An Introduction to Research Methodology 1

An Introduction to
RESEARCH METHODOLOGY
2 An Introduction to Research Methodology
 An Introduction to Research Methodology 3

An Introduction to
RESEARCH
METHODOLOGY

Editors
MKC Nair,
Remadevi S, Anish TS
Harikumaran Nair GS,
Ajith Kumar K, Leena ML

Kerala University of Health Sciences

Thrissur 680 596, Kerala
4 An Introduction to Research Methodology

An Introduction to
RESEARCH METHODOLOGY
Editors
MKC Nair,
Remadevi S, Anish TS
Harikumaran Nair GS,
Ajith Kumar K, Leena ML
Published by:
Kerala University of Health Sciences
Thrissur 680 596, Kerala
Layout & Typesetting
Ramya A.R.
Printed at:
Kerala Books and Publication Society
Kakkanad
Ist Edition:
January 2017
© Kerala University of Health Sciences
For private circulation only
 An Introduction to Research Methodology 5

FOREWORD

Research is vital for the progress of medical practice. It is a tool

for creating evidence which would form the basis for medical
practice and planning for the future. We become wiser from
the past, know from the present and learn from the future. The
evidences generated by researchers in 18th and 19th centuries
are still landmarks in the history of epidemiological research.
Demonstration of the effect of Vitamin C in the prevention of
Scurvy by James Lind (1747) through preliminary clinical trial
and discovery of cholera as a waterborne disease by John Snow
(1853) through natural experiment at community level are
excellent examples of this kind. Practice of medicine was an art
alone; knowledge was transferred from teacher to student and
the opinion of the senior most was the last word. However,
with the introduction of technology into medicine, the capacity
to confirm parameters have improved a lot and hence objective
outcome measurements became the corner stone of any medical
research.
Research is an integral part of academic training programmes
of all disciplines and this holds true with regard to all health
science streams under the Kerala University of Health Sciences.
Animal studies and basic science research, including molecular,
biochemical and genetic studies have thrown up new
opportunities for better understanding the functions of human
body. Some of these advanced research may be outside the reach
6 An Introduction to Research Methodology

of KUHS students during their study period. In general,

undergraduate students may focus on descriptive studies, the
post graduates on analytical studies, and doctoral candidates
on intervention studies. Up scaling community application of
proven interventions, its ecological and cost effectiveness are
the realm of post-doctoral candidates. Qualitative techniques
are a part and parcel of research at all stages. The role of
biostatistics in health research is very crucial and a biostatistician
should be consulted at the protocol development stage itself.
KUHS students could also take up health social science, health
economics and health policy related issues, as understanding
of social determinants of health is vital in improving the health
status of a population.
Research methodology is in fact a science that deals with the
step by step process of doing research in a systematic and
scientific manner. This book, primarily intended for students
and young researchers, cover the basic elements of research
methodology. The book is dear to me, not necessarily because
of contents, which are introductory in nature but because it is
written by faculty who have been certified by KUHS after
adequate exposure to methodological issues. We hope the book
will help the beginners to plan, conduct and report good
research. Some of the important ideas are repeated in subsequent
chapters in order to reemphasise the importance.
Prof (Dr). MKC Nair
Vice Chancellor, KUHS
 An Introduction to Research Methodology 7

CONTRIBUTORS

1. Dr Ajitha K,
Associate Professor, Govt Ayurveda College, Kannur
2. Dr Antony George,
Assistant Professor MES Dental College, Perinthalmanna
3. Dr. Anitha M A,
Lecturer, Dr Padiar Memorial Homeopathic Medical College,
Chottanikkara
4. Dr Balakrishnan V.K.V.
Associate Professor, Govt. Ayurveda College, Kannur
5. Dr Janaki Krishnan T S,
Lecturer, Dr Padiar Memorial Homeopathic Medical College,
Chottanikkara
6. Dr Jayasree A
Professor, Pankaja Kasthuri Arurveda Medical College, Tvpm
7. Dr Jayasree M.N.
Assistant Professor, Govt. Ayurveda College,Tripunithura
8. Dr Jeeja Sasi
Assistant Professor, Govt Ayurveda College, Tripunithura
9. Dr Jula Rani
Assistant Professor, Govt Ayurveda College, Kannur
10. Dr Madhu PM
Assistant Professor, Govt. Ayurveda College,Tripunithura
11. Dr Mithra Das M
Assistant Professor, Govt Ayurveda College, Kannur
12. Dr Mini P,
Assistant Professor, Govt Ayurveda College, Tripunithura
13. Dr Noble.T.M ,
Assistant Professor, Govt. Ayurveda College, Kannur
8 An Introduction to Research Methodology

14. Dr Prakash Mangalasseri,

Associate Professor, VPSV Ayurveda College, Kottakkal
15. Dr Raghunathan Nair,
Professor, Govt.Ayurveda College, Kannur
16. Dr Rajitha R Warriar
Assistant Professor, Govt Ayurveda College, Tripunithura
17. Dr Ramesh Kumar
Professor, Govt Dental College, Kozhikode
18. Dr Remadevi S
Project Officer, KUHS School of Health Policy & Planning,
Thiruvananthapuram
19. Dr Remani KK
Professor, Govt Ayurveda College, Kannur
20. Dr Rosamma Joseph V
Professor, Govt Dental College, Kozhikode
21. Dr. Sethuraj KS
Assistant Professor, Govt. Ayurveda College, Tripunithura
22. Dr Shaju MK
Assistant Professor, Govt. Ayurveda College, Tripunithura
23. Dr Shylamma TM
Associate Professor, Govt Ayurveda College, Kannur
24. Dr Shobha Sundareswaran
Professor, Govt Dental College, Kozhikode
25. Dr Smitha Ramadas
Associate Professor, Govt Medical College, Thrissur
26. Dr Sreeja Sukesan
Professor, Govt. Ayurveda Collge . Kannur
27. Dr Sreedivya
Assistant Professor, Govt. Ayurveda College, Kannur
28. Dr Subin VR
Assistant Professor, VPSV Ayurveda College, Kottakkal
29. Dr Vivek P,
Assistant Professor, VPSV Ayurveda College, Kottakkal
30. Dr Vinuraj S
Assistant Professor, Govt Ayurveda College, Tripunithura
 An Introduction to Research Methodology 9

CONTENTS

1 Literature Review 11
Sethuraj KS, Shaju MK

2 Research Question 24
Sreeja Sukesan

3 Aims and Objectives 30

Jayasree A

4 Hypothesis 34
Mithra Das M

5 Research Designs 42
Remadevi S

6 Descriptive Study 51
Ajitha K & Jula Rani

7 Cross sectional Study 61

Madhu PM, Anjali Sivaram, Jayasree M.N

8 Case Control Study 78

Rosamma Joseph & Shobha Sundareswaran

9 Cohort Study 92
Remani K.K & Shylamma T.M

10 Randomised Controlled Trial 105

Raghunathan Nair & Noble T.M
10 An Introduction to Research Methodology

11 Diagnostic Test Evaluation 117

Rajitha R Warriar & Vinuraj S

12 Systematic Review and Meta Analysis 131

Prakash Mangalasseri, Vivek P & Subin V. R

13 Questionnaire Development 158

Sreedivya

14 Ethics in Medical Research 182

Antony George & Smitha Ramadas

15 Data Mangement and Analysist 201

Balakrishnan V.K.V

16 Qualitative Research 207

Janaki Krishnan T S & Anitha M A

17 Protocol Development 230

Mini P & Dr Jeeja Sasi

18 How to Write a Thesis 243

Ramesh Kumar
 An Introduction to Research Methodology 11

Chapter 1

Literature Review
Sethuraj KS, Shaju MK

Literature review is probably the most important preparatory

work for doing a research work and the three critical points
in doing a good research involves; (i) good literature review,
(ii) identifying appropriate research question and
methodology and (iii) optimal use of data. As the first step,
the literature review provides the primary basis on which
further intensive explorations into the subject at hand can
be taken up. It is an overview and evaluation of the writings
in a specific area of interest. The ultimate purpose of the
literature review is to bring together, analyze and thus utilize
the significant background knowledge on a topic. A good
literature review shows the scholar’s intellectual grasp of a
topic and evaluating the knowledge. It will help to familiarize
with important theorists, research groups and writings in
the area of interest as well as the history, vocabulary,
methods, and key variables used in the field of study.
When a researcher wants to know something, in an
understandable way, he proceeds to search previous works
done on the same subject. Earlier researchers were completely
dependent on written books and other manuscripts. With
the advent of the advanced technologies like internet and
12 An Introduction to Research Methodology

other sources of information, the process of literary review

has become extensive and easier. But the manuscripts and
hand written documents and text books have their own
importance even today. Systematic and extensive search of
the literature in an ordered and structured fashion from all
sources, where data or information is available in ones area
of interest is indispensable.
When a scholar decides to study and review a subject, he
has to proceed in an orderly fashion, keeping the research
question in mind. It should be the most important guide in
determining the relevant studies and the not so relevant ones.
Literature review should be written in sections, each one
focusing on an issue like historical aspects, definitions,
burden/prevalence of the problem, etiological aspects
including risk factors, interventions, cost effectiveness
aspects in a chronological manner. The extent of review
would also depend on the purpose of research; (i)
undergraduate dissertations, mostly descriptive studies
would need only simple review, (ii) post graduate thesis
mostly analytical studies would need critical appraisal of key
articles, (iii) doctoral research mostly intervention studies
would require minimum systematic review and if possible
meta analysis and (iv) post doctoral research a combination
of all modes of literature search.
The key elements of literature review are;
 A collection of reports of primary or original studies.
 Description, evaluation and integration of the materials
found in this collection.
 The writing of the findings in an appropriate manner
Steps in Literature Review
A literature review should not be a mere annotated
 An Introduction to Research Methodology 13

bibliography, where every article surveyed is summarized

briefly. The focus in a literary review is essentially on the
specific topic of interest and indispensably includes a critical
analysis of the relationship among different research works,
and their relation to the study at hand1. It should also help
provide a theoretical framework and rationale for the research
study. The essential steps involved in the development of a
literary review can be summarized as follows.
1. Deciding on a topic: After having arrived at the research
question, the central topic upon which the entire process of
research is based becomes clear. This decision on the central
topic initially reveals the current understanding of the
researcher on the same and the dark areas which needs to be
surveyed further.
2. Identification of the literature to be reviewed: This is
one of the tedious and time consuming areas of literature
review. The researcher should correctly identify the area of
his own interest and should be familiar with the online data
base and other sources which are relevant in the topic. Using
relevant databases, search for genuine literature sources using
Google Scholar, PubMed, Medscape etc. Some tips for
identifying suitable literature and narrowing search are start
with a general description from the database thesaurus or
one that you know is already a well defined descriptor based
on past work that you have done in this field. One needs to
experiment with different searches, such as limiting search
to descriptors that appear only in the document titles, or in
both the document title and in the abstract.
As you start searching, you will quickly find out if the topic
that you are reviewing is too broad and then redefine your
topic if needed. Try to narrow it to a specific area of interest
within the broad area that you have chosen. It is a good
14 An Introduction to Research Methodology

idea, as part of your literature search, to look for existing

literature reviews that have already been written on this topic.
As part of your search, be sure to identify landmark or classic
studies and theorists as these provide you with a framework/
context for your study. Do not forget to document the
reference properly as you select each article for review.
3. Analysis of the literature: The articles that have been
identified and located for the review have to be analyzed
and organized before beginning the process of writing.
Initially the articles are skimmed to get an idea of general
purpose and content. It is better done by reading through
the abstract, introduction and first few paragraphs and
conclusion of the article, which will give an over view of the
article. The articles thus selected are now grouped into
categories by topics and sub-topics. Along with reading the
articles, relevant notes should be taken down in defined
format. But one has to be consistent in taking down the
notes as one go along with the process of reviewing the
articles.
If you copy the exact words from an article, be sure to cite
the page number as you will need this, in case you decide to
use the quote when you write your review. Direct quotes
must always be accompanied by page references. If you are
accessing the article in a format that allows this; you can
copy and paste using your computer “edit —> copy —>
paste” functions. Although you may collect a large number
of quotes during the note taking phase of your review, when
you write the review, use quotes very sparingly.
4. Summary of literature and recording: The findings of
the survey of the articles should be summarized. In this
process, tabulated findings can be used effectively as they
help to over view, organize and summarize the findings in a
 An Introduction to Research Methodology 15

simple and easy way. If tables are included as a part of the

review, each must be accompanied by an analysis that
summarizes, interprets and synthesizes the literature that has
been charted in the table. The tables can be created using
the table feature within Microsoft Word, or can create it
initially in Excel and then copy and paste/import the Excel
sheet into Word once you have completed the table in Excel.
5. Synthesizing the literature prior to writing the review:
The findings of the literature review should be synthesized
to develop an outline of the final review. The review notes,
summary tables etc. help this process. Before starting the
process of writing down the review, the purpose of the same
should be reconsidered. The writing should provide an
overview of topic of interest, the understanding of the key
works and the concepts within a chosen area of focus. In
this process the researcher is gradually acquiring the skills
in reviewing and writing. Ultimately it provides a
foundation, upon which a final project is built up. It
demonstrates the domination of the researcher on his field
of study.
6. Writing the review: The broad problem area should be
identified and it is safer to avoid global statements in the
review. In the initial part the importance of the topic being
reviewed is indicated. Distinction between research findings
and other source of information are clearly stated. The
importance of certain studies are highlighted along with the
comment on the timeliness of the topic. The classical or
landmark studies are identified and cited as such. Replicated
landmark studies are mentioned and the results of replication
are indicated. Earlier literature reviews on the topic are also
discussed. The reviews on the topics that will not be
discussed in details should be referred to the reader.
16 An Introduction to Research Methodology

Comments on the inability to find related studies should be

justified. The inconsistent or widely varying results in the
previous studies should be mentioned. The point of view
which serves as the thesis statement is proposed early in the
review. The general structure of the literary review is such
that it integrates the key details of the literature and
communicates the point of view of the researcher. In
extremely long reviews sub headings may be used. The
judicious use of transitions helps to trace the arguments.
Sources of Evidence in Literature Search
There are two ways to approach the identification of
appropriate materials:
1. Citation pearl searching 2
In this process, we begin with an article on the particular
topic of interest we are exploring. This could be reading
from a class, a mention in a text book or something we are
found searching the library catalog or data base. The
reference found in that work is surveyed and a body of
literature that is relevant to our topic is gradually built up.
Using citations we can access a number of writings related
to the original work including articles cited by the original
author. After viewing these articles, go back to the database
and catalogs. More materials can be found by searching the
names of the authors as well as the technical term used in
the study.
2. Standard hierarchal search3
In this process we may begin with an essay in an encyclopedia
and explore the references found at the end of the essay.
Again taking the terms and authors we find to explore topic
using a variety of catalogs and databases. The other sources
 An Introduction to Research Methodology 17

include journal articles, books, government documents,

statistical data bases, newspapers, achieves, hand books,
dictionaries and annual reviews.
On the basis of the nature of the sources of information
they are generally classified as primary, secondary and tertiary
sources4.
 Primary sources: Primary sources are the original
materials on which subsequent researches are based. They
are from the time period involved and have not been
filtered through interpretation or evaluations. They
represent original printing and reported discovery or a
case or share new information. They are usually the first
time appearance of the results in physical print or
electronic format. Some of the examples are artifacts like
coins, fossils etc. audio recordings, internet
communications, interviews, articles published in peer
reviewed journals, original documents like birth
certificates, marriage certificates etc. photographs,
proceedings of meetings, conferences, Government
records, survey documents etc.
 Secondary sources: They are accounts written after the
fact with the benefit of hindsight and are generally
interpretations and evaluations of the primary sources.
Secondary sources are not evidences but commentary in
the discussions of evidences. Examples are bibliographies.
biographical works, commentaries and criticism,
histories, magazines and newspaper articles etc.
 Tertiary sources: They are information which are derived
and collected from the primary and secondary sources.
They include text books, chronologies, dictionaries,
directories, guide books, abstracts, manuals, indices etc.
18 An Introduction to Research Methodology

On the basis of the nature of the source of information they

are further classified as
 Print or physical sources: These include printed materials
like original manuscripts research journals, original
documents etc. These were the earlier sources of
information regarding research topics. The published
research materials served as primary sources the greatest
limitation of these sources were that there was no easy
access and easy transferability. The cost involved in
procuring them also often served as an impediment in
their access.
 Electronic sources: The advent of electronic sources have
revolutionized the process of literary review to a great
extent. The access to vast treasures of knowledge became
literally as simple as the click of a button. This includes
video recordings, audio recordings, magnetic tapes
compact disc, internet etc. The use of the electronic
media as a popular means of mass communication
immensely contributed to the popularization and
enhanced accessibility to the research topics. Now a days
the internet with its wide variety of information transfer
such as PubMED, Medscape, Google scholar etc. have
made the process of access to the original research data
simple, easy, effective and free from manipulations.
PubMED is the US national library of medicines premier
search system for health information. It is available free on
the internet with over 24.5 million citations. Sophisticated
search capabilities including spell checker, advanced search
builder and special tools for searching for clinical topics
assistants in finding search terms using MeSH data base etc.
are some of the features of PubMED.
 An Introduction to Research Methodology 19

Elements of Literary Review

In the literature review, details of the following aspects should
be reviewed and critically commented.
Importance of the problem: One of the indented goals in a
literary review is to highlight the core problem and the
assessment and acknowledgement of the background in the
community that has lead to the generation of the problem.
Rationale: After having identified the problem the next step
involved is the statement of the hypothesis where the
rationale of the proposed study is stated. It states methods
proposed in research for tackling the identified problem with
the statement of logic behind it. Rationale behind the
proposed research gives an idea of the way of approaching a
problem and subsequently devising an effective way out of
it. It anchors the thought process of a researcher in the
background that has created the core problem. The rationale
thus serves as a frame work upon which the research has
been conceived in order to overcome the identified problem
at hand.
Background: The back ground gives an idea about the
environment that has generated the core problem and the
methods and methodologies that have to be adopted in
winning over the same. The scholar must be familiar to the
topic and work he is going to do. Here we find the
importance of sufficient compilations giving detailed
information and analysis of the proposed argument or
problem and the methods and methodologies adapted to
get the result. It requires proper interpretation of the research
done before and citation backing the research done.
Literature review is a way of ensuring that one is not
repeating what has already been done but that you are
20 An Introduction to Research Methodology

improving on what has already been obtained.

Mentioning prior studies on the topic: The prior studies are
real leads to the works that has already gone in to the topic
under consideration. It throws light in to the current state
of the topic with regards to possibilities as well as limitations.
It serves as a guideline providing direction in to the path
that has to be adopted and the one that has to be avoided in
the light of the experience acquired. The ways and means
adopted by the predecessors may be adopted in word and
spirit but that has to be acknowledged by properly citing
them.
Relating with research question: Mere mentioning of the
articles associated to the topic could not serve the purpose
of a literary review. They have to be analyzed and synthesized
to the research question. This is one of the critical
contributions of the researcher from the literary point of
view. How the present study is related to the previous works
in terms of theoretical grounding and how the practical
methods utilized earlier can help improvise the situation at
hand are some of the questions that are answered by relating
the findings to the research question.
Statement of hypothesis and null hypothesis: The hypothesis
devised from the detailed study of the literary review helps
as a touchstone for the further steps in the study. The
hypothesis generation is an essential component of every
research. Usually there is a null hypothesis which is a
statement of null association and an alternate hypothesis
which is a statement of a positive association. It is the null
hypothesis that is tested and in terms of research findings
using appropriate statistical methods. On occasions of true
associations, the null hypothesis is rejected and the alternate
hypothesis becomes valid.
 An Introduction to Research Methodology 21

Preparation of a standard question: The literary review helps

in molding a problem at hand into a true research question,
it is a skill that presents ones problem at hand in a form
that is self-explanatory and self-revealing. The settings, the
premises, methods proposed, way of administration, the
variable measured, statistics used and the methodology
adopted are essentially mentioned in the research question.
Writing the aims and objectives: The aim of a study indicates
the main goal attempted there in the study whereas the
objectives usually represents the associated achievable gains
of the study. The statement of the aim and objectives in the
literary review helps focus the further course of the research.
Outcome measurements: The statement of outcome
measurement reveals the research statistics and the
methodology that has to be adopted to attain the result. It
focuses the researcher’s attention to the measureable as well
as non measureable variables involved.
References 5: The authenticity of the literary review is
enhanced by the use of references citing their source author,
time period etc. Without proper references even the most
reliable piece of information becomes invalid in the terms
of research. There are various accepted modes of referencing
the data of which the most accepted ones are the Harward
and Vancuover referencing styles. The two styles differ
mainly in the way references are presented in the running
text: in the Vancouver style, references are identified by
Arabic numerals; in the Harvard system, references are
identified by the name of the author(s) and the year of
publication.
The Vancouver referencing style: It is mostly used in the
health and physical sciences and is numerical and each piece
of work cited within the text is identified by a unique Arabic
22 An Introduction to Research Methodology

number. The numbers are assigned in the order of citation.

If a piece of work is cited more than once, the same citation
number must be used. This number is commonly enclosed
in round brackets (1), but it can also be written inside square
brackets [1], as a superscript1, or as a combination of brackets
and superscript [1]. The name of the author(s) may or may
not be present. A comma is used to separate non – inclusive
(1, 6, 11) numbers and a hyphen to indicate a series of
inclusive (4-7) numbers while citing multiple references. The
page numbers are included while citing specific ideas or
direct quotes. A list of references must be provided at the
end of the scientific text. This list must include the full
information for all the works cited in the running text. The
entries in the reference list are placeed in the same order in
which they were cited in the text.
Harvard referencing style: The Harvard referencing style is
used mostly in the humanities and social sciences. The works
cited in the running texts are identified by the name of the
author(s) and the year of publication. For in-text citations
that are part of the sentence use brackets to enclose only the
year of publication. If you put the entire in-text citation
within brackets, no second pair of brackets is needed for
the year. For works with more than two authors, use et al.
after the name of the first author. Citations on different
articles by the same author in the same year are distinguished
by using the small Latin letter. Unlike in the Vancouver
system, the author-date in-text citations are part of the text
and are therefore placed before any commas and periods.
Different journals have their own styles for citations and
references. Always ensure the journal style before submitting
a paper for review. This makes a good impression and
improves the chances that your paper be accepted.
 An Introduction to Research Methodology 23

Plagiarism6: The term to plagiarize literally means to steal

or pass it off as one’s own or to use without crediting the
source to present as a new and original idea derived from an
existing source. The blame of plagiarism is often encountered
by a researcher who does not go about his work according
to the standard stipulations of referencing. This can be
overcome by adopting standard referencing techniques,
acknowledging the author, source and time etc.
Bibliography: Here the sources of information used in the
literary survey are presented in the chronological order of
the appearances in the literary survey. The source, author,
time of publishing etc. are cited here. The use of bibliography
gives greater authenticity to the data utilized.

References
1. Galvan, J. Writing literature reviews: a guide for students of
the behavioral sciences, 3rd ed., 2006. Glendale, CA: Pyrczak
Publishing.]
2. h t t p : / / o n l i n e l i b r a r y. w i l e y. c o m / d o i / 1 0 . 1 0 0 2 /
9781444303599.ch10/summary
3. https://en.wikipedia.org/wiki/Hierarchical_and_recursive_
queries_in_SQL
4. http://www15.uta.fi/FAST/FIN/RESEARCH/sources.html
5. http://site.uit.no/english/writing-style/citationstyles/
6. http://www.plagiarism.org/plagiarism-101/what-is-
plagiarism/
24 An Introduction to Research Methodology

Chapter 2

Research Question
Sreeja Sukesan

Formulation of a valid research question is a challenging

task that a researcher encounters when he initiates a research
project. There are various factors which influence the
researcher to develop a research question. Probably it is the
experience from day to day clinical practice from where he
get inspired about an unexpected outcome or a quick relief
compared to acceptable treatment protocol, which in turn
encourage the researcher to step into a related research and
formulation of research question. The above said situation
which stimulate the researcher to proceed with his plan can
be the motivation to identify a research problem. It is the
eagerness of the researcher which helps to identify such
issues.
Research problem
Research problem is a situation in need of solution,
improvement or alteration or a discrepancy between what is
existing or partly known and the truth.
Sources of ideas of research problem
 Clinical experience
 An Introduction to Research Methodology 25

 Peer interaction
 Literarature review
 Theoretical knowledge.
These are the common sources from where the idea of
research or a research problem ‘embarks in the mind of
researcher. For eg, There is a huge rush of elder age group in
the ophthalmology department of Govt. Ayurveda college
Tripunithura for application of Elaneerkuzhambu in the eye
(topical ayurvedic medicine for improving vision),and most
of the patients are regular visitors for more than 5 years. A
lot of queries arise in the mind of researcher in this regard.
 Why these elder age groups regularly use this
combination for many years?
 Whether this medicine is effective to prevent the
development of cataract, which is a common cause of
blindness in old age?
 Is this combination effective in blindness due to other
reasons?
 Is this combination more effective than standard topical
modern medicine drops?
For the conduct of good research, these questions should be
modified and converted into research questions that can be
reasonably answered.
Definition of Research question
1. Research question is an interrogative statement, that
focus on what variables or concepts are to be described
and what relationships might exist among them.
2. Research question is an uncertainty about a problem that
can be challenged, examined and analysed to provide
26 An Introduction to Research Methodology

useful information. A well formulated research question

is the first step while initiating a quality research project.
There are well-known strategies to develop research question.
Before developing an refined research question or just after
a primary question strike your mind, it could be thoroughly
analysed using the FINER criteria, so that it will enhance
the practical implementation of research work and may
catalyze the whole research process.
The FINER criteria
F- Feasibility
I- Interesting
N- Novel
E- Ethical
R- Relevant
Feasibility: The feasibility of research is the first important
thing to assess while initiating a research. For a beginner
the feasibility can’t be properly assessed without the help of
a guide especially when the work is novel. In this regard the
researcher can depend on preliminary pilot research work.
The feasibility of research should be checked in terms of
availability, accessibility and affordability of different kind
of resources like, intended number of samples from the
population of interest, infrastructure and facilities, number
of research staff needed, funding for the project etc.
Interesting: It is the right of the researcher to select a topic
of his/her interest, at the same time the research topic may
attract colleagues who work in the same field. An important
research question may not be always interesting. Hence the
skill of the researcher is pivotal in generating a research
question which may be of interest to others.
 An Introduction to Research Methodology 27

Novel: Before generating a research question, the researcher

should try to understand what already has been done related
to his proposed work. A good literature review will surely
help the researcher in this regard. The researcher should try
to maintain ‘novelty’ in the research topic but that doesn’t
mean a work which is already done can’t be repeated. The
same work can be repeated in different setting, using
different methodology so as to bring out any novel outcome.
Ethics: A research work can only be proceeded when it is
ethically right. While planning and executing a research
project, ethical issues that can arise during the work should
be cleared so that it is easy to get the ethical sanction for the
project.
Relevant: All relevant issues related to the problem should
be considered while selecting a research question, particularly
the outcome of research. Otherwise it will be wastage of
resources like money, human resources, time etc.
PICOT
There are 5 different components for a research question
and a well oriented formulation of these components ensure
a very precise and specific research question which enhance
the conduct of a good research.
P - Population
I - Intervention
C - Comparison
O - Outcome
T - Time
Population: This means the subjects the researcher addresses
in his work. For eg. if the researcher wish to focus the effect
of drug in preventing the progress of cataract the elderly
28 An Introduction to Research Methodology

(>60 yrs) having some diminished vision due to cataract

and having no other pathology in the eye is the population
of interest.
Intervention: Can be a treatment, a diagnostic method, an
‘educative questionnaire’, anything which the researcher uses
in his research to produce the outcome variable. Here the
researcher wishes to estimate the quality of vision by
providing ‘elaneerkuzhambu’ an Ayurvedic combination. So
topical usage of ‘elaneerkuzhambu’ is the intervention here.
Comparison or control: The outcome of new drug should
be compared with a standard drug and then only it is possible
to estimate the role of new a drug in reducing burden of
problem, in the example a modern medicine eye drops which
is the standard drug in use by ophthalmologist will be the
control.
Outcome: It is what the researcher is expected to get during
his research process, here the improvement/non progression
of vision impairment is the outcome.
Time: It is the expected period within which the researcher
wishes to get the outcome, here a minimum of 1year is
necessary to get the difference in outcome; so one year can
be considered as the time.
When we analyse the above questions it is clear that all of
them lack the criteria of PICOT. Rearranging these questions
by adding these component it is possible to generate an
adequate research question, which ultimately help the
researcher to develop a correct research design so as to
conduct a quality research.
Let us try to formulate a research question from the above
example.
 An Introduction to Research Methodology 29

Formulation of Research Question:-

Consider the above question “Whether ‘Elaneerkuzhambu’
is more effective than Modern medicine drops”? Let us
dissect this question to find out whether PICOT criteria is
followed or not?
 Intervention is mentioned but no specificity maintained
regarding the dose of drug, how many times in a day it
should be administered, how long it should be continued
etc.
 No indication of population
 Control group is mentioned but again no information
about quantity, timing and duration of drug.
 Outcome and Time not clearly mentioned.
Converting the above question an adequate research question
using PICOT would be;
Does application of one drop of ‘Elaneerkuzhambu’ once in
a day (6 am) regularly for one year is more effective than
appropriate dose of catlol in preserving the vision of patients
with immature cataract in age group between 60–70 years
who presented at the ophthalmology department of govt.
Ayurveda college Tripunithura?
Here all components of PICOT criteria are well represented
and now it’s a good research question; it clearly reflects the
study design and the methodology of research to be carried
out. A well-thought-out and focused research question leads
directly into our hypotheses.
30 An Introduction to Research Methodology

Chapter 3

Aims and Objectives

Jayasree A

Formulation of aims and objectives in an appropriate manner

is most important in a study because it determines the scope,
depth and overall direction of the research. The development
of a research question, hypothesis and formulation of aims
and objectives are the key step for the successful design of a
research study.
Aim and objectives should be closely related to the research
problem. In a research protocol the position of the aim and
objective is in between research question and Hypothesis.
Aim is the overall goal to be achieved by the study. Objectives
are the steps to achieve the aim. Both should be interrelated.
There would be one research aim and several research
objectives to facilitate the achievement of this aim. The aim
is what one want to achieve and the objectives describe the
steps how one is going to achieve that aim.
Objectives can be both general and specific. The general
objective is the first level of specification of the aim from
which it derives. General objectives are broad statements
which point out what one hope to accomplish at the end of
research. It should reflect the aspirations and expectations
 An Introduction to Research Methodology 31

of the research topic.

Specific Objectives are derived from a given general
objective. It emphasize on recent outcomes or immediate
effect of the study. These are the practical steps and tasks
implemented to meet the general one, numbered so that
each one stands out as a separate character and should be in
measurable terms. It is advisable and possible to break down
a general objective into smaller logically connected parts.
SMART
One way to develop well-written objectives is to use the
SMART criteria. Specific objectives should systematically
address the various research questions. Specific objectives
should be SMART which stands for specific, measurable,
achievable, realistic and timely. They should specify what
one will do in the study and where and for what purpose.
Specific target a specific area
Measurable quantify or at least suggest an indicator of
progress.
Achievable feasibility of the study in achieving the
outcome
Realistic state what results can realistically be achieved
Time bound stipulated time period should be clearly
mentioned
Specific objectives can be presented as primary and secondary
objectives. Main objective to be fulfilled is primary and the
other associated ones are secondary. Often secondary
objectives explore and interpret the outcome of the primary
objectives.
Significance: Objectives always direct the researcher to
32 An Introduction to Research Methodology

follow a correct and proper path throughout the study. It

helps to focus on the study and assists in avoiding the
collection of data which are unnecessary for the research.
Besides, it can help to organize the study in clearly defined
parts or phases. Further, the objectives will facilitate the
development of a good research protocol, study design and
help out to decide the sample size, collection, analysis,
interpretation and utilization of data. Based on objectives,
items are framed in questionnaire, tables are generated,
results are interpreted and discussed. Poorly focused
objectives can result in failure of proper evaluation of a
research.
Procedures for framing aim and objectives
Objectives should be brief and concise so that the research
can be completed within the stipulated period of time. These
must be interrelated, based on realistic goals, framed on the
resources and scope of research. They should give a clear
indication to others about the orientation and approach to
the given subjects. They should also provide the plans to
encounter the ethical and practical problems wherever
necessary. Objectives should be indicated by using action
verbs such as to determine, to compare, and to calculate etc
which are specific enough to be evaluated. Clearly mentioned
objectives are essential for the appropriate evaluation of the
research. Objectives are essential factors in the protocol,
synopsis, in the introductory chapters of thesis as well as in
the conclusion session.
Objectives of various research designs
The study designs selected depends on the objectives
prepared. Hence the objectives of each design have certain
characteristics. Pure or basic research aims at achieving deep
 An Introduction to Research Methodology 33

and new knowledge and insights. Objective of a descriptive

research is to describe the characteristic features of a
phenomenon. Hypothesis testing studies (analytical and
experimental studies) are aimed at determining the
association between the variables. The main objective of a
cohort study is related to incidence or prognosis. Treatment
effect is evaluated in controlled trials. In a descriptive cross-
sectional study the objective is to estimate the prevalence
and to describe the characteristics of disease in time, place
and person. In analytical cross-sectional study the objective
is to study the association between the variables.
34 An Introduction to Research Methodology

Chapter 4
Hypothesis
Mithra Das M

Research generally starts with a problem. Research problem

usually in the form of research question is the precursor of
hypothesis. Hypothesis provides a specific restatement and
clarification of research question. In simple words,
hypothesis is a research question/objective which has been
re-worded in a testable form. Hypothesis needs to be
structured before the data gathering and interpretation
phase. It is very important to frame the right hypothesis.
Hypothesis is not used for all types of research. It is used
only in quantitative research. In qualitative research, the
researcher states research question and objectives and not
hypothesis.
Definition and meaning
Hypothesis means a supposition or proposed explanation
made on the basis of limited evidence as a starting point for
further investigation. By definition, hypothesis is a tentative
explanation that accounts for a set of facts and can be tested
by further investigation. Hypothesis states the researcher’s
assumptions on the relationship between the study and the
outcome variables.
 An Introduction to Research Methodology 35

Nature
1. Hypothesis is a clear statement of what is intended to be
investigated.
2. It is specified before the conduct of research and openly
stated in reporting the results. Finding data first and then
formulating the hypothesis is like match fixing.
3. It expresses relation between two or more measurable
variables.
4. It carries clear implications for testing the stated relation
5. It is testable – verifiable or falsifiable
6. It should be specific and not too general.
7. It is a prediction of consequences
8. It is considered valuable even if proven false
9. It states expected relationship between variables
10. It is stated in simple and concise form
Uses
1. At the start of an investigation, hypothesis stimulates
critical thinking thereby clearing confusions. Towards
the end of research, it again assumes significance as the
preposition to be accepted or rejected in the light of
finding.
2. A well-grounded hypothesis indicates that the researcher
has sufficient knowledge in the area to undertake the
investigation.
3. It provides a tentative explanation to the phenomena
and facilitates the extension of knowledge in an area.
4. It enables the researcher to objectively investigate new
areas of discovery. Thus, it provides a powerful tool for
the advancement of knowledge.
36 An Introduction to Research Methodology

5. Hypothesis gives direction to the research study process

– collection and interpretation of data
6. It provides investigator with a relational statement that
is directly testable in a research study
7. It provides clear and specific goals to the researchers.
These clear and specific goals provide the investigator
with a basis for selecting sample and research procedures
to meet these goals
8. It provides a framework for reporting conclusion of the
study.
Variables and its types
Testable hypothesis states the expected relationship between
the independent variable and the dependent variable in a
study. It is tested through statistical procedures. Variables
are defined in terms of qualities, properties, characteristics,
behaviour etc of individuals or groups/objects/situations etc
that changes or vary. As far as hypothesis is concerned, two
categories of variables that matter are dependent variable
(outcome variable) and independent variable (study
variable).
1. Dependent variable (outcome variable or responding
variable) – It is generally the effect, that is measured or
observed as a variable.
2. Independent variable or manipulated variable – variable
that is being examined or tested against the dependent
variable.
Example – If you spend more time participating in research,
then your research skills will improve.
Dependent variable – research skill
Independent variable - participating in research
 An Introduction to Research Methodology 37

Categories of hypothesis
1. Simple hypothesis / Complex hypothesis
2. Non Directional / Directional
3. Causal/ Associative
4. Null hypothesis / Alternative hypothesis (for statistical
testing).
 Simple hypothesis
Predicts the relationship between a single independent
variable and a single dependent variable.
Ex: Persons with uncontrolled diabetes have more chances
to develop retinopathy than those with controlled diabetes.
 Complex hypothesis
Predicts the relation between two or more independent
variables or two or more dependent variables.
Ex: Persons with uncontrolled diabetes have more chances
to develop retinopathy, neuropathy, nephropathy than those
with controlled diabetes.
 Non directional hypothesis
It is a type of alternative hypothesis in which no definite
direction of the expected findings is specified. The researcher
may not know what can be predicted from past literature. It
may read “…. There is a difference between…”
Ex : There is a difference in the anxiety level of children of
high IQ and low IQ.
 Directional hypothesis
It is a type of alternative hypothesis that specifies the
direction of expected findings. Sometimes directional
hypothesis are created to examine the relationship among
variables rather than to compare groups. Directional
hypothesis may read “…is more than…”, “… will be less
38 An Introduction to Research Methodology

than…”
Eg: Children with high IQ will exhibit more anxiety than
children with low IQ
 Associative hypothesis
Predicts an associative relationship between independent
variable and dependent variable. When there is a change in
any one of the variable, change occurs in the other.
Ex: Persons practicing yoga have more concentration than
those who do not practice yoga.
 Causal hypothesis
Predicts a cause and effect relationship or interaction
between independent variable and dependent variable. This
hypothesis predicts the effect of independent variable on
the dependent variable.
Ex: Persons practicing yoga daily have more concentration
than those who practice yoga less than 3days a week.
 Null Hypothesis
Null hypothesis is the statement that there is no actual
relation between variables. It is designated as HO or HN. It
states the opposite of what the experimenter would expect
or predict. The final conclusion of the investigator will either
accept a null hypothesis or reject a null hypothesis in favour
of alternative hypothesis. Not rejecting null hypothesis does
not mean it is true. It is just that there might not be enough
evidence against null hypothesis. It is also called the statistical
hypothesis because this type of hypothesis is used for
statistical testing and statistical interpretation.
Ex: There is no significant difference in the concentration
level of children in kindergarten, with refractive errors, using
spectacles and those not using spectacles.
 An Introduction to Research Methodology 39

 Alternative Hypothesis
Alternative Hypothesis is the statement that there is
relationship between two variables. It suggests a potential
outcome that the researcher may expect. It is designated by
H1 or HA. It comes from prior studies or literature review. It
is established only when null hypothesis is rejected. Thus
alternative hypothesis relates to the statement to be accepted
if / when null hypothesis is rejected.
Ex: There is significant difference in the concentration level
of children in kindergarten, with refractive errors, using
spectacles and those not using spectacles.
Framing a hypothesis
A clearly stated hypothesis includes the variables to be
manipulated or measured, identifies the population to be
examined and indicate the proposed outcome of study. So,
to be complete, the hypothesis must include three elements,
1. Variables
2. Population
3. Relation
Ways to express hypothesis:
1. Suggest possible events. Eg: Concentration will improve
after practicing yoga
2. Suggest relationship between specific exposure and health
related event
Ex: A high cholesterol intake is associated with the
development of coronary heart disease
3. Suggest cause effect relationship
Ex: Cigarette smoking is a cause of lung cancer
40 An Introduction to Research Methodology

Characteristics of a good hypothesis

1. Conceptual clarity – It should consist of clearly defined
and understandable concepts. All definitions should be
commonly acceptable and easily communicable
2. Empirical referents – Variables used should be empirical
realities. If they are not, it would not be possible to make
observations and ultimately affects testing.
3. Objectivity – must be objective. It should facilitate
objectivity in data collection and keep the research
activity free from researcher’s value judgment.
4. Specificity – it should be specific, not general. It should
explain expected relation between variables- place,
situation, and operation.
5. Relevant – relates to the problem being studied as well
as objective of study
6. Testability – Should be testable and should not be a moral
judgment. Must be directly or indirectly observable,
measurable and verifiable. The researcher can set up a
situation that permits one to assess if it is true or false.
7. Consistency – should be consistent with an existing body
of theories, research findings and other hypothesis. It
should correspond with existing knowledge
8. Simplicity – should be formulated in simple and
understandable terms. Should require fewer conditions
and assumptions.
9. Availability of techniques – researcher must make sure
that the methods are available for testing their proposed
hypothesis. Either techniques are already available or
researcher should be in a position to develop suitable
techniques.
 An Introduction to Research Methodology 41

10. Purposiveness – Must formulate only purposeful

hypothesis which has relevance with research problem
and objectives.
Testing a hypothesis
Hypothesis tests are procedures for making rational decisions
about the reality of effects. It is a 4 step procedure that
involves
1. Stating the hypothesis ( null or alternative)
2. Setting the criteria for a decision
3. Collecting data
4. Testing the null hypothesis
Negative findings
Even if hypotheses are not confirmed, they have use.
Negative findings are as important as positive ones. It acts
as a guiding factor for future research in that field.
Hypotheses cannot be proved or disproved; but only
supported or not supported.
Critical evaluation of hypothesis
1. Is it formally stated?
2. Does it clearly emanate from the problem and purpose?
3. Does it link to the study frame work?
4. Is it testable in this study?
42 An Introduction to Research Methodology

Chapter 5

Research Designs
Remadevi S

Epidemiology is “the study of the distribution and

determinants of health related states or events in specified
populations and the applications of this study to the control
of health problems”( John M Last 1988). As the definition
implies the study methods in epidemiology are used to obtain
valid data to answer research questions pertaining to the
(i)frequency of a problem (quantification in the form of rates
and ratios),(ii) distribution (answer questions of who, when
and where), (iii) determinants (relevant factors contributing
to the problem) and (iv) also to test the efficacy/effectiveness
of interventions to control the health problems. As these
questions are concerning human health, it can only be
answered by research on human beings, which is not easy as
in laboratory science. Despite the limitations on human
research, epidemiologists have developed methods and rules
for obtaining evidence about a research question. There is a
hierarchy of evidence from simple description of what
happened, to carrying out controlled trials. The selection of
a particular design depends on the question that one wants
to answer. From an epidemiological point of view the type
of research study is very important since there are a number
 An Introduction to Research Methodology 43

of advantages and disadvantages to various types.

Research designs can be broadly divided into two:
Observational and Experimental
Types of Research Designs
Two main categories of research design are observational
studies and experimental or intervention studies. In
observational studies, the investigators stand apart from
events taking place in the study - simply observe and record.
In experimental or intervention study, investigators
introduce an intervention and observe the events or outcome.
Observational studies
Observational studies can be descriptive or analytical
Descriptive studies
This is the first phase of an epidemiological study.
Observations are done on various aspects of illness among a
population like time of occurrence, symptoms, duration of
illness, place of residence and individual characteristics (age,
education, occupation etc). This will lead to formulation of
hypotheses regarding disease pattern, aetiology and prognosis
which can be tested later on using an analytical type of
design. Census and clinical records are important sources
of information in this type of study. Case reports, case series,
cross-sectional and ecologic studies come under this category.
 Case reports: It is a careful detailed report of the profile
of a single patient or case. Occurrence of an unusual case or
rare event is studied as case report e.g. a case of 40 year old
woman developing pulmonary embolism within 5 weeks of
oral contraceptive usage in the year 1961.
 Case series: The individual case report can be expanded
to a case series. Here a series of homogenous cases are studied
44 An Introduction to Research Methodology

and the characteristics described. No etiological relationship

can be found out as there is no comparison group. For
example, one can review the case records of all patients
diagnosed as endometrial carcinoma and look for the history
of hormone replacement therapy (HRT). You may find that
majority of women who developed endometrial cancer had
a history of HRT. Still the result is not conclusive as we do
not know the HRT use among those who do not have
endometrial carcinoma. Hence further studies are needed
to prove the relationship.
 Cross-sectional studies: Cross-sectional studies provide
information concerning a situation at a given time. In these
studies, the status of an individual with respect to the
presence or absence of exposure and outcome/disease is
studied at the same point in time. Usually involve collection
of new data and these studies are also known as prevalence
studies. They are not feasible for studying rare conditions
and also not ideal for studying rare exposures.
Cross-sectional studies can be descriptive or analytical in
nature. If the information collected is purely of a descriptive
nature, not involving comparison of groups based on
exposure or outcome, it is a descriptive cross-sectional study.
E.g. A study to find out the proportion of obese children in
high school classes in a city. If the data are analyzed to
demonstrate differences either between exposed and non-
exposed groups with respect to the outcome or between those
with and without outcome with respect to the exposure, then
it is an analytical cross-sectional study. E.g. Factors associated
with obesity among high school students in a city.
 Ecologic studies: Here the unit of study is not individuals,
but populations. Data for exposure and outcome obtained
from different groups or populations are compared. Since
 An Introduction to Research Methodology 45

ecological studies refer to the population rather than

individual, it is not possible to link an exposure to the
occurrence of disease in the same person. Ecological studies
also can be descriptive or analytical.
Analytical studies
These are studies used to test hypotheses concerning the
relationship between a suspected risk factor and an outcome
and to measure the magnitude of the association and its
statistical significance. The cardinal feature in an analytical
study is the concern whether an exposure (risk factor,
independent variable or predictor variable) causes the
outcome. Analytical studies look for causal associations. It
includes two observational studies; cohort and case control
studies.
 Cohort studies: These studies are also known as
longitudinal studies or incidence studies. Here a group of
persons (cohort) are identified based on their exposure status,
who are free of the disease or outcome of interest, but differ
on a certain exposure, are selected. They are followed up in
an identical manner until they develop the disease or
outcome of interest. The incidence of the disease can be
measured and relative risk calculated. Relative risk is the
incidence of disease in the exposed / Incidence of disease in
those not exposed. Eg. Risk of passive smoking among
pregnant women on low birth weight of babies.
The non-exposed or comparison group may come from
within the initial group of subjects studied or may be a
separate population altogether. The cohort study can be
historical in which the group, e.g. employees in a factory
are enrolled in the study from a particular period in the
past, from company records, and outcome measured in the
46 An Introduction to Research Methodology

present. This may be referred to as a retrospective cohort

study.
In cohort study, exposure can be measured without bias and
absolute risk established. But often it is not time and cost
efficient as it requires subjects in large numbers for
uncommon outcome or rare diseases and follow up over a
long time.
 Case-control studies: In this type of study group of
subjects with disease/ outcome of interest (cases) and group
of subjects without disease/outcome of interest (controls)
are identified . Cases are compared with controls for the
presence or absence of hypothesized risk or exposure factors.
Since the design involves looking retrospectively for the
presence of risk factors after the disease has occurred, it is
sometimes called as retrospective study. Here we are getting
similar information to a cohort study, but much more
efficiently. The fundamental difference is that subjects are
defined by disease status and not by exposure status. Case
control studies are often used in situations where the disease
is a rare one or has long latent period. There are two major
potential problems in case control studies which may
produce substantial bias. They are related to (i) selection of
cases and controls and (ii) assessment of exposure which
has occurred in the past. As a general rule both cases and
controls should be selected independent of exposure and
the controls should come from the same source population
as the cases. Analysis in a case control study is done by
calculating the Odds Ratio (OR); by comparing the odds of
exposure among cases with that among controls. An example
of a research question suitable for a case control design would
be – ‘Is post- menopausal hormone replacement therapy
associated with increased risk of endometrial cancer’?
 An Introduction to Research Methodology 47

Case control studies are valuable for studying rare conditions

and relatively inexpensive. It yields odds ratio which is
usually a good approximation of relative risk. However, the
design is subjected to potential biases from different sources.
Experimental studies
Experiment- “Set of observations, conducted under
controlled circumstances, in which the scientist manipulates
the conditions to ascertain what effects such manipulation
has on the outcome “(Rothman, 1986). An experimental
study may be controlled or non-controlled. But, for a
definitive answer, we need a control group who do not get
the intervention under study. A controlled experiment can
be randomized or non-randomized. In a non-randomized
trial, we are not sure whether the difference observed in the
outcome is because of the intervention or because of the
difference in the characteristics of the two groups. With
randomization, only chance determines the assignment of
subjects to study groups. Whenever it is ethical and practical,
a randomized design should be considered in controlled
intervention studies. A before-after study is a method of
control in which the measurements are taken from
experimental subjects before and after intervention and
evaluation of outcomes are made by comparing the pre and
post intervention measurements. A crossover study is a
special design of controlled experimental study that is
sometimes used in clinical trials. In this design half the
participants are assigned to new treatment and then switch
to standard treatment/placebo, while the other half does the
opposite.
Trials may be done for various purposes;
(i) therapeutic trials (e.g. drug, surgical procedures)
(ii) prophylactic trials (e.g immunization, contraceptives)
48 An Introduction to Research Methodology

(iii)safety trials (e.g. side effects)

(iv)risk factor trials (e.g. oral contraceptives).
The interventions in experimental studies may include drugs,
surgical procedures, risk factors, lifestyle habits, medical
counselling, communication approaches etc. Experimental
studies are conducted as clinical trials, field trials or
community intervention trials. Field trials are intervention
studies usually done on healthy subjects e.g. testing the
effectiveness of a vaccination or health education package.
Community intervention trials are similar to field trials, but
allocation is made to groups rather than individuals. In
clinical trials, interventions are usually among patients in a
controlled setting e.g testing efficacy of a drug.
A clinical trial conducted in a strictly randomised manner
using comparable controls is termed as Randomized
Controlled Trial (RCT). This is a standard method of
answering questions about the effectiveness or efficacy of
different interventions. In this study design, the subjects of
experiment are allocated randomly into experimental and
control group, followed for a specified period of time under
strict conditions and outcome of intervention is carefully
determined. RCTs provide the best chance of obtaining
strong evidence of a cause and effect. Randomisation
ensuring comparability of two groups, standardisation of
eligibility criteria, the intervention, and outcome are other
highlights of this design. However many research questions
are not suitable for this design due to ethical issues or rarity
of outcome.
Hierarchy of Research Designs
Different types of research designs are not considered equal
in strength. The following figure shows the hierarchy of
 An Introduction to Research Methodology 49

different study designs based on the strength of evidence

they produce.

Even though the strength of evidence differs for different

designs, the choice of a particular design depends more on
the research question or objective of the study. The following
table summarises the study designs in relation to the
objective/s.
Objectives and Study designs
Objective/s Study Design
Describing clinical features,
prognosis, disease pattern Case series
Burden of disease/ Prevalence Cross-sectional study
Incidence/determine prognosis Cohort/Longitudinal study
Association between exposure Cross-sectional
and outcome
Causal relationship/identify
risk factors Case control or Cohort
Efficacy/Effectiveness of an Randomized Controlled
intervention Trials (RCT)
50 An Introduction to Research Methodology

References
1. Last, J.M (ed) A Dictionary of Epidemiology. Second Edition.
Oxford University Press, 1988
2. Rothman, K.J. Modern Epidemiology . Little Brown &
Co.1986
3. World Health Organization. Health Research Methodology.
A guide for training in research methods.2nd edition. , Manila.
WHO regional Office for the Western Pacific,2001
4. Fathalla MF, Fathalla MMF . A Practical Guide for Health
Researchers. Cairo, WHO Regional office for the Eastern
Mediterranean Series; 30, 2004
 An Introduction to Research Methodology 51

Chapter 6

Descriptive Study
Ajitha K, Jula Rani

Epidemiology is the study of the distribution and

determinants of health-related states or events (including
disease), and the application of this study to the control of
diseases and other health problems1.
Epidemiology has three main aims 2:
 To describe disease patterns in human populations.
 To identify the causes of diseases (also known as etiology).
 To provide data essential for the management, evaluation
and planning of services for the prevention, control and
treatment of diseases.
Research studies are designed in a particular way to collect
information for answering the research question. Pursuing
the ways and techniques outlined in the research protocol
helps to meet accurate results. Following a sound research
protocol and appropriate study design is necessary to get
similar results by other researchers and only reproducible
results are more likely to have acceptance by researchers and
public.
An observational study is one in which information is
52 An Introduction to Research Methodology

collected without changing the environment (i.e., nothing

is manipulated) 3. Office of Human Research Protections
(OHRP) defines an observational study as “Any study that
is not truly experimental.” In human research, an
observational study can provide information about the
naturally occurring health status, behavior, attitudes or other
characteristics of a particular group. Or it is concerned with
observation of the distribution of a disease in a community,
with reference to time, place and person and identifies the
associated characteristics of the disease to formulate an
etiological hypothesis 4.
Descriptive study is one form of observational study, which
collects analyses and present data on a specific topic. Key
features of descriptive study are the absence of a comparison
group and non-intervention.
Procedure 5
Defining the population under study: This means specifying
the type of population under study, i.e., whether the entire
population of an area or a representative sample of the
population like children, pregnant mothers, industrial
workers etc are involved. The population must also be
defined in terms of place and time. For example, children
under 5 years, in an area under the jurisdiction of PHC, in
the year 2015. Thus the study population becomes
population at risk i.e. it becomes the denominator and helps
in calculating the rates, i.e. in measuring the disease
frequency.
Defining the disease under study: That means the disease
which is taken up for study has to be defined in such a way
that the epidemiologist should not only be able to identify
those with disease from those without the disease, but also
 An Introduction to Research Methodology 53

be able to measure it with accuracy. This is called operational

definition4.
The definition of the disease must be precise and valid. If
not it becomes the source of error. For example, leprosy is
not simply a case with hypo-pigmented patch/patches but
is defined as a case with hypo-pigmented patch/patches with
partial or total loss of sensation, with thickening of nerves
and demonstration of acid base bacilli in the skin smear
examination. Here the needs of a clinician and
epidemiologist may diverge. The clinician may not need a
precise definition of the disease, for immediate patient care.
But the epidemiologist needs a precise and valid definition
to obtain an accurate estimate of disease in a population to
identify those who have disease (say migraine) from those
who do not (any other head ache). If the definition is not
precise and valid, it would be a powerful source of error in
collection and presentation of data. Once established, the
case definition must be adhered to throughout the study4.
Describing the distribution of the disease with reference to
time place and person:
Time distribution: describing the time of occurrence/ onset
of the disease with reference to year, month, week, day, hour
of the day, season etc. This study often gives a clue about
the etiology of the disease or the predisposing factors, so
that preventive measures can be adopted.
There are 3 kinds of time trends or fluctuations:
a) Short term fluctuation: sudden occurrence of a disease
in a given area, and lasting for a short period, eg, an
epidemic disease.
b) Periodic fluctuation: occurrence of disease in a
community during a definite period, either in a particular
54 An Introduction to Research Methodology

season (measles and chickenpox in the early spring

season) or periodically in a cyclic form.
c) Long term fluctuation: changes in the occurrence of the
disease over a long period of time, several years or decades.
Place distribution: the pattern of occurrence of disease in
different places. This helps to compare the disease occurrence
from one country to another country and within the same
country from one state to another state, from rural to urban
areas and local areas.
Person distribution: Describing the distribution of a disease
in the community with reference to the host characters of
the persons affected. Eg. age, sex, occupation, marital status,
social class, behavior etc.
Age: Certain diseases occur more frequently in certain age
group. Example: Measles and Diphtheria among pre-school
children, Cancer in the middle age, Atherosclerosis among
elderly.
Sex: Certain diseases are more common among men,
Example: Lung cancer and coronary heart disease and some
are more common among women. Eg. Diabetes, obesity,
myxedema. Such a difference may be because of either
genetic constitution or cultural and behavior factors like
smoking, drinking etc.
Marital status: Cancer cervix is rare in unmarried compared
to married woman. Similarly mortality rates are high among
unmarried than married persons.
Occupation: Persons working in particular occupations are
exposed to particular types of risks, example: tetanus is
common among agricultural workers.
Social class: Diseases like hypertension, diabetes, coronary
 An Introduction to Research Methodology 55

artery disease are common among people of higher socio

economic class and diseases like malnutrition, rheumatic
heart disease and communicable diseases are common among
low socio economic class. Thus social factors like poverty,
ignorance, illiteracy play a very important role in the
development of disease in the community.
Behaviour: Behavior such as alcoholism, smoking, over
eating, multiple sexual partners etc. influence the
development of the disease. Thus the study of these risk
factors helps the epidemiologist to formulate an etiological
hypothesis.
Stress: Stress increases the susceptibility of an individual to
the disease and often exacerbates the symptoms.
Migration: Movement of people from rural to urban areas
has resulted in the spread of diseases from one place to
another. Migration is a challenge for the prevention and
control of the disease.
Measurement of disease
This means estimating the disease load or magnitude of
the problem in terms of morbidity, mortality, disability etc.
Mortality is measured directly in terms of death rates.
Morbidity is expressed in terms of incidence (longitudinal
study) and prevalence (cross-sectional study) rates.
 Incidence rate is defined as the number of new cases
occurring in a defined population during a specified period
of time.
Incidence rate=
Number of new cases of specific disease
during a given time period X 1000
Population at risk during that period
56 An Introduction to Research Methodology

 Uses of incidence rate: 1. To control disease 2. For

research into aetiology and pathogenesis, distribution of
diseases and efficacy of preventive and therapeutic
measures.
 Increase in incidence rate indicates failure or
ineffectiveness of current control programs. It suggests
the need for a new disease control or preventive program.
 A change or fluctuation in the incidence of disease may
also mean a change in the etiology of disease.
 Prevalence: the term disease prevalence refers specifically
to all current cases (old and new) existing at a given point
in time, or over a period of time in a given population.
Prevalence is determined by incidence and duration. That
is if the duration is long prevalence is high. Likewise if
the incidence rate is high prevalence rate is also high.
 Point prevalence: is the number of all current cases (old
and new) of a disease at one point of time in relation to
a defined population.
Point prevalence=
Number of all current cases of a specified disease
existing at a given point in time X 100
Estimated population at the same point in time
Making comparison with known indices
The observations are compared with different groups. By
making comparison between different populations and
subgroups of the same population; it is possible to arrive at
clues to disease etiology. This helps to find out the etiological
factors and also helps to identify the risk group so that
preventive measures can be adopted.
 An Introduction to Research Methodology 57

Formulation of etiological hypothesis

By studying the distribution of a disease in the community,
with reference to time, place and person the epidemiologist
can formulate an etiological hypothesis. A hypothesis is a
supposition arrived at from observation or reflection. It can
be accepted or rejected, using the techniques of analytical
epidemiology. The hypothesis must be correct and complete.
Eg chronic alcoholism causes cirrhosis of liver is an
incomplete hypothesis. Better statement would be drinking
200 to 300 ml of alcohol per day cause cirrhosis of liver
among 20 percent of drunkards after 25 years of exposure.
Thus a correct and complete statement of the hypothesis
helps the epidemiologist to test the hypothesis.
Types of descriptive study
 Case report
 Case series
 Cross-sectional study
Case report
It is a report on a single patient with an outcome of
interest6 or it is the detailed presentation of a single case
which is unusual or without usual presentation. A clinician
may come across a particular clinical manifestation or an
unusual laboratory finding or may find an interesting finding
in imaging technique. It can be documented and presented
as a case report. Case reports are extremely useful to share
the unusual experiences of a clinician with peers. They can
represent the first clues in identification of new diseases or
adverse effects of an exposure7. They are useful to public
health as they can give an interface between clinical medicine
and epidemiology7.
58 An Introduction to Research Methodology

Case series
It is a collection of individual case reports. A case series is a
report on a series of patients with a common outcome of
interest8. In this study design, a group of individuals with a
particular disease is studied and documented. No control
group is involved. Case study can be used to identify the
presence or beginning of an epidemic. A classic example is
case series study of five homosexuals who presented a rare
type of pneumonia which lead to the discovery of HIV.
An advantage of case series over case report is that a case
series can help formulate a new and useful hypothesis rather
than merely documenting an interesting medical oddity7.
However, its disadvantage is that it cannot be used to test
for the presence of a valid statistical association7.
Cross sectional study
Cross sectional studies may be descriptive, analytical or both.
At descriptive level, it yields information about a single
variable or about each of a number of separate variables in a
study population. Here the researcher explains the
phenomenon observed. The data observed is not compared
with another set of data. Cross sectional studies are used to
assess the burden of a particular disease (magnitude of disease
load) in a defined population. Eg: prevalence of myopia
among school children.
Cross-sectional studies are carried out at a point or over a
period of time. They are usually conducted to find out the
prevalence of the outcome of interest in a given population
at a given time. It is carried out as a population study (census)
or in a representative sample of population. In this way cross-
sectional studies give a ‘snapshot’ of the outcome and the
characteristics associated with it, at a specific point. The
 An Introduction to Research Methodology 59

information other than the outcome of study gathered at

the time of data collection helps to formulate hypotheses
about the risk factors of the disease. For example in a study
of hypertension, along with the measurement of blood
pressure, other data like age, sex, food habits, occupation,
body weight can be collected. Then we can find out how
prevention of hypertension is related to certain variables
simultaneously collected. Such a study tells about the
distribution of a disease in population than its etiology.
Although the cross-sectional study provides information
about disease prevalence, it provides very little information
about the natural history of disease or about the rate of
occurrence of new cases.
Uses of descriptive epidemiology4:
 Helps to know the extent or magnitude of the disease in
the community in terms of morbidity and mortality rates.
 Helps to know the distribution of the disease with
reference to time, place and person.
 Helps to formulate an etiological hypothesis.
 Helps to plan, organize and implement curative and
preventive services.
 Helps in doing research.
In short, descriptive studies are observational studies that
describe the patterns of disease occurrence in relation to
variables such as person, place and time. They are the first
step of an epidemiological study. Case report, case series,
cross sectional study are the types of descriptive studies. Case
report is the detailed presentation of a single case, where in
a case series a larger group of patients with a particular
disease have been studied. Cross sectional studies are
60 An Introduction to Research Methodology

observation of a defined population at a single point or time

interval.
References
1. http://www.who.int/topics/epidemiology/en/
2. http://www.aea.asn.au/about-us/what-is-epidemiology
3. https://ori.hhs.gov/education/products/sdsu/res_des1.htm
4. A.H Suryakantha.Community MedicineWith Recent
Advances: New Delhi: Jaypee Brothers Medical Publishers
(P) Limited; 2009.
5. J.K Park. Park’s Text Book of Social and Preventive
Medicine: 21st ed. Jabalpur: BanaridasBhanot Publishers;
2011.
6. http://www.medicine.ox.ac.uk/bandolier/booth/glossary/
Caserep.html
7. http://www.drcath.net/toolkit/descriptive.html
 An Introduction to Research Methodology 61

Chapter 7

Cross Sectional Study

Madhu PM, Anjali Sivaram, Jayasree MN

Cross sectional study is a type of observational research study

design in which a representative sample of defined
population is investigated to determine the level of a
particular attribute, such as a specific exposure, disease or
any other health-related event, at a particular point in time.
In this type of study, subjects are contacted at one point in
time and relevant data is obtained from them. On the basis
of this information, they are then classified as having or not
having the attribute of interest. A cross-sectional study
examines the relationship between disease or other health
related state and other variables of interest (social &
behavioural attributes and personal & demographic
characteristics) as they exist in a defined population at a
single point in time or over a short period of time. Since it
provides information on disease prevalence, it is also called
prevalence study. Here a range of variables are measured on
a single occasion as also the outcome of interest; hence also
called Instantaneous study.
Frequency of disease and risk factors can be examined in
relation to age, sex and ethnicity etc. Studies of this type are
able to provide rough idea about association between the
62 An Introduction to Research Methodology

study variables and outcome but cannot establish a cause-

effect relationship. The results from these surveys give clues
of the extent of the problem in a particular population at a
particular point in time, and also provide a basis for
designing proper public health measures. Survey is a valuable
method of obtaining information on the patterns of
morbidity of a population, and also on the utilization of
preventive and curative health services. Thus it helps to
establish the health priorities. On the basis of the report
generated after survey study, institutions or government can
take actions on planning and implementing health projects.1
When we arrange the research designs in to a hierarchy,
according to their relative significance in the evidence they
provide, Randomized controlled trial (RCT) and Meta
analysis of RCT are placed at the highest level followed by
cohort , case control studies and cross sectional studies.
Classification
The cross sectional study can broadly be divided in to two
main groups namely, Descriptive and Analytical2.
Descriptive cross sectional study
A cross-sectional study may be purely descriptive when used
to assess the frequency and distribution of a particular disease
in a defined population. In this type, the researcher tries to
detail the phenomenon existing at a particular point of time
in the defined population. Let us consider an example.
Recent researches have already proved that the atmospheric
air in Mangattuparamba of Kannur district is highly
polluted. As air pollution is a major triggering factor for
many diseases, the people living in this area are at the risk
of development of such diseases. Here, a population survey
that focuses on the health issues of the inhabitants of
Mangattuparamba certainly helps to assess the magnitude
 An Introduction to Research Methodology 63

of the problem. Here the researcher is not making

comparison of any particular attribute of this population
with that of any other group, but only assess the burden of
a particular problem in a defined population. Assessing the
burden or prevalence of asthma among children of 12-14
years selecting a random sample of schools across Kerala
may be another example.
Analytical cross sectional study
In analytical cross sectional studies data on prevalence of
both exposure and a health outcome are collected and health
outcome differences are compared between exposed and non-
exposed. For example, suppose a researcher collected data
regarding health status from a sample drawn from Anthoor
municipality (Mangattuparamba region is in this
municipality); then he categorized the subjects into two
groups based on their health status and compared with their
place of habitat (polluted Vs Non-polluted),he could
establish an association between poor health status of
subjects and habitation in Mangattuparamba where
atmospheric air is highly polluted. Such a study is an
analytical cross sectional study. Thus, analytical cross-
sectional studies may also be used to investigate the
association between a putative risk factor and a health
outcome. However this type of study is limited in its ability
to draw valid conclusions about any association or possible
causality since the presence of risk factors and outcomes are
measured simultaneously. It may therefore be difficult to
find out whether the disease or the exposure came first, so
causation should always be confirmed by more rigorous
studies. A cross sectional study provides preliminary
information which can help create a hypothesis for further
study.
64 An Introduction to Research Methodology

Important steps in a Cross-sectional study

• Framing a Research Question and objectives
• Deciding the study setting and population
• Sample Size calculation
• Selecting a representative sample
• Study and outcome variables – Definition and
measurements
• Data collection :usually by questionnaires or Interview
Schedules
• Data management: Analysis and interpretation
• Reporting
Sample size and Sampling methods
Sample size can be calculated using sample size tables or
statistical packages. It should be sufficiently large enough
to estimate the prevalence of the conditions of interest with
adequate precision. A detailed survey of the population of
interest (target population) may not be feasible in all settings.
So, a true representative group of people is to be selected
from the desired population. This is termed as sampling in
biostatistics. The data collected from the sample is considered
as the data of the target population and hence the outcome
of the study can be generalized to that particular population.
In order to draw a representative sample from the population
by making use of suitable sampling method, we need to have
a sampling frame, ie. a complete enumeration of the
sampling units in the study population. The sampling unit
may be an individual person, a household, or a school.
Electoral registers may be a suitable sampling frame for adults
but not for children. If the sampling frame is based on official
 An Introduction to Research Methodology 65

statistics, such as village register, the possibility of under-

representation of some groups like recent immigrants, the
homeless, and slum dwellers becomes high. A perfect
sampling frame may not exist for some groups such as ethnic
and other nomad groups. So, the sample frame should be
comprehensive, complete and up-to-date, to maintain
selection bias to a minimum.
In cross sectional studies, sample can be selected from the
population using any of the following sampling techniques3.
Simple random sampling
If the source population (a subset of the target population
or the target population itself, from whom the sample is
drawn using the sampling frame) is formed by 5000
individuals, each one should be assigned a unique number
between 1 and 5000. Random number tables can then be
used to select a random sample out of the total sampling
units who make up the source population. The first step is
to select a random starting place in the table. A random
direction also is selected. Then all the sequential digits found
on the table are to be recorded, stopping only when the
required sample size is reached. Alternatively, sequences of
random numbers can be generated by a calculator or a
software program. The sample will be constituted by the
sampling units which correspond to these random numbers.
The main feature of simple random sampling is that it is
relatively simple and most fundamental type of random
sampling. Its main limitation is that it can be opted only
when the population is relatively small and concentrated in
a small geographical area and where the sampling frame is
complete.
66 An Introduction to Research Methodology

Systematic sampling
This is an organized way of selecting the sample by arranging
the whole population in to some kind of sequence as in a
directory or in a series of index cards, or houses along a lane
or patients as they arrive at a clinic. Then we need to decide
sampling fraction. For example, we want a sample of 50
from a population of size 500. This will be a 1 in 10 sample
and the sampling fraction is 10. We choose a number at
random between 1 and 10—let us suppose it was 3. Then
starting at the beginning of the list, we select the sampling
unit number 3 and then every tenth subsequent unit. The
sample will include units 3, 13, 23...and so on.
The systematic sampling is convenient method of sampling.
It generally offers a good approximation to simple random
sampling provided that the intervals do not correspond with
any recurring pattern in the source population. Bias will be
introduced if some particular characteristics arise in the
sampling frame at regular intervals. Suppose, the population
were made up of a series of married couples with the husband
always listed first. Choosing every fourth person would result
in a sample constituted exclusively by women if one started
with the second or fourth subject or exclusively by men if
one started with the first or third. Similarly, if there are 10
houses in a lane, every 20th house on a list of addresses or
houses might be a corner house which does not share
common characteristics with the other houses.
Stratified random sampling
While doing a cross-sectional study, it can be noticed that
important subgroups of people show different views or life
experiences or health related behaviors. For example, males
and females differ in their health issues and may have
 An Introduction to Research Methodology 67

different views on how health services should be delivered.

In such a condition, a stratified random sampling becomes
useful. It is done by dividing the population into distinct
subgroups on the basis of some important characteristics,
such as sex, age or socioeconomic status, and selecting a
random sample out of each subgroup. Each subgroup thus
selected is known as a stratum, and a separate random sample
(simple or multi-stage) can be drawn from each stratum.
Multi-stage sampling
Many a times, it is not feasible or practical to draw a simple
random sample or a systematic sample from the whole source
population. This is either because of the unavailability of a
sampling frame and the effort involved in drawing one up
would be too great, or because the population is dispersed
over a very large geographical area. For example, it would
be too idealistic to try to draw a simple random sample of
1000 people from the whole population of India in order to
study the nutritional status of Indian Population. Even if a
proper sampling frame did exist, most of the sample would
live in different communities far away from each other, and
the time and expense involved in contacting them would be
too expensive. The solution is to use multi-stage sampling
as follows:
1. The population is first divided into regions and a list of
these first-stage units is taken. (In the above mentioned
example, we can consider the list of different states in
India)
2. A random sample of first-stage units is selected from this
list. (We select 4 states randomly)
3. In each of the selected first-stage units, a sampling frame
of the second-stage units is drawn up, and a random
68 An Introduction to Research Methodology

sample of these selected. (We select randomly 4 districts

from each state that are selected in the first stage)
4. This plan can be extended to several stages. (Later we
select 4 villages randomly from each district)
If no frame of households exists and it is not practical to
create one, some selection method has to be used which
ensures that the sample is as representative as possible.
Cluster sampling
Similar to multi-stage sampling, cluster sampling is another
feasible sampling technique which may be followed if wide
geographic area needs to be covered in the study. First of all
divide the area into a number of smaller non-overlapping
areas eg, wards coming under a block. Then randomly
select a number of these smaller areas (usually called clusters),
with the ultimate sample consisting of all (or samples of )
units in these or clusters. Cluster sampling, no doubt,
reduces cost by concentrating surveys in selected clusters.
But certainly it is less precise than random sampling.
Based on the setting, most appropriate sampling method
can be adopted. All the time, not everyone in the selected
random sample will end up participating in the study. Some
subjects may refuse to participate despite all reasonable
efforts; some others might have died or migrated from the
area. All possible efforts should be made to ensure a high
level of response and participation to minimize selection bias.
In cer tain conditions, sampling method becomes
impractical. This occurs when the population is so small or
when the countable disease is very rare. The sample selected
may not be a true representative in these situations. Scientific
sampling method may miss the real patients at times. So, it
is better to perform census studies in such situations. For
 An Introduction to Research Methodology 69

example,
1. While assessing the health issues among geriatric patients
above the age of 60 years in Pariyaram panchayath.
2. While evaluating the prevalence of sickle cell anemia
among the tribal population of Kannur district.
Selection of respondents is also very important in survey
study
Data Collection Tool
In a cross sectional study, questionnaire is the main tool in
collecting the data from their respondents. Questionnaire
is the list of survey questions asked to respondents. A
questionnaire is designed to extract specific information. It
serves four basic purposes to (i) collect the appropriate data
(ii) make data comparable and amenable to analysis (iii)
minimize bias in formulating and asking question and (iv)
make questions engaging and varied. Asking the right
questions in the right way to the right people in the context
of an apt research framework generates relevant, useable
information.
Questionnaire design is only a single step in the process that
ultimately leads to generating answers to research questions
of interest. After designing the questionnaire, it is good to
run a pilot test of the questionnaire to make sure it is
understandable and acceptable to the intended audience.
This can be done by administering the questionnaire to a
small group of persons from the intended target group and
then following up to obtain feedback on the questions (eg.,
how they were worded, whether the respondents could
understood them, whether the respondents felt comfortable
answering them) and on the questionnaire itself (eg. whether
it was too long, was there any potential barrier in getting
70 An Introduction to Research Methodology

good responses). Pilot testing also involves evaluation of

other attributes, namely, precision (reliability) and accuracy
(validity). These attributes are critical to developing a
questionnaire whose results are reproducible and that
provides the researcher with a good measurement of the
phenomenon or phenomena of interest. After incorporating
feedback from the pilot test, the questionnaire can be
finalized and it is ready to be administered to a sample from
the target population.
Conducting the survey
Field procedures for carrying out a survey not only comprise
the methods of data collection but they also include deciding
how to introduce the study to respondents, when to schedule
the initial and follow-up contacts, how to keep track of the
results of those contacts, how to follow up with non-
respondents, whether to provide remuneration to
respondents, whether to allow respondents to serve as proxies
for other people, and what documentation to provide for
these procedures. The survey can be conducted as personal
interview, telephonic contact, or mail interaction.
Personal interview: Research to determine the times that
yield the highest rates of interviews with respondents suggests
that weekday evenings and weekends, especially Sundays,
are the best times to contact households. Researchers may,
however, encounter difficulty in meeting with respondents
who are too busy, unavailable, or physically or linguistically
unable to participate whenever contacted. Special strategies
may then be needed to encourage them to respond to the
survey, especially in urban, high-density areas and in low-
income communities4.
Telephonic contact: Many a times, direct personal interview
 An Introduction to Research Methodology 71

may not be practical in all cross sectional surveys. Then the

researcher can contact the respondent through telephone
following a prior intimation through ‘sms’ regarding the
study. The growth in number of cell phone users makes this
method more reliable in many study settings.
Mail survey: Following a general announcement in
professional newsletters or journals, mail surveys to special
populations, such as groups of health professionals, can be
conducted in many research settings5. But we cannot expect
each and every one, to whom the mail was sent, respond to
the mail. But the chance of answering a personal or sensitive
question is much high in mail survey.
Analysis of data
Analysis or coding of survey data involves translating answers
into numerical codes that can then be used in computer-
based data analyses. Data may be entered and stored on
computerized media, or other machine-readable storage
media. Before analyzing the data, researchers must decide
how to deal with questions for which responses are missing
because the respondent neglected to provide the information
or the interviewer failed to ask the question properly.
Generally, reduction in non-response bias after data
collection is accomplished through an adjustment in the
weight assigned to respondents. This is done on the basis
of an assumption that respondents are true representatives
of non respondents6.
In Descriptive cross sectional study, data analysis is done
mainly to measure the level of a particular attribute, such as
a specific exposure, disease or any other health-related event,
at a particular point in time. The data collected can be
presented as frequency tables, histograms, pie charts or line
72 An Introduction to Research Methodology

graphs, depending on the nature of attributes under study.

The main outcome measure of a cross sectional study is
prevalence of some health issues.
Prevalence=No. of cases in defined population at one point
in time / No. of persons in the defined population at the
same time
For continuous variables such as blood pressure or weight,
the obtained data is divided into those values that fall below
or above a particular pre-determined level and prevalence is
calculated on the basis of this.
Let us consider an example. In a survey of 1150 women
who got married in Pariyaram grama panchayath in the year
2015, a total of 115 reported attending premarital
counselling sessions. The proportion (prevalence) of subjects
attending premarital counselling in this group can be
calculated as follows
Numerator= 115
Denominator= 1150
Prevalence of premarital counselling = (115/1150) x 100
=10% .
In Analytical cross sectional study, analysis is done to
compare two or more than two groups with respect to some
particular attribute. Detailed statistical analysis becomes
relevant at this particular setting. Association between study
variables and outcome can be calculated using appropriate
tests of association, depending on whether the variable is
continuous or categorical. Odds ratio can be used to assess
the strength of an association between risk factor or
protective factor and health outcome of interest. Odds ratio
can be calculated from a 2 x 2 table (Details are available
 An Introduction to Research Methodology 73

in the module on case - control study)

Disease+ Disease-
Exposure+ A B
Exposure- C D
Odds ratio = AD/BC
For example, in a survey, data regarding tobacco usage and
educational status of 500 samples were obtained. Then the
subjects may be grouped into 2 based on usage of tobacco
as well as educational status, i.e., those who completed
secondary education and those who did not complete
secondary education. The odd ratio can be calculated from
the following 2 x 2 table
Tobacco + Tobacco -
Up to 10th standard 100 50
10th and above 50 300

So Odds ratio = (100 x300)/ (50 x 50) = 30000/ 2500 =

12.0. As the obtained Odds ratio is more than one, it can
be inferred that low level of education is a risk factor for the
usage of tobacco. The confidence interval also must be
considered while interpreting Odds ratio.
In analytical cross sectional study, there may be more
than one explanatory variable (risk factor) and outcome
variable. Most often the relationship between these
explanatory variables and outcome variable are not so simple.
This is because, in addition to their effect on outcome
variable, explanatory variable simultaneously influence the
effect of each other. Such complex relationship can be
accounted only by using multiple regression analysis7.
74 An Introduction to Research Methodology

Report generation
Interpretation refers to the task of drawing inferences from
the actual research work. It also means drawing of conclusion
of the study. It would bring out relations and processes that
underlie the findings. The utility of the outcome of the
research greatly lie on proper interpretations and is the
hardest part of solving a scientific problem. Interpretation
of results is important because; it
• links the present work to the previous,
• leads to identification of future problems,
• opens avenues of intellectual adventure and stimulates
the quest for knowledge
• makes others understand the significance of the research
findings and
• often suggests a possible experimental verification.
The basic rule in preparing results and conclusion is to give
all the evidences relevant to the research problem and its
solution5. Some key elements of a report are as follows8;
• The research question
• Background information from existing literature
• Details about the research method
 Who the respondents are and how they were selected
 Description of the survey and its development
process
 The response rate
• Details about the data collection and analysis procedures
• Results
 An Introduction to Research Methodology 75

• Discussion of the findings

• Limitations of the study
• Conclusion and recommendation
STROBE (Strengthening Reporting of Observational studies
in Epidemiology) checklist can be used as a guideline while
preparing the report of a cross sectional study. This checklist
consists of 22 items
Use of Cross sectional study
The findings obtained from the study may help the
researcher as well as decision makers to have an idea about
the burden of the situation among a defined population.
This data may be useful to plan and promote the health of
the population studied. Sometimes, this provides awareness
about some unknown factors associated with prevalence.
Apart from the single directional burden analysis, it may
fulfill more than one purpose. This may give clues about
the health status of a community, determinants of health
and disease, association between variables, identification of
groups requiring special care, surveillance, community
education and community involvement and evaluation of
community’s health care.
Advantages and Disadvantages
Advantages
• Relatively quick and easy to conduct (no long periods of
follow-up).
• Data on all variables is only collected once.
• Able to measure prevalence for all factors under
investigation.
76 An Introduction to Research Methodology

• Multiple outcomes and exposures can be studied.

• The prevalence of disease or other health related
characteristics are important in public health for assessing
the burden of disease in a specified population and in
planning and allocating health resources.
• Good for descriptive analyses and for generating
hypotheses.
Disadvantages
• Difficult to determine whether the outcome followed
exposure in time or exposure resulted from the outcome.
• Not suitable for studying rare diseases or diseases with a
short duration.
• As cross-sectional studies measure prevalent rather than
incident cases, the data will always reflect determinants
of survival as well as etiology.
• Unable to measure incidence.
• Associations identified may be difficult to interpret.
• Susceptible to bias due to low response and
misclassification due to recall bias9.
The benefit of a cross-sectional study design is that it allows
researchers to compare many different variables at the same
time. However, this may not provide definite information
about cause-and-effect relationships as they are snap shot
studies and do not consider what happens before or after
the snapshot is taken. So, following this basic survey study,
further longitudinal studies, cohort studies or case- control
studies may help in formulating valid and more reliable
theories regarding the causal relationships or incidence rate
of some diseases.
 An Introduction to Research Methodology 77

References
1. https://www.iarc.fr/en/publications/pdfs.../epi/cancerepi/
CancerEpi-10.pdf
2. Dr.A.K. Manojkumar-Lecture notes on Clinical Research
3. Chris Olsen, Diane Marie M. St. George-Cross-Sectional
Study Design and Data Analysis
4. Couper & Groves, 1996; Stussman, Taylor, & Riddick,
2003; Weeks, Kulka, & Pierson, 1987.
5. Lu Ann Aday,Llewellyn J. Cornelius- designing and
conducting health surveys- a comprehensive guide
6. Thomas Krenzke, Wendy Van de Kerckhove and Leyla
Mohadjer Westat. Identifying and Reducing Non response
Bias throughout the Survey Process
7. www.oxfordjournals.org>ma_chap6 on 23/02/2016
8. Kelley K, Clark B, Brown V et al. Good practice in the
conduct and reporting of survey research. International
journal for quality in health care 2003)
9. www.healthknowledge.org.uk/ Introduction to study
design- cross sectional study on 23/02/2016
78 An Introduction to Research Methodology

Chapter 8

Case Control Study

Rosamma Joseph, Shobha Sundareswaran

Clinical research broadly falls into two general categories,

observational and experimental. Observational studies can
be descriptive or analytical. In analytical studies, there is
always a control group, whereas descriptive studies do not
have controls. In analytical epidemiology the focus is on
the individual within the population. It is designed primarly
to establish the cause of the disease by investigating the
association between ‘exposure to a risk factor’ and the
‘occurrence of the disease’. Cohort and case control study
designs come under the group of analytical studies (Fig.1).
CLINICAL EPIDEMIOLOGIC STUDY

OBSERVATIONAL EXPERIMENTAL

DESCRIPTIVE ANALYTICAL CLINICAL

TRIALS
Case Report,Case Cohort Study
Series Randomized
Case Control Study
controlled clinical
Ecologic Study
Cross Sectional trial
Cross Sectional Study
NonRandomized
Study
clinical trial

Figure 1:-Epidemiologic study designs

 An Introduction to Research Methodology 79

Both Cohort and Case-control studies are undertaken to

investigate the association between exposure to a risk factor
and occurrence of disease. The basic difference is that in a
cohort, the group under study has been exposed to the risk
but the ‘outcome’ or ‘disease’ has not yet occurred at that
point in time. In a case control study, the disease has already
occurred in the group under observation and we need to
investigate for possible risk factors. Cohort studies are thus
prospective and ‘look forward in time’ for occurrence of
outcome whereas case control studies are retrospective
(outcome has already occurred) and ‘look backward in time’
for risk factors.
Cohort studies can be fraught with difficulties when;
• the outcomes being investigated are rare.
• there is a long time period between the exposure and
the development of the disease.
• it is resource intensive (time, money and effort)
• it is difficult to obtain accurate exposure information.
An alternative which avoids these difficulties is the case-
control design. Case Control studies contribute greatly to
the research tool box of an epidemiologist. Case control
studies are observational because no intervention is
attempted and no attempt is made to alter the course of the
disease. The goal is to retrospectively quantify and compare
the risk factor of interest using two groups of individuals:
cases (with outcome of interest) and controls (without
outcome of interest). Case control studies are also known as
case-referent studies1.
A case–control study starts with outcome. It involves the
identification of individuals with a particular disease (‘cases’)
80 An Introduction to Research Methodology

and without a particular disease or condition (‘controls’).

Through interviews and other means, the amount of
exposure to the risk factor is ascertained in each group. Based
on this, case group & control group are further divided into
those who are exposed to the risk factor under study & those
who are not.Then odds (chance quantified as a ratio) of
exposure is estimated among cases and controls 1. If the
exposure of risk factor among cases and controls is different,
it is possible to infer that the outcome (disease) may be
associated with an increased (risk factor) or decreased
(protective factor) exposure to the factor under study. For
eg, in a case control design, to study the association of oral
cancer with tobacco chewing, the odds of being a tobacco
chewer among oral cancer cases will be compared to the
odds of being a tobacco chewer among the control group
(Table 1).
Exposure Case Control
(tobacco (Oral cancer (Oral Cancer TOTAL
Chewing) present)150 absent)300
Chewers A (100) B (100) A+B=200
Non Chewers C (50) D (200) C+D=250
TOTAL A+C=150 B+D=300 450
Table 1: Frame work of analysis of odds Ratio
Odds of exposure among case= Ratio of exposure to non-
exposure in cases (A/C) =100/50
Odds of exposure among control = Ratio of exposure to non-
exposure in controls (B/D) =100/200
Odds of exposure among cases/Odds of exposure among
Controls=ODDS RATIO
A/C÷B/D = AD/BC = OR (Odds Ratio)=100x200/
 An Introduction to Research Methodology 81

100x50=4. It implies that those who are chewing tobacco

have four times greater Odd’s (chance) in contracting oral
cancer than those who are not chewing tobacco. Apparently,
oral cancer is four times more among tobacco chewers than
those who are not chewing tobacco.
From the above example,
If OR=1 i.e., No Risk [it implies that there is no association
between tobacco chewing and oral cancer].
If OR>1 there is Risk. For eg, if OR=4, those who are
chewing tobacco have four times greater Odd’s (chance) in
contracting oral cancer than those who are not chewing
tobacco.
If OR<1, there is no risk and the exposure may be protective.
History of case control study
Case-control studies were historically borne out of interest
in disease etiology. The conceptual basis of the case-control
study is similar to taking a history. It is like questioning and
examining a patient for valuable information regarding the
disease, and such important points are knitted together to
provide the factor which predisposed the patient to that
disease. Such practices dates back to the Hippocratic writings
of the 4th century B.C 2. This study design was first
recognized in Janet Lane-Claypon’s study of breast cancer
in 1926, revealing the finding that low fertility rate raises
the risk of breast cancer 3, 4. Case control study methodology
crystalized with the landmark publication of Richard Doll
& Bradford Hill in 1950, which reported that smoking is a
factor in the production of carcinoma of lung5. Since that
time, case-control studies have become very prominent in
the biomedical literature with more rigorous methodological
advances in design, execution, and analysis.
82 An Introduction to Research Methodology

Note: 1) Both exposure and outcome have occurred before

the start of the study, 2) The study proceeds backward from
effect to cause, 3) There is a comparison group or a control
group.
Case control design
The specific hypothesis under investigation must be clearly
stated before a case–control study is designed. Case
control studies start with an outcome. Individuals with a
particular disease (case) from a given population are selected
and individuals without diseae (control) are also selected
from the same population. The same diagnostic criteria
should be followed to assign as a case or as a control. The
basic steps involved are; (i) definition of cases and controls,
(ii) selection of cases and controls, (iii) matching if needed,
(iv) measurement of exposure and (v) analysis.
Definition of a case: The definition of a case needs to be
very specific and it may involve two specifications.
Diagnostic criteria:-Diagnostic criteria of the disease must
be specified before the study is undertaken. Once diagnostic
criteria are established, they should not be altered or changed
till the study is over.
Eligibility criteria:-Case-control studies identify subjects by
outcome status at the outset of the study. Outcomes of
interest may be a disease, procedure or complication.
Eligibility criteria will ensure that the cases selected in the
study group are representatives of those in the target
population.
Ideally incident cases (newly occurred cases) should be
selected for a case control study. However there are situations
where prevalent cases meaning those who already have the
disease prior to the study, may have to be used particularly
 An Introduction to Research Methodology 83

in very rare diseases or where rapid information is required.

One problem for using prevalent cases is that they are usually
more informed about the disease and hence may have altered
behaviour and attitude since the diagnosis is made6. Incident
case design is preferred as it reduces recall bias and over-
representation of cases with long standing disease.
Source of cases:- Cases may be selected from a variety of
sources like hospitals, clinics, OP registers, hospital records
& the community. The study may be ‘hospital-based’ if all
the cases are hospital patients fulfilling the eligibility criteria.
Definition of controls:-Controls must fulfil all the eligibility
criteria defined for the cases apart from those relating to
diagnosis of the disease. Controls should come from the same
population at risk of disease, should not have the disease
and should be representative of the target population.
Selection of Controls: Selecting the appropriate group of
controls can be one of the most demanding aspects of a case-
control study. It is much harder and requires great care to
prevent bias. Important points to be kept in mind while
selecting controls are:-
1) Both case and control should stem from the same source
population.
2) The control should have a similar risk to become a case
if exposed (with potential to develop the outcome/
disease).
3) Both Case and Control should meet the same inclusion
criteria and should be similar in all aspects except in the
outcome status.
Types of controls
1) Hospital controls - Have similar quality of information
84 An Introduction to Research Methodology

and are convenient to select, but they may have

characteristics or other diseases that led to
hospitalization.
2) Population controls - Random digit dialing eg:-
telephone directory
3) Best friend or neighborhood controls - May share similar
characteristics
In order to minimize bias, controls should be selected to
be a representative sample of the population which produced
the case.
For example, if cases are selected from a defined population
such as OP register, then controls should comprise a
sample from the same OP register. Hospital registers are the
most preferred sampling frames for a case control study as
community based disease registries are rarely available in
our setting.
Matching
This is the process of selecting controls in a case-control
study, so that the controls are similar to the cases with regard
to certain key characteristics—such as age, sex, and race.
Matching ensures comparability between cases and controls
and reduces variability and systematic differences due to
background variables (confounders) that are not of interest
to the investigator. The factor upon which cases and controls
are matched cannot be studied as a risk factor7.
Matching can be performed at an individual or group level.
Individual matching (matched pairs):-The exposure to the
risk factor of interest is compared between the cases and the
controls. Each case is individually paired with a control
 An Introduction to Research Methodology 85

subject with respect to the ‘matching’ variables. This

matching strategy is called individual matching. Age, sex,
and race are often used to match cases and controls because
they are typically strong confounders of disease.
Group matching(frequency matching):- In group matching,
the focus is on matching a group of cases with a group of
controls. eg, age: if 20% of cases are 50-54 years old, then
controls are selected in such a way that 20% of controls are
also 50-54 years old.
Multiple Controls—Investigations examining rare outcomes
may have a limited number of cases to select from, whereas
the source population from which controls can be selected
is much larger. In such scenarios, the study may be able to
provide more information if multiple controls per case are
selected. This method increases the “statistical power” of
the investigation by increasing the sample size. The precision
of the findings may improve by having up to about three or
four controls per case8,10.
Sometimes, the effect of exposure of interest measured is
distorted because of the presence of some extraneous factor.
Presence of this factor may either inflate or deflate the effect
under study. This factor, aside from those risk factors of
interest, might contribute to the development of disease and
confound the causal association under investigation and
these are called as confounders.
Confounders: Confounder is another variable other than
exposure variable. It should be a risk/protective factor for
that disease and it should affect the outcome of the disease.
Even though confounder is an independent risk factor for
disease, they can affect the disease directly (most of the time)
or indirectly.
86 An Introduction to Research Methodology

Exposure Disease

Confounder

Figure :-3 Schematic representation of pathway of

confounder

Note: Confounding factors must be identified prior to the

start of the study as far as possible.
How do you identify a confounder? Investigator must be
through with the literature to identify possible confounders
related with the exposure. For eg: In a study of esophageal
cancer, where alcohol consumption is a risk factor, smoking
can act as a confounder. This is because smoking is associated
with both exposure (alcohol consumption) and disease
(cancer). If there is no association between either smoking
or esophageal cancer or between smoking and alchohol
intake, there will not be any confounding effect.
How are the confounders controlled?
Confounders can be controlled at study design, conduct
or analysis levels.
Study level
1) Restriction – Resistricted to one group: eg:- esophagial
cancers among non-smokers are compared with non-
smoker controls (in above example)
2) Matching:- Cases and controls are commonly matched
by age and sex.
 An Introduction to Research Methodology 87

Analysis level:-An alternative method of overcoming

confounding is to collect relevant information on the
confounding factor during the course of the study and adjust
for this at the analysis stage (stratification and multivariate
adjustment).
Measuring exposure status
Exposure status is measured to assess the presence or level
of exposure in each individual for the period of time prior
to the onset of disease or condition under investigation when
the exposure would have acted as a causal factor. In case
control studies the measurement of exposure is measured
after the development of disease and as a result of this it is
prone to recall bias and interviewer bias.
Various methods can be used to ascertain exposure status.
These include:
1. Standardized questionnaires.
2. Biological samples
3. Interviews with the subject, spouse or other family
members
4. Medical records, Employment records, Pharmacy record.
Note: The procedures used for the collection of exposure
data should be the same for cases and controls .
Bias in case - control studies
Due to the retrospective nature of case - control studies,
they are particularly susceptible to the effects of bias, which
may be introduced as a result of a poor study design or during
the data collection of exposure and outcome. Because the
disease and exposure have already occurred at the outset of
88 An Introduction to Research Methodology

a case control study, there may be differential reporting of

exposure information between cases and controls based on
their disease status. For example, cases and controls may
recall past exposure differently (recall bias). Similarly, the
recording of exposure information may vary depending on
the investigator’s knowledge of an individual’s disease status
(interviewer/observer bias). Therefore, the design and
conduct of the study must be carefully considered, as there
are limited options for the control of bias during the analysis.
Selection bias in case - control studies: The aim of a case -
control study is to select study controls who are
representative of the population which produced the cases.
Controls are used to provide an estimate of the exposure
rate in the population. Therefore, selection bias may occur
when those individuals selected as controls are not
representative of the population that produced the case.
Selection bias is a particular problem inherent in case -
control studies, where it give rise to non-comparability
between cases and controls11.
Berksonian bias:—It is a form of selection bias, related to
differential selection of samples from hospitals. Cases from
the hospital sample may represent a more severe form of the
disease than those in the community 12.
Analysis
1) Odds ratio and its confidence interval
An odds ratio (OR) is a measure of association between an
exposure and an outcome. The OR represents the odds that
an outcome will occur in the presence of a particular
exposure, compared to the odds of the outcome occurring
in the absence of that exposure13.
 An Introduction to Research Methodology 89

Note: An odds ratio is the ratio of the odds of an exposure in

the case group to the odds of an exposure in the control group.
It is important to calculate a confidence interval for each
odds ratio. A confidence interval that includes 1.0 means
that the association between the exposure and outcome could
have occurred by chance and that the association is not
statistically significant. An odds ratio without a confidence
interval should be interpreted with caution.
OR is the relative difference of the exposure in case group,
compared to that of the control group. ie an OR of ‘4’ in
the case group means this group has a risk of ‘4’ assuming a
risk of ‘1’in the control group.
OR=1: Exposure does not affect odds of outcome——.>No
risk
OR>1: Exposure associated with higher odds of outcome—
-à Risk
OR<1: Exposure associated with lower odds of outcome—
—àProtection
Eg:-OR=1.59 (95% CI : 1.49-1.67) Risk & Significant [CI
not crossing1]
OR=1.2 (95% CI: 0.89-1.96) Risk &Not Significant [CI
crossing1]
OR=0.79 (95%CI: 0.74-2.1) Protective &Non Significant
[CI crossing1]
OR=0.72(95%CI: 0.71-0.85) Protective &Significant[CI
not crossing1]
90 An Introduction to Research Methodology

Strength and weakness of case control study:

Strength Weakness
easy to carry out cannot generate incidence data
cost effective subject to bias
efficient for rare difficult if record keeping is
diseases either inadequate or unreliable
can study multiple selection of controls can be
exposures difficult
generally requires few limited to examining one out
study subjects come

Variants of Case Control

1) Traditional case control (Cumulative case control): All
the cases are selected from the population and the rest is
considered as control. Controls are disease free people
throughout the study period (“survivors” at the end of
the follow-up)
2) Case Cohort design: In a case-cohort study, all
incident cases in the cohort are compared to a random
subset of participants who do not develop the disease of
interest
3) Nested Case Control14: A nested case-control study is
one where the cases and controls are selected from
individuals within an established cohort study. The case-
control study is thus said to be ‘nested’ within the cohort
study. Cases of a disease that arise within the defined
cohort during the follow up period are identified, then
a specified number of matched controls who have not
 An Introduction to Research Methodology 91

developed the disease are selected from the same cohort.

The main advantage of nested case-control studies is that
certain exposure data would already have been collected
for both cases and controls which limits the potential
for recall bias.
Conclusion
In the hierarchy of evidence based medicine (EBM), though
well-designed randomized controlled trials (RCTs) hold the
pre-eminent position as level I evidence, well-designed
observational studies, recognized as level II or III evidence,
can also play an important role in deriving evidence. The
case control design is efficient and cost effective, can be done
in a shorter time with lesser number of subjects and may be
used to study multiple exposures. Hence it is a very powerful
tool in the research armamentarium of the investigator.
92 An Introduction to Research Methodology

Chapter 9
Cohort Study
K.K Remani, Shylamma T.M

The word Cohort has its origin in the Latine ‘Cohors’, that
refers to warriors and gives notion of a group of persons
proceeding together in time. In research it means a group of
people who share common exposure/experience/ character/
condition within a defined time period. In ancient Roman
military, a band of warriors marching to a battle was named
as a Cohort. In epidemiology, it is a group of persons who
are followed over a time.
The Cohor t study is defined as an observational
epidemiological design, which attempts to study the
relationship between a purposed cause (exposure) and the
subsequent risk of developing disease. This is the most
powerful observational study for identifying an association
of risk factor and a disease.
Examples:
 Occupational cohort – miners ,chemical plant
workers etc
 An exposure cohort – radiologist, smokers etc
 Community cohort of specific age , sex etc
 Diagnosed or treated cohort – surgery , radiotherapy
etc
 An Introduction to Research Methodology 93

 Birth cohort
 Marriage cohort etc
Basic character of the study
 Starts with people free of diseases
 Assesses exposure at ‘base line’.
 Assesses disease status at ‘follow-up’
Indication of the study
1. When there is good evidence of association from clinical
observation and descriptive studies.
2. When exposure is rare, but incidence of disease is higher
among exposed.
3. To find out the outcome of a well-defined exposure.
4. If there is sufficient money and time.
5. When follow up is good, cohort remains stable.
Closed (fixed) cohort: A cohort is called closed; i) when a
group of individuals recruited and enrolled at a uniform
point in the natural history of a disease or by some defining
event, ii) study participants are included from the beginning
to the end of the cohort and iii) do not adopt new members
after it is assembled. For example: survivors of Hiroshima
bombing.
Open cohort: A cohort is called open; i) where a group of
individuals recruited and enrolled through a mechanism that
allows for in and out migration of people and ii) defined
characteristic other than disease (e.g.: geographic location),
dynamic population etc.
Types of cohort
Cohort study can be done in three ways, based on time of
occurrence of disease in relation to the time at which the
investigation is initiated.
94 An Introduction to Research Methodology

(1) Prospective Cohort study

(2) Retrospective cohort study
(3) Ambispective cohort study (Combination of
prospective& Retrospective study)
(1) Prospective Cohort Study
It can also be called as concurrent study, forward looking
study etc. the study is usually forward looking or planned
in advance and carried out over a future period of time.
Here the researcher first select a research question, then frame
a hypothesis about the potential outcome of an exposure.
The researcher then observe a group of people (cohort), over
a period of time (often several years), collecting data that
may be relevant to the disease .The study begin in present
and continue to future.

Develop
Exposure to Disease
risk factor
Do not
Develop
Reference
Sample Disease
population
Develop
Not Exposure Disease
to risk factor
Do not
Develop
(2) Retrospective cohort Disease
It is also known as historical cohort study or non con current
prospective study. Retrospective cohort studies look at data
that already exists and tries to identify risk factor for
particular condition. Here the outcomes have occurred
before starting the study.
 An Introduction to Research Methodology 95

Develop
Exposure to Disease
risk factor
Do not
Develop
Population
Sample Disease
at risk
Develop
Not Exposure Disease
to risk factor
Do not
Develop
Disease
(3) Ambispective study
Here the exposure has occurred and the outcome may or
may not have occurred. The follow up for outcome
estimation is started in the past, and will continue in the
present and to the future. Some of the outcome might have
been measured and the search is continuing for newer
outcome during further follow-up. The follow up period
can be fixed according to the nature of the disease.

Diseased
Diseased
Exposed Non Diseased Non
diseased
Population
Future
(Sample)

Diseased
Unexposed Diseased
Non Diseased Non
Diseased
96 An Introduction to Research Methodology

Prospective cohort Retrospective cohort

Primary data collection Secondary data (from records)
Higher authenticity Lower authenticity
Time consuming Time saving
Expensive Economic
New outcomes can be New outcome cannot be
measured measured

Advantages of cohort study

 We can find incidence rate and risk.
 Assess multiple outcome of a single exposure.
 Establish temporal relationship between exposure and
outcome.
 Good when exposure is rare.
 Less chance of encountering the problem of recall bias,
selection bias and survival bias.
 Can be done with original data or secondary data.
Disadvantages
 Need of recruiting large number of people at risk.
 Loss to follow up is a problem.
 Often requires long duration of follow-up
 Ineffective for rare diseases.
 Drop outs, deaths and other lose.
 Long time to complete the study.
 They are very costly in time, personnel, space, money
and patient follow up
 Ethical issues are more compared to a case control study
 Multiple exposures cannot be compared in a single study
 Loss of interest or job change of the investigator.
 An Introduction to Research Methodology 97

Potential bias
Bias is an inevitable issue in research. The common bias
associated with cohort study are confounding bias,
interviewer bias, information bias, non-response bias, losses
to follow up bias and analytic bias.
1. Confounding bias
In cohort study two groups are selected not by randomization
and hence imbalance in baseline characteristics and
prognostic factors occur in cohort studies. So external factors
directly or indirectly associated to the exposure or outcome
can influence the study results and this is called confounding
bias.
2. Interviewer bias
If the interviewer knows the expected outcome under study
there are chances that the exposed group will be investigated
more intensively. This may bring in interviewer bias.
3. Information bias
If the study participants are aware of the outcomes of a
particular exposure, it can result in differential reporting. A
few participants may report more information on details of
exposure and outcome. It can bring in a bias in the study
results.
4. Non –Response and losses to follow up bias
Some participants may not respond properly during the
course of the study. Causes for lose to follow-up include
migration, loss of interest and death.
5. Analytic bias
If the identity of the group is disclosed to the evaluator it
may produce unintentional bias in analysis and
interpretation.
98 An Introduction to Research Methodology

A comparison with case control

ASPECT COHORT CASE
CONTROL
Logistic of the study
1. Time Long Quick
2. Cost Expensive Relatively cheap
3. Sample size Larger Small
4. Ethical problem present Minimal
Types of investigation
1. Rare disease Not preferred Preferred
2. Disease with long Not preferred Preferred
latency period
3. Rare exposure Preferred Not preferred
4. Multiple outcome Preferred Not preferred
5. Several risk factors Not preferred Preferred
Problems
1. Selection bias Some times Always present
2. Recall bias Some times Present
3. Temporal Present Not certain
relationship
Problematic Not a problem
4. Attrition problem
Calculation of
measurements
Possible Not possible
1. Incidence rate
Possible Odds ratio-an
2. Relative risk
Possible approximation
3. Attributable risk
Can be Not possible
4. Dose-response
determined Approximation
relationship
 An Introduction to Research Methodology 99

Study design
An exposed and non-exposed group is observed over a time
period and outcome of the 2 group are estimated and
compared.
Methodology - Consists
1. Define exposure status
2. Selection of study subjects (Exposed cohorts)
3. Selection of comparison subjects (Non exposed cohort)
4. Examination of both groups over a period of time and
obtain data on exposure variables and explanatory
variables
5. Follow up
6. Measurement of outcome in each group
7. Analysis
Study plan
 Define population at risk using inclusion and exclusion
criteria (individuals with outcome of interest at time of
screening and enrollment are not eligible in study).
Cohorts can be selected from
(a) General Population- Cohorts may be selected from
general population with well defined parameters and
frequent exposure.
For example:
1. Whole population in an area (area surrounding a
factory)
2. A representative sample. (for example if the cohort
is panparag using plus two level students, we select
the cohort only from a batch that is the representative
sample)
100 An Introduction to Research Methodology

(b) Specific groups (Special groups)

i) Select group E.g.: Doctors, Nurses etc. (Occupational
/ Professional groups)
ii) Exposure group: In rare exposure condition and if
huge findings are needed, select Cohorts of people
known to have experienced the exposure. E.g.: Mine
workers.
All subjective and objective parameter are used for selection
and strictly based on diagnostic criteria.
Selection of Comparison Group (Controls)
It is better to select controls from same population but not
exposed. Other variables should be similar to the study
group. It is of two types, internal and external comparison
groups. Non exposure and exposure subjects of same cohort
are selected -internal comparison whereas comparison group
is selected from outside the cohort - external comparison. If
comparison group is not available we can compare the rate
of study subjects with the existing rate in the general
population. e.g.: cancer rate in uranium miners with cancer
in general population.
Data collection and Follow up
It is the most important step in cohort study .But it will
take a long time. So appropriate measures should be taken.
 Personal interview / mailed questionnaire
 Physical measurements- BP checkup, height, weight,
measurements etc
 Review of records –dose of drugs, surgery type etc
 Medical and lab tests- USG, Serum Cholesterol etc
 Environmental survey etc
Measurement of exposure of cohort is important. It should
 An Introduction to Research Methodology 101

be done carefully with accuracy and defined in advance.

Follow up: Can be passive and active.
Passive follow-up
 Hospitals
 Disease registers
 Clinics / physician offices
 Surveillance systems e.g.: NDI(National Death Index)
Active follow-up
 Systematic evaluations for outcome of interest
 Regular time intervals
 In all study subjects
Strategies for minimizing losses during follow up
1) During design and enrollment, exclude those likely to
be lost.
2) Obtain information to allow future tracking complete
address, telephone and mobile number(s) from subject.
3) Address and telephone of one or two close friends or
relatives who do not live with the subject.
4) Name, address and telephone number of primary
Physician.
5) Periodic contact with the subjects by telephone.
Repeated mailing with stamped self-addressed envelopes.
6) For those who are not reached by phone or mail, request
forwarding addresses from postal services.
Analysis
Data are analysed in terms of
1) Incidence rate of outcome among exposed and non-
exposed.
102 An Introduction to Research Methodology

2) Estimation of Relative Risk

3) Estimation of Attributable Risk.
Incidence rate - is the proportion/percentage of new cases
per population at risk in a given time period.
Relative risk- is the ratio of the probability of an event
occurring (developing a disease) in exposed group to the
probability of the event occurring in a comparison, that is
non-exposed group.
Attributable risk- is the difference in the rate of a condition
between an exposed population and an unexposed
population.
Calculations of IR, RR and AR
Exposure Oral Oral Total Incidence
Cancer + Cancer - rate
Use of a b a+b a/a+b
Panparag 70 6930 7000 70/7000
Non Use of c d c+d c/c+d
Pan Parag 3 2997 3000 3/3000
73 9927 10000

1. Incidence Rate (IR)

Incidence (Ie) of Oral Cancer among Panparag user
Ie (Incidence among exposed) =a/a+b =70/7000 = 10 per
thousand (10/1000)
Incidence of Oral Cancer among Non-users of Panparag
Io (Incidence among non-exposed)=c/c+b = 3/3000 = 1 per
ten thousand (1/1000)
2. Relative Risk (R R)
RR = Incidence of Oral Cancer among Panparag Users
Incidence of Oral Cancer among Non Panparag Users
 An Introduction to Research Methodology 103

=10/1 = 10, oral cancer is 10 times more common among

panparag users than non Panparag users.
3. Attributable Risk
Incidence of oral cancer among panparag users minus
incidence of oral cancer among non-users of panparag. AR
represented as a percentage of incidence among exposed
group is called Attributable Risk Fraction
AR% = 10-1x100 =90%
10
Therefore, 90% of the cases of oral cancer among panparag
users are attributed to their habit of panparag use.
Measures of frequency
1. Cumulative incidence (Incidence proportion)
This index of disease frequency is based on the total
population at risk which was, at entry to the study, free of
the disease under investigation. The incidence of the disease
is calculated for each stratum of exposure to the risk factor,
and is the proportion of the number of new cases or events
in a specified period of observation, to the total population
at risk during that period.
This incidence measure provides an estimate of the
probability or risk of developing disease among all members
of the group who were included in the study at its initiation,
and were at risk of disease. Because cumulating all the new
cases in the total population at risk derives the measure, the
term cumulative incidence has been applied. Cumulative
incidence is a proportion that can vary from 0 to1 ,that is,
no less than 0% and no more than 100% of the population
at risk can acquire the disease.
2. Incidence density (incidence rate)
This approach is an improvement over the conventional
104 An Introduction to Research Methodology

measure of incidence, because it takes into consideration

both the number observed and the duration of observation
for each individual. Thus, if 30 individuals were observed
as follows : 10 for two years, 5 for three years, and 15 for
four years, they would contribute (10*2)+(5*3)+(15*4)=95
person-years of observation, which would become the
denominator. The numerator is the number of new cases
observed in these groups over the specified period of time.
This gives an incidence rate per person-year, called the
incidence density.
Conclusion
The cohort study design is the best available scientific
method for measuring the effect of a suspected risk factor.
Cohort studies are also good at finding relationship between
health and environmental factors such as chemicals in the
air, water and food issues that the WHO helps in researchers
to investigate with large scale cohorts studies. Thus cohort
studies are considered are level 2 evidence within the
hierarchy of medical evidence that is just one rank below of
RCT.

Bibliography
1. Research methodology by C.R Kothari
2. Research methodology by Dr. A.K Manoj
3. Health research methodology ; a guide for training in research
methods by WHO
4 . Research methodology in health system by APJ association
of Govt. Ayurveda College Trivandrum.
5. www.wikipedia.com
 An Introduction to Research Methodology 105

Chapter 10

Randomised Controlled Trial

Raghunathan Nair, Noble TM

Trials are true experiments in medicine. New diagnostics

and therapeutics are adapted to medical practice in a step
wise manner based on different phases of evaluation and
experimentation. WHO classification of trials during course
of development of drug or device testing is as follows:
Phase 1 studies for safety: This involve healthy volunteers
who receive a fraction of the anticipated dose and monitored
the effect on cardiovascular, hepatic, renal, and endocrine
functions. The metabolism of the drug is also investigated
Phase 2 studies for dosage and efficacy: This involves patients
to assess the efficacy of the drug or device to determine the
appropriate dosage and safety
Phase 3 studies for efficacy and effectiveness: This is the
classical phase of clinical trial performed on patients with
strict criteria for inclusion and exclusion. The subjects are
randomized into two groups for comparability.
Phase 4 studies for community effectiveness: This phase
involves post marketing surveillance to re- assess the
effectiveness, safety acceptability and continued use of the
drugs or devices.
106 An Introduction to Research Methodology

In the clinical setting most randomized controlled trials are

phase 3 trials for efficacy (result under ideal trial conditions)
and effectiveness (result under real life clinical situations).
Usually clinicians, have to choose between two or more drugs
/ therapeutic regimens for treatment with an uncertainty or
whether an intervention (or drug) is efficacious or not. If
there is such a clinical “equipoise”, it is ideal to have a clinical
trial to evaluate the efficacy of the drugs. A Clinical trial
conducted in a strictly randomised manner, using
comparable control is usually termed as randomised
controlled trial 1 . Here one group will be given- New
treatment/intervention and the other, the control will be
given standard treatment/ Placebo and treated same as the
intervention group except that they don’t get the intervention
under study.
Importance of control group in RCT is because of following
reasons;
1. Regression to the mean-If a disease is at its extreme,
subsequent evaluation may show improvement which
may be statistically significant, but not occurred as a
result of the intervention as such.
2. Unpredictable outcome- Clinical course of disease is
usually varying and the improvement or worsening may
not be because of the treatment given.
3. Placebo effect- Placebo effect is the improvement due to
the feeling of being treated. Subjects in both intervention
and control group (treated with placebo, an inert
substance) may show some improvement due to placebo
effect. The effect observed in the treatment group is a
combination of real clinical effect and placebo effect and
changes in control group will help us to identify placebo
effect from treatment effect.
 An Introduction to Research Methodology 107

4. Hawthorne effect- The phenomenon where subjects tend

to improve when they know that they are being observed.
Randomisation - Rationale
Randomisation is a key characteristic of RCTs.
Randomization is the process by which every subject gets
an equal chance to be in the intervention or control group.
At the end of the process the two groups are expected to be
identical and comparable in all aspects before intervention.
This means that variables that may influence the outcome
of clinical trial are equally distributed among the two groups.
Randomisation eliminates selection bias and confounding
bias, the confounder being an external factor which has got
independent association to the exposure and the outcome
under study.
Randomization is known as the heart of an RCT. It include
two steps;(i) the sequence generation and (ii) allocation
concealment. The generation of an unpredictable allocation
sequence is the first step of randomization process. The
sequence would be generated using a random number
generator like tossing a coin, random number table and
softwares like WinPEPEI. The best technique is to use a
specialized computerized database to randomize and to
ensure that the randomization sequence is not available to
clinicians (allocation concealment) 2.
Sequence generation
Simple Randomisation: This method is equivalent to tossing
a coin for each subject that enters a trial, such as Head=
intervention, Tail=placebo. A random number generator is
generally used to assign the subject individually to either
group. The process is simple, easy to implement and
treatment assignment is completely unpredictable. However,
108 An Introduction to Research Methodology

it can result in imbalance of treatment assignment, especially

in smaller trials. This imbalance can result in reduced
statistical power of the study.
Block randomisation: The imbalance that might have
occurred by simple randomization procedure can be avoided
using block randomization. Here the subjects are divided
into block of size of 4, 6 or 8 and they are allocated into two
groups in the required proportions. Equal numbers are
ensured in both the groups at the end of enrollment of every
block, thus balancing treatment assignment, especially in
smaller trials.
Stratified randomisation: Stratified randomization is opted
if the different strata (like male-female, mild-moderate-
severe) are represented unequally in groups under study.
Here equal number is ensured in every stratum between the
treatment and control groups. As the number in some strata
(eg, severe group) is likely to be small, block randomization
may have to be resorted to.
Allocation Concealment
A practical issue in experimentation is non-concealment in
allocation. The solution is allocation concealment,
implementation of which prevents foreknowledge of
treatment assignment. Usually Sequentially Numbered
Sealed Opaque Envelope (SNOSE) can be used to conceal
the allocation sequence. If allocation is not concealed,
research staff may knowingly or unknowingly assign subjects
to a particular group. Preferably randomisation should be
completed by an independent person or group not directly
involved in the study and usually a statistician assumes
responsibility for performing randomisation. In multi-site/
centric trials randomisation usually occurs at a centralized
location.
 An Introduction to Research Methodology 109

Blinding
Blinding is the process of avoiding bias in outcome
measurement by masking the intervention/control status to
various stakeholders directly involved in the study. It may
be the tester, the investigator, the subject or even the
biostatistician. The major objective of blinding being
avoidance of measurement bias, the person/s assessing the
outcome must be blinded.
Single blind: One of the major stakeholders who are directly
involved in the study will be masked. It is important to blind
the investigator, who is measuring the outcome, but
conventionally the patient is blinded in a single blinded trial
where the outcome measured is more subjective and reported
by the patient, eg, pain.
Double blind: Two key stakeholders directly involved in the
study, usually the researcher measuring the outcome and
the subject are masked.
Triple blind: Here the patient, researcher measuring the
outcome and statistician are blinded
Benefits of blinding:
1. Individuals: (i) Psychological – no prejudice, (ii) More
likely to comply with trial regimens, (iii) Less likely to
seek additional interventions and (iv)Less likely to leave
trial.
2. Trial investigators: Less likely to (i) transfer their
inclinations or attitudes to participants, (ii) differentially
administer co-interventions, (iii) differentially adjust
dose, (iv) differentially withdraw participants and (v)
differentially encourage or discourage participants to
continue trial.
110 An Introduction to Research Methodology

3. Assessors: (i) Less likely to have biases affect their

outcome assessments, especially with subjective
outcomes, (ii)to avoid bias on outcome after
randomization and (iii) avoids measurement bias.
Limitations of blinding3
The conduct of a study will become more complicated with
the levels of blinding from single blinded to triple blinded
trials. Participants and investigators may get curious and
interfere with blinding. At the participant’s level, they may
try to identify the drug by assessing size, weight, colour,
coating, odour, taste, etc and they may exchange information
about the efficacy of drug when meet each other in clinic.
The trial drug may get unblinded due to its known effects
or side effects- eg-Aspirin- G I bleeding. Participants and
investigators may guess the intervention out of curiosity and
it may be correct by chance in 50 % of times.
Types of RCTs:
1. Parallel Group Trial
Each participant is randomly assigned to a group, either the
intervention arm or the control arm. Nowadays in drug trials,
almost all people in the control arm may also receive some
kind of existing treatment, so this group is known as active
control group. If active drugs are not given to the people in
control arm, the investigator may give them some inert
materials with the same size, shape and colour of the
preparation given to the interventional group, known as
placebo. These groups are followed for a fixed period of time
parallel and the outcome will be measured.
2. Cross over Trial
Trial group and control group will be crossed over their
interventions after following for certain period of time. The
 An Introduction to Research Methodology 111

advantage of this design is that all individuals will get an

opportunity to be in both the arms. A person can act as his/
her own control in two different time point. A time lag is
given between the two follow-up times to completely cancel
the effect of first intervention, known as washout period.
The design is very powerful but can use only in limited
occasions where the intervention will not completely cure
or permanently modify the clinical condition under study.
3. Factorial RCT:
The effect of two or more independent interventions
(factors) is studied in a single experiment. The major purpose
of the research is to explore their effects independently but
in same time points. Factorial design produce efficient
experiments as it can also give the interaction between the
interventions on a common effect.
Each participant is randomly assigned to a group that receives
a particular combination of interventions or non-
interventions. For example in a study to find out the effect
of two vitamins A and D in a certain outcome,
Group 1 receives both vitamin A and vitamin D
2-vitamin A and placebo resembling vit D capsule
3-placebo of vit A and vitamin D
4-placebo of vit A and placebo of vit D
Advantage: Save cost, time and resources
4. Cluster RCT:
Randomisation is not done for individuals rather it is done
for groups or clusters from where the individual study
subjects are enrolled.
Clusters (e.g., communities, hospitals, or other aggregates
112 An Introduction to Research Methodology

of people (e.g., workplace, bars, brothels) are randomized

and all consenting persons enrolled from the selected clusters
to the selected groups
 Easy to carry out and large numbers can be studies
 Less individual-level self-selection
 The power of the study could be affected if the
individuals selected from the clusters have common
attributes (intra-cluster correlation)
RCTs according to Hypothesis:
Superiority trials – in which one intervention is hypothesized
to be superior to another in its effect
Non-inferiority trials-To determine whether a new treatment
is no worse than a reference treatment in its primary effect
(but it can have some other kind of superiority like less side
effect, less cost, etc)
Analysis
Intention to treat Analysis:
An intention-to-treat analysis is where the groups are
analysed based on the initial treatment to which they were
randomised (not by treatment eventually administered or
whether they finished the trial).
This type of analysis is generally considered optimal for most
RCTs.
Per protocol analysis
Groups are categorised based on the treatment received (an
individual randomized to control group can somehow receive
the active treatment and vice-versa)
 An Introduction to Research Methodology 113

Statistical Tests:
Different statistical tests can be used in RCT
 To compare dependent samples(pre and post values) we
use paired t test
 To compare independent sample(post values in two
groups) we use unpaired t test
 In large samples(more than 30) we use Z test
 If more than 2 groups involved Anova
 If more than 2 assessment stages RM Anova
 Advanced statistical analyses like ANCOVA (Analysis of
co-variance) is also very popular in RCTs
Effect measures in RCT
For categorical outcome (Cure + / -)
Relative Risk = Outcome rate in Rx group/Outcome rate
in control group
Absolute Risk Difference = Change in Rx group – Change
in control group
The appropriate type of test statistic presented will depend
on what type of outcome was measured. For dichotomous
outcomes, the results can be expressed in many ways,
including
Relative Risk (RR) = risk of the outcome in the treatment
group / risk of the outcome in the control group
Absolute Risk Reduction (ARR) = risk of the outcome in
the control group – risk of the outcome in the treatment
group
114 An Introduction to Research Methodology

Relative Risk Reduction (RRR) = absolute risk reduction /

risk of the outcome in the control group.
Number Needed to Treat (NNT) = inverse of the ARR and
is calculated as 1 / ARR.
The study should present a confidence interval or P value
so that you can judge the statistical significance of the result.
People People Total
died not died
Use of statins 10 90 100
Use of placebo 15 85 100
Total 25 175 200

This is an example in which risk of dying is the outcome

measure. Statin and placebo are the treatment and control
arms.
Relative risk= Risk of dying with statins = 10%/15% =.67
Risk of dying with placebo

Absolute risk reduction (Risk Difference) =15%-10% = 5%

Relative risk deduction= ARR/risk of outcome in control
group= 5%/15% =.33
 Number Needed to Treat (NNT):
Number of persons who would have to receive an
intervention for one to benefit.
 1/ARR (where ARR is proportion)
 1/5%=20
 To save life of 1 patient we must treat 20
 An Introduction to Research Methodology 115

Help to administrators to decide priorities

Limitations of RCTs:
Most criticized for ethical concerns. Many argue that it is
unethical to expose patients to a treatment that may be
inferior to other treatments; also it is unethical by giving
placebo
 Some RCTs require long period time and financial
burden
 Finally RCTs lack generalizability.
Ethical concerns:
One should follow guidelines listed for an ethical RCT
 ICMR guidelines
 Good Clinical Practice.
 Get informed consent
 Get cleared by Ethical Committee
Reporting RCT:
To improve reporting of RCTs, an international group of
scientists and editors published CONSORT in 2010 and
have become widely accepted.

References
1. A Primer on Research Methodology. M.K. C. Nair,
Harikumaran Nair, Leena Sumaraj, Babu George, Child
Development Centre, Medical College Campus,
Thiruvananthapuram – 695 011, Kerala, India, 2014 page
no 44
116 An Introduction to Research Methodology

2. Clinical research made easy –A guide to publishing in medical

literature, Mohit Bhandari, Parag Sancheti, Jaypee brothers
medical publishers (p) ltd, New Delhi, 2010 page no 215
3. ABC of research methodology and applied Biostatistics- A
primer for clinicians and Researchers. M N Parikh, Nithya
Gogtay, Jaypee brothers medical publishers (p) ltd, New
Delhi, 2009 page no 17
4. Health Research Methodology –A Guide for Training in
Research Methods, Second Edition World Health
Organization, Manila 2001 page no.62
 An Introduction to Research Methodology 117

Chapter 11

Diagnostic Test Evaluation

Rajitha R Warriar, Vinuraj S

Diagnostic tests have become part & parcel of our clinical

practice. They help physicians to study the disease
probability for their patients in the following ways,
 To establish a diagnosis in symptomatic patients. For
example, an ECG to diagnose myocardial infarction in
patients with chest pain.
 To screen for disease in asymptomatic patients. For
example, pap smear for cervical malignancy in a female.
 To provide prognostic information in patients with
established disease. For example, a CD4 count in patients
with HIV.
 To monitor therapy by either benefits or side effects.
For example, HbA1c into monitor glucose control in
patients with diabetes
 To confirm that a person is free from a disease (to rule
out a disease). For example, a pregnancy test to exclude
the diagnosis of ectopic pregnancy in irregular bleeding.
Every clinical assessment begins with an initial clinical
impression, to arrive at a pre-test probability of disease. Each
diagnostic test results in a change in the physician’s
118 An Introduction to Research Methodology

probability of disease, leading to the post test probability.

The degree to which a diagnostic test increases or decreases
the probability of disease from pre-test to post test, represents
the clinical utility of the test as measured by certain
characteristics.
1) Reliability ( Precision, Reproducibility, Repeatability)
It is the ability of a test to give consistent test results even
after doing repeatedly on the same individual or material
under the same conditions.
2) Validity (Accuracy)
It is the ability of test to measure accurately what it is
intended to measure and separate those who have the disease
from those who do not. Accuracy refers to the closeness with
which measured values agree with true values either as disease
present or disease absent with respect to truth.
3) Sensitivity and Specificity- Measures of Validity
The measures of sensitivity and specificity describe how well
the proposed test performs against an agreed ‘Gold Standard’
test. A ‘gold standard test’ or ‘criterion standard test’ is a
diagnostic test that is regarded as definitive and it shows
the reality. But, in reality a ‘Gold standard’ is the test, is one
which shows the situation ‘closer’ to reality with minimum
errors (false negative or false positive) or very rarely a 100%
accurate test with maximum sensitivity and specificity. In
clinical practice, however such a gold standard is often
lacking. For assessing various new diagnostic tests,
histopathology, angiography and CT/MRI scan are chosen
as gold standard according to the context. The actual gold
standard test may be too unpleasant for the patient, too
impractical or too expensive to be used widely or may not
be freely available. Alternate tests are needed in this venture
 An Introduction to Research Methodology 119

and validity of these tests must be studied.

Assessment of performance of such a test is usually
presented in a ‘two X two’ table, the disease status assessed
by the test (and cross verified using the Gold Standard) is
conventionally put in rows and the Gold Standard test
results in columns.
The total number of subjects is ‘n’, each subjects were given
both the diagnostic test, as well as the gold standard test for
final diagnosis.

Disease status as determined by ‘Gold Standard’

Disease No Disease
(as by Gold Std.) (as by Gold Std.)
Test True positives False positives Total test
positive positives
(a) (b) (a+b)
Test False negatives True negatives Total test
negative (c) (d) negatives
(c+d)
Total with disease Total without Total studied
(a+c) disease (b+d) (a+b+c+d)=n

The cells in the above table can be explained as follows:

Cell ‘a’ represents those subjects diagnosed as diseased by
the gold standard test as well as positive by the diagnostic
test. They are called ‘true positives’.
Cell ‘b’ represents those subjects diagnosed as diseased by
the diagnostic test and not diseased by the gold standard
test. They are called ‘false positives’.
Cell ‘c’ represents those subjects diagnosed as diseased by
120 An Introduction to Research Methodology

the gold standard test and not diseased by the diagnostic

test. They are called ‘false negatives’.
Cell‘d’ represents those subjects diagnosed as not diseased
both by the gold standard test as well as by the diagnostic
test. They are called ‘true negatives’.
(a+b) is the total number of subjects labelled as positive for
disease by the diagnostic test, irrespective of whether they
had disease or not. i.e, it is the sum of true positives and
false positives.
(c+d) is the total number of subjects labelled as negative for
disease by the diagnostic test, irrespective of whether they
had disease or not. i.e, it is the sum of true negatives and
false negatives.
(a+c) is the total number of subjects really having the disease
with gold standard test, irrespective of whether they were
recognized as positive or not by our diagnostic test. i.e, it is
the sum of true positives and false negatives.
(b+d) is the total number of subjects really not having the
disease with gold standard test, irrespective of whether they
were recognized as positive or not by our diagnostic test.
i.e, it is the sum of true negatives and false positives.
Sensitivity (true positive rate, positivity in diseases)
Is the ability of the test to identify correctly those who have
the disease (a) from all individuals with the disease (a+c)
(test positives in disease positive).
Sensitivity is a fixed characteristic of the test and calculated
as
Sensitivity = Number of true positve a
Total with disease a+c
 An Introduction to Research Methodology 121

Specificity (true negative rate, no disease rate, negativity

in health rate)
Is the ability of the test to identify correctly those who do
not have the disease (d) from all individuals free from the
disease (b+d) (test negatives in disease negatives).
Specificity is also a fixed characteristic of the test.
Number of true negatives d
Specificity=
Total without disease b+d
The primary purpose of a ‘screening test’ is to detect early
disease or risk factors for disease in large numbers of
apparently healthy individuals. A more sensitive test must
be used for screening purpose without much compromise
on the specificity. The purpose of a ‘diagnostic test’
(confirmatory test) is to establish the presence (or absence)
of disease as a basis for treatment decisions in symptomatic
or screen positive individuals. Test must be more specific
here. The ‘diagnostic test’ is performed on screen positives,
after a positive screening test, to establish a definitive
diagnosis.
The mnemonics SnOut and SpIn provide some guidelines
on how to interpret sensitivity and specificity for an
individual patient. SnOut helps physicians to remember that
a highly Sensitive test (have a high negative predictive value)
with a negative result is good at ruling-out the
disease. SpIn reminds physicians that a highly Specific test
(have a high positive predictive value) with a positive result
is good at ruling- in the disease.
Different diagnostic tests for the same disease often trade
sensitivity for specificity or vice versa. In general, the more
sensitive a test is for a disease, the higher its false-positive
122 An Introduction to Research Methodology

rate, lowering its specificity. A test with a higher specificity

will usually sacrifice sensitivity by increasing its false-negative
rate. This makes a highly sensitive test without much
compromise on specificity is ideal for screening and highly
specific tests without much compromise on sensitivity, are
best for confirmation.
Sensitivity contains no information about false-positive
results and specificity does not account for false-negative
results. This limits the applicability of sensitivity and
specificity in predicting disease. For example, a positive test
result with 90% sensitivity does not predict a 90%
probability of disease in a patient.
4) Predictive Values- Diagnostic Power of the Test
Sensitivity and specificity are test specific. They describe
how well the test performs against the gold standard
independent of prevalence of disease. PPV and NPV however
are sample specific. Positive predictive value and negative
predictive value depends on prevalence of the condition
under study. PPV and NPV give information on how well a
test will perform in a given sample with known prevalence
(disease prevalence dependent).
Predictive values are horizontally calculated operating
characteristics, which incorporate both false positive and
false negative results into disease probability. The positive
predictive value PPV is the probability of a patient actually
having the disease if the test results are positive (disease
positive in test positive). The probability of the patient being
free of the disease after a negative test result is given by the
negative predictive value NPV (disease negative in test
negative).
Unfortunately, predictive values are not stable characteristics
 An Introduction to Research Methodology 123

of diagnostic tests. The predictive values are dependent on

the prevalence values in different samples. Without knowing
the disease prevalence in the sample of interest, predictive
values cannot be accurately estimated.
The positive predictive value (PPV) describes the
probability of having the disease given a positive test result
in the sample. This is expressed as the proportion of those
with disease among all test positives.
number of true positives
PPV = = a / (a+b)
total test positives
The negative predictive value (NPV) describes the
probability of not having the disease given a negative test
result in the sample. This is expressed as the proportion of
those without disease among all test negatives.
number of true negatives
NPV = = d / (c+d)
total test negatives
Generally a highly sensitive test have better negative
predictive value (NPV), a highly specific test has better
positive predictive value (PPV).
Prevalence
Prevalence is the total number of diseased in a sample. This
is the probability of disease ascertained by clinical or other
simple tests before administering the proposed test. This is
also the pre-test probability.
If the pre-test probability is high, a test has relatively high
positive predictive value. If the pre-test probability is low, a
test has relatively high negative predictive value.
124 An Introduction to Research Methodology

Accuracy
Is the ability of test to identify the conditions correctly as
disease positive or disease negative out of total sample. This
is an oblique measure.
Example- Assume a sample of 1,000 people, 100 have a
disease, 900 do not have the disease, A test is used to identify
the 100 people with the disease
The results of the screening appears in this table
Screening True Characteristics in the Sample Total
Results Disease No Disease
Positive 80 (a) 100 (b) 180(a+b)
Negative 20 (c) 800 (d) 820 (c+d)
Total 100(a+c) 900 (b+d) 1,000

Sensitivity = a/ a+c × 100 = 80/ 100× 100 = 80%

Specificity = d/ b+d × 100 = 800/ 900× 100 = 89%
PPV = a/ a+b × 100 = 80/ 180 × 100 = 44.4%
NPV = d/ c+d × 100 = 800/ 820 × 100 = 98%
Accuracy = a+d/a+b+c+d = 880/1000 x 100 = 88%
Prevalence = a+c/a+b+c+d =100/1000x100 =10%
5) Likelihood Ratio
Likelihood ratios combine the qualities of sensitivity,
specificity and predictive values. It tells us how much we
should shift our suspicion for a particular test result. Tests
can be positive or negative; there are two likelihood ratios
for each test. The “positive likelihood ratio” (LR+) tells us
how much to increase the probability of disease if the test is
 An Introduction to Research Methodology 125

positive. On the other hand, the “negative likelihood ratio”

(LR-) tells us how much to decrease it if the test is negative.
Likelihood ratios reflect the change in odds favouring the
particular test result.
The general formula for calculating the likelihood ratio is:
probability that an individual with disease has a
particular test result
LR =
probability that an individual without disease has
the same test result
Thus, the positive and likelihood ratio are:
probability that an individual
with disease has a positive test
Positive LR=(LR+) = probability that an
individual without disease has a
positive test
= Sensitivity/(1-Specificity)
probability that an individual with
disease has a negative test
Negative LR= (LR-) =probability that an individual
without disease has a negative test
= (1-Sensitivity)/(Specificity)
Thus, Likelihood ratios are proportions of probabilities. It
can be described in terms of sensitivity and specificity. A
likelihood ratio for a positive test result (LR+) is the ratio of
the true positive rate (sensitivity) divided by the false-positive
rate (1 - specificity). LR+ explains how much more likely
the patient is to actually have the disease after a positive test
result.
126 An Introduction to Research Methodology

true positive
Likelihood ratio of Sensitivity rate(truth)
=
A positive test result (LR+) 1- specificity False Positive
Rate(error)
Dividing the false-negative rate (1 - sensitivity) by the true
negative rate (specificity) gives the likelihood ratio for a
negative test result and provides the strength of a negative
test result in convincing the physician the patient is free of
disease.
Likelihood ratio of = 1- sensitivity false negative rate
a negative test result (LR-) Specificity true positive rate
Since likelihood ratios are calculated from sensitivity and
specificity, LRs are stable operating test characteristics,
unaffected by prevalence of disease. A LR of 1.0 is a useless
test because this result fails to change the opinion of
probability of disease from pre-test to post-test. LR+ are
always greater than 1.0; the larger the number, the more
likely is the patient to have the disease after a positive test
result. LR- are always less than 1.0, with the smaller numbers
signifying a lower risk for disease than pre-test estimates.
For example, In a study of the ability of a test to diagnose a
disease condition, 80% of patients with disease have a
positive test, while 95% of those without disease have a
negative test. Thus, the sensitivity calculated is 80% and
the specificity is 95%.
The LR+ for the ability of test to diagnose disease is:
LR+ = 80% / (100%-95%) = 80% / 5% = 16
The negative likelihood ratio is: LR- = (100%-80%)/95%
= 20/95 = 0.21
 An Introduction to Research Methodology 127

Multiple Testing
Multiple testing is very commonly done in medical practice.
Choices of tests depend on cost, invasiveness, volume of
test, presence and capability of lab infrastructure, urgency,
etc. Multiple tests can be done sequentially (serial) or
simultaneously (parallel).
Sequential Testing (Two-Stage Screening) - After the first
(screening) test was conducted, those who tested positive
were brought back for the second test to further reduce false
positives. Consequently, the overall process will increase
specificity and PPV but with reduced sensitivity.
Post Test Odds = Pre Test Odds X Likelihood Ratio
In simultaneous testing two or more tests are conducted in
parallel. The goal is to maximize the probability that subjects
with the disease true positives are identified (sensitivity) and
NPV increased. Consequently more false positives are also
identified which decreases specificity. Here disease positives
are those become positive by either one test or by both tests.
In short, in simultaneous testing, there is a net gain in
sensitivity but a net loss in specificity, when compared to
either of the tests used. In sequential testing when positives
from the first test are re- tested, there is a net loss in
sensitivity but a net gain in specificity, compared to either
of the tests used false positive results produced by high
sensitivity of the screening test can easily be excluded by a
confirmatory test with high specificity. Physicians use parallel
testing when rapid assessment is necessary as in hospitalized
and emergency patients.
Cut off Point
Many test results have a continuous scale (are continuous
128 An Introduction to Research Methodology

variables) as test result. Cut off points are invariably needed

to predict the disease condition as categorical (as Yes/No).
Different cut-points yield different sensitivities and
specificities. The cut-point determines how many subjects
will be considered as having the disease.
The cut-point that identifies more true negatives will also
have more false negatives
The cut-point that identifies more true positives will also
have more false positives
So to utilize this cut off value as-
If the second test or the diagnostic (confirmatory) test is
expensive or invasive then we have to minimize false positives
or use a cut-point with high specificity.
If the penalty for missing a case is high (e.g., the disease is
fatal and treatment exists, or disease easily spreads) maximize
true positives, then we have to, use a cut-point with high
sensitivity.
The use of receiver operating characteristic (ROC) curves
The two most common uses of ROC curves in medicine
are:
- to set a cut-off value for a test result (for continuous
diagnostic variables)
- to compare the performance of different tests measuring
the same outcome (test validation)
In order to set the cut-off value for a continuous diagnostic
variable, the proportion of true-positives (sensitivity) and
false-positives (1-specificity) are calculated for possible values
of the test result. The ROC curve is a graphical display of
the how the proportions of true positives and false positives
 An Introduction to Research Methodology 129

change for each of the possible pre-determined values with

true positive rate (sensitivity) in y- axis and false positive
rate (1-specificity) in x- axis.

The choice of a particular cut-off value for a test is essentially

derived by maximum sensitivity and specificity. Generally,
there is a trade-off between sensitivity and specificity, and
the decision must be based on their relative importance. The
prevalence of the outcome, which is the pre-test probability,
must also be known to fix the cut off along with cost of
therapy, harm of therapy, availability of therapy etc..
In situations where there are multiple tests for particular
condition, the area under each respective ROC curve (AUC)
can be used to compare the overall performance of those
tests. The perfect test would have an AUC of 1 whereas a
test with no diagnostic capability would have an AUC of
0.5. An AUC of 0.5 indicates that a test based on that
variable would be equally likely to produce false positive or
130 An Introduction to Research Methodology

true positive results. This equality is represented by a

diagonal line from (0, 0) to (1, 1) on the graph of ROC
curve. AUC is usually calculated with statistical packages.
A ROC curve can demonstrate several things:-
1. It shows a trade- off between sensitivity and specificity.
Any increase in sensitivity will be accompanied by a
decrease in specificity and vice versa.
2. The closer the curve follows the left hand border and
the top border of the ROC space, the more accurate the
test.
3. The closer the curve comes to the 45- degree diagonal
of the ROC space, the less accurate the test.
4. The area under the curve (AUC) can be used to assess
test accuracy, and compare the performance of different
tests.
 An Introduction to Research Methodology 131

Chapter 12

Systematic Review and

Meta Analysis
Prakash Mangalasseri, Vivek P, Subin V. R.

The support of medical decisions comes from several sources.

These include individual physician experience as case reports
or case series, well conducted observational studies,
pathophysiological constructs, pivotal clinical trials,
qualitative reviews of the literature etc. Well conducted
Randomized clinical trials (RCT) are considered as gold
standard for medical evidence. When so many clinical trials
or observational studies deal a similar problem with similar
methodology, still if the results do not agree for unanimous
outcome it becomes practically impossible to conclude.
Further it makes lot of difficulties in health care decision
making. Here the importance of systematic review (SR) and
meta-analysis is evident as it hopefully put all the results
together and reaches a definitive conclusion, as positive,
negative, or inconclusive impact of the particular
intervention.
Historical Background
The technique for compiling various studies and data
pooling was introduced by Karl Pearson in 1904 1. He
132 An Introduction to Research Methodology

examined the correlation coefficients between typhoid and

mortality by inoculation status among soldiers in various
parts of the British Empire and calculated the arithmetic
mean of them across five two-by-two contingency tables.
Tippett in 1931, described a method for estimating the
likelihood of a significant effect from the ordered p-values
observed in various studies 2 . Subsequently in 1932,
renowned statistician Fisher also established a procedure for
combining p-values from several studies postulating a similar
question3.
The term “meta-analysis” was coined by Gene V. Glass in
1976 as a final outcome of a great debate over the efficacy
of psychotherapy4. In 1952, Hans J. Eysenck had concluded
that there were no favorable effects of psychotherapy and
this caused starting a hot debate. Two decades of on-going
researches and hundreds of studies failed to resolve the
debate. Glass (and colleague Smith) statistically aggregated
the findings of 375 psychotherapy outcome studies to prove
Eysenck’s finding as wrong. Glass concluded that
psychotherapy definitely works. Glass called his method
“meta-analysis” and stated that, “Meta-analysis refers to the
analysis of analyses, the statistical analysis of a large collection
of analysis results from individual studies for the purpose of
integrating the findings” 5. By time, the Meta-analytic
techniques are continuingly evolving and get refined with
originative advances such as DerSimonian and Laird’s
development of the random-effects model and the
advancement of meta-regression methods to explain across
study heterogeneity6.
There are classical examples for findings of initial trials found
to be concluded as wrong after the conduction of systematic
reviews and meta-analysis. Beneficial usage of Tamoxifen is
 An Introduction to Research Methodology 133

a good example. It is currently used for the treatment of

both early and advanced estrogen receptor positive breast
cancer in pre- and post menopausal women. Most of the
individual studies had a clinically significant but statistically
insignificant treatment effect. Several trials had shown
unfavorable results also even though statistically
insignificant. The beneficial mortality effect of tamoxifen
therapy was definitively established only after 1998 as the
patient-level data was combined from all the trials in a meta-
analysis of the Oxford-based Early Breast Cancer Trialists’
Collaborative Group7.
Chondroitin is a polysaccharide derived from cartilage. It is
commonly used by people with arthritis in the belief that it
will reduce pain, but clinical studies of its effectiveness have
yielded conflicting results. Reichenbach et al. (2007)
performed a meta-analysis of studies on chondroitin and
arthritis pain of the knee and hip. The initial literature search
yielded 291 potentially relevant reports. But after rigorous
inclusion, exclusion criteria to avoid bias in randomization,
allocation and control selection only 20 trials were eligible
for meta-analysis. The statistical analysis of all 20 trials
showed a large, significant effect of chondroitin in reducing
arthritis pain. However, the authors noted that earlier
studies, published in 1987-2001, had large effects, while
more recent studies showed little or no effect. In addition,
trials with smaller standard errors (due to larger sample sizes)
showed little or no effect. In the end, Reichenbach et al.
(2007) analyzed just the three largest studies with what they
considered the best designs, and they showed essentially zero
effect of chondroitin.
Another famous case is described by The Cochrane Library:
a single research paper, published in 1998 and based on 12
134 An Introduction to Research Methodology

children, cast doubt on the safety of the mumps, measles

and rubella (MMR) vaccine by implying that the MMR
vaccine might cause the development of problems such as
Crohn’s disease and autism. The paper by Wakefield et al
has since been retracted by most of the original authors
because of potential bias, but before that it had triggered a
worldwide scare, which in turn resulted in reduced uptake
of the vaccine. A definitive systematic review by Demicheli
et al on MMR vaccines in children concluded that exposure
to MMR was unlikely to be associated with Crohn’s disease,
autism or other conditions8.
Types of Literature Review
1. Narrative Review
2. Systematic review
Narrative reviews are usually written by experts in the field.
It is otherwise called as traditional review. In such reviews
the author use informal and subjective methods to collect
and interpret information in the public domain and becomes
the narrative summaries of the evidence. Narrative reviews
are done according to authors’ inclination (bias), he may
search any databases and the methods not usually specified.
On the other hand Systematic review has explicit and
reproducible protocol to locate and evaluate the available
data. Criteria for inclusion and exclusion are determined at
outset and the search will be comprehensive. There is a
more rigorous and prospectively defined objective process
for the collection, abstraction, appraisal and compilation of
the data. Systematic review is replicable also, which is not
possible in traditional review. In narrative review, vote count
or narrative summary is used for conclusion. It relies only
on statistical significance of a study which may be erroneous
since it is highly influenced by the size of the study and
 An Introduction to Research Methodology 135

magnitude of the effect. Unimportant effects can be

statistically significant, and important effects can be
statistically insignificant. But in systematic review meta
analysis is often used to combine data and to statistically
analyse it. Here the precision and accuracy of estimates can
be improved as more data is used. This, in turn, may increase
the statistical power to detect an effect.
Even though systematic review has various advantages like
reduced bias, replicability, resolving controversy between
conflicting studies and provide reliable basis for decision
making it also has few limitations. Results may still be
inconclusive in some situations. Sometimes the trials may
be of poor quality or there may not be any trials or evidences
for synthesis. The intervention may be too complex to be
tested.
Sources commonly used for Systematic review include
MEDLINE (United States Library of Medicine database),
EMBASE (medical and pharmacologic database by Elsevier
publishing), CINAHL (cumulative index to nursing and
allied health literature), CANCERLIT (cancer literature
research database) and Cochrane Collaborative . The
Cochrane Collaboration was established in October 1992.
It is an international non-profit organisation that prepares,
maintains, and disseminates systematic up-to-date reviews
of health care interventions. It was named in honour of
Archie Cochrane, a British researcher (1979).
Definitions
“A systematic review attempts to collate all empirical
evidence that fits pre-specified eligibility criteria to answer
a specific research question. It uses explicit, systematic
methods that are selected with a view to minimizing bias,
136 An Introduction to Research Methodology

thus providing reliable findings from which conclusions can

be drawn and decisions made.”9 It is also defined as a “a
review that is conducted according to clearly stated, scientific
research methods, and is designed to minimize biases and
errors inherent to traditional, narrative reviews.”10 Meta-
analysis is the statistical counterpart of Systematic review.
Meta-analysis is the process of combining the quantitative
results of separate (but similar) studies by means of formal
statistical methods in order to increase the precision of the
estimated treatment effect. It is generally used when
individual trials yield inconclusive or conflicting results. It
may also be used when several trials asking similar questions
have been conducted and an overall conclusion is needed.
“Meta-analysis is the use of statistical techniques to integrate
and summarize the results of included studies. Many
systematic reviews contain meta-analyses, but not all. By
combining information from all relevant studies, meta-
analyses can provide more precise estimates of the effects of
health care than those derived from the individual studies
included within a review9.
Methodologies of Systematic Review and Meta-analysis
Systematic review methodology is the central part of meta-
analysis. Presenting a balanced and impartial summary of
the existing research, enabling decisions on effectiveness to
be based on all relevant studies of adequate quality are the
objectives of systematic reviews. Frequently, such systematic
reviews provide a quantitative (statistical) estimate of net
benefit aggregated over all the included studies. Such an
approach is termed a meta-analysis11.
Systematic review
The need for rigour in the production of systematic reviews
 An Introduction to Research Methodology 137

has led to the development of a formal scientific process for

their conduct. Understanding the approach taken and the
attempts to minimize bias can help in the appraisal of
published systematic reviews, which should help to assess if
their findings should be applied to practice. The overall
process should, ideally, be directed by a peer-reviewed
protocol8.
Briefly, developing a systematic review requires the following
steps.
1. Defining an appropriate healthcare question
2. Searching the literature
3. Assessing the studies
4. Combining the results
5. Placing the findings in context
Performing a rigorous systematic review is far from easy. It
requires careful scientific consideration at inception,
meticulous and laborious searching, as well as considerable
attention to methodological detail and analysis before it truly
deserves the badge ‘systematic’. The quality of a systematic
review can be assessed by using a standard checklist. Example
checklists are available from the NHS Public Health
Resource Unit via the Critical Appraisal Skills Programme
(CASP) 12 or from the Centre for Evidence-Based Medicine
at the University of Oxford13.
In alternative medicine systems such as Ayurveda, it is
difficult to find systematic reviews and meta-analyses.
Narrative reviews are more common in the field of Ayurveda.
However, a preliminary attempt was found published by
Institute of Applied Dermatology in Ayurveda cikitsa for
switra (Vitiligo) (Protocol)14.
138 An Introduction to Research Methodology

Meta-analysis
Requirements
The main requirement for a worthwhile meta-analysis is a
well-executed systematic review15. However competent the
meta-analysis, if the original review was partial, flawed or
otherwise unsystematic, then the meta-analysis may provide
a precise quantitative estimate that is simply wrong.
The main requirement of systematic review is a complete,
unbiased collection of all the original studies of acceptable
quality that examine the same therapeutic question 11.
Conducting meta-analyses
1. Location of studies
Meta-analysis requires a comprehensive search strategy which
interrogates several electronic databases (for example,
MEDLINE, EMBASE and Cochrane Central Register of
Controlled Trials). Hand-searching of key journals and
checking of the reference lists of papers obtained is also
recommended16. The search strategy – the key terms used
to search the database –needs to be developed with care.
The strategy is written as a sequence of requirements: include
papers with specified terms, exclude papers that do not meet
certain criteria (for example, age or diagnostic group), only
include studies that follow certain research designs (for
example, randomized controlled trials)11.
2. Quality assessment
Once all relevant studies have been identified, decisions must
be taken about which studies are sufficiently well conducted
to be worth including. This process may again introduce
bias, so good meta-analyses will use explicit and objective
criteria for inclusion or rejection of studies on quality
 An Introduction to Research Methodology 139

grounds17. There is a bewildering array of scales for assessing

the quality of the individual clinical trials18.
Perhaps more important than the scale used is whether a
scale has been used at all. Once a quality score has been
assigned, the impact of excluding low quality studies can be
assessed by sensitivity analysis 11.
3. Calculating effect sizes
Clinical trials commonly present their results as the
frequency of some outcome (such as a heart attack or death)
in the intervention groups and the control group. For meta-
analysis these are usually summarized as a ratio of the
frequency of the events in the intervention to that in the
control group. In the past the most common summary
measure of effect size was the odds ratio, but now the risk
ratio (relative risk) can be given. Thus, a ratio of 2 implies
that the defined outcome happens about twice as often in
the intervention group as in the control group; an odds ratio
of 0.5 implies around a 50% reduction in the defined event
in the treated group compared with the controls11.
The findings from individual studies can be combined using
an appropriate statistical method19. Separate methods are
used for combining odds ratios, relative risks and other
outcome measures such as risk difference or hazard ratio.
The methods use a similar approach in which the estimate
from each study is weighted by the precision of the estimate.
4. Checking for publication bias
A key concern is publication bias, as clinical trials that obtain
negative findings (that is, no benefit of treatment) are less
likely to be published than those that conclude the treatment
is effective20. One simple way of assessing the likely presence
of publication bias is to examine a funnel plot21. Funnel
140 An Introduction to Research Methodology

plots display the studies included in the meta-analysis in a

plot of effect size against sample size (or some other measure
of the extent to which the findings could be affected by the
play of chance).
The funnel plot has some limitations; for example, it can
sometimes be difficult to detect asymmetry by eye. To help
with this, formal statistical methods have been developed
to test for heterogeneity. Egger’s regression test has been
widely used to test for publication bias21. However, care is
needed in the interpretation of the findings whatever test
has been used. There is currently no clear direction in recent
literature to indicate when to use each test.
5. Sensitivity analysis
Because of the many ways in which decisions taken about
selection, inclusion and aggregation of data may affect the
main findings, it is usual for meta-analysts to carry out some
sensitivity analysis. This explores the ways in which the main
findings are changed by var ying the approach to
aggregation11. In meta-analyses without sensitivity analyses,
the reader has to make guesses about the likely impact of
these important factors on the key findings.
Analysis
In systematic reviews and meta-analysis, collection
(extraction) of data and analysis is from the primary previous
studies rather than individuals. This can be narrative; like a
structured summary and discussion of studies, characteristics
and findings or can be quantitative involving statistical
analysis. The statistical analysis mainly deals with meta-
analysis. There are mainly three scholar organized
international, interdisciplinary groups to promote and to
give direction for systematic review and meta-analysis. They
 An Introduction to Research Methodology 141

are Cochrane collaboration, Campbell Collaboration and

Society for Research Synthesis Methodology. of which
Cochrane Collaboration is the most famous and accepted
one.
1. Systematic reviews
Assessing the studies: Systematic review and meta-analysis
are done by more than one reviewer, usually two. If there is
confusions exist between these two, opinion of a third
reviewer should be accepted. Once all possible studies have
been identified, they should be assessed in the following
ways.
(a) Eligibility - Each study needs to be assessed for eligibility
against inclusion criteria and full text papers are retrieved
for those that meet the inclusion criteria.
(b) Methodological quality - Following a full-text selection
stage, the remaining studies are assessed for methodological
quality using a critical appraisal framework. Poor quality
studies are excluded but are usually discussed in the review
report.
(c) Extraction of data - Of the remaining studies, reported
findings are extracted onto a data extraction form. Some
studies will be excluded even at this late stage. A list of
included studies is then created. Assessment should ideally
be conducted by two independent reviewers.
Combining the results 22. The findings from the individual
studies must then be aggregated to produce a ‘bottom line’
on the clinical effectiveness, feasibility, appropriateness and
meaningfulness of the intervention or activity. This
aggregation of findings is called evidence synthesis. The type
of evidence synthesis is chosen to fit the type (s) of data
within the review. If a systematic review inspects qualitative
142 An Introduction to Research Methodology

data, then a meta-synthesis is conducted. Narrative

summaries are used if quantitative data are not
homogenous23.
Placing the findings in context22. The findings from this
aggregation of an unbiased selection of studies then need to
be discussed to put them into context. This will address issues
such as the quality and heterogeneity of the included studies,
the likely impact of bias, as well as the chance and the
applicability of the findings. Thus, judgement and balance
are not obviated by the rigour of systematic reviews – they
are just reduced in impact and made more explicit.
2. Meta-analysis
Meta-analysis can be done as a continuation of systematic
review if the variables are not much heterogeneous. Statistical
analysis is usually an essential part of meta-analysis and the
recommended software by Cochrane collaboration is
RevMan24.
The analytical part mainly deals with the following four
points
a) Direction of effect
b) Effect size
c) Consistency of effect across studies and
d) The strength of evidence for the effect
The plan of analysis mainly depends on the aim of the study.
Some reviews have more than one type of comparison. First,
may be to identify and collate all studies in a field and second
may be to identify the best intervention. Some reviews aim
to investigate the relationship between the size of an effect
and some characters of the studies. In this area meta-analysis
is having a comparatively less role.
 An Introduction to Research Methodology 143

Meta-analysis can be used in the review to increase the power

and precision, to answer questions not possible by individual
studies, to settle controversies arising from apparently
conflicting studies and to generate new hypothesis. But it is
not having much role in entirely different studies as the
heterogeneity increases poor or false results. The result of
meta-analysis can be misleading if the studies are of poor
quality and/or with serious publication bias.
Type of Analysis
The type of analysis can only be decided by considering the
following points
(a) Type of data – for dichotomous data usual analysis are
Risk Ratio (Relative Risk), Odd’s ratio, Risk difference and
Number needed to treat. In the case of continuous data,
mean deviation, standard deviation and other statistical tests
are done. In the case of ordinal data, usually proportional
Odd’s ratio is used. Hazard ratio is the analysis for time to
event (Survival) outcome data25.
(b) Study design and identifying the unit of analysis – Type
of study design and the randomization pattern are the main
issues to be dealt here. Care must be taken to avoid bias
while selecting the data26.
(c) Intention-to-treat (ITT) analysis or available case
analysis27
Steps for analysis
The following steps should be followed for doing analysis
in meta-analysis
1. Data Extraction
After deciding the type of analysis the data for individual
studies can be extracted. For this special data extraction forms
144 An Introduction to Research Methodology

or programmes like RevMan can be used. It also depends

on the type of data and nature of the study28.
2. Data Summarizing
After data extraction overall statistical test can be carried
out to get pool data of Standard deviation, mean or t value
using RevMan software. It involves two steps, in the first
step a summary statistic is calculated for each study which
describes the treatment effect of the study. In the second
stage, a summary (pooled) treatment effect is estimated in
the individual studies. If the pooled effect of each studies
are not at all the same but follow a distribution across studies,
then one can perform a random meta-analysis. If it is
assumed that each study is estimating exactly the same
quantity a fixed effect meta-analysis can be performed. The
methods for meta-analysis mainly depend on the nature of
data, type of meta-analysis, consistency of estimates of
treatment effect across trials & sub-groups and ease of
interpretation of summary statistics. For example Mantel-
Haenszel, Peto and Inverse Variance are the three fixed
method and DerSimonian and Laird is a random method
for dichotomous data 29. Mantel-Haenszel method is mainly
based on the pooling of odds ratio of the individual studies
whereas Peto method is mainly based on the calculation with
inversing the variance. RevMan programme is mainly based
on inverse variance method.
Occasionally meta-analyses use “vote-counting” to compare
the number of positive studies with that of negative studies.
Vote-counting is limited to answering the simple question
“is there any evidence of an effect?” Two problems can occur
with vote-counting, which suggest that it should be avoided
whenever possible. Firstly, problems occur if subjective
decisions or statistical significance are used to define
 An Introduction to Research Methodology 145

“positive” and “negative” studies. To undertake vote counting

properly the number of studies showing harm should be
compared with the number showing benefit, regardless of
the statistical significance or size of their results. A sign test
can be used to assess the significance of evidence for the
existence of an effect in either direction. Secondly it takes
no account of the differential weightage given to each study.
3. Handling heterogeneity
Even though heterogeneity is the main attraction and key
point by which the result of meta-analysis is extrapolated in
the population heterogeneity, the results will be wrong
directional if a large heterogeneity exist or if the
heterogeneity is not properly handled. The reasons for
heterogeneity may be difference in the patient population
studies, intervention used, co-interventions, outcome
measured, study design, quality of studies or it may be due
to random error. This can be identified by two methods (a)
Graphically by forest plot i.e. if the individual studies are
more scattered in the forest plot from the pooled value then
more heterogeneity will be there and (b) statistically by Chi-
square test for heterogeneity (Mantel-Haenszel test or
Cochrane Q test). Heterogeneity is usually represented as
I2. Thresholds for the interpretation of I2 can be misleading,
since the importance of inconsistency depends on several
factors. A rough guide to interpretation is as follows, 0% to
40%: might not be important; 30% to 60%: may represent
moderate heterogeneity; 50% to 90%: may represent
substantial heterogeneity; 75% to 100%: considerable
heterogeneity30.
Usually there are certain criteria for tackling heterogeneity
like checking the data thoroughly, exploring heterogeneity,
ignoring heterogeneity, performing a random effect meta-
146 An Introduction to Research Methodology

analysis, changing the effect measure or excluding studies

with high heterogeneity31.
4. Examining the quality of the study and publication bias
The quality of the studies can be determined by detailed
narrative discussion of impact of quality on result; display
quality and result of each study in tabular form, sub group
analysis and by statistical method by include quality as a
covariate in meta-regression.
Not only the studies with negative results or with adverse
effects but also some other problems in the language,
citation, lack of proper funding etc. can also make
publication bias. This can be identified graphically by funnel
plot and statistically by Rosenthal’s Fail-safe N and Egger
test.
Funnel plot - A funnel plot is a simple scatter plot of the
intervention effect estimates from individual studies against
some measure of each study’s size or precision and to tackle
the publication bias. In common with forest plots, it is most
common to plot the effect estimates on the horizontal scale,
and thus the measure of study size on the vertical axis. This
is the opposite of conventional graphical displays for scatter
plots, in which the outcome (e.g. intervention effect) is
plotted on the vertical axis and the covariate (e.g. study size)
is plotted on the horizontal axis. The name ‘funnel plot’
arises from the fact that precision of the estimated
intervention effect increases as the size of the study increases.
Effect estimates from small studies will therefore scatter more
widely at the bottom of the graph, with the spread narrowing
among larger studies. In the absence of bias the plot should
approximately resemble a symmetrical (inverted) funnel. If
there is bias, for example because smaller studies without
 An Introduction to Research Methodology 147

statistically significant effects remain unpublished, this will

lead to an asymmetrical appearance of the funnel plot with
a gap in a bottom corner of the graph. In this situation the
effect calculated in a meta-analysis will tend to overestimate
the intervention effect (Egger 1997a, Villar 1997). The more
pronounced the asymmetry, the more likely it is that the
amount of bias will be substantial32.
Symmetrical funnel plot Asymmetrical funnel
plot with publication bias

Trim and fill method - The ‘trim and fill’ method aims both
to identify and correct for funnel plot asymmetry arising
from publication bias (Taylor 1998, Duval 2000). The basis
of the method is to (i) ‘trim’ (remove) the smaller studies
causing funnel plot asymmetry, (ii) use the trimmed funnel
plot to estimate the true ‘centre’ of the funnel, then (iii)
replace the omitted studies and their missing ‘counterparts’
around the centre (filling). The trim and fill method requires
no assumptions about the mechanism leading to publication
bias, provides an estimate of the number of missing studies,
and also provides an estimated intervention effect ‘adjusted’
for the publication bias (based on the filled studies).
However, it is built on the strong assumption that there
should be a symmetric funnel plot, and there is no guarantee
that the adjusted intervention effect matches what would
have been observed in the absence of publication bias, since
148 An Introduction to Research Methodology

we cannot know the true mechanism for publication bias.

Equally importantly, the trim and fill method does not take
into account reasons for funnel plot asymmetry other than
publication bias. Therefore, ‘corrected’ intervention effect
estimates from this method should be interpreted with great
caution. The method is known to perform poorly in the
presence of substantial between-study heterogeneity.
Additionally, estimation and inferences are based on a dataset
containing imputed intervention effect estimates. Such
estimates, it can be argued, inappropriately contribute
information that reduces the uncertainty in the summary
intervention effect33.
Rosenthal suggested assessing the potential for
publication bias to have influenced the results of a meta-
analysis by calculating the ‘fail-safe N’, the number of
additional ‘negative’ studies (studies in which the
intervention effect was zero) that would be needed to increase
the P value for the meta-analysis to above 0.05. However
the estimate of fail-safe N is highly dependent on the mean
intervention effect that is assumed for the unpublished
studies, and available methods lead to widely varying
estimates of the number of additional studies. The method
also runs against the principle that in medical research in
general, and systematic reviews in particular, one should
concentrate on the size of the estimated intervention effect
and the associated confidence intervals, rather than on
whether the P value reaches a particular, arbitrary threshold,
although related methods for effect sizes have also been
proposed. Therefore this and related methods are not
recommended for use in Cochrane reviews34.
5. Sub group analysis
Subgroup analyses involve splitting all the participant data
 An Introduction to Research Methodology 149

into subgroups, often so as to make comparisons between

them. It is a method to investigate heterogeneity in the
studies. Findings from multiple subgroup analysis are
misleading. They are observational in nature and are not
based on randomized comparison. Hence false positive or
false negative results may increase if the number of subgroups
increased irrationally. So proper selection of subgroups and
performing meta-regressions are very much important in
sub-group analysis to avoid misleading conclusions35.
6. Presenting the Data
Several methods like figures and tables are available for the
presentation of results in a meta-analysis as per Cochrane
collaboration. Most of them can be created by RevMan
software.
(a) As text – The results of individual studies and meta-
analyses in a Cochrane review are displayed in Figures and
Tables. Each Figure and Table should be referred to in the
results section of the review text. The results section should
summarise the findings in a clear and logical order, and
should explicitly address the objectives of the review. The
section should be organised to follow the order of
comparisons and outcomes specified in the protocol, and
used as the data structure in RevMan. Findings for the most
important comparisons and/or outcomes should be
prominent in the text of the review, even when little relevant
data is available. Answers to post hoc analyses and less
important questions for which there happen to be plentiful
data should not be over emphasised. Post hoc analyses should
always be identified as such. The analytic methods that are
used in a review should be described in the methods section.
The author should also make clear in the results section the
method of analysis used for each quoted result (in particular,
150 An Introduction to Research Methodology

the choice of effect measure, the direction of a beneficial

effect and the meta-analysis model used). Results should
always be accompanied by a measure of uncertainty, such as
a 95% confidence interval36.
(b) Figures – Main figure for representing the data is forest
plot and funnel plot.
Forest plot - It is a graphical representation of a meta-analysis
which itself is the emblem of Cochrane collaboration. It is
usually accompanied by a table listing references (author
and date) of the studies included in the meta-analysis. The
table also lists the mean scores and standard deviations of
these scores from each of the included studies; and it lists
the number of participants in each study (under ‘Total’).
Each study is represented by a block at the point estimate of
treatment effect with a horizontal line extending either side
of the block. The area of the block indicates the weight
assigned to that study in the meta-analysis while the
horizontal line depicts the confidence interval (usually with
a 95% level of confidence)37.

In the example, the first line represents the SMD of scores

from the Evans 2011 study, (SMD=-0.50); the second line
represents the SMD for scores from the Jones 2010 study
(SMD=-0.30); and the third line represents these scores from
the Smith 2013 study (SMD=-0.50).The black diamond at
the bottom of the graph shows the average effect size of the
three studies37.
 An Introduction to Research Methodology 151

(b) Tables - RevMan supports three types of tables of results

that can be linked to the Results text of the review. First one
is about pertaining to the data of forest plot, second table of
Comparisons allows an outcome type of ‘Other data’ and
the third table provided by the Additional Tables feature,
allowing presentation of results of both trials and meta-
analyses, and other meta-analytical investigations38.
Results - Interpretation
While preparing the final result, the following areas should
be clearly defined such as strength of evidence, applicability,
other relevant information, adverse effects and trade off. A
common mistake when there is inconclusive evidence is to
confuse ‘no evidence of an effect’ with ‘evidence of no effect’.
When there is inconclusive evidence, it is wrong to claim
that it shows that an intervention has ‘no effect’ or is ‘no
different’ from the control intervention. It is safer to report
the data, with a confidence interval, as being compatible
with either a reduction or an increase in the outcome. When
there is a ‘positive’ but statistically non-significant trend
authors commonly describe this as ‘promising’, whereas a
‘negative’ effect of the same magnitude is not commonly
described as a ‘warning sign’. One way of avoiding errors
such as these is to consider the results blinded; i.e. consider
how the results would be presented and framed in the
conclusions if you reversed the direction of the results38.
Another common mistake is to reach conclusions that go
beyond the evidence that is reviewed. Often this is done
implicitly, without referring to the additional information
or judgements that are used in reaching conclusions about
the implications of a review for practice. Even when
conclusions about the implications of a review for practice
are supported by additional information and explicit
152 An Introduction to Research Methodology

judgements, the additional information that is considered

is rarely systematically reviewed and implications for practice
are often dependent on specific circumstances and values
that must be taken into consideration. Authors should always
be cautious about reaching conclusions about implications
for practice and they should avoid making
recommendations38.
Checklists
There are many checklists for the assessment of the quality
of systematic reviews39; however, the QUOROM statement
(quality of reporting of meta-analyses) is particularly
recommended40.
Reporting
The PRISMA Statement - Preferred Reporting Items for
Systematic Reviews and Meta-Analyses is an evidence-based
minimum set of items for reporting in systematic reviews
and meta-analyses. A 27-item checklist, PRISMA focuses
on randomized trials but can also be used as a basis for
reporting systematic reviews of other types of research,
particularly evaluations of interventions. PRISMA may also
be useful for critical appraisal of published systematic
reviews, although it is not a quality assessment instrument
to gauge the quality of a systematic review41.
MOOSE Guidelines - Meta-analysis of Observational
Studies in Epidemiology checklist contains specifications for
reporting of meta-analyses of observational studies in
epidemiology. Editors will expect you to follow and cite this
checklist. It refers to the Newcastle-Ottawa Scale for
assessing the quality of non-randomized studies, a method
of rating each observational study in your meta-analysis42.
 An Introduction to Research Methodology 153

Conclusion
Appropriately and precisely accomplished Systematic review
and Meta-analysis are milestones in the development of
pertinent scientific knowledge. Therefore these enjoy the
supreme position in the hierarchical pyramid of clinical
studies. Hence utmost care must be taken regarding
collection of scientific data, detecting bias in publications,
eliminating unsuitable analysis and interpretations to bring
forth definite conclusive facts.

References
1. Pearson K. Report on Certain Enteric Fever Inoculation
Statistics. Br Med J 1904; 2 doi: http://dx.doi.org/10.1136/
bmj.2.2288.1243 (Published 05 November 1904): Br Med
J 1904; 2:1243.
2. Tippett, LHC. in: The Methods of Statistics. Williams &
Norgate, London; 1931.
3. R.A. Fisher. Statistical methods for research workers
[electronic resource] - 4th ed., rev. and enl. Edinburgh Oliver
and Boyd - Biological monographs and manuals, 1938.
4. Dr. Anish T S. Lecture on Meta-analysis. Research
Methodology Training Program for Medical Teaching Faculty,
Kerala University of Health Sciences. 2015.
5. Glass GV. Primar y, secondary, and meta-analysis of
research. Educational Researcher. 1976; 5:3–8.
6. DerSimonian, R.; Kacker, R. Random-effects model for meta-
analysis of clinical trials: An update, Contemporary Clinical
Trials, Volume 28, Issue 2, February 2007, Pages 105-114.
7 A.A. Bartolucci and W. B. Hillegass F. Chiappelli et al. (eds.),
Evidence-Based Practice: Toward Optimizing Clinical
Outcomes, 17 DOI: 10.1007/978-3-642-05025-1_2, ©
Springer-Verlag Berlin Heidelberg 2010.
154 An Introduction to Research Methodology

8. Tamoxifen for early breast cancer: an overview of the

randomized trials. Early Breast Cancer Trialists’ Collaborative
Group. [No authors listed]. Lancet. 1998 May
16;351(9114):1451-67.
9. Pippa Hemingway, Nic Brereton. What is a systematic review?
Second Ed. April 2009 www.whatisseries.co.uk. Hayward
Medical Communications.
10. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC,
Ioannidis JPA, et al. (2009) The PRISMA Statement for
Reporting Systematic Reviews and Meta-Analyses of Studies
That Evaluate Health Care Interventions: Explanation and
Elaboration. PLoS Med 6(7): e1000100. doi: 10.1371/
journal.pmed.1000100.
11. Margaliot, Zvi, Kevin C. Chung. Systematic Reviews: A
Primer for Plastic Surgery Research. Plastic & Reconstructive
Surgery: December 2007 - Volume 120 - Issue 7 - pp 1834-
1841. doi: 10.1097/01.prs.0000295984.24890.2f
12. Iain K Crombie, Huw TO Davies. What is meta-analysis?
Second Ed. April 2009 www.whatisseries.co.uk Hayward
Medical Communications.
13. Public Health Resource Unit. Critical Appraisal Skills
Programme (CASP). Making sense of evidence: 10 questions
to help you make sense of reviews. www.phru.nhs.uk/
Doc_Links/S.Reviews%20Appraisal%2 0Tool.pdf (last
accessed 19 November 2008)
14. Centre for Evidence-Based Medicine. Critical Appraisal.
www.cebm.net/index.aspx?o=1157 (last accessed 23 January
2009)
15. Narahari SR, Aggithaya GM, Suraj KR. Ayurveda cikitsa for
switra (Vitiligo) (Protocol). 2008; Institute of applied
dermatology. www.indiandermatology.org. (last accessed 3
February 2009)
 An Introduction to Research Methodology 155

16. Bailar JC 3rd. The promise and problems of meta-analysis.

N Engl J Med 1997; 337: 559–561.
17. Higgins JPT, Green S. Cochrane Handbook for Systematic
Reviews of Interventions 4.2.6. http://cochrane.org/
resources/handbook/Handbook4.2.6Sep 2006.pdf (last
accessed 28 October 2008)
18. Jüni P, Altman DG, Egger M. Assessing the quality of
randomized controlled trials. In: Egger M, Smith GD, Altman
DG (ed). Systematic Reviews in Health Care: Meta-analysis
in context, 2nd ed. London: BMJ Publishing Group, 2001.
19. Cook DJ, Sackett DL, Spitzer WO. Methodologic guidelines
for systematic reviews of randomized control trials in health
care from the Potsdam Consultation on Meta-Analysis. J Clin
Epidemiol 1995; 48: 167–171.
20. Deeks JJ, Altman DG, Bradburn MJ. Statistical methods for
examining heterogeneity and combining results from several
studies in meta-analysis. In: Egger M, Smith GD, Altman
DG (ed). Systematic Reviews in Health Care: Meta-analysis
in context, 2nd ed. London: BMJ Publishing Group, 2001.
21. Thornton A, Lee P. Publication bias in meta-analysis: its
causes and consequences. J Clin Epidemiol 2000; 53: 207–
216.
22. Egger M, Davey Smith G, Schneider M, Minder C. Bias in
meta-analysis detected by a simple, graphical test. BMJ 1997;
315: 629–634.
23. Pippa Hemingway, What is a systematic review, obtained from
net file http://www.medicine.ox.ac.uk/ bandolier/painres/
download/whatis/syst-review.pdf on 05/02/2016
24. Sandelowski M, Barroso J. Toward a metasynthesis of
qualitative findings on motherhood in HIV-positive women.
Res Nurs Health 2003; 26: 153–170.
156 An Introduction to Research Methodology

25. Higgins JPT. Cochrane Handbook for Systematic Reviews

of Interventions. Text, 2006 September,97-99 obtained from
the net file http://doi.org/10.3233/PRM-2010-0129
26. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,101-109 obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
27. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,110,111 obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
28. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,112,113 obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
29. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,115 obtained from
the net file http://doi.org/10.3233/PRM-2010-0129
30. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,126-135 obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
31. From net http://handbook.cochrane.org/chapter_9/
9_5_2_identifying_and_measuring_ heterogeneity.htm on
22/03/2016
32. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,136-140 obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
33. Obtained from net file http://handbook.cochrane.org/
chapter_10/10_4_1_funnel_plots.htm on 21/02/2016
34. Obtained from net file http://handbook.cochrane.org/
chapter_10/10_4_4_2_trim_ and_fill.htm on 21/03/2016
35. Obtained from net file http://handbook.cochrane.org/
chapter_10/10_4_4_3_fail _safe_n.htm on 21/03/2016
36. Word to word copy, Higgins JPT. Cochrane Handbook for
Systematic Reviews of Inter ventions. Text, 2006
 An Introduction to Research Methodology 157

September,147, obtained from the net file http://doi.org/

10.3233/PRM-2010-0129
37. Centre for evidence based intervention, Department of Social
policy and inter vention, from the net file http://
www.cebi.ox.ac.uk/for-practitioners/what-is-good-evidence/
how-to-read-a-forest-plot.html on 21/02/2016
38. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,150, obtained from
the net file http://doi.org/10.3233/PRM-2010-0129
39. Higgins JPT. Cochrane Handbook for Systematic Reviews
of Interventions. Text, 2006 September,167-171, obtained
from the net file http://doi.org/10.3233/PRM-2010-0129
40. Shea B, Dubé C, Moher D. Assessing the quality of reports
of systematic reviews: the QUOROM statement compared
to other tools. In: Egger M, Smith GD, Altman DG (ed).
Systematic Reviews in Health Care: Meta-analysis in context,
2nd ed. London: BMJ Publishing Group, 2001.
41. Moher D, Cook DJ, Eastwood S et al. Improving the quality
of reports of meta-analyses of randomized controlled trials:
the QUOROM statement. Quality of Reporting of Meta-
analyses. Lancet 1999; 354: 1896–1900.
42. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group.
Preferred reporting items for systematic reviews and meta-
analyses: the PRISMA statement. PLoS Med. 2009 Jul
21;6(7): e1000097. Epub 2009 Jul 21. PubMed PMID:
19621072.
43. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD,
Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-
analysis of observational studies in epidemiology: a proposal
for reporting. Meta-analysis Of Observational Studies in
Epidemiology (MOOSE) group. JAMA. 2000 Apr
19;283(15):2008-12. PubMed PMID: 10789670.
158 An Introduction to Research Methodology

Chapter 13

Questionnaire Development
Sreedivya

A questionnaire is a research instrument consisting of a

concise series of pre-planned set of questions for the purpose
of gathering information from respondents about a pertinent
topic. It is a standard method for data collection in clinical,
epidemiological, psychosocial and demographic research.
Questionnaire was invented by the Statistical Society of
London in 1838. The questions in a questionnaire are also
known as items.
Purpose:
A questionnaire can help in assessing an individual’s
attitudes, beliefs, experiences, behaviour or attributes.
 Attitude-. what people say what they want
 Belief -. what people think is true
 Behaviour-. What people do
 Attribute -. What people actually are
Overall considerations
As a questionnaire relies much on its own merits to elicit
accurate and complete information, it should be designed
 An Introduction to Research Methodology 159

so as to obtain information relevant to the purpose and to

do this with maximum reliability and validity.
Before designing a questionnaire, start by asking yourself a
few important questions and decide what you are measuring:
 What do I need to know?
 Why do I need to know it?
Consider which of the following you are aiming to measure:
 Attitude
 Goals, intentions, aspirations
 Experiences
 Behaviors and practices
 Perceptions of knowledge, skills
 Attributes or characters
Be clear about the purpose of your questionnaire, the
information you will need and how it will contribute to the
research question. Develop a scientific research question and
be certain that questionnaire is the best way to obtain the
information. Each question should have an explicit rationale
for being in the questionnaire. Questions that do not directly
address the research problem should be eliminated. While
preparing a questionnaire, consider the nature of your topic,
restraints of time and money and the respondents with regard
to their age, education level, familiarity with tests and
questionnaires. Also consider cultural bias/language barrier.
The quality of information will be maximized if wording
and sequence are designed to motivate the respondent,
facilitate recall, if questionnaires are interesting, and not time
consuming, embarrassing or threatening. It should show
respect to respondent’s dignity and privacy. Confidentiality
must never be breached.This is crucial in attaining honest
160 An Introduction to Research Methodology

and complete answers from respondents.

Types of questionnaire
The type of questionnaire should be selected according to
the purpose of research, the topic area and the respondents.
The questionnaire may be Interviewer administered or Self-
administered and Open-ended or Closed-ended.
Interviewer administered questionnaire (also known as
inter view schedules): This involves use of trained
interviewers to administer the questionnaire. The interview
is usually done in a nonthreatening environment.
It has many advantages.
 The Interviewer is sure who is responding and who is
not.
 He can reduce the number of items omitted by
respondent.
 The interviewer can determine if subject is having
difficulty in understanding items.
 It allows the interviewer to repeat questions if not
understood by respondent.
 It can avoid confusions about skip patterns.
 The presence of interviewers can help reduce errors.
 If the interviewer is present, he can have a combination
of both closed and open ended questions.
The disadvantages are;
 Increased cost in time and money.
 Uniform training of interviewers is must. Otherwise it
may lead to false results.
 An Introduction to Research Methodology 161

 Number of possible interviews in a day may be limited.

 If there are several different language sites, there arises a
need to have bilingual interviewers.
 Attributes of the interviewer may affect responses e.g.
biases of interviewer.
Self-rated questionnaire: Here the interviewer is not present
at the spot but the respondents read instructions and fill
out at their leisure. Or else for a group of respondents, a
single interviewer will be present. The presence of interviewer
can ensure completion rate.
It has the following advantages;
 Reduced cost is involved.
 Can reach a wide population at low cost if sent out by
mail.
 It is a good means of collecting data from specialized
and highly motivated groups.
 Absence of interviewer can minimize social desirability
bias. Social desirability bias is a social science research
term that describes the tendency of survey respondents
to answer questions in a manner that will be viewed
favorably by others. It can take the form of over-reporting
“good behavior” or under-reporting “bad” or undesirable
behavior.
There are disadvantages too.
 It will not work in the case of non-literate persons.
 Less educated persons will have difficulty following
instructions.
 Poor return rate in case of mailed questionnaires.
162 An Introduction to Research Methodology

 The respondents may omit some items

 The interviewer cannot be sure who actually filled the
questionnaire
 No assurance can be given that subjects will read in the
desired sequence
Closed ended questionnaire: limit the answers of the
respondents to response options provided in the
questionnaire.
 Advantages: time-efficient, responses are easy to code
and interpret; ideal for quantitative type of research
 Disadvantages: respondents are required to choose a
response that does not exactly reflect their answer; the
researcher cannot further explore the meaning of the
responses
Open-ended questionnaire: there are no predefined options
or categories included. The participants are free to provide
their own answers
 Advantages: participants can respond to the questions
exactly as how they would like to answer them; the
researcher can investigate the meaning of the responses;
ideal for qualitative type of research
 Disadvantages: time-consuming; responses are difficult
to code and interpret.
Steps in designing a questionnaire
Devising the items in a questionnaire
Certain aspects have to be considered while devising a
questionnaire.
 An Introduction to Research Methodology 163

The Sources of items

Items of a questionnaire usually derive from theory, work
experience/clinical observation, research finding, or expert
opinion. First of all, list out all broad areas and then write
specific questions.
Wording of items
Interpretability depends on the questions and thus wording
of items have to be given due importance. The wording of
questions should be simple and comprehensible to the
respondent population. The reading level should be that of
a 12year old. The following type of questions should
generally be avoided
Ambiguous questions : The definition of an Ambiguous
Question within a survey is a question in which there is
more than a single way to interpret it. When a question is
vague or generalized, this creates an ambiguous inter-
pretation. Eg. Where do you like to shop? In this ambiguous
question it is not exactly clear as to what kind of shopping
you are referring to. Are you asking where you like to do
your regular grocery shopping? Or perhaps where do you
shop for dresses? Or maybe it was even asking what online
website you like to shop at. The reason why ambiguous
questions are problematic is because they allow the
respondent to interpret the meaning of the question however
they like. They will then respond to the question according
to their interpretation. When ambiguous questions are
present in survey they will create false data and cause the
researcher to draw conclusions based upon one meaning
while the respondents actually responded to a different
interpretation.
 Double barreled questions: These are the questions where
164 An Introduction to Research Methodology

two or more questions are combined into one. E.g. How

satisfied or dissatisfied are you with the pay and work
benefits of your current job? In the case of the example
above, it makes sense to break the question into two;
satisfaction with pay and satisfaction with work benefits.
Otherwise, some of your respondents will be answering
the question while giving more weight to pay, and others
will answer giving more weight to work benefits. Hence,
the question may be split in to two;
How satisfied or dissatisfied are you with the pay of your
current job?
How satisfied or dissatisfied are you with the work benefits
of your current job?
 Double negatives (e.g., do you favor or oppose not
allowing gays and lesbians to legally marry)
 Questions that are too demanding and time consuming.
Eg. Rank the following 15 items according to the
importance.
 Embarrassing/ Threatening questions.
 Questions about experiences or practices which are too
difficult to remember.
 Leading and value laden questions.
Sequencing of items: Questions should be sequenced in such
a way that they follow a logical order. General questions
should precede specific questions. Going from general to
specific is called the funnel approach, because you begin
with broader (more general) questions and then ask narrower
(more specific) questions, reflecting the shape of a funnel
The opening questions can introduce the topic, attempt to
gain the confidence and cooperation of respondents, or
 An Introduction to Research Methodology 165

establish the legitimacy of the study. These questions should

be interesting, simple and non-threatening. A few easy to
answer neutral questions may be helpful to build a
relationship with respondents. Transition statements and
filter questions (skip patterns) may be added if necessary to
maintain the flow.
Transition statements are a sort of signal. They are words,
phrases, or sentences that connect one topic or idea to
another in a smooth, coherent way. They essentially let the
reader know that you are about to change directions. A skip
pattern is a question or series of questions associated with a
conditional response. Although all respondents answer most
questions on a survey or questionnaire, some questions
pertain only to certain respondents. For instance, a typical
question on a survey might inquire about a respondent’s
gender: “a” for male, “b” for female. A possible skip pattern
for this question would read “If you marked ‘a’, skip to
question 4; if you marked ‘b’, continue to the next question.”
The questions skipped by male respondents might refer to
issues that are relevant only to females; for instance,
questions related to the physical act of giving birth.
Formatting the Response
Response format depends on the nature of questions. In
order to collect answers from the participants, the response
format is utilized. The response format is comprised of
response options found in the questionnaire. Basically,
response formats are divided into two types: structured and
unstructured response formats.
Structured Response Formats: Structured questions ask
respondents to select an answer from a given set of choices.
They are more beneficial in surveys because respondents can
answer in a more efficient and easier manner. Researchers
166 An Introduction to Research Methodology

find it easier to summarize participant responses through

structured response options. There are many types of
structured response formats. Some examples are given below:
1. Fill-in-the-Blank
D Put a check mark on the blank which applies:
Do you use the Internet for research purposes?
_____Yes ____No
D Another type of fill-in-the-blank response format is rating
the options. For instance:
In a scale of 1 to 5, please rate the following according to
your preferred car brand, where “1” as your most preferred,
and “5” as your least preferred.
____ Ford ____Honda
____Nissan ____Toyota
____Others (please specify: ________________)
2. Multi-Option Format
As the name suggests, you present a question to the
respondent and he will answer it based on the multiple
options available. Here is an example:
How many hours do you usually spend using Facebook per
day?
i less than 1 hour
ii 1-3 hours
iii more than 3 hours
 Unstructured Response Formats: While researchers utilize
structured response formats for greater efficiency, many use
unstructured response formats to gain more understanding
about what the respondent really means when he/she answers
 An Introduction to Research Methodology 167

a particular question. These response formats are used in

qualitative research studies, especially those which try to
explore the feelings, experiences and perceptions of the
respondents. Unstructured response formats simply require
the respondent to write his answer in detail. An example of
this is: How do you use the Internet in your daily life?
Some survey questionnaires contain both structured and
unstructured response formats. First, the researcher asks
questions with a structured response format, and then the
questionnaire is ended by an unstructured format with
“Please add any thoughts regarding [the subject]” and
placing blank spaces to accommodate the respondent’s
answer.
Response categories
The nature of the question asked mostly decides the response
options. The kinds of possible responses to questions/items
in structured response formats is based on two related
features: the nature of response (continuous or categorical)
and level of measurement (nominal, ordinal, interval or ratio
variables). Haemoglobin and blood pressure measurements
are examples of continuous measurements, and measurement
of religion and marital status are categorical in nature.
Similarly religion and marital status are nominal variables
also as they have named categories (Eg. Hindu, Christian,
Muslim). Ordinal variables have responses in ordered
categories eg. stages of breast cancer I, II, III & IV. Interval
level measurements are quantitative, with no meaningful zero
point and equal interval between responses eg. IQ, body
temperature. Ratio variables (continuous variables) are also
quantitative with meaningful zero point and the ratio of
two responses having some meaning eg. height, weight,
haemoglobin. These distinctions are important in
168 An Introduction to Research Methodology

determining the appropriate method of analysis.

Based on the number of categories in the response, It can be
a dichotomous or multi-chotomous response format, a
numerical answer or take the form of a rating scale.
Dichotomous response format : This include a dichotomous
choice of ‘yes’ or ‘no’, ‘approve’ or ‘disapprove’, or ‘effective’
or ‘not effective’. Questions of this type generally include a
‘don’t know’ response category also.
Multi-chotomous response format: Include more than two
options to choose from. Eg. The options for a question on
‘Source of information about contraceptives’ could be family
members, friends, hospital, mass media ,or others (specify)
If the respondent is allowed to select all the options that
apply, the multiple response formats is used. Response
categories should be mutually exclusive and all inclusive
Numerical answer: Quantitative variables are often elicited
in numbers; eventhough categorization is possible for these.
Eg. How old are you? , What is your monthly family income?
Rating scales: A rating scale is a set of categories designed
to elicit information about a quantitative or a qualitative
attribute. Structured response formats that utilize scale
formats also have a dichotomous, a three-point, a five-point,
a seven-point or a multiple-step scale as in semantic
differential scale. Each of these response scales has its own
advantages and disadvantages, but the rule of thumb is that
the best response scale to use is the one which can be easily
understood by respondents and interpreted by the researcher.
Dichotomous Scales: A dichotomous scale is a two-point
scale which presents options that are absolutely opposite each
other. This type of response scale does not give the
respondent an opportunity to be neutral on his answer in a
 An Introduction to Research Methodology 169

question.
Examples:
• Yes - No
• True - False
• Fair - Unfair
• Agree - Disagree
Three-point, five-point, and seven-point scales are all
included in the umbrella term “rating scale”. This rating
scale provides more than two options, in which the
respondent can answer in neutrality over a question being
asked.
Examples:
D Three-point Scales
• Good - Fair - Poor
• Agree - Undecided - Disagree
• Extremely- Moderately - Not at all
• Too much - About right - Too little
D Five-point Scales
• Strongly Agree - Agree - Undecided / Neutral -
Disagree - Strongly Disagree
• Always - Often - Sometimes - Seldom - Never
• Extremely - Very - Moderately - Slightly - Not at all
• Excellent - Above Average - Average - Below Average
- Very Poor
D Seven-point Scales
• Exceptional – Excellent – Very Good – Good – Fair
– Poor – Very Poor
170 An Introduction to Research Methodology

• Very satisfied - Moderately satisfied - Slightly satisfied

- Neutral - Slightly dissatisfied - Moderately
Dissatisfied- Very dissatisfied
The scale measurements, often having more than two
options, are not an interval level measurement. But for all
practical purposes they are considered as interval and
analysed accordingly.
Commonly used scales in Health Measurements
• A Likert-type scale assumes that the strength/intensity
of experience is linear, i.e. on a continuum from strongly
agree to strongly disagree, and makes the assumption that
attitudes can be measured. Respondents may be offered a
choice of five to seven or even nine pre-coded responses with
the neutral point being neither agree nor disagree.
Eg. Mobile phone should be banned in school campuses.
I———————————————————I
Strongly Agree No opinion Disagree Strongly
Agree Disagree
• Visual analogue scales
It is a line of fixed length with anchors like ‘no pain’and
‘pain as bad as it could be ‘at the extreme ends. Respondents
are required to place a mark, usually an ‘X’ or a vertical line
, on the line corresponding to their perceived state.
Eg. How severe is your arthritic pain today?
Pain as bad
as it could be No _____ Pain
• Semantic Differential Scales
The Semantic Differential (SD) measures people’s
 An Introduction to Research Methodology 171

reactions to stimulus words and concepts in terms of ratings

on bipolar scales defined with contrasting adjectives at each
end. This is a rating scale that can measure respondent
attitudes towards ideas, concepts, items, people, and events.
(Sometimes referred to as an attitudinal study). Semantic
differential questions simply ask where the respondent’s
position is on a scale between two bipolar adjectives, such
as “Happy-Sad,” “Creamy-Chalky,” or “Bright-Dark.”A
semantic differential scale is only used in specialist surveys
in order to gather data and interpret based on the connotative
meaning of the respondent’s answer. It uses a pair of clearly
opposite words, and can either be marked or unmarked.
Examples:
1. Marked Semantic Differential Scale
Please answer based on your opinion regarding the product:
very slightly neither slightly very
Inexpensive [ ] [] [] [] [] Expensive
Effective [] [] [] [] [] Ineffective
Useful [] [] [] [] [] Useless
Reliable [] [] [] [] [] Unreliable

2. Unmarked Semantic Differential Scale

The central line serves as the neutral point:
Inexpensive _____________|______________ Expensive
Effective ______________|______________ Ineffective
Useful _________________|_________________ Useless
Reliable _______________|________________ Unreliable
172 An Introduction to Research Methodology

There is no assurance whatsoever that a subject choosing

the middle scale position harbors a neutral opinion A
subject’s choice of the scale midpoint may result from:
Ignorance—the subject has no basis for judgment.
Uncooperativeness—the subject does not want to go to the
trouble of formulating an opinion.
Reading difficulty—the subject may choose “Undecided”
to cover up inability to read.
Reluctance to answer—the subject may wish to avoid
displaying his/her true opinion.
Inapplicability—the question does not apply to the subject.
Translation of a Questionnaire:
Translating the questionnaire has to be done to minimize
the ethnocentric bias and to maximize its usefulness.
To bring cultural equivalence, it has to be translated not
only into the required language, but the culture as well.
E.g.: using an intelligence test with a cultural group other
than the one for which it was designed.
Conceptual Equivalence- the Conceptual Equivalence of a
questionnaire indicates that an item measures the same
concept in all languages into which this questionnaire has
been translated. It differs from linguistic equivalence. There
should be no differences in meaning & content between
two versions of an instrument E.g. Chocolates—Candy,
Biscuits —cookies
Approaches to translation:
1. Translation and Back translation- Here Original
instrument is translated into target language & then back
to source language
 An Introduction to Research Methodology 173

 Translated by people fluent in both language

 Translators should have knowledge about the content
 Back translated by independent people
 See the agreement
2. Problems in translation can be
 Translator not being sufficiently skilled
 Translator not being culturally sensitive
 Translator’s peculiarities of word usage
 Following formal dictionary translation
 Poorly written target language version
Importance of Pretest and pilot study in questionnaire
development
 Pretesting refers to testing the questionnaire on a small
sample of respondents, usually 15 to 30, to identify and
eliminate potential problems. It involves a series of activities
designed to evaluate a survey instrument’s capacity to collect
the desired data, the capabilities of the selected mode of
data collection, and the overall adequacy of the field
procedures. It tests whether the instrument would elicit
responses to achieve the research objectives. It gives an
evaluation of the questionnaire, including the question
content, wording, format, sequence, layout, question
difficulty and instructions should tested. Pretests also test
comprehensibility
 Ensure that items are unambiguous & ask only a single
question
 Ensure that the questions are interpreted similarly by all
respondents
174 An Introduction to Research Methodology

 Test the qualitative aspects like question structure and

sequence
Pre-testing should be conducted with a subset of the
respondent group. The pre-test groups should be similar to
the respondents in terms of their background characteristics,
familiarity with the topic, and attitudes and behaviours of
interest.
Pre-tests are best done by personal interviews, even if the
actual survey is to be conducted by telephone, mail, or
electronically, so that interviewers can observe respondent
reactions and attitudes. Pre-tests may be done among peers
(peer review), experts (expert review) and respondent
population (respondent review). Based on feedback from
the pre-test, the questionnaire should be edited, and the
identified problems corrected. After each significant revision
of the questionnaire, another pre-test should be conducted,
using a different sample of respondents. Pre-testing should
be continued until no further changes are needed.
 Pilot study
We use the term “pilot study” to refer to pretesting
procedures that employ all the procedures and materials
involved in data collection (regardless of how small of a scale)
before the actual data collection begins. Pilot studies are
also referred to as “dress rehearsals”, or “field tests” in the
survey literature as they are conducted in an environment
and context similar to that of the actual study. It is a ‘small
scale replica’ of the main study. They have a specific goal –
from estimating response rates under a particular recruitment
protocol to identifying an optimal design characteristic
through experimentation. Pilot studies are not optional, it
is crucial.
 An Introduction to Research Methodology 175

By a pilot study,
• Nature of population and variability can be identified
which are important in determining an efficient sample
design
• Shows available sampling frame which is adequate,
complete, up to date and convenient
• Reveals logistic issues and feasibility
• Helps to discover the nature of relationship between
variables
• Confirms the adequacy of tool for data collection and
relative suitability of alternative methods of data
collection
• After pilot study the researcher shows better approaches
to subjects, introduction, rapport etc.
• Helps in sample size calculation. The sample should be
large enough to fulfill the above functions.
• Pre-coding of questionnaire can be done
Note: Why is pilot study necessary?
It gives a better knowledge of the problem under study. It is
difficult to plan a major study without adequate knowledge
of subject matter, population to be covered, how long will it
take, how much money will it cost, what data collection
method is appropriate etc.
Considerations before administration of questionnaire
Before administration of questionnaire the researcher must
decide on
• Sample population and sample size should be decided at
the beginning itself
176 An Introduction to Research Methodology

• Uniform training of interviewers (if more than one)

should be done
• If mailed, an efficient system to keep track should be
there
• Coding- Assigning code numbers to response categories
• Questionnaires may be either pre-coded and post-coded
• Planning the analysis simultaneously with questionnaire
design is essential
Assessment of reliability and Validity
Reliability and validity are two essential requisites of a
measuring instrument.
Reliability refers to consistency of measurement, whether
the instrument (eg. a test, questionnaire) is measuring
something in a reproducible fashion. It shows whether the
instrument is consistent in producing the same or similar
results in measurements of individuals on different occasions
or by different observers or similar or parallel tests.
Reliability Estimates:
Reliability (Repeatability/Stability) is not measured, it is
estimated. The methods are
1. Test-retest reliability: To assess consistency of measures
on repeated applications
2. Intra-rater / inter-rater reliability: Intra-rater reliability is
the degree of agreement among repeated administrations
of an instrument by a single rater whereas Inter rater
Reliability is determining whether two observers are
being consistent in their observations. If the responses
are in categories, calculate percentage of agreement
between observers. If continuous, calculate correlation
between ratings of two observers.
 An Introduction to Research Methodology 177

3. Internal consistency reliability: Internal consistency

examines the inter-item correlations within an
instrument and indicates how well multiple items
combine as a measure of a single concept.( applicable in
measurement of subjective states)
Validity: A study is valid if it measures actually what it claims
to measure and if there are no logical errors in drawing
conclusions from the data. Validity not as easily quantified
as reliability
Types of validity are
1. Face Validity
2. Content Validity
3. Criterion Validity
4. Construct validity
Face validity is the least statistical estimate. It is simply an
assertion of researcher that he has reasonably measured what
he intended to measure. A colleague or expert may be asked
to review the measure and provide inputs on its relevance
and comprehensiveness.
Content validity: It is the extent to which the questionnaire
probes the various aspects of the area it is supposed to
measure. The Questionnaire should cover a representative
collection of items. It relates to the idea of completeness
and relies on reason and past experience regarding the
adequacy with which important content has been covered
and transformed to questions.
Criterion validity: It correlates with some other measure,
ideally a gold standard.
There are two forms of Criterion validity: Predictive and
concurrent
178 An Introduction to Research Methodology

 Predictive validity shows whether the questionnaire has

some value as an instrument of prediction, assessed by
follow-up study
 Concurrent validity assesses the practical usefulness of
the questionnaire as an instrument of classification and the
extent to which the questionnaire results agree with
independent external criterion. It is expressed in terms of:
a) coefficient of agreement b) Sensitivity and specificity
Construct validity assesses whether the questionnaire results
are in accordance with the present theories concerning the
area of research. It refers to circumstantial evidence of the
usefulness of a questionnaire. It is used to refer something
not observable but constructed by investigators to account
for relationships in observed behaviour. E.g. Relationship
between anxiety and migraine
 Factor analysis is used to establish construct validity. It
is a technique which enables us to determine whether the
variables we have measured can be explained by a small
number of underlying factors. It is a multivariate reduction
technique which groups the number of variables into a
smaller set of uncorrelated factors. Factors represent specific
underlying domains. It tells us what variables belong together
– which ones virtually measure the same thing.
Note: Reliability and validity checks mentioned above need
to be followed meticulously if a scale is used or developed as
part of your study.
Common issues in the preparation of questionnaire
• Lack of attention
During literature search, question construction and pilot
testing
 An Introduction to Research Methodology 179

• Survey Question format

It should be easy to follow and pleasing to the eye.
Respondents are more likely to answer a questionnaire when
they see it as interesting, of value, short, clearly thought
through, and well presented.
• Respondent characteristics
May influence questionnaire results- eg. Age, literacy, etc
• Response error
Sources like memory, motivation, communication and
knowledge
• Response bias
 Social desirability bias is a social science research term
that describes the tendency of survey respondents to answer
questions in a manner that will be viewed favorably by others.
 Response set bias, which is the tendency for a person to
respond to questions or statements in such a way that it
produces a certain image of the respondent, rather than
answering based on the respondent’s true feelings or
behaviors. Response sets may occur because a person is
purposefully trying to be deceitful, or they may be
unintentional
 End-aversion bias (end-of-scale or central tendency bias):
In questionnaire-based surveys, respondents usually avoid
ends of scales in their answers. They tend to try to be
conservative and wish to be in the middle.
• Return rate
Low return rate may be a source of bias, making results
unacceptable. It is a good idea to ask each potential subject
180 An Introduction to Research Methodology

to complete a very short pro-forma, confirming a few

essential details, including name and address. You will get a
higher response from this first approach; this could be done
by telephone and then followed up with a full questionnaire
by post. Introductory letters, reply paid envelopes and
follow-up telephone calls also help to raise the response rate
for self-administered questionnaires.
• Covering letter
It is an important aspect in mailed surveys to introduce the
study to respondents. Where the questionnaire is
administered by post, or e-mail or some other indirect means,
it is usual practice to provide a letter that explains what the
questionnaire is about and why its completion is of value.
• Ethics
Give importance to the respondent’s privacy and dignity. If
a questionnaire contains sensitive questions yet must be
identified for accomplishment of its purpose, the best policy
is to promise confidentiality. In this case a code number
should be clearly visible on each copy of the instrument,
and the responders should be informed that all responses
will be held in strict confidence and used only in the
generation of statistics. Informing the responders of the
future use of study results is also likely to be helpful. When
some of the questions elicit personal or confidential
information, it is better to locate them at the end of the
questionnaire.
Conclusion
Questionnaires have advantages over some other types
of data collection tools in that they are cheap and large
amounts of information can be collected from a large number
 An Introduction to Research Methodology 181

of people in a short period of time and in a relatively cost

effective way and often have standardized answers that make
it simple to compile data. The results can be easily quantified
and analysed.
182 An Introduction to Research Methodology

Chapter 14

Ethics in Medical Research

Antony George, Smitha Ramadas

Ethics is the philosophical discipline that pertains to our

notions to differentiate good from bad, right from wrong
i.e. our moral duties and obligations to the society (science
of morality).1 Bioethics is application of ethics in the field
of healthcare. It incorporates philosophy, theology,
anthropology, history, and law, with practice of health care
.1 The field of bioethics has grown to encompass among
others an ethical concern in clinical settings, stem cell
research, genetic cloning, and reproductive technologies, to
broader concerns of human research participants, healthcare
policy, and allocation of scarce resources. This
multidisciplinary interconnecting and overlapping evolving
field may be broadly categorized under three headings1
1. Academic Bioethics: theoretical and practical biomedical
aspects of the moral obligations and responsibilities of
healthcare clinicians, researchers, and/or scholars.
2. Law Bioethics and Public Policy: legal and extra-legal
regulation of clinical and research practices, and framing
world policies for the betterment of the universe as a
whole.
 An Introduction to Research Methodology 183

3. Clinical Ethics: incorporation of bioethics in clinical

practice to improve patient care and management. Eg:
Code of conduct in the relationship of patient-healthcare
provider.
Research Bioethics
Research or biomedical research is defined by International
Organizations of Medical Sciences (CIOMS) 2 as
scientifically designed activity or activities to develop,
contribute, and/or improve the theories, principles,
relationships, and accumulation of information pertaining
to betterment of human life and healthcare. Research
bioethics is the branch of bioethics that deals with ethics in
the field of research in medical and life sciences. It deals
with the code of conduct in the relationship of participant
(subject)-researcher, including conflict of interest of sponsor/
investigator, to supervision and monitoring of research.
Evolution of Research Bioethics3-10
All the religions of the world preach ethics and propose codes
of conduct for human beings to follow. The ancient Indian
books of Vedas, Sushrutha Samhita, Charaka Samhita,
Bhagavad Gita, among others promulgate various principles
of ethics. The Hippocratic Oath about selfless service is
traced back to the 3-5th century. With the rise of scientific
experimentation in the 17th and 18th centuries, the need for
self-regulation and ethics was first articulated by American
and British doctors John Gregory and Thomas Percival.
Code of Nuremberg 1947:3 As an aftermath of the inhuman
treatment of Jews by the Nazi doctors in the concentration
camps for various human experiments and research code of
Nuremberg was involved. The key outlines of the code are
the following
184 An Introduction to Research Methodology

1. Human subjects should be included in any study only

with their Voluntary consent.
2. The experiment should yield fruitful results for the good
of society, unprocurable by any other scientific methods.
3. Anticipated results should justify outcome of the
experiment.
4. All unnecessary physical and mental suffering or injuries
to the subjects are to be avoided.
5. No research is to be conducted if priori reason is present
to believe that death or disabling injury shall occur to
the subject.
6. Research should be conducted only by scientifically
qualified persons.
7. Subject should be at liberty to end or opt out of
experiment.
8. Scientist in charge must be prepared to terminate
experiment at any stage to safeguard the health and
wellbeing of the subject.
Declaration of Helsinki 1964:4 In the late 1950s the drug
thalidomide was not approved by the Food and Drug
Administration (FDA) for use in USA. This unapproved
drug was however used by American physicians to treat
nausea associated with pregnancy on unsuspecting mothers,
and thalidomide was subsequently found associated with
limb-defects in more than 12,000 new-born. Eventually this
lead to the Declaration of Helsinki (Finland) in 1964, and
this document was later revised in 1975, 1983, 1989, 1996,
2000, 2002, 2008 and 2013. This forms the basis of Good
Clinical Practices (GCP). The main issues addressed in the
 An Introduction to Research Methodology 185

declaration are
1. Research on humans should be based on the results of
studies in laboratory and animal experiments.
2. Research protocols should be reviewed by an independent
committee prior to the initiation of the trail.
3. Informed consent from research par ticipants is
mandatory .
4. Research should be conducted by medically/scientifically
qualified individuals.
5. Risks from a clinical trial should not exceed benefits.
Belmont Report 1979:5 As a consequence of the 40 (1932-
1972) year unethical Tuskegee clinical trials sponsored by
the Dept. of Public Health, Govt. of USA and conducted
on patients infected by syphilis in Alabama on economically
backward African-American people. The important outlines
of this report drafted by the National Commission for the
Protection of Human Subjects of Biomedical and Behavioral
Research are
1. Ethical principles for research on humans.
2. Conflict between medical practice and research.
3. Concepts and application of the principles of respect for
persons (informed consent), beneficence (assessing risks
and benefits), and justice (participant selection).
The Nuremberg code 3, Helsinki deceleration 4, Belmont
report 5, and United Nations Universal Declaration of
Human Rights of 19486, provided the foundation for the
more recent important and internationally accepted
guidelines on bioethics.
186 An Introduction to Research Methodology

1. Guidelines for Good Clinical Practice -World Health

Organization (GCP-WHO) 1994. 7
2. International Conference on Harmonization - Good
Clinical Practice (ICH-GCP E6 R1) 1996.8
3. Council for International Organization of Medical
Sciences (CIOMS) 2002.2
4. Indian Council for Medical Research (ICMR) Guidelines
2006.9
5. Declaration of Helsinki - World Medical Association
(WMA) updated 2013.4
6. European Group on Ethics in Science and New
Technologies (EGE) updated 2015.10
Principles of Research Bioethics1,11
The three cornerstones or pillars of healthcare research
bioethics are
1. Autonomy: respect for individual is the corner stone of
this principle. This involves , protecting people with
diminished autonomy (orphans, prisoners, people with
special needs, and such others) .Voluntary participation,
informed written consent, privacy and confidentiality,
respecting cultural characteristics of individuals are key
components of this principle.
2. Beneficence and non-maleficence: do research for the
benefit of others and never do harm. Calculation of risk-
benefit ratio, ensuring scientific validity of research are
important components of this principle..
3. Justice: equal distribution of risks and benefits when
scientists conducting clinical research are recruiting
volunteer.
 An Introduction to Research Methodology 187

The core values and concepts of research bioethics includes

among others
1. Justifying the inclusion and need of human participants
in research.
2. Ensuring the scientific value and validity of the
conducted research.
3. Bringing about more good than harm for the betterment
of the society.
4. Promoting and protecting the interests of research
participants before the interest of science and society.
5. Voluntary participation and the right of the participant
to discontinue participation.
6. Distributing the risks and potential benefits, and sparse
resources.
7. Showing respect for participant, and maintaining their
privacy and confidentiality.
8. Upholding transparency and veracity (truthfulness) of
research, institutional ethical or review board
certification, and registration in clinical trial registries.
9. The accessibility, storage, and re-use of data and remnant
biological cell, tissue or fluid of participant.
10. Ensuring the participant and the associated community
receives the benefit of the newer discovery, intervention
or invention.
Along with other countries India also has designed its own
regulatory mechanism for research based on the principles
evolved from time to time. “The ICMR Ethical Guidelines
for Biomedical Research on Human Participants” forms the
188 An Introduction to Research Methodology

basic document on which the regulatory mechanism in India

is evolved.
Indian Council for Medical Research (ICMR) Guidelines9
Over the last many decades, bioethical guidelines for
performing research involving humans have been developed
and further improved upon by multiple organizations and
study groups at the international and national levels.
Adherence to these guidelines by the researcher(s) is essential
to ensure that all the principles of bioethics in the conduct
of these healthcare researches are followed. It should be the
duty of the researcher to protect the rights, wellbeing and
autonomy of the research participants and the concerned
community.
The Indian Council of Medical Research (ICMR) issued
the ‘Policy Statement on Ethical Considerations involved
in Research on Human Subjects’ in 1980 and was revised in
2006.
The basic principles have been expanded into 12 general
principles described below, which are to be applied to all
biomedical and health research involving human
participants or research using their biological material or
data.
1. Principles of essentiality: Only the research which is
essential should be conducted on Human participants. The
institutional review board or ethics committee should
acknowledge the necessity/essentiality of the research,
scientific validity, benefit to the community, and need for
human participants.
2. Principles of voluntariness, informed consent and
community agreement: Only the research participants who
voluntarily give a written informed consent should be
 An Introduction to Research Methodology 189

included in the proposed research, and they should have

the right to abstain from further participation if so desired
at any time without any obligation or loss of benefit/
treatment to which they are normally entitled. When the
participant is deemed to have diminished autonomy or
belongs to a vulnerable population, the informed consent
shall be obtained from those who are empowered to or has a
duty to act on their behalf. If required for cultural
characteristics/appropriateness, adequate consent or
agreement should be obtained from the head of the family
or the community leader. Only the review board / ethics
committee has the authority to permit waiver of informed
written consent based on the degree of risk involved.
3. Principles of non-exploitation: Burdens and benefit of
research should be distributed without any discrimination
between communities, groups or countries. An adequate
mechanism should be established for providing required
aftercare and treatment, and if the need arises to immediately
provide compensation and/or rehabilitation either through
insurance or any other appropriate means for all foreseeable
and unforeseeable risks. Research participants may if desired
be provided with reasonable remuneration for their time,
loss of pay, and distance of travel, but shall never be undue
inducement for participation.
4. Principles of privacy and confidentiality: The identity,
all records/report/data, and personal matters of the
participant shall be kept confidential at all times, unless when
required by the law or is essential for providing intervention
or treatment. The participant should never suffer from any
discrimination, stigmatization or hardship as a consequence
of having participated in the research.
5. Principles of precaution and risk minimization: At all
190 An Introduction to Research Methodology

the stages of research due care and caution should be

maintained to ensure that the participant and community
are put to minimum risk, and the benefit obtained from the
research is partaken by the said participant and community.
The review board / ethics committee shall monitor the
research and if required give necessary directions and specific
guidelines to minimize all risks as the research/study/trail
progresses.
6. Principles of professional competence: All research shall
be conducted by competent and scientifically qualified
persons who can act with total integrity and impartiality.
7. Principles of accountability and transparency: All
research shall be conducted in a fair, honest, impartial and
transparent manner after full disclosure of any conflict of
interest by the researcher/sponsor. Complete records of the
research, data, and notes shall be retained for a prescribed
period for monitoring, evaluation, conducting further
research, and for scrutiny by any appropriate legal or
administrative authority.
8. Principles of maximization of the public interest and of
distributive justice: Research should be for the benefit of all
mankind and never for certain group of people or
community. The benefits should be shared with the
participant and the community from which they were drawn.
9. Principles of institutional arrangements: The
researcher(s) should comply with all the required procedures
and institutional arrangement made in connection with the
research and its subsequent use in a bonafide and transparent
matter. The researcher shall take adequate precaution and
appropriate steps that are necessary to ensure that the
research reports, materials and data are duly preserved and
archived.
 An Introduction to Research Methodology 191

10. Principles of public domain: All research result, data,

and evaluation shall be placed in public domain through
scientific publication and such methods, so as to benefit all
mankind and to avoid repetition of research and wasting of
sparse resources.
11. Principles of totality of responsibility: It shall be the moral
responsibility of the researcher, institution, sponsor, groups
or undertakings who perform or profit from the research or
product to ensure safety and scientific validity, and to subject
themselves to monitoring and to take all due remedial action
whenever required.
12. Principles of compliance: It shall be the duty of all the
persons associated with the research to ensure that both the
letter and the spirit of these guidelines or notifications
specifically prescribed for that area of research / experiment
are scrupulously observed and duly complied with.
Informed Consent Form (Written and Audio-visual)9
World Medical Association (WMA) defines informed
consent as “a freely-taken decision by a participant/research
subject whether or not to participate in a research project,
based on careful consideration of the advantages and
disadvantages of such participation for the person involved.”
As per ICMR code, for all biomedical research involving
human participants the investigator(s) must obtain an
informed written consent from every prospective participant.
If an individual is not capable of giving an informed written
consent the consent of a legal guardian or legally authorized
representative (LAR) along with assent of the participant
has to be recorded.
The process of obtaining informed consent has two
components.
192 An Introduction to Research Methodology

1. Information Sheet: For consent to be ‘informed’ the

potential research participants must be provided with all the
information they should know about the proposed research,
in order to make an objective and learned decision of whether
or not to participate in the said study or trail.
2. Certificate of Consent: A formal written document/
agreement on which the participant gives voluntary consent
by placing his signature and full name, after understanding
in full the provided Information Sheet. An illiterate
participant can place the left hand thumb print, which
should be witnessed by a literate legally authorized
representative (LAR).
Information Sheet is an invitation to participate in the
research and shall include all the details of the proposed
research. All scientific terms should be explained in layman’s
term, to be understood by a person educated or schooled
up to 7th class in a standard Indian school. The language
used should be non-technical and simple to understand and
shall be in the local language of the participant. A copy of
the information sheet should be given to the participant. If
the participant is illiterate then the investigator shall read
out the information sheet in the presence of an impartial
witness or have a literate person / LAR read it out in the
local language and witnesses the consent. The information
sheet should include
1. Name, nature and purpose of the research study.
2. Duration of participation, number of visits required,
number of participants, and centers of study.
3. Procedures to be followed are explained in layman’s term,
and not in confusing or unknown scientific terms.
 An Introduction to Research Methodology 193

4. Investigations, if any, to be performed for research

purpose.
5. Participation is voluntary and the participant has the
right to abstain from participation or to withdraw
consent to participate at any later time without reprisal.
6. No penalty or any loss of benefits/treatment for not
taking part, or on withdrawal from the study at a later
date.
7. Foreseeable risks and discomforts should be adequately
described, and whether this project involves more than
minimal risk.
8. Benefits to participant, community and/or medical
profession from this study or trail.
9. Policy on compensation in case of any foreseen or
unforeseen injury or adverse event.
10. Remuneration without undue inducement, if any, for
participant time, loss of pay, travel.
11. Availability of medical treatment for injuries / adverse
events arising out of the research.
12. Alternative treatments or standard form of treatment that
is already available at that period of time for that disease.
13. Steps taken for ensuring confidentiality of the personal
details and data collected, and privacy of the participant.
14. Method of benefit sharing with the participant in the
event of commercialization or invention of any new
diagnostic or such product, and the access to any newer
beneficial intervention identified in the study.
15. Contact details of principal investigator (PI) and co-
194 An Introduction to Research Methodology

investigators (Co-PI) for asking more information on

the research or to contact in case of injury/adverse event.
16. Contact details of Chairman of the IEC for appeal or
complain against violation of rights.
17. Full disclosure of any conflict of interest on part of
researcher/sponsor/institution.
18. Storage period of biological sample/fluid/cells and related
data, with choice offered to participant regarding future
use of remaining sample, refusal for storage, and
notification of results of such future studies.
19. Any other pertinent or additional information.
Audio-visual recording for Clinical and Drug Trails: 12
The honorable Supreme Court of India in its judicial order
dated 19/11/2013 has issued direction that all future clinical
and drug trials to be conducted in any part of India shall
have audio-visual recording of the process of obtaining the
written informed consent of each trial subject, adhering to
the principles of confidentiality and privacy. Prior oral
consent shall be obtained from participant for audio-visual
recording of the consent process, and only such consented
participant shall be included in clinical and/or drug trails.
The video camera and audio recorder used for the audio-
visual recording should be adequately positioned to
simultaneously capture clearly and eligibly the facial details
and audio of the participant, investigator, and LAR or
impartial witness if any. Such audio-visual recording and
related documentation should be preserved for five years (if
not permanently) after termination/completion of the study
and should be available for scrutiny by authorized personal.
 An Introduction to Research Methodology 195

Institutional Ethics Committee (IEC) / Institutional

Review Board (IRB)9
All contemporary, academic, and futuristic healthcare
research should always be subjected to ethical review by a
competent institutional review board or ethical committee
prior to the start of the study or trial. Such review and
monitoring by the committee / board ensures that the ethical
principles and practices put-forth in the national and
international guidelines are followed by the researcher. The
committee shall be multidisciplinary and multisectorial in
composition, and shall be independent and competent to
review all research proposals. ICMR describes the
responsibility of IEC as
1. To ensure a competent review of all ethical aspects and
if required the scientific validity of the research project
in an objective and systematic manner.
2. To protect the dignity, rights and wellbeing of the
potential research participants.
3. To provide advice to the researchers on all aspects of the
welfare and safety of the research participants, after
reviewing the risk benefit ratio.
4. To ensure the scientific soundness of the proposed
research, the essentially of research, and the need for
human participant.
5. To monitor and ensure that universal ethical values and
international scientific standards are followed in
accordance with the values and customs of the local
community.
6. To assist in the development and the education of a
research community responsive to local health care
requirements.
196 An Introduction to Research Methodology

In India, as per Schedule Y of Drugs & Cosmetics Act

amended in 2005, the ICMR code require that the IEC shall
be made / composed of:
1. Chairperson (reputed impartial and authoritative person
from outside the institution, and shall never be the head
of the institution).
2. One-two basic medical scientist(s) (preferably one
pharmacologist).
3. One-two clinician(s) from various institutes.
4. One legal expert / retired judge.
5. One social scientist / representative of non-governmental
organization.
6. One philosopher / ethicist / theologian.
7. One educated suitably knowledgeable lay person from
the community.
8. Member-Secretary (responsible for record keeping,
notification, managing the activities/duties of the
committee, and to represent the committee legally or
otherwise).
9. A subject expert invited by the committee to give expert
opinion if the proposed research is within his preview of
expertise.
The researcher should submit an application in the
prescribed format along with an all-inclusive study protocol
as prescribed in standard operating protocol (SOP) of the
concerned IEC. Prior to initiation of any scientific research
it is essential that all research protocol are subjected to
detailed peer-review by the institutional scientific or research
committee, followed by certification and monitoring by the
 An Introduction to Research Methodology 197

institutional ethical committee or review board (IRB). The

member-secretary shall initially screen/scrutinize all the
research proposals for their completeness and acceptability.
Depending on the risk involved the member-secretary after
scrutiny shall categorize the submitted research proposal into
three forms of review. The investigator cannot decide or
categorize his/her proposal’s review process.
1. Exempted from review: Research proposals which present
less than minimal risk and in which there is no disclosure of
personal identity of the participants. Such as research on
educational practices like curriculum modification,
instructional strategies or teaching technologies, class-room
management, etc.
Minimal risk is defined as the risk, harm or discomfort that
may be anticipated or encountered in the routine daily life-
activities of the general population or during the
performance of routine physical, clinical or psychological
examinations, tests or treatment pertaining to that given
period of time in that person’s life.
2. Expedited review: The Member-Secretary and
Chairperson along with a designated member of the
committee, or a subcommittee of the IEC can subject a
proposal to expedited (speed-up/quicker) review if the
proposals present no more than minimal risk to the research
participants. Such as
a. Minor deviations from originally approved research
protocol during the first year of receiving approval/
certification of the same IEC.
b. Revised proposal previously approved through full review
by the same IEC.
198 An Introduction to Research Methodology

c. Clinical studies on known interaction of drugs and

medical devices that are already been approved by the
Drug Controller General of India (except when on
vulnerable population).
d. Research involving clinical materials (data, documents,
records, or specimens) that have been collected for non-
research (clinical) purposes.
e. When a full review of the research proposal is not possible
due to time constraints in an emergency situation (like
sudden disease outbreaks / disasters).
3. Full review: All research proposals that may have more
than minimal risk, which do not qualify for exempted or
expedited review, and proposals that involve vulnerable
population / special groups shall be subjected to a full review
by the IEC.
IRB has the responsibility not only to review but also to
follow up the conduct of the study and ensure it is being
conducted ethically.
It is mandatory that any research on participants/subjects
in India should be performed within the guidelines/
principles of the Indian Council for Medical Research
(ICMR), which may get promulgated as a Bill on Biomedical
Research by the parliament of India in 2016-17. It is
mandatory since 2009 that all clinical trial protocols with
details of the researcher, duration and the methodology are
registered online in the Clinical Trail registry of India
(CTRI)13
Conclusion
An ethicist asks relevant questions pertaining to ethics rather
than providing definitive and accurate answers, as there are
 An Introduction to Research Methodology 199

times when there are no right answers to contentious issues.

Bioethicists attempt to answer significant questions related
to the purposes of life science research with regard to its
harm benefit ratio, meaning and implications of distributive
justice, in the interest of global healthcare. They explore
deeper issues of life and death, pain and suffering, and rights
and responsibilities.( As the multi-disciplinary field of
bioethics evolve to a full-fledged discipline, it has become a
prominent force supporting legislation, public policy, and
life sciences research.

Bibliography
1. Foster C. The Ethics of Medical Research on Humans. 1st ed.
Cambridge 2001. Cambridge University Press.
2. Council for International Organization of Medical Sciences
(CIOMS). ww.cioms.ch/ publications/layout_guide2002.pdf.
Accessed on Feb 2016.
3. Nuremberg Code. https://history.nih.gov/research/
downloads/nuremberg.pdf. Accessed on Feb 2016.
4. Declaration of Helsinki. www.wma.net/en/30publications/
10policies/b3/17c.pdf. Accessed on Feb 2016.
5. Belmont Report. http://videocast.nih.gov/pdf/ohrp_
appendix_belmont_ report_vol_ 2.pdf. Accessed on Feb
2016.
6. United Nations Universal Declaration of Human Rights.
www.un.org/en/universal-declaration-human-rights.
Accessed on Feb 2016.
7. Guidelines for Good Clinical Practice – World Health
Organization. apps.who.int/ medicinedocs/pdf/whozip13e/
whozip13e.pdf. Accessed on Feb 2016
200 An Introduction to Research Methodology

8. International Conference on Harmonization - Good

Clinical Practice (ICH-GCP). http://www.ich.org/
fileadmin/Public_Web_Site/ICH_Products/Guidelines/
Efficacy/E6/E6_R1_Guideline.pdf. Accessed on Feb 2016.
9. ICMR Ethical Guidelines for Biomedical Research on
Human Participants. http://icmr.nic.in/
ethical_guidelines.pdf. Accessed on Feb 2016.
10. European Group on Ethics in Science and New
Technologies. (EGE). https://ec.europa .eu/research/ege/
pdf/opinion-29_ege_executive-summary-
recommendations.pdf. Accessed on Feb 2016.
11. Guraya SY, London NJM, Guraya SS. Ethics in medical
research. Journal of Microscopy and Ultrastructure.
2014;2:121-126
12. DCGI guidelines on audio-visual recording of informed
consent process in clinical trial (Draft). http://
www.cdsco.nic.in/writereaddata/
Guidance_for_AV%20Recording_09. January.14.pdf.
Accessed on Feb 2016.
13. Clinical Trials Registry - India (CTRI). www.ctri.nic.in.
Accessed on Feb 2016.
 An Introduction to Research Methodology 201

Chapter 15

Data Mangement and Analysis

Dr. V.K.V.Balakrishnan

All types of research require data which serve as the base or

raw material for analysis. Data management is a
comprehensive plan of effectively dealing with the research
data all through the study.
What is data management?
Data management is the complete series of activities
necessary to make research data discoverable, accessible and
understandable today, tomorrow and into the future. It is
the process of collection, cleaning, and management of data
in compliance with regulatory standards. Primary objective
of data management is to provide high-quality data by
keeping the number of errors and missing data as low as
possible and gather maximum data for analysis.
Effective Data Management Practices
 Designating the responsibilities of every individual
involved in the study.
 Determining how data will be stored and backed up.
 Implementing the data management plan.
 Deciding how data will be appropriately dealt with
202 An Introduction to Research Methodology

Advantages
For the Researchers
• Replicated and verified
• Preserved for future use
• Linked with other research project
• Shared and reused
• Reduces the risk that data could be stolen, lost or misused
• Gaining easier access to raw and processed data
• Increases the research profiles
• Potentially finds new audiences & collaborators through
dissemination, citation and re-use of data
Data Management includes:
Data Entry
o Recording Data or Schedule
o Choice of Software
o Duplicate Data Entry
o Missing Data
Data Processing
· Data Cleansing
· Data Storage
Data Analysis
Steps of Data Management
Primary Step- Data Collection
§ Most important step
§ Tools for collecting Data will depend upon the nature
of relationships to be examined between two variables
 An Introduction to Research Methodology 203

and the type of study

§ We must select source of data, kind of data sought,
procedures for collecting and processing them
After collecting data, researchers have to look into—
o Checking questionnaires/Schedules/Case proforma or
clinical record format
o Sorting out and reducing information collected to
manageable proportions
o Summarizing the data in tabular form
o Analyzing facts so as to bring out their salient features
i.e. search for trends, patterns and relationships
o Interpreting the results or converting data into
statements, propositions or conclusions as answer to
Research Question
o Writing or presentation of the report
Data Processing
It involves various manipulations necessary for preparing
data for Analysis.
The process of manipulation could be manual or electronic.
The data are generally made computer sensible i.e. prepared
for computer processing by assigning codes to various
response categories. It includes editing, categorising the open
ended questions, computerisation, preparation of tables and
diagrams. The computer gives the desired computations and
comparisons including data for statistical analysis.
Checking and Editing Data
Information gathered during the stage of data collection
varies in nature and quantity from study to study. When
204 An Introduction to Research Methodology

sur veys are conducted and data obtained through

questionnaire and schedules the answers either may not be
ticked at proper places or some questions may be left
unanswered or may be given in a form which needs
reconstruction in a category designed for analysis as
converting daily/ monthly income into annual income or
identifying family structure (nuclear or joint) & so on. One
has to take a decision as to how to edit multiple answers or
complex answered questions and identify the right choice
E.g. for the question, Is the industry in which you are
working the largest or average in size or small respondent
ticks largest and average. Many of these data would require
categorisation, restructuring, additions or editing of
complexly answered questions and so forth.
Checking also needs that the data are relevant and
appropriate and errors modified (eg. impossible answers
mistakenly recorded by investigator). Such answers need
editing. Editing is required for proper coding and entering
in the computer (when not manually analysing). Editing
thus means the data are complete, error free readable and
worthy of being assigned a code.
Editing process begins in the field itself. Interviewers
(Investigators) soon after completing interview (for filling
the schedule) should check the completed forms for errors
and omissions and complete the incomplete responses and
reduce the number of ‘No responses‘ with rapid follow up
stimulated by field editing. If field editing is not possible in
house editing helps. Editing occurs simultaneously with
forming categories eg. age may be put in categories. Eg. age
may be put in categories like , 18 yrs(very young), 18-30
yrs(young), 30-40 yrs(early middle age), 40-50 yrs (late
middle age), 50+(old). Field supervisors can do editing at
field itself by re-contacting the respondents.
 An Introduction to Research Methodology 205

Editing also requires rearranging answers to open ended

questions {eg. Don’t know does not mean no response but
that respondent is not sure or is unable to formulate an
opinion and No response means respondent is not familiar
with situation or object…. of the question.}. Editing can
be done along with coding too.
Coding of Data
Coding is translating answers into numerical values or
assigning numbers to various categories of a variable to be
used in analysis.
Coding is generally done while preparing questionnaires,
interview schedule or case proforma and investigation done
with pre-coded questions. If pre-coding not done, coding
may be attempted after the investigation). Coding is done
based on instructions given in the code book which provides
a numerical code for each variable. It should be specific and
include every data item.
Data Entry and storage
Pre-coded items are directly fed to the computer for
processing and analysis. Entry errors like double entry,
misplacing of data and so on should be checked and editing
done if required. Data entered should be properly stored so
that it is easily transferable or used later for evaluations or
further interpretations
Tabulation of Data
After editing data are put together in some kind of tables to
make statistical analysis easy.
Data Distribution
A distribution is a form of classification of scores obtained
for various categories of a particular variable. There are 3
206 An Introduction to Research Methodology

types of distributions-Frequency Distribution, Percentage

Distribution, Cumulative Distribution
Data Analysis and Interpretation
 Analysis is the ordering of data into constituent parts in
order to obtain answers to research questions
 Interpretation: With results of analysis inferences are
made and conclusions drawn.
 After statistical analysis the contribution and significance
of research and relationship among variables are
established and generalisation attempted
 The interpretation of data are presented with the help
of or graphs, histograms bar diagrams, pie charts,
pictographs and so on and report prepared
Conclusion
Proper management of the data collected is an essential
requisite for every research. Unless it is promptly edited,
clarified and stored it will hamper analysis and thereby
damage the research project.
 An Introduction to Research Methodology 207

Chapter 16
Qualitative Research
Janaki Krishnan T S, Anitha M A

Scientific research consists of an investigation that seeks

answers to question/s and collect evidence systematically
using predefined sets of procedures to find the answers.
Qualitative research also shares these characteristics. It can
be used when;
 little is known about an area/problem
 you are unsure about what you are going to find,
 you are interested in studying the ‘ Human Side’ of an
issue – how people experience an issue
 you need to construct a theory, that reflects reality.
Definition
“Qualitative research is the systematic collection,
organization and interpretation of textual material derived
from talk or conversation. It is used in the exploration of
meanings of social phenomena as experienced by individuals
themselves, in their natural context”(Malterud, 2001,p.483).
It focuses on eliciting peoples stories in the form of
conversations, written texts, or visual forms like photographs,
drawings etc.
208 An Introduction to Research Methodology

Qualitative research is a process of enquiry with the aim

of understanding a given problem from multiple
perspectives. It is conducted in a natural setting with the
goal of building a complex and holistic view of the
phenomenon of interest. It seeks to understand research
problem from perspectives of the people participating in the
study i.e., ‘Emic’ perspective and has a background in
Anthropology. The researcher in qualitative research, tries
to find out what people think, act and feel and are concerned
more with meanings and processes rather than numbers.
People under study have some cultural parameter in
common, like a geographical area, religion, shared experience
(e.g. group of men in their early thirties, who have undergone
sterilization). Qualitative research provides insights into
meanings of decisions and actions and the information may
be subjective. The aim of this type of research is to gather
‘insight’, or to learn rather than generalization. There are
no a priori hypotheses or preconceptions and theories are
generated during the course of research. This relies on a
research strategy which is flexible and iterative and defines
the researcher as an instrument in the research process.
Differences between qualitative and quantitative research
Qualitative research basically is exploratory research, trying
to understand the how and why of a problem, rather than
the what, where, when and who as in quantitative research.
Qualitative research helps to synthesize a hypothesis when
compared to quantitative research which helps in the
verification of an already prepared hypothesis. The purpose
of quantitative research is to provide causal explanations for
phenomena. Meanwhile qualitative research helps to
understand perceptions and provide interpretations.
Quantitative research employs a fixed study design and use
 An Introduction to Research Methodology 209

closed ended questions for data collection. Qualitative

studies utilize flexible evolving designs and use open ended
questions for data collection. The sample size typically in
quantitative studies is large whereas in qualitative research
it is small. The data collected from quantitative studies are
measurable and composed of numerals which can be further
subjected to statistical analysis. But the data in qualitative
research is composed of opinions, thoughts and beliefs.
Statistical techniques are rarely used in the analysis of textual
data generated from qualitative studies.
Approaches of qualitative research
A qualitative “approach” is a general way of thinking about
conducting qualitative research and it describes (i) the
purpose of the qualitative research (ii) the role of the
researcher (iii) the stages of research, and (iv) the method of
data analysis.
Approaches of qualitative studies include:
1. Historical research: Historical research allows the
researcher to discuss past and present events in the context
of the present situation. It studies available data from past
records and other information sources with a view to
reconstructing the origin and development of an institution,
movement or a system and discovering the trends in the
past. Data are gathered from multiple authentic sources;
primary and secondary. Precise measurements, verification
and replication are not possible in historic research. However
they are of immense value in understanding the past and
drawing inferences for the present and future. Eg. Evolution
of social security system in India.
2. Ethnography: Ethnography comes from the discipline
of social and cultural anthropology and focuses on the
210 An Introduction to Research Methodology

culture of group of people. People under study have some

cultural parameters in common which guide their view of
the world and their experiences.
Elisa Sobo defines ethnography as the presentation of a
given group’s conceptual world, seen and experienced from
the ‘inside’. For eg. ‘what it was like for a family when a
child had cerebral palsy. Researcher as an instrument is
important in the role of an ethnographer in analyzing and
interpreting a culture and relies substantially on participant
observation
3. Grounded theory: Grounded theory is an inductive type
of research with its specific approach to theory development.
Theories are not applied to subjects under study, but are
discovered. Theory developed from research is based
(grounded) on the observations or data, from which it was
developed through careful analysis. Grounded theory
suggests a continuous interplay between data collection and
analysis. For eg.- A study on the religious coping by adults
who were diagnosed with cystic fibrosis after the age of 18.
4. Phenomenology: Phenomenology describes the subjective
reality of an event as perceived by the study population and
it is all about the search for meaning and experience. It is
the study of phenomena. For eg. How do mothers feel about
living with their teenager suffering from cancer, experiences
of mothers with suicidal adolescents? Phenomenological
studies often utilize in-depth interviews.
5. Action research: Action research assumes that complex
social phenomena are best understood by introducing
interventions or actions into that phenomenon and
observing the effects of those actions. This type of research
is initiated to solve an immediate problem. Eg, management
of a solid waste disposal programme in a locality. Researcher’s
 An Introduction to Research Methodology 211

choice of action may be based on existing theory. He or she

then observes the results of intervention, modifying it as
necessary and simultaneously learning from the process. New
theoretical insights may be drawn regarding the target
problem and the intervention.
Research designs in Qualitative research
Research design is a plan for collecting and analyzing
evidence so that the investigator can answer questions, posed
by him or her. It is always worth to discuss the question,
methods, techniques and data source issues with others in
the field. The research questions are usually oriented to
describe the states or processes of the phenomenon under
study. The design emerges as the data collection progresses
and an entire advance blueprint cannot be prepared.
Research designs in this type of study are selected for the
purpose of specifying an initial focus during the planning
stage. As the fieldwork progresses, the plan of study may
change. Controlled experimental designs are not suitable for
qualitative studies.
There are some basic study designs and these are;
Case study: The aim of case study in research is to
understand a case comprehensively using multiple sources
of evidence, employing methods like observation, interviews,
review of documentation and collection of physical evidence
or artifacts. It is characterized by detailed contextual analysis
of a limited number of events, conditions and their
relationships. Persons, families, organizations and
institutions can be subjects of a case analysis.
Comparative studies: Need multiple cases, with a
connecting event, whereby we study the impact of the event
in different people/societies. Eg. The effect of mass
212 An Introduction to Research Methodology

immunization program on urban/semi urban/ rural

population and the acceptance of the program through
PHC’s in these categories.
Snapshots: These are concerned with the description of
circumstances at the time of investigation and this design
may be called a qualitative cross sectional study.
Retrospective studies: These studies refer to reconstruction
of cases from the past, like research done to write a biography.
Longitudinal Studies: Analyze an interesting process or state
at a later period in time. Here, it is not necessary that
information is collected from the same set of people and
somebody who uses the same facility can give us the
information in the subsequent visit. Lot of planning is
necessar y to begin a qualitative study. In practice
combination of basic designs is possible, e.g. combining a
case study with retrospective study etc.
Sampling
Sampling is a part of the design that needs to be developed
before the beginning of the study, and sampling decisions
can be made during data collection based on what one finds
in data. While doing qualitative studies, we should always
have the following question in our mind; (1) why this or
that group or category is relevant to the study? and (2) how
will the relevance of the study or understanding the process
will change, if we include a particular group/category in the
study?
In Qualitative Research, usually purposeful sampling is done;
the goal here is to understand a phenomenon and not to
represent a population. Information rich cases are selected
for intensive study. The various kinds of purposeful sampling
are:
 An Introduction to Research Methodology 213

1. Criterion Sampling: This involves selecting cases that

meet some predetermined criterion of importance and these
criteria can be a score or a measure. This method is used
widely in quality assurance efforts. Critical incidents can be
a source of criterion sampling. Eg while studying diabetes
mellitus, subjects with uncontrolled diabetes (a higher
specific value of Hb A1c) could alone be included.
2. Maximum variation Sampling: The cases studied
represent maximum variations on specified aspects of the
phenomenon of interest or the subject matter of study. This
can identify common patterns and uniqueness among the
groups. There are some central themes that cut across
variation and these can be highlighted. Eg. evaluating the
effect of supplements in overcoming malnutrition. The
sample can be constituted by children under 5 years, the
participants selected representing different geographical or
economic variations.ie Urban, rural, semi urban, middle
class, poor.
3. Extreme or Deviant case sampling: This is learning from
highly unusual manifestations of the phenomenon of
interest, with an aim to improve typical cases. Outstanding
successes and notable failures are studied. For eg, in the early
days of AIDS research when HIV infections almost always
resulted in death, a small number of cases of people infected
with HIV who did not develop AIDS became crucial as they
provided important insights into directions that researchers
should take in combating AIDS.
4. Homogenous Sampling: Participants with similar
demographic characteristics like age, gender, education etc.
are selected and the purpose is to describe some particular
subgroup in-depth. This is used mainly in Focus Group
Discussions.
214 An Introduction to Research Methodology

5. Stratified purposeful sampling: People under study are

divided into different subgroups, each group comprising a
fairly homogenous sample. It Illustrates characteristics of
different subgroups of interest, and capture major variations
among them i.e. more importance to variations among
subgroups than similarities or what is common among them.
6. Snowball Sampling: This is an approach in locating
information-rich cases especially when the sampling frame
is hidden. The initial participants recommend others they
know and the snowball becomes bigger. Eg. HIV positive
cases.
Sample size
The aim in qualitative study is to understand a phenomenon.
Determining adequate sample size in qualitative research is
ultimately a matter of experience and judgment. Sample size
can be considered adequate when there is redundancy of
information. Theoretical saturation occurs when (i) no new
relevant data emerges regarding a category, (ii) the category
is well developed in terms of its properties demonstrating
variations, and (iii) relationships among categories are well
established and validated. Sample size should be large enough
to make meaningful comparisons in relation to research
questions and if comparison is done between groups, the
sample size should be larger. Small sample size is the norm
if the purpose is to collect details, complexity and depth of
a phenomenon. Care should be taken not to over-generalize
the data.
Data collection
Extensive fieldwork is necessary for the collection of data.
This is extremely time consuming and requires long periods
of time in the field. An ‘emic’ approach is adopted in the
 An Introduction to Research Methodology 215

analysis of data, i.e. it is interpreted from the point of view

of the population under study, as though being expressed
by the subjects themselves, using local language and
terminology to describe the phenomena. Qualitative research
requires that the researcher is familiar with the social norms
and local language of the participants. Interpretation from
an outsider perspective (etic) will lead to false results.
Informed consent should be obtained in local language.
Data collection Techniques in Qualitative Research
Data collection in Qualitative approach involves direct
interaction between the researcher and the participants of
the study. Major data collection techniques include; (i) In –
depth interviews, (ii) Observation and (iii) Focus Group
Discussions.
In-Depth Interview
This is a purposeful face to face interaction between two
people, where one person tries to collect information from
another. This permits the exploration of the participant’s
thoughts and emotions. The interview may be either formal
structured or informal unstructured type. In formal
structured type, there is a definite pre-structured group of
questions and all the participants are formally asked to
answer this set of questions. Informal unstructured interview
allows the researcher to obtain insights into a phenomenon
or a problem of which little is known. This can be an
informal conversational interview, with no predetermination
of questions and interview guide approach, where topics and
issues to be covered are specified in advance. Data during
interview can be collected either by writing notes during or
after the process or by tape-recording the whole interview.
All these have its own disadvantages like disruption of the
216 An Introduction to Research Methodology

process (if done during), loss of information (if done after

the interview). Audio recording can be disruptive and
intrusive of privacy; and transcription of audio records
involves cost and labor. However recording of interviews
will guard against loss of valuable information gained, from
the participants.
An interview guide can be prepared taking into consideration
the specific objectives of the research, the respondents and
their characteristics, the agency /body who is interested in
this information, and the purpose of this information i.e.,
if it is going to play a role in policy making etc. While
preparing an interview guide, ensure that your questions
are clear, unambiguous, simple, not answerable in
monosyllables like ‘yes’ or ‘no’; and within the experience of
the respondent. Along with each question ‘probes’ can be
constructed. Probes are devices to make respondent elaborate
their responses. They prompt the respondent to speak further
when an initial question fails to elicit the response.
How to conduct the Interview
Participants should be representative of the study population
and selected purposively. People who are well informed
about the issue may be selected for interview. Seek privacy;
assure confidentiality and make the participant relaxed by
establishing rapport. Voluntary participation of the subjects
should be assured by the consent form in local language.
Participants can refuse to answer any question, and can
withdraw from the interview, if they want to. Sit squarely,
maintaining eye contact. The researcher (interviewer) should
maintain neutrality in the tone of voice, gestures and
expressions. They should listen with an open mind, ensure
natural flow of the conversation and allow participants to
do most of the talking. Conclude the interview with general
 An Introduction to Research Methodology 217

questions and thank the respondent. After the respondent

has finished speaking, clarification questions, elaboration
questions, ‘show’ me (how you reacted) questions and
eliciting personal meaning questions can be asked by the
researcher. An interview can take as long as two hours or
more and analyzing the vast amount of information obtained
requires great skill. An in-depth interview is an explanatory
tool, more appropriate in rural setting and the responses are
more valid. But it is time consuming and extrapolating the
information to the society at large may also prove to be
difficult.
Observation
Observation is a technique of directly observing the behavior
with the purpose of describing it. This method of data
collection is employed to detect the discrepancies between
what the subjects express as they feel and what they actually
feel. Observation can be of two types: participant and non-
participant observation. In participant observation the
researcher becomes an actual part of the community/ process
under study by, “talking the talk and walking the walk”. By
directly observing the operations the evaluator can develop
a holistic perspective. In non-participant observation,
observer is an on-looker, observes the event / process as an
outsider. Non-participant observation may be structured,
using an observational guide or unstructured with no
predetermined guide.
Focus Group Discussion (FGD)
FGD is one among the most widely used tools in qualitative
research. It is a discussion among a small group of
participants which rely more on interaction among members
of the group, and each participant can build on what others
in the group have said. It can provide insight into how a
218 An Introduction to Research Methodology

group thinks and the range of opinions and ideas that exists
in the community about an issue. The number of
participants should be kept small and 6-10 participants form
a good group. It is best to have participants seated in a
circular fashion. Participants should be selected purposively
usually through informal networks. Other important factors
in a focus group are the presence of a moderator and a
discussion guide which is used by the moderator to conduct
the discussion. The FGD guide contains the issues to be
discussed and it may be kept to a maximum of 6-7. Usually
the whole discussion is tape recorded with prior consent
from the participants. A note taker/observer is a member of
the team who follows the discussion carefully and takes down
notes indicating the code or letter of identification of each
speaker and the first few words of each comment. Note takers
focus on documentation and should have mastery of an
efficient system for taking notes, ability to quickly identify
and take down individual quotes that capture the spirit of
the given point. An effective note taker knows the research
material well and is familiar with the focus group guide,
just in case the need arise to conduct the FGD in the absence
of the moderator in the last minute. Preparatory steps for a
note taker include becoming familiar with FG guide, practice
taking notes in a pilot or mock FGD, label all materials to
be used during the discussion including cassette tapes, note
taking forms etc.
The moderator should not interview individual
members, but introduce the topic and encourage the group
members to express their views on the topic. Informed
consent should be taken from the participant for discussion
and also recording the whole process. The participants are
made to understand that there is no right or wrong answers,
 An Introduction to Research Methodology 219

can talk freely and even express contradictory views.

Although the moderator’s role is really passive, it is important
in keeping the discussion relevant, without being dominated
by one participant and encouraging all participants to speak.
The moderator should be careful to ask effective questions
for leading focus group. He should initiate discussion by
suitably framing the issues as statements and avoid questions,
eliciting Yes/No answers. Each participant should be
encouraged to provide elaborate responses that elicit the
participant’s views and experiences. Eg: Have you taken your
child for the immunization program to the PHC? The above
question may be posed like ‘Would you tell me your
experiences when you took your child for vaccination at the
PHC’? Do not pose several questions at a time. All
participants in the group should be given opportunity to
respond, and in an active discussion, there may be responses
from one participant many times and also those who do not
respond at all. The moderator should take note of this and
encourage the silent person by identifying this participant
either by name or number, and ask for his/her opinion.
Participants may require clarification of the question, or at
times, rephrasing the question is needed. The moderator
should remain neutral throughout the discussion, and never
say comments like ‘that’s good, it is interesting ’etc. Instead,
he can use neutral comments like ‘I see’. Leading questions
are to be avoided. Asking follow-up or sub-questions would
ensure that the participants provide the complete set of
information that each main question was designed to elicit.
Probes are neutral questions, sounds or gestures, moderators
can use in focus groups to encourage participants to
elaborate, and explain ‘why’ or ‘how’. Probes can be direct
or indirect.
220 An Introduction to Research Methodology

Direct probes: eg
i. How did this happen?
ii. How do you feel about…..?
iii. What happened there?
Indirect probes : eg
i. Neutral verbal expressions such as ‘uh…uhh..’
ii. ‘Mirroring technique’ i.e. repeating what the participant
said, like ‘so you were just nine, when your father caught
you with a beedi’…?
Managing a Focus Group Discussion
1. Open with a general comment, like ‘you all know, what
has brought all of us here today’. You can also give a
general introduction to the topic.
2. A brief background information on participants may be
useful for later analysis and presentation of findings.
3. Invite wide responses from the participants.
4. Do not get bothered by silence, the participants need
time to think.
5. Limit researcher’s own participation once the discussion
begins.
6. Material in the guide should be covered i.e., all the
questions are to be addressed and sufficient time should
be given to the respondents
7. Keeping a checklist would make it easier to return to
the questions that were skipped during natural
progression of the discussion.
8. A good moderator shows flexibility, sensitivity, has a sense
of humor, links ideas together and encourages
 An Introduction to Research Methodology 221

participation from everyone. He tries not to dictate the

course of discussion, lose control over the conversation,
judge comments, inform or try to educate the
participants during the course of an FGD and never leads
a question answer session.
9. A skilled note-taker takes detailed notes on what they
observe and also take notes of the conversation regardless
of whether the FGD is tape recorded or not. In addition
to the notes on verbal process, the note taker also write
down non-verbal messages that have a bearing on the
discussion.
10. Towards the end, inform the participants that the session
is about to wind up. Reassure them of confidentiality
and provide refreshments. Thank the participants.

Fig 1.Focus group discussion

Sociogram: Sociogram is a diagrammatic representation of
the entire session of the FGD. It reflects on the conduct of
222 An Introduction to Research Methodology

the discussion, interaction among the participants, silent

participants and dominators. Thus it helps to understand
the group dynamics among the participants.

Fig 2.Sociogram
Qualitative data analysis
Qualitative analysis transforms data into findings. Data
analysis is time consuming and consequently expensive.
Before trying to analyse data, watch for disagreements among
knowledgeable informants, check for informant accuracy,
continue to look for alternate explanations for phenomena
as your understanding of the phenomena deepens. Also try
to fit negative cases into your theory.
Steps in qualitative data analysis
1. Reading
2. Coding
3. Choosing and using computer software
4. Displaying data
5. Developing hypothesis
6. Data reduction
 An Introduction to Research Methodology 223

7. Interpretation
1. Reading: Read all notes and transcripts. Start reading as
soon as the first interview is over. (Need not wait for the
whole data), and begin to identify emergent themes.
Look for potential gaps and incomplete responses to be
addressed in subsequent interviews.
2. Coding: A code is a word or a short phrase that
symbolically assigns a summative essence for a text of
the raw data. Coding begins after becoming familiar with
the data. The codes are like street signs inserted into
margins of notes or typed in after a segment of the text.
These help to identify emergent themes.
There are two approaches to coding;
a) deductive and
b) inductive approaches
Deductive approach uses categories constructed from
theories in order to develop a coding framework. In inductive
method, the coding scheme emerges from the data collected
in qualitative research. The researcher employs his intuitive
senses to determine similar characteristics. Similarly
categories are formed from coded data, i.e, codes are
developed into categories. Categories share common
characteristics. Themes (domains) should be developed from
categories by continuing revision and refinement. This is
by consolidating major categories which are comparable with
each other. Actual act of reaching a theory is more complex
and not always a necessary outcome in qualitative studies .
3. Choosing and using computer software: Modern
computer software programmes can facilitate analysis of
qualitative data. Common software packages used for
qualitative research are N VIVO, EZ Text, ATLAS.Ti,
224 An Introduction to Research Methodology

NUDIST, and SONAR. They help in:

a. Data storage and management
b. Data searching and retrieval
c. Coding
d. Developing and testing theory
e. Writing reports
4. Displaying data: Themes are displayed by laying out what
you know related to a theme capturing variations of each
theme.
5. Developing hypothesis: As the organization of
information associated with each theme proceeds,
hypothesis starts to form. Responses have to be constantly
interpreted, theoretically explained and then validated.
The following questions need to be constantly asked,
a. Do the categories developed make sense?
b. What pieces of information contradict the emerging
ideas?
c. What other opinions should be taken into account?
d. Does the researcher’s own bias influence the data
collection and analysis?
6. Data reduction: The mass of data has to be organized
and somehow meaningfully reduced. Data reduction is
described as the first step of qualitative data analysis by
Miles and Huberman (1994). This refers to the process
of selecting, focusing, simplifying, abstracting and
transforming the data that appear in written up field
notes or transcriptions. The process of data reduction
increases manageability and the data is made intelligible
in terms of issues being addressed.
 An Introduction to Research Methodology 225

7. Interpretation: It is the act of identifying and explaining

the data’s core meaning. This must reflect the attitudes
and responses of study participants and must answer the
research question.
Ethical considerations
Qualitative studies are unique in research goals, study design,
data collection and analysis. The design of these studies are
flexible, evolving and emergent and hence not restrict itself
to pre-determined questions or hypotheses. In qualitative
research usually the interviews and discussions are recorded
to guard against loss of information. Hence, in this context
many of the ethical issues would have distinct and unique
characters over and above the ones that characterize the
biomedical and epidemiological research, leading to distinct
research ethics challenges. The researcher should understand
the dignity of the participants/community and the cultural
contexts to avoid harms like (1)dignitary (not treated as
person with own values and preferences, but as mere means
to an end), (2)social (effects on one’s interaction with others
eg. stigmatization) and (3) informational harm (harm from
disclosing information about an identified research
participant)
The informed consent: The informed consent document
includes the participant information sheet (PIS) and the
consent form. Every participant should be told:
 The purpose of research
 What is expected of a research participant including the
probable amount of time
 Expected risks and benefits- both psychological and social
 Mention specifically if the interview or discussion is to
be recorded
226 An Introduction to Research Methodology

 The fact that the participation is voluntary and that one

can withdraw from the study with no negative
repercussions
 How confidentiality will be protected
 The name and contact information of the local lead
investigator to be contacted for questions or problems
related to research
In the consent form the participants should confirm that
they have understood all the information given in the
information sheet and agree to participate voluntarily by
giving dated signature on the document. The participant
should be provided with a copy of the PIS and consent form
signed by the investigator.
How to protect confidentiality
In qualitative research it is important for data collectors to
maintain boundaries between what they are told by
participants and what they tell to participants. People can
become upset and untrusting about even seemingly trivial
comments being shared especially if they have divulged very
personal information and grow concern that you will divulge
more. The participant might be compelled to give socially
correct responses in an in-depth interview. When disclosures
is perceived as stigmatizing, wrong information may be
provided and hence privacy and confidentiality are all the
more important. Prior appointments need to be taken for
interviews specifying the place and time. Community
consent or gate keeper consent, in addition to individual
consent, may be taken wherever needed.
Reliability and validity in qualitative research
Journal editors and funding agencies are using ‘check lists’
for assessing the reliability and validity of qualitative research
 An Introduction to Research Methodology 227

sent to them. There are also criticisms from researchers who

do not favor qualitative studies. Proving that your data
analysis is rigorous becomes important in these contexts.
Reliability
To test reliability, it is important that you use methods that
are reproducible and consistent. The approach and
procedures for data analysis should be made clear with
justifications as to why these are important and appropriate
within the contexts of the study. Document the process of
generating concepts and themes from the data. Make use of
evidence from previous qualitative and quantitative studies
to substantiate your conclusions.
Validity
Guba and Lincoln’s criteria
Credibility: Is equivalent to internal validity in quantitative
research. Since qualitative research studies a phenomenon
from the participant’s view, the participants can only judge
whether the results are credible or not.
Transferability: This can be equated to external validity and
transferability becomes the responsibility of the researcher.
At large, a qualitative researcher can best describe the context
of his research thoroughly and the assumptions on which
his research was based.
Dependability: In quantitative studies, it is seen, if the study
is replicable. In qualitative studies the researcher should
emphasize the ever changing contexts within which research
occurs.
Confirmability: It is similar to objectivity in quantitative
studies. Qualitative research tends to assume that each
researcher brings a unique perspective to the study. There
are a few strategies in use to increase confirmability. Like
228 An Introduction to Research Methodology

the researcher documents the procedures, for checking and

rechecking the data throughout the study, another researcher
can cross check the proceedings and document it and check
for negative instances that contradict prior observations.
Triangulation: Triangulation means gathering and analyzing
data from more than one source, to reach a closer perspective
to the research under test. Triangulation can be done on
data sources, data collection methods, or researcher
triangulation can also be done.
Strengths and weaknesses of qualitative research
Qualitative studies help to study issues in depth and detail.
The research technique, framework and guidelines are so
flexible that these could be changed anytime during the
research. Qualitative methods of open-ended questions and
probing which gives participants the opportunity to respond
in their own words, rather than forcing them to choose from
fixed responses, as quantitative methods do. The data based
on human experience are more powerful. However, research
is to a great extent dependant on the individual skills of the
researcher. It also depends upon the biases and idiosyncrasies
of the researcher. It requires much time and man power to
analyze data as it is voluminous.
Summary
Qualitative research is the type of investigation which is
employed more, when the topic under consideration was
not subjected to much previous studies. Here a hypothesis
is formulated after the systematic collection, organization,
and interpretation of textual material derived from talk or
conversation which involves people who are in one or the
other way related to the topic. Qualitative research has its
own methods for the collection of data. Even though it is
 An Introduction to Research Methodology 229

criticized as biased or small scale, if properly conducted

qualitative research is unbiased and it can produce valid,
reliable and credible results.

References
1. Miles, M B and Huberman A M(1994). Qualitative data
analysis, Sage publications : New Delhi
2. Richardson, JTE (Ed.)(1996), Handbook of Qualitative
research methods for psychology and social sciences. BPS
Books: Leicester
3. Handbook of qualitative research methods and analysis –
Samarth and BRTC CMC Vellore Collaboration
4. Textbook of qualitative research methodology ( PGD –HSR)
of Institute of distance education, University of Kerala
5. Claire Anderson , presenting and evaluating qualitative
research, American Journal of Pharmaceutical education
2010;74(8).
230 An Introduction to Research Methodology

Chapter 17
Protocol Development
Mini P. Jeeja Sasi

A research work is a systematic and scientific process carried

out to solve a problem or to answer a research question.
The process starts with identification of research problem,
which may come from several sources like personal
experience, critical appraisal of literature, previous researches,
existing theories, social issues, brain storming sessions,
intuition, folklore, exposure to field situation, consultation
with experts etc. it is a rough idea about a problem where
the investigator is interested to do the research work.
After identification of research problem, one must go
through the literature on recent researches to find out the
gap or lacunae in the existing knowledge about the problem.
Pick up those gaps which are feasible, meaningful and
relevant, where research can be done.
Then the research topic has to be refined and narrowed
down in to a well-defined research question. From this
research question, objectives and research hypothesis (if any)
in a refined form has to be generated. Then think about the
research design that will give the most definitive answer to
the research question and is feasible to conduct. After
planning all these things, it has to be written down as a
protocol and submit the same for detailed evaluation and
 An Introduction to Research Methodology 231

approval. It is to be followed by implementation of research,

data collection, description, analysis and interpretation. The
research is not complete without reporting or disseminating
the result. So the whole process can be concluded as follows:-

The key attributes of good research are proper planning,

accuracy in data collection and proper unbiased
interpretation. The most important aspect of a research is
to write a proper research protocol.
Research Protocol
The word protocol is derived from Greek word ‘protokollon’
means first page. So protocol is a format procedure for
carrying out a scientific research. It is a detailed written plan
of the study, which explains, what you are planning to do
and how you are planning to do it.
A written protocol forces the investigator to clarify their
thought and to think about all aspects of the study. It is
viewed as an agreement between the investigator and
scientific community.
Purpose of protocol writing is to:-
 Clarify the research question
232 An Introduction to Research Methodology

 Formulate hypothesis and objectives

 Decide about the study design
 Get ethical clearance
 Apply for funding
 Have a guideline and tool for research team
The protocol should explain the study in terms of answers
to the study questions viz. Why? How?, Who?, What? and
So What?.
WHY? Sets out the study questions and the relevant
background information
HOW? Describes the study design and the rationale
for choosing it. Also describes Instruments /
techniques to be used
WHO? Defines the targets and the study population
and sample size.
WHAT? Identifies the variables to be measured and
outcomes to be analyzed.
SO WHAT? Comments on the expected significance of
results and contribution to knowledge.
A good research protocol should be-
 Adequate to answer the Research Questions and achieve
the study objectives.
 Feasible in the particular set-up for the study
 Provide enough detail (methodology) that can allow
another investigator to do the study and arrive at
comparable conclusions.
Format for the Protocol
Format of a research problem will differ slightly according
to the nature of the study. Despite the difference, all protocol
should contain certain basic information. It is generally
 An Introduction to Research Methodology 233

written in the following format.

1. Title
2. Abstract/Summary
3. Introduction
4. Review of literature
5. Aim and objectives
6. Hypotheses
7. Methodology
8. Time line
9. Ethical consideration
10. Budget
11. Reporting format
12. References
13. Annexures
Title: Title should be simple, accurate, informative and
attractive. Specify the population to be investigated, and
make the central objectives and variable of the study clear
to the reader. It should not be too long or too short. Better
to be in one sentence with about 10-12 words. Avoid
abbreviations and unwanted words like ‘to study’, ‘to
evaluate’ etc in the title. The title should be revised after
completion of protocol to reflect more closely the sense of
the study. A title once approved should not be changed. A
short title or acronym can be given while publishing. Provide
the key words for the initial classification and indexing of
the project.
Abstract/Summary: It is a sketch plan of the study, which
provides quick details and general plan of the study before
getting into the main document. It should include the central
elements like research problem, rationale, methodology,
234 An Introduction to Research Methodology

population, time frame and expected outcomes.

Unstructured abstract in paragraph form/structured abstract
is to be given in a protocol. It is placed before the protocol,
but is often written after the protocol itself is completed.
Introduction: Introduction typically begins with a statement
of the research problem in a precise and clear term. Here
the background, problem statement and significance or
rationale of the proposed research should be explained. The
problem statement should be presented in a logical sequence
as follows and clearly specify the gaps in existing knowledge.
• Magnitude, frequency and distribution of problem
• Probable causes- current knowledge/controversies/
consensus/conclusive evidences
• Possible solutions-proposed/attempted
• Unanswered questions
It should make a convincing argument that there is not
sufficient knowledge available to explain the problem and
to give a possible solution. From the above background the
justification of the present study is to be given by explaining
the expected result in terms like (i) what knowledge and
information will be obtained, (ii) how will the result be used
and who will be the beneficiaries, (iii) how will the result be
disseminated,(iv) contribution to the public health policy
and (v)scientific importance of the study. Thereby make a
convincing argument that the knowledge generated will have
a practical value.
Literature Review: Review of literature is a very important
part of a research project. All available literature that supports
the research idea should be summarized. In this part
definition of the problem, facts already known, weaknesses
and gaps in the literature etc should be mentioned. It
familiarizes the reader to the problem under study and helps
 An Introduction to Research Methodology 235

to structure the methodology of the research work. It

describes the work done by others either at local or
international level on it or similar subject. It helps to prevent
the duplication of work. The researcher can understand the
difficulties faced by others and the corrective steps taken or
modifications made by them. The researcher can anticipate
similar or additional problems during the study. The relevant
information should be covered in about 300 words quoting
8-10 authentic, easily retrievable references. The source of
literature review can be specialized books, national and
international journals and internet search.
Aim: Aim is broad but concise statement of what the research
study hopes to prove. Objectives refer to what would you
actually do in this study to answer the research question.
Ideally in any study, both the aim and objectives have to be
stated clearly.
Objectives: Objective is a clear statement of the specific
purpose of the study. There may be more than one objective.
The study objectives emerge from the research question. The
objectives should be Specific, Measurable, Achievable,
Relevant and Time bound (SMART objectives). It may be
presented as primary and secondary objectives.
The primary objective reflects the main clinically
relevant goal of the study. Every study must have a primary
objective. Define the primary objective in terms of what
will be measured in a single, clear and concise statement. It
is hypothesis driven. The sample size calculation is based
on the primary objective.
A study may or may not have secondary objectives. It
includes more general non-experimental objectives and may
not be hypothesis-driven. The number of objectives should
be kept low (2-4), as too many objectives may make the
236 An Introduction to Research Methodology

study logistically difficult to perform. Hypotheses:

Hypotheses are research questions as a state merit form more
specific than objectives and are amenable to statistical
evaluation. Research hypotheses are the specific testable
statements made about the relationship between
independent and dependent variables in the study. In a
proper scientific methodology a research hypotheses is stated
not in order to prove it but to test it. The study starts with
the assumption that the research hypothesis is not true. So
the null hypothesis is to be stated first. If the results do not
support the null hypothesis then the research hypothesis is
more likely to be true. It should be stated as alternate
hypothesis.
Methodology: It is the most important part of research
protocol and forms the core of the research project. It should
contain sufficient information about all the procedures
planned to achieve the objectives. It shows that the proposal
is feasible and practical. It also addresses all potential
problems that may arise in the course of conducting the
study and suggest solutions to them. The scientific integrity
of the study and the credibility of the study data depend on
methodology. Methodology can be modified according to
the specific type of study. It should be written in full detail
including research design, the research population,
interventions, measurements and plan of analysis.
Study Design: Study design would vary according to the
research problem. Study design should be mentioned clearly
and should be appropriate to answer the research question.
Justification for selection of the particular study design is to
be given along with strengths and weaknesses of design and
that of alternative designs. The selection of study design
needs to be based on feasibility, resources, time frame and
ethical considerations.
 An Introduction to Research Methodology 237

Study Setting: This includes information about the location

where the study will be conducted eg community, school,
hospital etc and the facilities available.
Study Population: Selection of study population should be
done very carefully. We should describe all the steps and
decisions made in order to select the subjects to be studied
as this affects the external validity of the study. Selection
criteria should not be too lenient or strict. If the inclusion
criteria are loose, the population may be too heterogeneous
for the study question to be answered. If the criteria are too
stringent, few patients could be enrolled, and selected
population may represent only a fraction of the study
population, making it difficult to generalize the results.
Detailed information regarding the subjects should be
given i.e. reference population, source population, sampling
frame and study population and the rationale for selection.
Describe the methods by which subjects will be identified
and recruited for the study. Mention the inclusion and
exclusion criteria, retention strategies and withdrawal
criteria.
Inclusion criteria : Inclusion criteria are the
‘characteristics’ required for a participant to be included in
the study. The criteria may be based on factors such as age,
gender, the type and stage of a disease, previous treatment
history, and co-morbid medical conditions.
Exclusion criteria : Details of participants who are
considered ineligible to participate, even though they are
eligible as per inclusion criteria, and justification for their
exclusion.
Sampling: All steps adopted in the recruitment of study
subjects should be described. The samples should be
representative of the population. Steps taken to minimize
238 An Introduction to Research Methodology

sampling errors should be clearly mentioned.

Sample size calculation: Sample size should be calculated
based on each primary objectives and higher number should
be taken. It should be adequate to answer the research
questions and test statistical hypotheses. Provide data
necessary to calculate the sample size, state how the estimates
were obtained and include statistical power estimates.
Identification of variables: Variables are values that can
change from subject to subject. It is important to identify
the variables at the planning stage itself. Knowledge of the
variables- independent variables, dependent variables,
background variables and confounders in a research project
are essential in refining the objectives. They should be
defined and quantified with a measurable unit.
Interventions : Detailed description about the
intervention (drug/device/procedure/investigation etc) must
be given. For drugs and devices that are commercially
available, mention the proprietary names, manufacturer,
chemical composition, storage, compounding, dose,
frequency and method of administration. If they are in
phases of experimentation or are already commercially
available but used for other indications, information must
be provided on available pre-clinical investigations in animals
and/or results of studies already conducted in humans.
Measurement / Data collection : Instruments for
measurement and method of measurement should be given.
Refer to the validity and reliability of the measurement
instrument and quality control measures should be specified.
For validated questionnaires reference to published work
should be given and for a new questionnaire details about
preparing, pre-coding and pretesting of questionnaire should
be furnished. For established procedures like medical
 An Introduction to Research Methodology 239

examination, laboratory tests and screening procedures

reference of published work can be cited; but for a new or
modified procedure, an adequate description is necessary
with justification for the same.
Data Management and Statistical Analysis: This section
should describe information on data access, confidentiality,
editing, coding, classification and tabulation of data.
Statistical tests to be carried out in order to test each of the
stated hypotheses should be clearly outlined. Mention about
the computer programs and softwares that will be used.
Provide dummy tables to make sure that all variables
measured would ultimately be used for analyses.
Time Line
The protocol should specify the time that each phase of the
project is likely to take, along with a detailed month by
month timeline for each activity to be undertaken. In the
project-management timetable use an action plan or Gantt
Chart and write the exact dates on which following specific
actions will occur.
• Preparation of study questionnaires or formats
• Pilot testing
• Recruiting of subjects
• Logistics including personnel and training
• Data collection
• Follow-up procedures
• Data scrutiny and statistical analysis
• Reporting procedures
Ethical and Legal Considerations
This should explain how we are going to deal with ethical
constrains to the design and implementation of the study
and ICMR guidelines could be followed. It is not sufficient
240 An Introduction to Research Methodology

to state that we are going to respect these principles but is

to be explained how we are going to achieve that. In this
part it is important to specify the ethical issues involved in
the study like benefits of the study procedures, privacy of
participants, confidentiality of information, any possible
harm/ side effects/ consequences, right of discontinuation
at any time, alternative treatments/ approaches (in trials),
incentives/ rewards, any conflict of interest, and voluntary
consent etc. This session will help ethical committee to
scrutinize the protocol considering all the ethical issues
involved before approving it.
Budget/Resources
A brief outline of the budget requirement showing head-
wise expenditure for the study should be given in this session.
Eg: personnel, consumables, equipment, supplies,
communication, funds for patients, data processing,
transportation etc. Jusitify the use of each item, consider
the workplan of the study.
Dissemination activities/reporting
Outline the planned publication/reporting strategy. State
how the results will be reported and who will be given access
to the data.
References
In this section list of the various references quoted while
formulating protocol may be listed in a sequential manner.
Commonly used referencing systems are
Harvard Style
Name and publication year in text
Alphabetical bibliography
Vancouver Style
Numbered references
 An Introduction to Research Methodology 241

Continous referencing in text

Make use of softwares like Reference Manager or Endnote
software
Annexure
The following documents should be appended
o Case Record Form
o Data collection tools/Questionaires
o Patient information form (in required languages)
o Consent form (in required languages)
o CV of investigators
o Letters of approval
Conclusion
Planning is essential when preparing a research proposal.
On an average, 2 – 3 months will be needed to prepare and
get clearance of a proposal. In depth review of literature
and recent studies about the topic should be done before
drafting a protocol. It must be peer-reviewed and finalization
may be done only after incorporating the suggestions
obtained. A good protocol will give maximum information
in minimum words. A research proposal is intended to
convince others that you have a worthwhile research project
and that you have the competence and a work- plan to
complete it.

References
1. Fathalla MF, Fathalla MM. A practical guide for health
research. WHO Regional Publications, Eastern Mediteranean.
Series 30. In: WHO 2004.
2. How to Prepare a Research Proposal: Oman Medical Journal-
2008 Apr;23(2):66-69
242 An Introduction to Research Methodology

3. Silverman,Kwiatkowski. Research Fundamentals: 111.

Elements of Research Protocol for Clinical Trials
4. How to Design a (Good) Epidemiological Observational
Study: Epidemiological Research Protocol at a Dlance. Acta
Med Port 2013 Nov-Dec;26(6):731-736
5. Indian Journal of Dermatology,Venereology and Leprology
(IJDVL) http:ijdvi. Com/printarticle. Asp?issn=0378-6323;
year=2008;volume=74 issue:6, page:687-690 Accessed Feb.
21,2016
6. Kenneth A Eaton and Ario Santini. An Introduction to
Research for Primary Dental Care Clinicians.p Part 3: Stage
5. Writing a Protocol. Primary Dental Care 2011;18(2);91-
94
7. Enarsan DA, Kennedy SM & Miller DL: Getting started in
research – the research protocol, Int J Tuberc Lung Dis, 2004,
8, 1036-1040.
8. Writing Research Protocol – medIND http://www.
medind.nic.in/nac
9. Research Protocol Template – Mater Health Services http:/
/ www.mater.org.au/Home/Research
10. Developing a Protocol – http://www.cdc.gov/niosh/nas/
mining/pdfs/Protocol
11. WHO/ Recommended format for a Research Protocol –
http://www.who.int/rpc/research_ethics/format_rp/en/
12. Hilsden RJ, Verhoef MJ .Writing an effective research
proposal. Department of Medicine and Community Health
Sciences, University of Calgary. Calgary, Alberta, Canada
13. Kothary CR.Research Methodology- Methods and
Techniques.New Age International Publishers. Second
Revised Edition.
 An Introduction to Research Methodology 243

Chapter 18
How to Write a Thesis
Ramesh Kumar

Writing a thesis poses an immense challenge to new scholars.

An early introduction to the thesis writing process would
provide a context for planning scholarly writing,
demystifying the process, leading to the improvement of
writing skills 1. Preparing a thesis and publishing it is
paramount obligation of a researcher which is, necessary for
accumulating a professional body of knowledge and fully
developing its implications for theory and practice2.
Purpose and motives of Thesis
Doing a thesis is primarily a learning experience and is useful
to develop expertise in a selected topic through independent
research. It also helps to compile and present information
that is easy to retrieve. Quality thesis can later be further
refined for a publication.
Types of thesis
1. Research study with primary data collection: Data are
collected specifically for the thesis/ dissertation which
directly addresses the research question.
2. Research study with secondary data analysis: Here data
collected for a different project is used (secondary
analysis) i.e. analysis of data that have already been
244 An Introduction to Research Methodology

collected by others.
Structure of a thesis
A standard format helps readers to find expected
information. A proposed structure which can be used to
effectively write a thesis, is summarised below.
Chapter 0: Title and Abstract: Key words, Dedication,
Acknowledgements, List of abbreviations used, Table of
contents, List of figures, List of tables, List of graphs,
Statement of original authorship.
Chapter 1: Introduction: Rationale, ending with broad
objectives of the study
Chapter 2: Review of Literature
Chapter 3: Methods: Experimental technique, sampling
strategy, detailed methodology
Chapter 4: Results
Chapter 5: Discussion
Chapter 6: Summary and Conclusions: Limitations,
Suggestions for further research
References,
Annexures
Chapter 0: Title Page: It should contain Title of the thesis,
Author’s full legal name and Year of acceptance
a. Title of the thesis: The title of thesis/dissertation is your
first opportunity to let the reader know what your
dissertation is about. The title has to highlight the aim,
design and setting of the study in a few words. Title should
never contain any abbreviation. A title predicts the content
of the thesis, contains keywords that will make it easy to
access it by a search engine. Title must be short, informative,
attractive, specific and accurate subject matter of the thesis
 An Introduction to Research Methodology 245

and generally should not exceed 15 words. The title gives

the potential reader; concise and accurate first impression
of the article’s content and hence, should be a concise
summary of the topic.
Running title: Short title usually printed as a header at the
top of each or alternate pages.
b. Authors: Author’s full legal name. Authorship is the
currency of a scientist’s career and research experience. When
the thesis is converted to a research paper, the person who
did the work and wrote the paper is generally listed as the
first author, the guide as second author, co-guide or
statistician who actively participated in the research work
as third author. Other persons who made substantial
contribution to the work are also listed as authors.
c. Abstract: An abstract is a greatly condensed version of a
longer piece of writing that highlights the major points
covered, and concisely describes the content and scope of
the writing. Abstracts give readers a chance to quickly see
what the main contents are. Readers can decide whether
the work is of interest for them by looking at the abstract.
Generally structured abstract is used and the components
of a structured abstract are; introduction, the objectives of
study, the methods, results in brief with their statistical
significance and the main conclusions.
d. Dedication: It is used to mention those who supported
the study by all means during your research work. This is
not typically the place to recognize those who assisted you
in your academic research, which is done on the required
Acknowledgements page.
e. Acknowledgements: This page can be in the beginning
or at the end of the dissertation. Names of the colleagues
who supported the study including the computer assistant
246 An Introduction to Research Methodology

can be added while highlighting one’s gratefulness

f. List of abbreviations used: It is the explanation of
acronyms. This list provides topic specific abbreviations and
definitions that are not common in your field of study or
between disciplines. By looking at this alphabetized list, the
reader can easily locate defined abbreviations. Use
abbreviations accepted in the specific discipline only, not
abbreviations made by researcher for convenience.
g. Table of contents, List of figures, List of tables, List of
graphs
Chapter1: Introduction
a) Background of the research
Introduction is the part that gives the first impression about
the thesis. So the introduction is a key paragraph as far as
readers and writers are concerned. The reader will be more
inclined to read a paper and consider a position if the initial
paragraph is clear, organized, and engaging. The purpose
of introduction is to inform the reader about the background
and context of the study and how it relates to the previous
works done in the area. The introduction should outline
the broad field of study and then lead on to the focus of the
research problem. The introduction is used to (i) to establish
the overall field, (ii)summarise previous research (iii) indicate
the research gap and (iv) state the purpose of the study
Introduction introduces subject matter of the investigation
to the reader, the background of the study, current state of
knowledge regarding the subject (are there any existing
solutions for the problem being addressed, what is the main
solution in terms of published references) and gaps in
existing knowledge.
b) Rationale/Justification of the research
Ideally, this section should progress from the general to the
 An Introduction to Research Methodology 247

specific as done in journalistic writing (inverted triangle).

Start broadly laying out the background and then narrow it
down to the specific question that is being addressed . This
section should include, Background of the project, Short
history and current situation, context of the work, gaps in
the research, nature of the problem and its relevance, reason
for specific interest, rationale /justification in brief and ends
with broad objectives of the study.
Chapter 2: Review of literature
This section carries an overview and evaluation of the
writings in a specific area of interest. It is the compilation
of the essence of papers related to the subject,.
The literature review demonstrates an author’s knowledge
about a particular field of study and its methods. It also
provides a framework for relating new findings to the
available findings later in the discussion section of a
dissertation. To obtain useful and comprehensive literature
search the following steps may be adopted.
Step1: After deciding the central topic of the study, the
literature review leads to the dark areas to be surveyed further
It helps to relate new findings to existing knowledge.
Step2: Identification of the specific literatures to be surveyed
and relevant reviews to be collected especially on definition,
prevalence, etiological factors, previous interventions, cost
effective issues and up scaling problems.
Step3: Redefine or refine the research question, from a broad
area, to a narrower but specific area of interest.
Step4: Narrow down your search to get methodological
insights in other similar studies done elsewhere
Step5:Set a stage for our own study, based on strengths and
weaknesses of other studies.
248 An Introduction to Research Methodology

Step6: Identify problems encountered and possible solutions

from previous similar studies.
Procedure to be adopted include the following
1. Collection of original articles
2. Separate them into descriptive, analytical and
interventional studies.
3. Review of descriptive studies can give useful information
on definition; incidence/prevalence etc while analytical
studies give risk associations .Interventional studies
particularly RCTs give insight into effectiveness of
interventions.
4. Systematic reviews give broad statements on description,
evaluation, and integration of materials found in this
collection, meta analysis would give risk quantification
like attributable risk and a chance for comparison of
Odds ratios and their confidence intervals
5. Consolidation and writing the findings in an appropriate
manner under definition, incidence/prevalence, risk
associations and effectiveness of interventions.
6. For doctoral and post doctoral thesis it is a must to do
critical review of each article or reference included in
the review of literature.
Sources
1. Primar y sources: Original materials, on which
subsequent researches are based, not interpreted or
evaluated. E.g.: original documents, original articles
published in peer reviewed journals
2. Secondary sources: Interpretation and evaluation of the
primary source E.g.: bibliographies, biographical work,
criticisms, magazines, newspapers etc.
 An Introduction to Research Methodology 249

3. Tertiary sources: Information derived and collected from

primary and secondary sources E.g.: text books,
dictionaries, guide books
Apart from printed or physical sources like books, magazines,
journal etc. electronic sources particularly medical related
search sources like Pubmed, Medline etc can be very useful
if the correct MESH headings can be typed in.
Chapter 3: Methods
This section describes in detail how the study was conducted.
i. What and how it was done?
ii. How would the study answer the specific questions?
iii. Who are the study population and what was the sample
size and sampling strategies?
iv. Describe research procedure in detail. Give enough
information for another scientist to repeat your
experiment (that means method should enable
subsequent researchers to duplicate the study).
v. Describe Laboratory techniques used: You have to
describe what apparatus and materials was used and from
where you obtained them. A well made schematic
diagram is more helpful than a photograph of the
apparatus.
vi. Reproducibility: - key principle of science, study results
are valid if it can be consistently repeated.
The method section is expected to describe, all stages of
planning, composition of the study sample eg: patients,
animals, execution of the study, statistical methods planned.
Study design
The most important element of a scientific investigation is
a study design. This should be mentioned without ambiguity.
Study design is the method by which we try to answer our
250 An Introduction to Research Methodology

research question.

Purpose of investigation Study type

Investigating causation of rare diseases. Case-control
study
Investigation of exposure to multiple agents Case-control
e.g. Gastric cancers (multiple risk factors study
like food habits, smoking, stress etc.)
Investigating rare exposures e.g. industrial Cohort study
chemicals
Investigation of multiple outcoms e.g. the Cohort study
risk of bad neonatal outcomes from single
risk factor in pregnancy like birth asphyxia.
Estimation of incidence in exposed Cohort study
populations
Investigation of effectiveness of RCT
interventions (eg. A new drug)

The choice of the study design should be explained in

clear terms. Method of randomisation if any, should be
specified. In experimental studies, the precise depiction of
the design and execution is vital.
Steps in methodology
1. Describe the subjects/materials used
Animals – species, strains, rearing method, suppliers name,
animal diet given
Drugs – generic and trade name, schedule of drug, form of
giving and usage
2. Mode of selection of subjects: Sample size, Inclusion and
 An Introduction to Research Methodology 251

exclusion criteria, Specific procedures adopted (eg.

matching, randomisation, restriction.)
3. Description of control group (active historical or placebo
controls)
4. In case control studies, cohort studies and non-
randomised clinical trials, choice of the controls must
be described precisely.
5. Randomisation procedure (assigning patients to the
different arms of the study)
6. Blinding procedures: Procedure of blinding investigators
who measure the outcome, blinding study participants,
and blinding the biostatistician.
7. Measurements of data: Instruments and techniques, i.e.
the measuring devices, scale of measured values,
laboratory data and time point should be described in
details. So also standardisation of instruments, validity,
blinding etc.
8. Methods of data collection: Observation,
Questionnaire or Interview method.
9. The accuracy of a method i.e. its reliability (precision)
and validity of instruments used. Details of steps taken
to make sure that measurements are done under
standardised and thus comparable conditions in all
patients, should be mentioned. Details of measurement
procedures important for assessment of accuracy
(reliability, validity) are also described in this section.
10. Response rate of the participants. (e”80% is good, d”
80% no or only slight power)
11. Loss to follow up particularly in Randomised Control
Trial (RCT).
12. Plan of Data analysis: The choice of test used for
252 An Introduction to Research Methodology

statistical analysis to be mentioned.

Sample size should be adequate to ensure power of the study.
In the case of negative study results, power of the study
should be re calculated and mentioned.
Chapter 4: Result
In this section the findings should be presented clearly and
objectively without interpretation. The result section should
address directly aims of the study. Results should be
presented in a well structured readily understandable and
consistent manner. This section should include a
comprehensive description of the study population. Results
can be summarised and presented as texts, tables, charts,
illustrations/pictures. Tables and figures may improve the
clarity and the data therein should be self-explanatory and
to be interpreted under each table. Same data should not be
represented in more than one method.
Chapter 5: Discussion
In this section the author should discuss his or her results
with detailed interpretations.
1. Highlight the most significant results.
2. Explain how questions posed in introduction are
answered by the result section.
3. Discuss whether the objectives are fulfilled
4. Dscribes how hypothesis is tested and results are in favour
of or against null hypothesis
5. Explain whether the data support your hypothesis.
6. Compare the data with earlier studies and see how these
answers fit with the existing knowledge.
7. Discuss whether your results are consistent with what
other investigators have reported.
 An Introduction to Research Methodology 253

8. Explanation of unexpected results should be included

in this area.
9. Critical analyses of limitations of the study are an
essential part of this section.
10. Usually this section concludes with suggestions for
further research to answer the questions raised by the
result.
Chapter 6. Summary and Conclusion
This gives the brief and logical interpretation of the result
and the answer to your research questions in brief. In
assessing the results and formulating the conclusion, the
weakness of the study must be given due consideration.
The inclusion of non-significant results contributes to the
credibility of study. “Not significant” should not be confused
with “no association”. Significant results should be
considered from the view point of biological and medical
plausibility. All conclusions should be supported by “your”
study findings.
References
The authenticity of the literature review is enhanced by the
use of references citing their source, author, time period etc.
Modes of referencing should be in an accepted format
1. Vancouver referencing style for referencing
· Mostly in health and physical science
· List of references: numbered consecutively in
chronological order of quotation in the text
· Arabic numbers are assigned in the order of citation
· Numbers are commonly enclosed in round brackets,
e.g. (1) but it can also be written in square brackets,
254 An Introduction to Research Methodology

e.g. [1] or as superscripts, e.g.1 or as a combination

of bracket and superscript, e.g. (1)
· If more than once cited, the same citation number is
used
2. Harvard referencing style for referencing.
· Mostly in Humanities and Social sciences
· List of references placed in alphabetical order.
· References quoted in the text: surname of author(s)
and publication year are listed
A properly drafted thesis starts with evolution of a new
research question which evolves into a scientifically executed
research process, which in turn leads to new research
findings. These findings are discussed in the light of existing
knowledge to evolve new conclusions and further research
questions.