the
following:
i) Data collection instruments. These include all the tools to be used
in the study: questionnaires, interview schedules, interview guides,
observation guides, etc.
ii) Data analysis tools. These include largely formulae and tests to be
employed during data analysis.
iii) Time schedule. This is a detailed description in chronological order
of each activity to be undertaken in the conduct of the study, and the
length of time it is to take. The starting date and planned completion
date are indicated in year and month. The time schedule may be
presented by the use of software tools such as Microsoft Project, the
Gantt chart (found in most Management textbooks), or in a simple
tabular form. This schedule is a valuable guide for evaluation of the
research progress.
250 Social Research Methodology: An Introduction

iv) Budget proposal. This is the financial plan for implementation of

the research. It is a detailed/itemized breakdown of the total
research project costs and the sources of funds. It should be clear,
realistic and reasonable(affordable). This section is only relevant if
external funding has been or is to be provided, in which case a
detailed disclosure of granted and/or anticipated funding is
expected. Otherwise, if the costs are to be borne by the researcher
or covered by voluntary labour, then no budget may be required
other than a simple statement. The budget should be itemised
according to the following: equipment; stationery; travel; research
assistance; services (secretarial, photocopying, printing, binding
etc); research honorarium (about 30%); institutional administration
fees (about 15%), etc.
v) Explanatory notes. These include research approval letters, maps
and lists of areas to be visited.
vi) Research assistance. This specifies the number of staff needed to
rationalize the proposed schedule of activities and to conduct the
study.

The length of the proposal may vary from institution to institution.

At Makerere University, for example, research proposals for Masters
degrees are up to 15 pages including references but excluding
appendices, and for Ph.D up to 25 pages.
Chapter Seventeen

The Research Report: Thesis/Dissertation

Introduction
A research report is a detailed account of the study, which in academic
settings is commonly known as a thesis or dissertation. The words
“thesis” and “dissertation” are most commonly—and erroneously— used
interchangeably. However, a keen usage of the terms takes a thesis to be
a research report for a Masters degree and a dissertation for PhD
research. In this chapter, I am concerned neither with the semantics nor
the internal structure of the contents of an academic research report, but
with the physical format and content of theses and dissertations. Of
course this chapter is as much relevant to undergraduate students doing
or intending to do research and write their reports before they graduate.
Their works are variously referred to as research reports, extended
essays, student projects, etc; and their format more or less conforms to
that of theses and dissertations, but in some cases tends to
differ—sometimes significantly.
Theses and dissertation vary in size (number of words or pages)
from University to University. (I am using the term “University” in a
loose and generic sense to include all colleges, institutes, schools and
polytechnics that have University status, i.e. can design their
programmes, teach and award degrees). Some Universities restrict the
number of words or pages, others do not. Most Universities also
prescribe the fonts and font sizes to use, the margins to set, the citation
system, and other fine details. It is important that students acquaint
themselves with these regulations and guidelines as laid down in the
research manuals.
252 Social Research Methodology: An Introduction

In terms of format, theses and dissertations generally fall into two

categories—the structured format, and the integrated format. I will deal
with these in detail under the section on the substantive text of
theses/dissertation. What is important to point out here is that whether
structured or integrated, a thesis or dissertation comprises of four distinct
divisions—the Preliminaries, the Substantive Text, the Reference
List/Bibliography, and the Appendices.

The Preliminaries
The preliminaries are the items preceding the main chapters of the work.
They are normally numbered in Roman numerals, and include the title
page (not numbered), declaration, dedication, acknowledgment, table of
contents, list of tables, list of illustrations, abbreviations and acronyms,
and the abstract. Each item should be indicated on a separate page.

The Title Page

The title page indicates the title of the study, the full names of the
researcher and the qualifications already attained and their awarding
Universities. It also indicates where the thesis/dissertation is submitted,
the purpose of submission, and the date of submission. It should be noted
that different Universities have different templates, which students must
do well to adhere to. The title page is not page numbered.

Dedication
This item indicates the name(s) to whom the research work is dedicated,
and it is optional.

Declaration
This is a statement by the candidate that his or her work is a result of an
original study, and has not been published and/or submitted for any other
degree award to any other university before. The declaration must be
signed by both the researcher and the supervisors and the dates indicated.

Acknowledgment
This is a statement of recognition of all the people and institutions that
supported the research one way or the other. They normally include
supervisors, colleagues who critiqued, individuals who offered this or
that, sponsors and hosting institutions, etc.
 The Research Report: Thesis/Dissertation 253

Table of Contents
This is the list of headings and subheadings starting with the declaration
and ending with the last appendix.

List of Tables
This is a systematic listing of all the tables used in the thesis/dissertation
and the corresponding pages. Tables are a useful medium of presenting
statistical data for analysis, and should thus not be distorted. They are
listed by their headings, which are themselves numbered.

List of Illustrations
This is a systematic listing of the figures and/or plates used in the
thesis/dissertation and their corresponding page numbers. These are
listed by their captions, which are also numbered.

List of Abbreviations and Acronyms

All the abbreviations and acronyms (abbreviations that sound like names,
e.g. UNICEF) used in the thesis/dissertation should be listed, indicating
what they are in full. It is important to note that the abbreviated words or
acronym should first be written in full when first used in the report, and
the abbreviation or acronym placed in parentheses immediately after.
Thereafter, they are used directly.

Abstract
The abstract is sometimes indicated as the Executive Summary. It entails
a brief summary of the report, indicating the problem, objectives,
methodology, findings and recommendations. Abstracts permit readers
to get a quick overview of the research in order to decide whether to read
the thesis/dissertation or not. An abstract should be anywhere between
120 and 300 words.

The Substantive Text of a Thesis/Dissertation

This text comprises the main chapters of the research work, and is page-
numbered in Arabic numerals starting from 1. There are normally two
formats of theses/dissertations—the structured format and the integrated
format.
254 Social Research Methodology: An Introduction

The Structured Thesis/Dissertation

Structured theses or dissertations are divided into five and sometimes six
chapters. These are 1) introduction, 2) literature review, 3) methodology,
4) presentation and discussion of findings, and 5) conclusions and
recommendations. Sometimes the scope and magnitude of the study may
lead to the generation of a large amount of data, which calls for a
decision to present findings as chapter 4, and discussion of findings as
chapter 5. I shall now highlight the contents of each chapter, then turn to
the integrated theses and dissertations.

Chapter One: Introduction

This chapter presents an overview of the thesis/dissertation, highlighting
the background, statement of the problem, purpose/aim of the study,
specific objectives, hypotheses/research questions, scope,
significance/justification, brief methodology, major findings and
recommendations. Most of these items were discussed in chapter sixteen.

Chapter Two: Literature Review

The literature reviewed in the proposal should be reexamined to see
whether it should be discarded, modified, or supplemented. Any relevant
literature that has been identified since the writing of the proposal should
be reviewed and integrated. This chapter therefore builds on the literature
review presented in the research proposal to expand as well as add
details. Remember that while any relevant literature may be cited in any
of the chapters, only related literature on on-going or concluded research
is reviewed. In some cases where there is scant or no research on the
problem under study, a general literature review and theoretical analysis
are done.

Chapter Three: Research Methodology

This chapter describes in details the methodology that was used in the
study. If the methodology used in the research differs from the originally
proposed one, the proposed one should be modified to what was actually
used. Remember that the language of the proposal dwelt on future simple
tense but that of the research report has to dwell on past simple tense.
Remember too, that you need to append the data collection instruments,
even formulas or statistical tests you used during research.
 The Research Report: Thesis/Dissertation 255

Chapter Four: Presentation and Discussion of Results

This is the most important chapter of the thesis or dissertation because it
presents the researcher’s contribution. While the contents of the first
three chapters are what is called secondary data, the contents of this
chapter are primary data because it is data that is directly collected from
the respondents. The presentation of data should be systematic, following
the sequence of the objectives or research questions, and indeed of the
research instruments. The findings can be presented togther with the
discussion as chapter four, but if the quantity of data is enormous, the
discussion of results should be separated and presented as chapter five.
The discussion of the findings is extremely important—the findings are
compared to those in previous studies, and positions are taken and
conclusions drawn accordingly.

Chapter Five: Summary, Conclusions and Recommendations

This chapter presents the summary of the major findings, the conclusions
derived from such findings and appropriate recommendations and the
way forward. The recommendations should focus on specific actions
related with the conclusions, administrative and/or policy issues, as well
as further research.

The Integrated Thesis/Dissertation

Under this format, the number of chapters a thesis or dissertation has is
determined by the number of objectives or research questions that were
addressed in the study. Unlike in the structured format where all the
findings are presented in one chapter, here the findings of each objective
are reported and discussed in a separate chapter.
Like in the structured format, chapter one of the integrated thesis or
dissertation presents the background to the problem, statement of the
problem, purpose/aim of the study, specific objectives,
hypotheses/research questions, scope, and significance/justification for
the study. Unlike in the structured thesis/dissertation however, this
chapter also includes the literature review and methodology, and an
overview of the contents of the subsequent chapters. Basically, the
research proposal is refined and turned into this chapter.
The subsequent chapters are based on the specific objectives of the
study, which normally should not exceed five. There should be no such
objective as “to make recommendations,” “to propose the way forward,”
256 Social Research Methodology: An Introduction

etc, because this is already implied in the problem and is the subject of
the last chapter. The major attraction to this approach is that each chapter
is independent enough and may easily be published as a scholarly paper.

The Bibliography/Reference List

The bibliography or reference list is displayed at the end of the
substantive text of the thesis or dissertation, and provides the information
necessary to identify and retrieve each source. Unlike a bibliography
which includes all relevant works, a reference list includes only the
works that were cited in the text. There are different modes of citation
and presentation of the reference list. Whichever adopted should be
strictly followed to ensure consistency and ease of reference. A detailed
discussion of modes of citation is made in the next chapter.

The Appendices
These comprise of instruments, maps, introductory letters and some
research approval letters as well as areas visited.
Chapter Eighteen

Scholarly Writing and Citation Methods

Introduction
What is the ultimate end of a research project, whether it is qualitative,
quantitative or one that combines aspects of both?; and whether it is
field-based or purely desk research? The findings must be published. For
commissioned research such as a commission of enquiry, action or
market research, the findings are published in a report in conformity with
the commissioning body’s recommended style. For research leading to
an academic award, the findings are published as a thesis or dissertation
in conformity with the format or style of the respective college or
university. For scholarly or intellectual research, the findings are reported
in a scholarly paper, normally intended for publication in a specialized
scholarly journal. Thus, every kind of writing is governed by a specific
format or stylistic regime. This addresses not only the general purpose of
writing and the audience the writing is targeting, but also the features that
define a specific kind of writing.

Defining Scholarly Writing

How does a scholarly paper differ from other papers such as those
presented at a business forum, strategic planning meeting, or articles
published in newspapers and popular magazines? What constitutes
scholarly writing? There are basically four features that differentiate
scholarly writing from other types of writing (generally referred to as
popular writing). The first one is the content of the work, which is
characterized by critical thinking about a subject, building on existing
knowledge base, synthesis and integration of existing knowledge in order
to form new insights and draw new conclusions, and an attempt to apply
to a real life situation. The second regards the format of the paper.
258 Social Research Methodology: An Introduction

Scholarly writing must conform to a style of presentation that is

acceptable among the scholarly community. The third area is the sources
of information. A scholarly paper is built on a solid foundation of
scholarly references. The fourth one is the writing process itself. Writing
a scholarly paper is an iterative process: it involves drafting and self-
assessment, peer review, and revision and reassessment. Let me return to
these features.

Content of a Scholarly Paper

The content of the paper is arguably the most important factor in
determining if it meets the scholarly requirements. Rossman (1995:69)
identified five elements of a good scholarly paper that are embedded in
the four aforesaid components of content. These are comparison to
determine similarities, contrast to determine differences, analysis to
determine relationships, synthesis to combine parts to make something
new, and integration to unite the parts to form a new whole. To fully
appreciate these elements, let us revisit the components of content.
Critical thinking about the social world and outstanding social
problems, and the analysis of the existing knowledge base about these
issues is the roadmap to building on that base. A scholar curves an
argument or builds a discourse against a foundation of previous work.
This is done by scanning what is already known about a subject or issue,
and identifying the common themes in the existing literature. Then an
analysis (comparisons and contrasts) is done of the various contributions
to the discourse in order for the scholar to take a guarded position, offer
new insights and draw veritable conclusions.
Synthesis and integration are yet another important element about
the content of a scholarly paper. The scholar always seeks to look at
issues and problems in relation to other disciplines and wider contexts.
Boyer (1990:18) has coined the concept “scholarship of integration” to
explain the scholar’s role of synthesizing and integrating isolated facts
and putting them in perspective. He discusses integration to involve
making connections across the disciplines, placing the specialties in
larger contexts, illuminating data in a revealing way, and often educating
 Scholarly Writing and Citation Methods 259
*
non-specialists. Note that some writers tend to include in the paper
numerous lengthy direct quotes; and sometimes these quotes are longer
than the writer’s own contribution to the discussion. The key to limiting
the content and length of quoted material is to synthesize and summarize
the material and concepts from the sources. And the ultimate question:
how and where will this new knowledge be used? A scholarly piece
attempts to build bridges between theory and application. Unless the
conclusions made in the paper can be translated into practical
information that can be implemented to help a situation in the real world,
they are not worth making.

Format and Style

Format and style are critical characteristics of the scholarly paper.
Standard formatting makes it easier for the intellectual community
(scholars in the same or allied field) to follow and understand the
scholar’s thoughts and the basis for them. All academic and professional
disciplines have specific systems of citation, which provide sufficient
information to allow a reader to locate the source of a quotation or
reference. Most citation styles share a common two-part structure: (1) an
indication in the text that acknowledges another scholar’s words, facts,
and ideas; and (2) a corresponding full source of information on the
reference list. There are three approaches to the in-the-text indication of
sources:
i) footnotes—using a symbol, digit or letter in the text and the same
symbol appears at the bottom of the page indicating the source of
information and its bibliographic description.
ii) endnotes—using a symbol, digit or letter in the text and the same
symbol appears at the end of the paper indicating the source of
information and its bibliographic description.
iii) indication of the source of information in parentheses, and a
corresponding full bibliographic citation on the reference list at the
end of the paper.

* The other areas of scholarship he discusses are scholarship of discovery,

scholarship of application, and scholarship of teaching.
260 Social Research Methodology: An Introduction

There are a number of systems of citation of sources, some of which

are applied to specific disciplines while others are more general. Here
below are some citation systems.
i) Modern Language Association (MLA), which guides citation and
referencing in language research. Reference can be made to MLA
Handbook for Writers of Research Papers, 6th ed. (2003).
ii) The American Psychological Association (APA) that is used in the
fine arts (art, art history, dance, music), classical studies, history,
philosophy, social sciences, library and information studies,
physical and natural sciences (astronomy, biology, chemistry,
computer science, geology, physics, etc.). Reference can be made
to Publication Manual of the American Psychological Association,
5th ed. (2001). This system is the most widely used across
disciplines.
iii) The Uniform System of Citation (USC), which is the standard
system used in legal journals. Reference can be made to The
Bluebook: A Uniform System of Citation, 17th ed. (2000).
iv) The Chicago system that is used in humanities, social sciences, as
well as scientific and technical writing. Reference can be made to
The Chicago Manual of Style, 15th ed. (2003).
v) There is a style manual published by the Council of Biology Editors
(CBE), which is widely used not only in the life sciences and
medicine but also in many engineering disciplines. Reference can
be made to Scientific Style and Format: The CBE Manual for
Authors, Editors, and Publishers, 6th ed. (1994).
vi) The Institute of Electrical and Electronics Engineers (IEEE)
publishes a short guide, Information for IEEE Transactions,
Journals, and Letters Authors (1996) for citation and referencing in
that area.
vii) For electronic resources, reference can be made to Electronic Styles:
A Handbook for Citing Electronic Information, 2nd ed. (1996); and
The Columbia Guide to Online Style (1998).
The discussion in this chapter is based on the APA style, this being the
most commonly used across disciplines.

Scholarly References
In order to build a solid bridge that others can walk across you will need
to start with a solid foundation of scholarly references. It is important to
 Scholarly Writing and Citation Methods 261

differentiate between practitioner writing and scholarly writing, that is,

the nature of the underlying knowledge that is being built upon—what
Atkinson (1992:52, cited in Latham, 2002) refers to as “facts amidst
appearance.” One of the most commonly accepted definitions of
scholarly references is that they are articles from refereed journals and
research studies including dissertations (Latham, 2002). Although
practitioner writing is often based on facts, it is also often full of
appearances and anecdotal rather than theoretical evidence. Let me
illustrate this with an example of well-known magazines. Magazines
such as Parenting, City Beat, Monitor Gossip or True Love are not
considered scholarly references, but The Administrator, The Banker or
The Courier are. This is not to say, however, that such practitioner
references should form the foundation of the scholarly paper but they
can—and actually do—play a significant role in supplementing and
completing the bridge between theory and application.

The Iterative Process of Scholarly Writing

Scholars, scholarly journals and academic institutions use an iterative
process to write, evaluate, make feedback, and improve papers until they
meet standards for content, references, and format. This is a great
learning process and by no means one that is fun for the writer. The
iterative process refers to the steps taken in the actual writing of the
paper, including identifying and obtaining the materials.
The first step is to create a vision of the end product, and this is
called conceptualization. After this, one should identify the tangible
deliverables such as research notes, a reference list with solid scholarly
references, a structure of the paper, etc. Thereafter, one should identify
the activities required to accomplish each of these deliverables. After this
stage, one should estimate the time it will take to accomplish the
individual activities. Finally, starting from the desired end date, one
should then work backwards to schedule the activities to fit within this
time line. This enables the writer to track his or her progress and self-
assessment.
262 Social Research Methodology: An Introduction

Scholarly versus popular writing: A summary

Difference Scholarly writing Popular writing

Subject The subject matter is serious Subject matter is mainstream,

matter and academic: Philosophy, such as fashion, news, sports,
Nursing, Science, Literary politics, gossip column stories,
Criticism, Law, etc. financial news, consumer
goods, entertainment, etc.

Purpose of To share information and Inform or entertain the reader;

writing research. Scholars report on sell or promote fashion,
original research, theory, consumer products; promote an
and/or methodology. opinion or belief; etc.

Authorship Articles, mostly in scholarly Journalists and other freelance

journals, are written by writers, who may even lack
professionals and scientists in expertise in the subject.
the academic discipline. These
include professors, researchers
and graduate students.

Readership The intellectual community Readership is general.

(professors, researchers and
scholars) as well as policy and
decision makers.

Publisher Many scholarly journals are Popular literature is published

published by universities and by commercial, profit-driven
professional organizations such firms such as newspaper
as the American Medical publishers.
Association, the Uganda Law
Society, the Organization for
Social Science Research in
Eastern and Southern Africa
(OSSREA), etc.

Language The writing style is formal, The writing style is informal

and style dry, and technical. Quite often, and easy to read. This is
terminology may be unfamiliar because it is aimed at general
to someone outside the readership.
discipline.
 Scholarly Writing and Citation Methods 263

View point Generally objective; arguments Often the ideology of the

and opinions are backed by publisher is evident in the
factual information from political, religious, ethical,
research. Case studies are corporate and/or social views
common in these sources of and values expressed. Biased,
evidence. emotional, and/or moralistic
arguments and editorials are
clues.

Cross Footnotes, end notes and/or None of these.

referencing bibliographies are found in all
scholarly journals. Articles are
sometimes preceded by an
abstract.

Quality Papers are reviewed and edited Only the magazine editor
assurance by experts in the field or selects and edits the submitted
discipline—refereed or peer articles, or writes them.
reviewed

Plain illustrations such as Often has slick, glossy photos,

Illustration graphs, tables, and lists are and other eye-catching
used. Less often, there are material on the cover and
limited adverts, and these are throughout the magazine.
of publications and services
within the discipline covered.

Reward for The authors generally receive Popular writing is published to

writing little or no payment for their make money and entertain.
work. Their prestige as The authors make some or all
scholars rests on how well they of their living from their work.
present or argue their They may become well-know
information and research. and well-paid as a result of
their work and how they
influence their audience.

Citation of Sources under the APA System

The APA style uses a “within-text” citation system instead of footnotes
or endnotes. The briefly cited sources are then given as full bibliographic
references on the reference list at the end of the paper. The style of
presentation on the reference list is the subject I now embark on.
264 Social Research Methodology: An Introduction

Specific material content

When a writer wants to cite specific material from a source, he or she
indicates in brackets the author, date of publication, and page number.
This may be done in two ways. The first is where citation is short and
remains part of the sentence. This may be done either by reference to the
area in the source generally (see the Katebire citation below) or by
paraphrasing the author (see the Okello citation). Here are the citations:

On the complexity of the theory of utilitarianism in polarized

societies (Katebire, 1999: 47), it has been argued that the
proponent has “himself not soared above the smudges of the duo
nature of racial and class conflict” (Okello, 2001: 12).

The second way is where the writer wants to make use of a long,
verbatim quotation. In that case, he or she should isolate the specific
material and double-indent it in a freestanding block, and then supply the
author, date and page number. For example:

I don’t know what is behind the mentality of the

Australian leaders, but I don’t think it is right. A country
so desperate with its economy, and you try to dangle a
carrot in front of them, of course, just like a prostitute ...
if you dangle money in front of her, you think she will not
accept it. Of course she will, because she is desperate
(Matherell & Grattan, 2002: 2).

Such direct quotations are normally about 40 words or more, page

number must be indicated, and quotation marks are not used. Note also
that initials are not used in the citations, but are included in the complete
bibliographic record on the reference list.

General material content

If the writer is referring to the content of a work generally, he or she
indicates only the date after the author, but not the page number. Take
the following citation for example.

Copyright law is basically a territorial concept (Tabaro, 2003). On

the premises, a conflict of laws is likely to arise where
 Scholarly Writing and Citation Methods 265

enforcement of rights is preferred in a territory whose copyright

regime differs from the broadcaster’s home of origin.

Compare the above example with the following:

Tibatemwa (1998) has extensively investigated the phenomenon

of domestic violence, including homicide by women, among the
highly polygamous Basoga people of Eastern Uganda.

Here only the date is bracketed because the author’s name is supplied in
the text as part of the sentence. It should be noted that when two or more
authors are cited, but their names are part of the sentence, the word “and”
is used. However, if the citation is within parentheses, the ampersand (&)
is used instead of “and.” For example:

Tamale and Asiimwe (2005) investigated the problem of electoral

violence in national politics in post-1995 Uganda. ...

Compare this example with the following:

The problem of electoral violence in post-1995 Uganda has been

thoroughly investigated (Tamale & Asiimwe, 2005).

Multiple authors
The above principles relate to works with one or two authors. When a
work has three to five authors, the Latin abbreviation “et al.” (meaning
“and others”) is used in referencing as follows: Cite all authors’ last
names the first time the reference occurs.

Kanyangoga, Wasswa, Oburu, Tindifa and Guweddeko (2001)

found out that cattle rustling by the Karimojong was only a proxy
game.

Subsequently, cite only the first author, followed by “et al.” with the
date.

Kanyangoga, et al. (2001) found . . .

266 Social Research Methodology: An Introduction

However, if the authors are above five, the first author is cited with “et
al.” straight away. And in subsequent citations within the same
paragraph, the year is also be omitted. Note, of course, that the use of “et
al.” is only in citation but not on the reference list, except where there are
more than five authors.
Besides “et al.” there are other two commonly used Latin
abbreviations: “ibid” and “op. cit.” The first one, “ibid” is used if there
is another reference to a work, without any other work being mentioned
in between. The second one, “op. cit.” is used if reference is made to a
preceding work, but it is not the immediately preceding one. While
“ibid” is used directly in parentheses (ibid), “op. cit.” is used following
the author’s name (Tindifa, op. cit.: 18). The name may be inside or outside
the parenthesis.

Multiple citations
If more than one work is cited for reference, the authors of the works
should be arranged alphabetically, not chronologically, and the citations
should be separated with semicolons. For example:

A number of studies (Batte, 1999; Fabusha, 1992; Twino & Twine,

1995) sought to test the theory of . . .

Sometimes, works may be by the same author(s). These are arranged by

year of publication, giving surnames once, and then separating dates by
commas. For example:

A number of studies (Twino & Twine, 1992, 1994, 1997) have

been conducted on various aspects of . . .

If any of the works by the same author(s) have the same year of
publication, the works may be identified by the suffixes a, b, c, etc., and
these must be indicated as such on the reference list. For example:

A number of studies (Twino & Twine, 1992, 1994a, 1994b, 1997)

have examined the problem of . . .

Non-retrievable material
 Scholarly Writing and Citation Methods 267

Personal letters, telephone conversations or interviews, and other

material that cannot be retrieved are cited in the text, but are not listed
among the references. Examples:

The Dean of Law (telephone conversation, September 11, 2001)

confirmed the formulation of a tough sex harassment policy to
govern student-student, student-staff, and staff-staff relations at
the Faculty.

It is now confirmed that Bakibinga (2004) relied entirely on the

Internet for sources for his ground-breaking article on . . .(personal
letter, July 28, 2003).

Secondary sources
A writer may cite material that is quoted or paraphrased elsewhere,
without himself or herself accessing the original source. This is
acceptable and may be indicated as follows:

Kazibwe and Kembuzi (1998) pointed out: “It is possible for any
lay person to witness an event without knowing that the event is
occurring. It takes an engineer to know the existence of a valley
dam on a given ground”(cited in Sematimba, 2003).

Online sources
For a specific part of an Internet source, indicate the page number,
chapter, figure, table, or equation in the text. The words “page” and
“chapter” are normally abbreviated in such text citations. For example:

In an independent qualitative enquiry, Shokoro and Senyonga

(2004, p. 19) found out that . . .

Following the controversial report, an independent qualitative

study established the nature of the abuse (Shokoro & Senyonga,
2004, chap. 3).

If the electronic source does not provide page numbers, use the paragraph
number, if available, preceded by the paragraph symbol or the
abbreviation para. For example:
268 Social Research Methodology: An Introduction

An independent enquiry revealed the extent of the abuse (Shokoro

& Senyonga, 2004, ¶ 5).

Or:

An independent enquiry revealed the extent of the abuse (Shokoro

& Senyonga, 2004, para. 5).

If the source indicates no author, use the title in an abbreviated form,

then date. For example:

A similar study was done in the context of student learning

behaviour, which revealed strikingly similar patterns (Controlled
Behaviour, 2003).

The Reference List

The reference list is a list of all the works you have cited in your text (by
the author’s name and year of publication, and page number in case of
works within works—e.g. an article within a journal). The list is
organised alphabetically by the names of the authors (or originators) of
the work. During the course of your reading you may use material for
extending your knowledge of the subject, but which you may not cite. A
bibliography lists all these works, again alphabetically by author (or title
if author is not indicated). A bibliography may be included after the
reference list, especially if you used footnotes or endnotes, but it is
generally not always necessary to include it.
All the sources cited in the text, except the non-retrievable, must
therefore be listed in alphabetical order at the end of the paper or thesis
or dissertation. The listing should give the full bibliographical details of
the works; that is, names and initials of the author, date of publication,
title of work, title and editor(s) of larger work, place of publication,
publisher, volume and issue numbers, etc–depending on the type of
work. Please take note of capitalization in titles in all cases. I now turn
to the different types of works.

Books by Individual Authors

The presentation of books begins with the last names of the authors,
followed by initials and the date of publication in parentheses. This is
 Scholarly Writing and Citation Methods 269

punctuated by a period. Then the title of the book follows in italics, also
punctuated by a period. After this follows the place of publication,
normally the city, punctuated by a colon and then the publisher. If the
book has a sub-title, it should follow the title after a separating colon.
The items use a hanging indentation. A pattern for such works appears
as follows:

Author’s last name, initials (date of publication). Title of the book.

Place of publication: Publisher.

Note that a book may be authored by a single author, two authors, or

three and above authors. The pattern is basically the same. For a book
that is jointly authored by two to five authors, the authors are listed in the
order they appear on the book cover, the last name separated by the
ampersand. For a book authored by over five authors, only the first
author is taken, followed by “et al.” in the place of the rest of the authors,
but everything else remains the same. Here below are examples:

Bailey, K.D. (1994). Methods of social research. New York: The

Free Press. [single author]

Katebire, D.A. (2005). Research methodology: An introduction.

Kampala: Makerere University Press. [book has title and sub-
title]

Allport, G.W., Olupot, P., & Best, J.W. (1967). Age and
performance in mental tests. London: Longman. [joint authors,
two to five]

Munno, G.W. et. al. (1978). Growth and performance as

personality indicators. London: Heinemann. [over five authors]

Books by Corporate Authors

When the work belonging to an organization is not attributed to any
specific individual, it is a case of corporate authorship. Corporate authors
therefore include institutions, associations, and other organizations. In
such cases, use the name of the organization as the author. For examples:
270 Social Research Methodology: An Introduction

Makerere University. (2005). Sex harassment policy. Kampala:

Makerere University Press.

Department of State. (1999). State violation of human rights:

Country reports. Washington DC: Official Records
Corporation.

If the corporate author is also the publisher, use the word “Author” as the
name of the publisher. For example:

UNDP. (2001). The human development report. New York:

Author.

Subsequent Edition (Republished Book)

This is presented in the same format as in the above examples, except
that the writer should endeavour to list the most recent date of
publication he or she can find. Then after full presentation, the date of the
first edition is stated in brackets. Below are three examples:

Moser, C.A. & Kalton, G. (1986). Survey methods in social

investigation. London: Gower. (Original work published 1971).

Bryman, A. (2000). Quantity and quality in social research.

London: Routledge. (First published in 1988).

American Psychological Association (2001). Publication manual

(4 ed.). Washington D.C.: Author.

Part of a Book
If the material in reference is a prologue, foreword, preface, introduction,
or epilogue, list the name of the author of that particular material, the
date, name of the material, the title of the book, and, in parentheses, the
page number or page range on which the material appears (with p. or pp.
to indicate page or pages). Example:

Mapapa, L.N. (2003). Preface to the second edition. Modern

thinking on politics of transition in African states (pp. iv–xi).
Nairobi: East African Political Books.
 Scholarly Writing and Citation Methods 271

Edited Books
There are two types of reference to edited works. The first is to list a
work in its entirety, attributing it to the editor. In this case, the work is
treated in the same way as an authored one, except that after the name
and initials is added “Ed” in parentheses. The second is to list a signed
chapter in an edited volume. In this case, the name and initials of the
chapter author are given, the date and title of the chapter. Then follows
the bigger work where the chapter is published. This is given by putting
“In,” followed by initials and names of editors, followed by “Eds,”title
of the work, page range of the chapter, place of publication and
publisher. Below are the respective examples:

Afunaduula, F.O. & Asiimwe, D. (Eds.). (1998). Interdisciplinary

teaching of human rights at Makerere University. Kampala:
HURIPEC. [editors treated as authors]

Mbireba, R. E. (2002). State militias in electioneering: The

Kalangala Action Plan. In D.A. Katebire & M.A. Fernando,
(Eds.). Election violence in Museveni’s Uganda (pp. 213-27).
New Delhi: Indo-Afri Books. [the editors begin with initials; page
numbers are not italicised]

No Author or Editor
For a book where the author or editor is not indicated, the title is taken
and listed alphabetically. The rest proceeds normally. To alphabetize
such a work on the Reference page, the first significant word of the title
(i.e. ignoring the articles, “a,” “an,” and “the”) is used. Below are two
examples:

Analysis of the U.S. military and diplomatic maneuvres in post

Sadam Iraq. (2005). Warsaw: Polish Scientific Press. [NOT
An Analysis…]

British foxy games in the Gulf crisis. (2004). London: Blueman.

[NOT The British…]

Dictionary
272 Social Research Methodology: An Introduction

Simply state the title in full, the edition in parentheses, and the date.
Then proceed with place and publisher as in the following example:

Shorter Oxford English dictionary (5th ed.).(2002). New York:

Oxford University Press.

Multi-volume Works
To refer to a single volume of a publication, include only the relevant
date and volume number. When you are referring to more than one
volume, include all the relevant volume numbers. In this case, the date
should include the range of years of publication (if for instance the work
appeared over a number of years). Take the following respective
examples:

Gitau, K.K. (2003). Studies in ephemeral values in Eastern African

societies: The Kisii of Kenya. (Vol. 13). Nairobi: Nairobi
Institute of Cultural Anthropology.

Omoding, B. M. & Mwisho, S. (Eds.). (1999–2002). Studies in

intermarriage and racial relations (Vols. 5–8). Harare:
Mugabe Academy of Race Relations.

Thesis or Dissertation
For such works, state name and initials, date and title. Then state what
sort of work and of which university. Note that these works are
unpublished works and they should thus be stated as such. Take the two
examples below:

Matovu, C. W. (2004). Intrigue in interlacustrine state relations at

the dawn of colonialism: A study of Buganda and Bunyoro
kingdoms. Unpublished doctoral dissertation, Kyambogo
University, Kampala.

Muthoni, W. (2003). Female genital mutilation: The rise and

demise of the Gikuyu. Unpublished M.A. thesis, Nairobi
Institute of Cultural Anthropology.

Periodicals
 Scholarly Writing and Citation Methods 273

These are works that come periodically, normally in volumes or series,

under the same title. They are thus of different categories. They include
scholarly journals, magazines and newspapers. Basically, the reference
pattern is the same, with slight variations as I show below.

Articles in scholarly journals

Many scholarly journals number their pages consecutively throughout a
given volume. A journal may publish four times a year, and those four
issues normally constitute one volume. If the first issue of one volume
ends at page 145 the following issue begins at page 146. The listing order
of an article is: name, initials and date; then title of article, title of
journal, volume number, and page range. Note that for consecutive
journals the issue number is not indicated on the Reference list. Here is
an example of a consecutive journal:

Abuya, E. (2003). The “Pacific Solution”: A durable answer or

political stopgap option? East African Journal of Peace and
Human Rights, 9, 144-49. [note that the volume number is also
italicised, and the journal title is capitalised]

If that article appears in a non-consecutive journal (where each issue

begins at page 1), the issue number is given in parentheses but not in
italics after the volume number. An example of a non-conservative
journal:

Kagame, C.S. (1998). Vagina monologues: Pervasive feminism?

African Feminism Review, 9(3), 21–29.

Note that page numbers are indicated without the use of p. or pp. as in
edited books.

Magazine articles
These are somewhat similar to scholarly journals in form, the major
difference being that the date is more detailed since such publications are
more frequent than journals. The date includes a month, and even the
day. Some examples below:
274 Social Research Methodology: An Introduction

Wheatcroft, G. (2004, June). The tragedy of Tony Blair. The

Atlantic, 293, 56–72.

Thomas, E. & Hosenball, M. (2004, May 31). Bush’s Mr. Wrong:

The rise and fall of Chalabi. Newsweek, 143, 22–32.

Newspaper articles
If the author of the article is indicated, begin with that author’s name,
followed by full date and title of article. If the author’s name is not
available, begin the reference with the headline or title in the author
position, followed by the full date and full title of the newspaper and
page numbers. (Do not omit the “A,” “An” or “The” from the title of a
newspaper as it is done with books). Examples:

Dhakaba, M.M. (2004, January 22). FIFA’s Mamelodi in town. The

Monitor, 32.

Controversial rat-eater of Teso dies (2005, March 12). The New

Vision, 3.

If the text being cited is from an editorial or letter to the editor, it should
be indicated in square brackets after the title and date. Example:

Silverman, P.H. (2004, June). Genetic engineering a threat to

social heredity [Letter to the editor]. The Atlantic, 293, 14.

Book Review
Ssebagala, N.N. (2005). All are indeed the king’s men [Review of the
book Not all king’s men]. Ugandan Cultural Journal, 7, 130.

Internet Material
For such material, list the author and initials or, if not available, the title
of the work. Then indicate the date or, if not available, “n.d.” meaning
“not dated.” List the title of the work, followed by the date the
information was retrieved and the Uniform Resource Locator (URL). The
date is important because web sites are updated, shifted or removed
anytime. Two examples:
 Scholarly Writing and Citation Methods 275

Tibaijuka, S. & Beebwa, T. (n.d.). Sleep disorders among children.

Retrieved June 12, 2005, from http://www.commnet.edu/apa/.

Eating disorders among adolescents. (n.d.). Retrieved March 12,

2006, from www.nutritionmasters.org/

Internet articles based on print source

Today, most articles retrieved from online publications have exact print
versions. For these, the same basic primary journal reference can be used,
but if you have viewed the article only in its electronic form, you should
add in parentheses after the title of article “Electronic version.” For
example:

Mugerwa, S., & Nuwe, J. (2005). Consequentialism and

deontology [Electronic version]. Journal of Contemporary Ethics,
5, 117-123.

If you suspect or know however, that the format of the online article
differs from the print version (page numbers are not indicated or it
includes additional data or commentaries), you should add the date you
retrieved the document and the URL. For example:

Kamugisha, J., Ngaiza, J. & Atim, P. (2005). Quality assurance in

the teaching of law at Makerere University. Rethinking Curriculum,
3, 23-45. Retrieved September 12, 2006 from
http://cur_rev.org/articles.html.

Electronic databases
Universities, colleges and public libraries subscribe to electronic
databases such as Datad, EBSCO, LexisNexis, OCLC, WilsonWeb,
SIRS, etc. There are also online databases such as ERIC, PsychINFO,
and Public Affairs Information Service (PAIS). These databases contain
full-text articles and article abstracts, which should be listed as normally
done, but the retrieval date and database should be indicated. For
example:

Maguru, J.B., Mikono, G.W., Maraka, P., & Mutwe, S. (2003). The
role of early supervisory experience in supervisor
276 Social Research Methodology: An Introduction

performance. Journal of Applied Psychology, 12, 120-132.

Retrieved October 20, 2005, from PsychINFO database.

Nonprint Resources
The role is shown in brackets after the name while the source is
identified in square brackets after the title. Some examples below:

Film
Bagenda, S. (Director). (2002). Agony and ecstasy [Film].
Kampala: Local Talent Ltd.

Cassette
Mugenyi, P.A. (Author and speaker). (2004). The welfare needs
of Makerere University support staff [Cassette]. Kampala:
MUWU.

TV programme
Nagenda, J. & Kibenge, O. (Narrators). (2003). The lies on NRA
abuse of female child soldiers peddled by China Keitetsi.
[Television broadcast]. Kampala: UTV.

Musical recording
Kafeero, P. (1999). Nantabulilirwa. On Embaga ya Kafeero [CD].
Kampala: Ganda Records Ltd.

Non-recoverable Material
Material such as personal interview, phone conversation, personal letter,
etc. is not easy for any third party to see or hear. It should thus not be
included on the list of References. It can, however, be cited
parenthetically within the text as we have already noted.

Classroom Lecture
Like personal interviews and phone conversations, material presented in
a classroom lecture is regarded as non-retrievable data because it is not
published. A lecture, therefore, should not be included in references
unless the lecturer distributed material at the lecture. In that case, it
should be presented as follows:
 Scholarly Writing and Citation Methods 277

Asiimwe, D. (2004, April). Salient issues in participant

observation. Outline presented in a classroom lecture at
Faculty of Law, Makerere University, Kampala.
278 Social Research Methodology: An Introduction

References
Adcock, R. and D. Collier (2001). Measurement validity: A shared
standard for qualitative and quantitative research. American
Political Science Review 95: 529-546.
American Psychological Association (APA) (2001). Publication manual
of the American Psychological Association. (5th ed.). Washington
D. C.: APA.
Amrozik, A. & Nocella, L. (1998). The sociology of social problems:
theoretical perspectives and methods of intervention. Cambridge:
Cambridge University Press.
Bacon, J. and Olsen, K. (2003). Doing the right thing. Dept. for Work
and Pensions – Research Working Paper No.11, London: HMSO.
Babbie, E.R. (1995). The practice of social research. Belmont:.
Wadsworth
Bailey, K.D. (1994). Methods of social research. 4th ed. New York: Free
Press.
Barbour, R.S. & Kitzinger, J. (Eds) (1999). Developing focus group
research – politics, theory and practice, London: Sage.
Belmont Report (1979). Ethical principles and guidelines for the
protection of human subjects of research. The National Commission
for the Protection of Human Subjects of Biomedical and Behavioral
Research, April 18.
Benchmark Institute (2006). The Legal Services Practice Manual.
California.
Bourner, T (1996). The research process: four steps to success. In T.
Greenfield (ed) Research methods: guidance for postgraduates (pp.
7-11). Arnold: London,
Bulmer, M. (2001). The ethics of social research. In Gilbert, N. (ed.)
Researching social life, London: Sage.
Busha, C. and Harter, S.P (1980). Research methods in librarianship:
techniques and interpretations. New York: Academic Press.
Boyer, E.L. (1990). Scholarship reconsidered: priorities of the
professoriate. San Francisco: Josey-Bass, Inc.
Bryman, A. (1988). Quantity and quality in social research. London:
Routledge.
Bryman, A. (2001) Social research methods, Oxford: Oxford University
Press.
 References 279

Campbell, D.T. and Julian C.S. (1963). Experimental and quasi-

experimental designs for research. Boston: Houghton Mifflin.
Campbell, M.J. & Machin, D. (1999). Medical statistics: A commonsense
approach. 3rd ed., Chichester: JohnWiley & Sons.
Carley, K. (1992). Coding choices for textual analysis: A comparison of
content analysis and map analysis. Unpublished Working Paper.
Clark, P.A. (1972). Action research and organizational change. London:
Harper and Row.
Chambers, R. (1992). Rapid but relaxed and participatory rural appraisal:
towards applications in health and nutrition. In Rapid assessment
procedures: qualitative methodologies for planning and evaluation
of health related programmes. Boston, MA: INFDC.
Cooper, H. (1998). Synthesizing research: a guide for literature reviews.
Thousand Oaks, CA: Sage.
Coulehan, J.L. & Block, M.R. (2005). The medical interview: mastering
skills for clinical practice. New York: Davis.
Creswell, John W. (1998). Qualitative inquiry and research design:
choosing among five traditions. Thousand Oaks, CA: Sage
Publications.
Creswell, John W.(1994). Research design: qualitative and quantitative
approaches. Thousand Oaks, CA: Sage.
Crisp, D. & Pedry, C. (1993). Researching sensitive issues: a report of
the sra one-day seminar on “Asking the Unaskable.” SRA
Newsletter.
Dudeley, R.M. (1999). Uniform central limit theorems. New York:
Cambridge University Press.
Dunn T. (1994). Rapid rural appraisal: a description of the methodology
and its application in teaching and research at charles stuart
university. Wagga Wagga, Australia: Rural Society.
Folz, David H. (1996). Survey research for public administration.
Thousand Oaks: Sage Publications, 1996.
Foster, C. (2001). The ethics of medical research on humans, London:
Cambridge University Press.
Freedman, D.A. (1999). Ecological inference and the ecological fallacy.
Technical Report No. 549, October (Prepared for the International
Encyclopaedia of the Social and Behavioural Sciences).
280 Social Research Methodology: An Introduction

Freed-Taylor, M. (1994). Ethical considerations in European cross-

national research. In International Social Science Journal, 142, 523-532.
Frey, J.H. & S.M. Oishi (1995). How to conduct interviews by telephone
and in person. London: Sage.
Galvan, J.L. (1999). Writing literature reviews. Los Angeles: Pyrczak
Publishing.
Glaser, B. (1978). Theoretical sensitivity. Mill Valley, CA: Sociology
Press.
Glaser, B. and Strauss, A. (1967). The discovery of grounded theory.
Chicago: Aldine.
Gorard, S. (2002). Ethics and equity: pursuing the perspective of non-
participants. In Social research update, 39. Dept of Sociology,
University of Surrey.
Gorard, S. (2003). The role of numbers in social science research.
London: Continuum.
Greene, J.C., Caracelli, V.J., & Graham, W.F. (1989). Toward a
conceptual framework for mixed-method evaluation designs.
Educational Evaluation and Policy Analysis, 1989 – JSTOR.
Guilford, J. P. & Fruchter, B. (1978). Fundamental statistics in
psychology and education, 6th ed. New York: McGraw-Hill.
Hamel, J. (1993). Case study method. Beverly Hills, CA: Sage.
Heaton, Janet (1998). Secondary analysis of qualitative data. In Social
research update, 22. Autumn.
Heffner, C.L. (2004). Allpsych online: the virtual psychology classroom.
<http://allpsych.com/socialresearchmethods/index.html>
Heise, David R. (1970). The semantic differential and attitude research.
In Summers, G.F. (ed). Attitude measurement (pp 235-253).
Chicago: Rand McNally.
Hinton, P.R. (1993). The psychology of interpersonal perception.
London: Routledge.
James, W. (1907). Pragmatism: a new name for some old ways of
thinking.New York: Longman Green & Co.
Johnson, R.B. & Onwuegbuzie, A.J. (2004). Mixed methods research: a
research paradigm whose time has come. Educational researcher,
33(7), 14–26.
Kashyap, P. (1992). Rapid rural appraisal (RRA) methodology and its
use in nutrition surveys. Rapid assessment procedures - qualitative
methodologies for planning and evaluation of health related
 References 281

programmes. Boston, MA: International Nutrition Foundation for

Developing Countries (INFDC).
Keats-Rohan, K.S.B. (2003). Progress or perversion? Current issues in
prosopography: an introduction.
Khun, T.S. (1970). The structure of scientific revolutions. 2nd ed.
Chicago: University of Chicago Press.
King, Nigel (2004). Using interviews in qualitative research. In C.
Cassell & G. Symon (Eds.) Essential guide to qualitative methods
in organizational research. London: Sage.
King, N.M.P. et al. (Eds) (1999). Beyond regulations – ethics in human
subjects research. Chapel Hill: University of North Carolina Press.
Krippendorf, K. (1980). Content analysis: An introduction to its
methodology. Beverly Hills, CA: Sage Publications.
Kumar, Krishna (1993). An overview of rapid appraisal methods in
development settings. In K. Kumar (Ed.) Rapid appraisal methods.
Washington D.C.: World Bank.
Latham, J.R. (2002). Thoughts on scholarly writing (mimeo).
Ledbetter, C.E. (2001). Research Design and Methodology Types. After
Gay, L. & Airasian, P. (2000). Educational research: Competencies
for analysis and application (6th ed.). Columbus, OH.:Merrill.
Lee, R.M. (1993). Doing research on sensitive topics. London: Sage.
Lee, R.M. (2000). Unobtrusive methods in social research, Buckingham:
Open University Press.
Macauley, P. (2001). The literature review. Geelong: Deakin University.
Mauthner, M., Birch, M., Jessop, J., and Miller, T. (eds) (2002). Ethics
in qualitative research. London: Sage.
May, M. et al (2001). Understanding social problems: issues in social
policy. Blackwell.
May, T. (2001). Social research: issues, methods and process,
Buckingham: Open University Press.
McNamara, C. (1998). Basic guide to program evaluation. Retrieved
October 27, 2004 from
http://www.mapnp.org/library/evaluatn/fnl_eval.htm
Meier, Kenneth J. & Brudney, J.L. (1993). Applied statistics for public
administration, 3rd ed. Belmont, CA: Sage.
Miles, M. & Huberman, A. (1994). Qualitative data analysis: An
expanded sourcebook (2nd ed.). Thousand Oaks: Sage.
282 Social Research Methodology: An Introduction

Miller, D. (1991). Handbook of research design and social

measurement. CA: Sage.
Mills, C.W. (1959). The sociological imagination. New York: Oxford
University Press.
Mt. Elgon Integrated Conservation and Development Project (1999).
Participatory rural assessment training notes.
Muskal, F. (2000). Guidelines for the literature review paper. University
of the Pacific.
Neuman, L.W. (2000). Social research methods: qualitative &
quantitative approaches (4 th ed.), London: Allyn & Bacon.
Nichols, Paul (1991). Social survey methods. Oxford: Oxfam.
O’Brien, R. (2001). An overview of the methodological approach of
action research. In Roberto Richardson (Ed.), Theory and practice
of action research. João Pessoa, Brazil: Universidade Federal da
Paraíba.
Onwuegbuzie, A.J. & Teddlie, C. (2003). A framework for analzing data
in mixed methods research. In A. Tashakkori & C. Teddlie (eds.),
Handbook of mixed methods in social and behavioral research (pp.
351-383). Thousand Oaks, CA: Sage.
Palmquist, M.E. et al. (1997). Two applications of automated text
analysis: Analyzing literary and non-literary texts. In C. Roberts
(Ed.), Text analysis for the social sciences: methods for drawing
statistical inferences from texts and transcripts. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Picciotto, R. (1993). Foreword. In K. Kumar (Ed.) Rapid appraisal
methods. Washington D.C.: World Bank.
Pole, C. & Lampard, R. (2002). Practical social investigation:
qualitative and quantitative methods in social research. New York:
Prentice Hall.
Ragin, C.C. & Becker, H.S. (1992). What is a case?: exploring the
foundations of social inquiry. Cambridge: Cambridge University
Press.
Renner, Tari (1988). Statistics unraveled: a practical guide to using
statistics in decision-making. Washington, DC: International City
Management Association.
Rossman (1995). Negotiating graduate school: a guide for graduate
students. Thousand Oaks: Sage.
 References 283

Rowley, J. & Slack, F. (2004). Conducting a literature review.

Management research, 27(6), 31-39.
Sarantakos, S. (1998). Social research (2nd ed). Macmillan.
Sarantakos, S. (2000). Research stats without maths. Harvard Press.
Schlabach, T.F. (1996). Ten Commandments of Good Historical Writing.
<http://personal2.stthomas.edu/gwschlabach/courses/10commnd.
htm>
Sleeper, R.W. (2001). The necessity of pragmatism: John Dewey’s
conception of philosophy. University of Illinois Press.
Stake, R.E. (1995). The art of case study research. Thousand Oaks:
Sage.
Stone, L. (1972). Prosopography. In F. Gilbert & S. Graubard (Eds).
Historical studies today (107-140). New York: Norton.
Strauss, A. & Corbin, J. (1990). Basics of qualitative research:
Grounded theory procedures and techniques. Sage Publications.
Strauss, A. & Corbin, J. (1994). Grounded theory methodology: An
overview. In N.K. Denzin & Y.S. Lincoln (Eds), A handbook of
qualitative research (1-18). London: Sage Publications.
Susman, G.I. (1983). Action research: A sociotechnical systems
perspective. London: Sage.
Sutherland, E.H. (1949). White collar crime. New York: Dryden Press.
Taylor, D. & Procter, M. (2005). The literature review: a few tips on
conducting it. University of Toronto.
Trochim, W. (2000). The research methods knowledge base (2nd ed.).
Cincinnati, OH: Atomic Dog Publishing.
Tufte, E.R. (1983). The visual display of quantitative information.
Cheshire: Graphics Press.
Tufte, E.R. (1997). Visual and statistical thinking: displays of evidence
for decision making. Cheshire: Graphics Press.
Tufte, E.R. (1997). Visual explanations: images and quantities, evidence
and narrative. Cheshire: Graphics Press.
Weber, R. P. (1990). Basic content analysis (2nd ed.). Newbury Park,
CA: Sage Publications.
Welch, S. & Comer, J. (1988). Quantitative methods for public
administration: techniques and applications (2nd ed.). Chicago, Il:
Dorsey Press.
Wimmer, R.D. & Joseph, R.D. (1997). Mass media research: an
introduction. Belmont, MA: Wadsworth.
284 Social Research Methodology: An Introduction

Winter, R. (1989). Learning from experience: principles and practice in

action-research. Philadelphia: The Falmer Press.
Worsley, P. (1972). Problems of modern society. Bristol: Penguin Books.
Yin, R. K. (1981). Case study as a serious research strategy. Knowledge
(3), 97-114.
Yin, R. (1994). Case study research: design and methods (2nd ed.).
Beverly Hills, CA: Sage Publishing.
Yin, R. K. (2002). Case study research, design and methods (3rd ed.).
Newbury Park: Sage Publications.
 References 285

Back Cover

Denis Asiimwe Katebire lectured research methods at the National College of Business
Studies (now Makerere University Business School). He was also lecturer in computer
literacy and information systems at Uganda Martyrs University. Currently, he is a
lecturer in social research methods at Makerere University’s Faculty of Law.

Mr Asiimwe Katebire was educated at Rushasha Primary School in Rukungiri district,

Apostles of Jesus Bukinda Missionary Seminary in Kabale district and Makobore High
School in Rukungiri district. He proceeded to Makerere University where he graduated
with an honours degree in social sciences and a postgraduate diploma in information
sciences. He also obtained graduate diplomas in development, law and social justice form
the Institute of Social Studies in The Hague, The Netherlands; and peace and conflict
research from the European Peace University in Schlaining, Austria. He returned to
Makerere where he graduated with an M.Sc in information science. He is currently
studying for a doctorate at the University of KwaZulu-Natal, South Africa.