Business Research Methods
Business Research Methods
Page No. 1
5.6 Formulation of Research Plan
5.7 Ethical Issues in Conduct of Research
Page No. 2
UNIT 1: INTRODUCTION TO BUSINESS RESEARCH
Page No. 3
1.2 Business Research
Page No. 4
market-men and other persons associated with manufacture, distribution.
and sale of its product.
Certain scientific methods are adopted in business research, on the basis of certain
basic assumptions. These assumptions include:
(a) Business problems have causal relationships with certain parameters and
phenomena prevalent in the planning and execution of business operations.
(b) Various business activities do not occur in a haphazard manner. There is
some system and trend behind these.
(c) It is possible for a human, being part of business-society interaction system,
to study the business related systems in a detached manner.
(d) In a large business system comprising of producers-marketers-distributors-
retailers-customers‟ system, study may be made on a representative sample
and results may be applied to the whole system.
The field of business research can be classified into number of ways as described
below:
Page No. 5
There are two basic approaches for social science research: qualitativeapproach
and quantitativeapproach. Both differ in the nature of problem and problems
parameters or research variables, and the type of data associated with the
phenomena being investigated.
Qualitative approach can have two further sub-groupings: inferential approach and
experimental approach where data generated is in quantitative form. In inferential
approach, survey research is done on a sample population and then it is inferred that
the population has same characteristics as the sample. Experimental research
involves study on an experimental group selected from the population. Here greater
control is possible over the research environment
Data for business research can be obtained from two types of sources: documentary
sources and field sources. Documentary sources include material already collected
which may be published or unpublished. Such sources are called
secondarysources, as information has been complied already and is being
conveyed under second step.Fieldsources may include living persons, scholars,
Page No. 6
research workers, leaders of social groups who have first-hand information regarding
the phenomena being researched. Such information is to be compiled / gathered by
the researcher himself/herself, and therefore such sources are called
primarysources.
Various experts and researchers have described their views and findings on different
issues and phenomena in terms of books, research articles and other
documents. Such secondary sources are a rich source of information for any new
researchers. Many governmental and private organisations carry out studies and
surveys as part of their regular activities. Some of these findings are published in for
of reportsandmonographsforpublicinformation. However, a large part of their
study reports are not published due to lack of funds or other reasons, but may be
made available to new researchers on special requests.The reliability and accuracy
of secondary data cannot be assumed without verifying the reliability of the source.
Research methodology refers to the systematic way used to solve the research
problem. The methodology describes the steps adopted for solving the
problem, and also the logic behind use of such steps. The methodology
includes:
Page No. 7
the research methods used;
the details of allied steps used data capturing,
interaction of data source(s),
steps for data interpretation and analysis,
finding answers to the research questions, and
interpretation of the research findings etc.
A research plan or a research proposal covers, from start to finish, various steps
used for carrying out the objectives and the sequence in which these are
undertaken. The activities carried out under each step are also described in the
sequential manner.
Page No. 8
formulated in terms of the research related „questions‟ whose answer will be
available after the proposed research. This can also be stated in terms of
working hypotheses which can be tested on the basis of the results of the
research.
(d) Development of research design:Research design refers to development
of conceptual structure under which the research will be carried out. It
also describes the research method to be adopted for the research. Summary
of methods to be used for data collection and data analysis is also presented.
Sampling method adopted in the research for data collection is also
presented.
(e) Data collection method: Primary data is generally collected through
experimentation, observation, and surveys. Under each of these methods,
there can be several specific methods in the micro sense. Therefore, the
researcher must describe the research methods, and its details, adopted for
collection of data for the research.
(f) Data analysis & interpretation methods and techniques: There can be
many methods and techniques for analysing the data. Therefore, the
researcher should describe the specific techniques used for data analysis and
interpretation of results. Methods of hypotheses testing should also be
presented. Similarly, the method and procedure used to find answers to the
research questions should also be described under this section.
(g) Estimates of resources (time, cost, and manpower) involved:As the
research plan /proposal document may be used for seeking formal approval
for carrying out the proposed business or social research, it must present the
resources required for physical conduct of the research work. It must present
realistic estimated of resources namely money, time, personnel, building
space, transportation/logistics facilities required etc.
Page No. 9
The research methods and procedures should be finalised and recorded at
the beginning.
Detailed research design should be prepared and strictly followed.
The analysis of data should be sufficiently elaborate, and its validity and
reliability should be checked out properly.
Testing of hypotheses and finding answers to research questions should be
done carefully and strictly in accordance with established practice adopted by
various researchers in the past.
Conclusions should be confined to the research evidences brought out
through interpretation of the findings and the results of the research.
Business research, being a scientific study follows the basic principles of scientific
research. It has certain characteristics which are common to scientific research. It
involves data measurement, which has to have accuracy, reliability and repeatability
of measurement. This is necessary as data collection and data measurement are the
foundation pillars of any good useful research. There are number of other
characteristics necessary in any business research to make it a scientific research,
based on which results obtained from a small sample of population may be
generalised to be true for the population on the whole. Some of such characteristics
of a good business research are presented and discussed below.
A researcher‟s personal values, belief, work habits and attitude towards work
have a direct influence of his research work, Therefore, the researcher has to
plan his actions and search appropriate research methodologies to ensure that
the observations and the results are not influenced by his beliefs, but are purely
the outcome of research outputs from the selected business/social group.
If the study is influenced by personal views and beliefs of the researcher, then the
results reflect his/her views and not the true views of the persons being
Page No. 10
researched. This is called influenced caused by subjectivity of the researcher.
This lowers the correctness value of the research. The research study is
called subjectivewhen it is influenced by researcher‟s personal views and beliefs;
and is not purely based on the true observation of the facts
Page No. 11
instruments are so designed that these do not add errors due to their inbuilt
features, and eliminate responder‟s bias and errors entering the measurement.
Predictive Validity: The predictive validity reflects the accuracy by which the
current measurement of a variable can be used to predict its value at a later
stage. For example, if a test is used to measure achievement of athletes in a
tournament, it can also be used to predict their achievement at a later stage. If
we are able to make accurate predictions or if our predictions come true, we
may say that the measuring instrument is valid.
Page No. 12
attributes like „customer satisfaction‟ and „competitive advantage‟ are
measured through sum of other sub-attributes which are called „contents‟ of
the main attribute to be measured. Content validity is the representativeness
of the content, as described above. Here validity represents the degree to
which the selected content represents the total content or universe of the
attribute being measured. Content validity is guided by the question: Is the
content of the measure representative of the content or the universe of
content of the property/attribute being measured?
Page No. 13
UNIT 2: IDENTIFYING THE RESEARCH PROBLEM
First task for solving a problem is to study all symptoms arising from various
individuals and or groups of persons. Study of the symptoms generally raises a
pointer in direction of some definite aspects which are usually associated with
particular problems. Thus, examination of symptoms leads to certain set of
problem issues. The probability of occurrence of various problem issues are
assessed and suggestions are evolved on existence of certain problems.
Page No. 14
definition helps to focus attention on the correct aspects of the problem. For defining
the problem, it is necessary that the boundaries of the problem are stated clearly on
the basis of detailed study of symptoms. This will help a researcher to study the
problem with a predetermined set of objectives.
The research problem should be defined in a systematic manner giving due weight
age to all points relating to different types of symptoms.
Page No. 15
2.2 Relevance of Research
The proposed research study should be relevant to the, business organisations,
academicians, society, social groups, and researchers of current time and of the
near future as well. The researcher should study and analyse the current and
emerging future environment, and establish the relevance and usefulness of the
proposed research. Such potential relevance and usefulness should be established
and recorded in the beginning of the research. It may be reviewed and updated, if
required, on completion of the research study.
The objective can also be stated in terms of specific questions whose answers have
to be found through the proposed research. Such “questions” whose answers are to
be found on the basis of proposed research are called as the “research questions”.
A hypothesis is a proposition which the researcher wants to verify. Often there may
be several competing hypotheses, either specified or implied. One objective of
research is to select among the possible hypotheses and to test them empirically
with the help of statistical tools in order to ascertain whether they are true or false.
Hypothesis for mulation deals with assumptions made by the researchers. There
are two kind of hypotheses: Null and Alternative hypothesis. Null hypothesis is an
assumption that there is no difference between the sample and population
phenomena. And alternative hypothesis states that the difference between the
sample and population phenomena is statistically significant.
Page No. 16
The symbol used for null hypothesis is Ho and for alternative hypothesis is denoted
The formal definition of the research problem clearly brings out the issues proposed
to be studied and analysed through the proposed research. This brings out the
scope of the proposed research.
The issues not covered, directly or indirectly, in the definition of the research problem
are not likely to be studied under the proposed research, and therefore are outside
the scope of the proposed research.
Page No. 17
UNIT 3 PREPARING FOR RESEACH STUDY
A researcher may have some personal knowledge regarding the research problem
and may also collect some information from the persons connected with the problem.
Before starting he research, it is essential for the researcher to develop strong
theoretical knowledge-base regarding the background and all aspects of the
problem situation.Authentic published literature available in reputed books and
research journals provide rich and reliable source of background information on the
subject matter.
Page No. 18
Published reports / articles are not used as the data for the proposed research
to be carried out, but only serve as material useful for building ‘theory’ about
reasons of the problem and identifying the inter-relationship between various
problem issues.
When to stop the study of literature? The research evidences provided by each
source should be recorded theme wise, subject wise, and alphabet wise for quick
reference and cross checking. Different sources may support the same evidence, or
may bring out more material, or may even provide counter evidence. When
sufficiently large amount of literature has been surveyed, a stage is reached that with
more studies of journals or books/reports, very less amount of new research
evidence is available, and no new point emerges. This is the stage where the
researcher may decide about closing/completing the study of literature.
The figure below shows a problem issue in a specific environment, say society or
business organisation. When the problem is understood, discussed and analysed, a
clear picture emerges.
The problem has some root causes due to which the problem is generated and is
reported. The problem, if left unattended, leads to certain adverse impact on the
society or the organisation.
Page No. 19
(b) The „problem’ which results into „adverse impact‟, and
(c) It can also be stated that the „causes‟ ultimately lead to the „impact’ om
society or om business organization.
Environment
Impact of the
Causes of the The Problem Problem Issue
Problem Issue Issue on Society or
Organisation
Environment
Fig: Conceptual Representation of the Problem, showing Independent and Dependent Variables
Page No. 20
issue results into a change in the impact of society/organisation. The variable
corresponding to the middle or intermediate stage is called as the intermediate
variable.
Like the figure given above in section 3.2, a pictorial representation of the
problem issue facilitates in clearer understanding of key issues related to the
problem. The background issues relate to the independent variables and their inter-
dependence can be represented on left side and problem issues can be represented
on the right side.
Such a conceptual model helps in understanding the variables involved in the study;
and also the relationship between the dependent and independent variables. This
helps in developing deeper insights into the research problem.
Another issue to be understood before launching the research study is to identify the
sources from where research data can be collected. All units providing data for the
study constitute the „universe ’for the study, and all persons providing data are said
to constitute the „population’ for the study.
In some cases, the universe is not fully known, though a broad understanding
may be there. Indirect methods have to be developed to access the population for
carrying out the research study. Some time, the persons providing the data are not
directly known to the researcher.
For example, a research study is to be carried out among the school students of the
country. Here the names of students are not known to the researcher. But, the
country can be seen to be divided into number of provinces and each province can
be seen to be composed of number of districts. Now the district becomes a
manageably smaller unit (as compared to the country). The district has known
number of schools. Thus, students can be identified through the school where they
Page No. 21
are studying. Here respondent unit is the student. But for purpose of convenience,
students can be identified through the name of the school where they are studying.
Therefore, school can be called as a basic „research unit‟ where research is carried
out, and student becomes the source of data for the research.
Sampling Theory: Physically, it may not be always possible to collect data from
each member of the universe or population of very large size. Further, the research
has to be completed within a reasonable time and cost. Statistically, it is observed
that if the data collected from a smaller group called as the sample, which is
representative of the larger group i.e. universe, it will be representative of the
data collected from the whole universe. The smaller sample group is to be
carefully selected so that its properties are true representative of the properties of
the universe or population. In view of the above stated statistical property, research
data may be collected from the carefully selected smaller group called sample group,
and the results so obtained can be generalised to be true for the whole universe of
the population.
Sampling design refers to the plan for obtaining a sample from a given population.
Sampling design is decided before data collection is commenced. Elements of the
sample are decided in a manner that they yield the same information about the
universe. The more approximately the sample presents the true characteristics of the
population, the more representative and dependable it is.
Sampling has number of advantages: (i) it cuts the costs involved;(ii) it increases
the accuracy as in small sample the accuracy of data can be checked more easily;
and (iii) the administrative issues become simpler due to smaller size. However,
some care is also essential. Sampling must be carefully designed and followed
otherwise the sample may become a wild sample.
Page No. 22
Various sampling techniques can be divided in two broad groups namely
probabilistic sampling and non-probabilistic sampling.
There can be two type of probability sampling: simple random sampling method, and
stratified random sampling method.
(a) Under simple random sampling method, all elements of the universe are
arranged in a random manner, and a criteria of selection is decided, say every
tenth or hundredth item is selected for inclusion in the sample.
(b) In stratified sampling, the universe is defined in terms of different groups of
homogenous nature called stratum. Therefore, elements of the population are
placed under its proper stratum as per the criteria for stratification. Then,
random sample is selected from each stratum.
Page No. 23
(a) Nominal Scale: This scale provides a system of assigning numbers like
players of a team are assigned unique numbers. Such numbers serve the
purpose on only identification of object. This scale provides a system of
identifying and tracking of persons, objects etc, but provides no quantitative
information.
(b) Ordinal Scale: This scale also has a system of assigning number. But with a
difference over nominal scale. It has rank ordering system. Two parameters
are assigned numbers in a manner that larger number signifies higher value
of the parameter being measured. Thus “less than” significance and “greater
than” significance is conveyed by the number assigned.
(c) Interval Scale: This scale has an additional feature over the ordinal scale.
The difference in the number(s) assigned to the two parameters denote
difference in values of two parameters. This scale can have an arbitrary zero,
but there is no concept of absolute zero.
(d) Ratio Scale: This scale has absolute zero for measurement. Another
important feature is that ratio of numbers assigned to two parameters
indicates the ratios of values assigned to these parameters. Ratio scale
represents the actual parameters‟ value like weight, length, volume etc.
Page No. 24
Some typical scales used in business research and in social science research for
measuring of qualitative parameters are the (i) Summated Scales, called “Likert-
Type Scale” and (ii) Differential Scale or Thurstone Scale.
This scale is used for measuring qualitative variables. It makes an assumption that a
respondent is able to clearly discriminate between „low‟ and „high‟ scores/values for a
particular variable. The statements which meet such discrimination criteria are
included in the measuring instrument.
In the above mentioned situation, the respondent is asked to give his views on the
statement: “the management encourages risk-taking by the employees as part of
discharge of their duties” The respondent is given number of options to choose from,
which may be as under:
i. I strongly agree;
ii. I agree;
iii. I neither agree nor disagree;
iv. I disagree; and
v. I strongly disagree.
Each statement on the Likert Scale is assigned a quantitative number; say least
score (I strongly disagree) is assigned a score of 1 (one) and highest possible score
(I strongly agree) is assigned a score of 5 (five). Number of respondents give their
response to the statement, then the average value of all scores is taken and is given
Page No. 25
as final score. Say, the sum of score for response given by 10 respondent comes out
to be 34. Then the average score is 3.4 divided by 10 i.e. 3.4.
A three-point scale can also be used, which will have response options as: (i) I
agree; (ii) I neither agree nor disagree, and (iii) I disagree. Similarly, a seven-point
scale can also be designed and used. In practical research, use of “Five-point Likert
Scale” is more common and provides reasonable level of accuracy / reliability.
Likert type scale is easy to design and also easy to use. Most of the researchers use
this scale in business research and also in social science research. There is a
limitation of the Likert Scale. The various intervals like between „agree‟ and
„strongly agree‟ may not be equal to spacing between „neither agree nor disagree‟
and „disagree‟. Thus, to this scale is an approximation the reality.
This scale is generally eleven-point scale, where value for each item is not assigned
randomly as in Likert Scale, but by a panel of judges. It follows a detailed procedure,
which is summarised as under:
a) Large number of statements are gathered expressing various points of view
about a phenomenon.
b) These statements are submitted to a panel of judges. They examine the
statements, identify which are relevant and which are not relevant. They
select 11 most relevant statements and arrange them in order of being
favourable.
c) Based on selection and sorting by the judge, each selected statement is
assigned a unique „median value‟ between 01 to 11.
d) Judges make a final selection of statements whose median scores are spread
equally from one extreme to another extreme. These statements constitute
the final scale to be administered to the respondents for the research study.
Like the Likert type scale, respondents are asked to identify the statement to which
they agree. Then median value of each statement is used as the score.
Page No. 26
UNIT 4.DATA COLLETION METHODS
Data collected for business research may generally be of two types: primary and
secondary data. Primary data are collected fresh by the researcher, whereas the
secondary data has already been collected by others. Methods of collecting these
two types of data differ because of the nature of these two types of data. Researcher
has to personally identify sources of primary data, make an access to them for
obtaining the data, and use appropriate methods for collecting the desired data.
Whereas for secondary data, researcher has to access the source where such data
collected earlier has been kept/stored. The researcher has to search and find out the
data relevant for his study.
Page No. 27
4.2 Observation Method
In this method, the researcher himself observes the business or social phenomena
himself either through participation as a member of the group, or by observing in a
detached manner without involving himself in the phenomena. This is one of the
traditional and simple methods, but suffers from the criticism of being subjective by
nature. Here the researcher should observe a particular phenomenon in a realistic
situation, by personally visiting the particular place at particular time.
To serve the purpose of a scientific research method, it must meet two conditions:
(a) it should serve a formulated scientific purpose, and (b) it should be applied as a
systematically planned and recorded method, and should be subjected to checks
and controls of reliability and validity. In this method, respondent is not giving
response as the information is being directly collected by the researcher who acts as
the observer. Therefore,respondent’s biases do not enter the research data.
However, it has certain limitations. This method provides very limited information and
is time consuming as well as very expensive.
Page No. 28
In non-participant observation method, the researcher observes the situation and the
phenomena in a detached manner without involving himself as member of the group.
The information recording is carefully planned. The observer watches the
phenomena „from an outside position‟ and thus ensures that observations involve
lesser degree of personal bias and subjectivity. One limitation of this method is that
the members of the social group,under study, are aware that they are being watched
and their behaviour is being recorded. Being aware, they do not give their natural
behaviour, thus bring error in the observation.
In this method, the researcher does not record the observations in real-time as and
when these are occurring, but subsequently by gathering information from sources
who may have been part of the event, or phenomena. As the name suggests, this
method relates to extensive study of a case or an event, or a process that has
occurred in the past. The case study is basically an „in-depth study‟ of a particular
„case‟ involving a unit or an organisation. As the name suggests, case study refers to
intensive examination and analysis of a single unit, or number of parts of one single
unit.
Several methods are common for data collection under case study method of
business research. Most common method is holding extensive discussion and
interviews with the personnel of the „case-unit‟ to seek all information in a sequential
manner on time scale i.e. covering both the past information and the information
about the current status. All events, activities, behaviours, conflicts, and benefits
occurring to the „case-unit‟ are discussed first in unstructured manner and then
followed by planned and structured discussions and interviews.
Page No. 29
The „behaviour interview‟ starts with general questions and gradually focuses on the
areas of interest, narrowing the questions to obtain more and more specific
information. Here the data is collected in chronological order. The investigator has to
selectively collect pertinent information according to the problem under investigation.
The next important task is organise, summarize and interrelate the information
collected for finding explanation regarding past sequence of events. Final phase of
actions relates to identification and diagnosis of causal factors as a basis for
remedial or development of treatment strategy.
In this method a list of persons, who have the first-hand information regarding the
research subject (event, attribute, behaviour, or phenomena) is prepared. Number of
questions are also prepared for seeking information from the selected respondents.
Further, the „situation‟ for research and the set of questions are explained to certain
selected persons, who may act as the „interviewers‟ for gathering the research data.
Page No. 30
higher interviewing skills on part of the interviewers. However, use of structured
interview facilitate economy of operation and also save time.
Interview method has some merits over observation method and the case study
method. It facilitates collection or more information, and even more in-depth
information. A skilled interviewer can win support of the respondent and may be able
to overcome his initial resistance. The interviewer has greater flexibility as compared
to observation method and case study method. The language and personal style of
the interviewer can be changed to match with educational and social background of
the respondent, thus facilitating better information, both in terms of quantum and
quality of information.
Page No. 31
can get a feeling of impersonal interview being forced on him or her. It has major
advantages of being cheaper, easy and quick method of collecting desired
information. Replies can be recorded without making the respondent aware about it.
However, limited time is given to respondent for answering the questions. There are
greater chances of bias on the part of respondent, who may not be giving compete
and true information.
In this method, the researcher carries out detailed study of all issues relating to the
research problem and prepares a detailed questionnaire covering all aspects of the
problem to be investigated. The questionnaire is sent to the respondents by post with
a request to answer to various questions given in the questionnaire. Generally, a
„self-addressed stamped envelope‟ is also enclosed for ease of sending the
response without any cost to the respondent.In current time when internet facility has
become common in most countries, the questionnaire can also be sent by e-mail.
Page No. 32
reasonable confidence is achieved that respondents will be able to understand the
questions in similar meaning as understood by the researcher.
The advantages of this method is its low cost when the universe is large and
widely spread. Further,it is free from the biases of the interviewer. The
respondent replies in his own language reducing the inaccuracy in answering. But it
has a problem that rate of replies is very low. Further, it is useful only if the
respondents are educated and are likely to be cooperative for the research.
Page No. 33
h) There should be some control questions to check the reliability of answers
provided by respondent.
The qualitative methods offer economy in both cost and time, it provides rich
data and preliminary insights in to a phenomenon. However, the common
problem with the data collected by using these methods is that it lacks the ability to
get generalized, reliability and validity. Some of the common qualitative methods
employed by the researchers are described below.
Page No. 34
The researcher can determine respondents‟ motivations and resistance
towards a particular product, process or event
Useful in case of complicated and comparatively new research
subjects about which adequate information is not available
Ability of the interviewer helps him to develop mutual trust with the
respondents and this helps in collecting most personal information also
The reason behind selecting participants from the same kind of age, income, and
gender and occupation etc. background is that, it helps them to get comfortable with
each other. It has been observed that participants from similar background feel more
comfortable in expressing their opinions. However, if a researcher is aiming for a
wider perspective, a diverse group of participants with different demographic
characteristics can also be selected. This is an extremely important issue as it is
hard to control group dynamics when more than 12 people are involved in a
discussion.
Moderator as discussed above is the leader who guides this group of participants.
Moderator initiates the discussion by introducing the discussion topic and
Page No. 35
encourages every participant to discuss and debate. The role of the moderator is not
to intervene or influence the opinion of participants, he is supposed to have some
control over the discussion and to ensure that the participants don‟t drift away from
the research objectives.
The main aim of any focus group is to give adequate information to the decision
maker regarding the issue at hand. The group dynamics is the crucial feature of any
focus group discussions. The success factors for focus group discussions are:
Group dynamics;
Members‟ willingness to contribute; and
Ability of the moderator to keep the discussion relevant to the research
objectives.
There are several variations possible in focus group discussion such as small groups
or larger group size, use of single or multiple moderators, conducting the interview in
neutral setting or direct organizational involvement.
There are several advantages of focus group technique. The major advantage of
focus group is the quality of information provided by them.
The moderator can use focus group to support his or her own pre-
conceived notions and he may guide the discussions in that direction
At times this technique is not specific and it tries to capture too many
things which may dilute the information collected
As the focus group sample is drawn based on convenience and it is
smaller in size, the results of focus group research can‟t be
generalized.
Moderator‟s biases and limitations will influence the overall
effectiveness of the research results.
Page No. 36
The selection and recruitment of focus group participants is also very
challenging, especially if research requires to form a large number of
groups.
Projective Techniques
Page No. 37
UNIT 5: DEVELOPMENT OF RESEARCH DESIGN/PLAN
Research design refers to description of the plan for collection of required research
data,and further analysis of data and interpretation of results. It describes the
research planning, and step-by-step methodology of research operations. In
other words, it presents detailed planning for collection, measurement and analysis
of data.
Page No. 38
5.2 Purpose of a Research Design
The purpose and benefits of evolving a research design before actually starting the
business research are mentioned below:
Every major and complex task needs advance planning regarding its
activities, the processes & strategies to be followed for completion of task by
consuming minimum amount of resources.
Secondly, research design answers questions regarding the research issues
namely objectivity, validity, accuracy and empirical evidences.
It also provides information regarding methods used for finding answers to the
research questions.
The research problem can be stated in terms of research hypothesis. Another
objective of the research design is to control variance.
Research design has significant bearing on reliability of the results achieved.
Such a design, prepared before commencing the research, helps the
researcher in organising his/her efforts and thereby identifying and removing
the flaws and inadequacies,if any.
A good research design basically has (i) adequate degree of flexibility, (ii) proper
selection of research techniques, (iii) and minimum bias, but (iv) maximum reliability
of data.
Page No. 39
(f) It brings out reasonably accurate estimates for resources required for the
study, including time, money, persons, and logistics requirements;and
(g) It provides for specific research design depending on type of research method
and the particular data collection methodologyused.
(h) It gives due weightage to describing particular research design such as
survey design, exploratory design and sample design etc.
A research study has to be so designed that the effect on the dependent variables of
the study is attributed only by the independent variables of the study, and the effect
of extraneous variables is eliminated. In a real situation, if elimination is not possible
then effect of extraneous variables should be minimised.
Control:It refers to the extent by which the effect of extraneous variables can be
controlled during the study. Ideally, the extraneous variables should not affect the
values taken by the dependent variables. But in reality, this may not always be
possible. Therefore, the attribute ‘control’ refers to the degree by which the effect of
extraneous variables can be reduced or minimised during the research study.
Page No. 40
Research Questions& Research Hypothesis:The objective of the research study
is to investigate inter relationships between the research variables. This purpose of
research enquiry can also be stated in terms of question framed regarding effect of
certain parameters on other parameters. For example, the study may involve
studying effect of „rise in average environmental temperature‟ on „availability of water
in the rivers originating from mountains‟.
The above question can also be transformed into relational statement expressed in
terms of causal effect of change in the independent variable on the value taken by
the dependent variables. Then such a relational statement can be tested, during the
research’ for being true or false. Such relational statements are called hypothesis
(meaning assumed relationship).
The hypothesis for the above example can be:“Higher the „average environment
temperature‟, lesser is the „availability of water originating from mountains‟”.
Based on the data collected during the research, effects of change in independent
variables on the dependent variables are studied and analysed. This analytic
information is then used to establish whether the proposed/hypothesised relationship
between independent variable „average environment temperature‟ and the
dependent variable „availability of water originating from mountains‟ is „true‟ or „false‟.
A hypothesis is a proposition which the researcher wants to verify. Often there may
be several competing hypotheses, either specified or implied. One objective of
research is to select among the possible hypotheses and to test them empirically
with the help of statistical tools in order to ascertain whether they are true or false.
Page No. 41
Hypothesis formulation deals with assumptions made by the researchers. There are
two kind of hypotheses: Null and Alternative hypothesis. Null hypothesis is an
assumption that there is no difference between the sample and population
phenomena. An alternative hypothesis states that the difference between the sample
and population phenomena is statistically significant.
The symbolused for null hypothesis is Ho and for alternative hypothesis is denoted
by Ha. The hypotheses are always formulated for population parameters or
characteristics rather than sample parameters. Null hypothesis always states that
there is “no significant difference.” It means that there is no significant difference
between the population mean and the sample mean. Whereas, alternative
hypothesis states that the difference is real.
Research design has to be evolved for the specific research technique/method used
in the study. There can be three different type of research situation depending on the
nature of the study. For example, aim of some study may be: (i) exploratory research
i.e. to explore the unknown aspects of the research situation, or (ii) descriptive and
diagnostic research, or (iii) study of causal relationship among variables based on
specifically framed research hypothesis. For three different objectives, different type
of research design is required.
Exploratory research studies are conducted when deep insights are not available
regarding the main problem of research interest. The problem, as such, has not been
studied and even the variables have not been established earlier. The proposed
study aims to explore into the unknown aspects of the research problem. Such a
study provides a foundation for further empirical study at a later stage of time.
The exploratory research design is very flexible for consideration and studying
different aspects of the specified problem. It generally has three stages of
investigations:
Page No. 42
extensive study of related literature,
discussions or interview of persons having sufficient knowledge and
experience about the research situation, and
synthesis and analysis information gathered from above two steps.
Study and survey of related published literature provides a very useful method for
gaining basic insights into the aspects of research problem. Views and theories
developed earlier by other persons are very useful for evolving framework for further
study and investigation.
Analysis of information collected from above described steps provides clearer picture
about the research problem and may lead to formulation of research hypothesis.
This can set the stage for conduct of empirical research study at a later stage.
Page No. 43
In both types of studies, researcher must clearly define the variables to be
measured, and also the measurement methods. The population for data collection
must also be identified properly. Further, the procedure to be used should be
carefully planned so that bias may be minimised and reliability be maximised.
In hypothesis-testing based research studies, the researcher has to collect data for
testing causal relationship among variables stated in terms of hypothesis.Here
procedure for reducing bias and increasing reliability are important for
success of the study.
Page No. 44
5.6 Formulation of Research Plan
Research design can be restated in simple terms to serve as the guidelines and
directions for the researcher for planning and carrying out the research study. Such a
systematic description of research steps which helps the researcher in
organising his ideas in an orderly manner in document format is called as the
research plan. The plan-document provides all information regarding what steps are
to be taken during the research work, and also the reason thereof.
‘Research plan document’ provides for the following information / guidance to the
researcher:
Problem definition
Need for the study
Relevance of the study to the business organization and/or to the society
Plan for conducting literature survey
Objectives of the research study
Research questions/hypothesis to be tested
Scope of the study
Conceptual model of the research problem (presenting the independent
and dependent variables for the study, and the hypothesis for testing)
Research methodology (presenting research method/techniques, sampling
design, data collection method, data processing plan, data analysis
techniques etc.)
Plan for presentation of results, and
Estimate for the resource required: timeframe, physical resources
required, and the funds requirement.
The researchers face different ethical issues while collecting and interpreting the
data. The most common of them is the issue of maintaining the confidentiality
while collecting the primary data from the respondents. However, the professional
research organizations have developed a serious code of conduct to ensure
Page No. 45
that the data collected by using survey methods will be maintained for its
confidentiality.
The researcher should take the consent of the author or the article writer while
dealing with their sensitive data from secondary sources. It is not always
possible for the researcher to obtain manyconsents; therefore, the researcher needs
to make a professional judgement regarding the re-use of the data. If it violates the
contract between the respondents and the primary researchers, then it should be
avoided.
Page No. 46
UNIT 6: PROCESSING AND ANALYSIS OF DATA
After collection of data, it has to be processed and analysed in accordance with the
procedure laid down in the research design.The processing of data refers to
converting data in a form that facilitates easy handling and analysis. The
processing operations include: editing, coding, classification, and tabulation.
6.1.1 Editing
The objective of this stage is to examine the raw data collected and to identify errors
and omissions made by the respondents in giving response, and also to incorporate
corrections where ever possible. Editing is carried out with a purpose to ensure
that the data is accurate, consistent with other facts gathered, uniformly
entered and as complete as possible.
Editing can be done at two occasions. Firstly, it can be done when the „data form‟
is being collected by the researcher or by the research staff, when the respondent
has not yet left. This is called fieldediting. Secondly, editing can be done when all
„response forms‟ are received in the central office, where the researcher or the
„research team staff‟ may carry out the editing operation. This is called centralediting.
Field Editing:The research staff while collecting the response form, can review the
entries made in the form for completeness of response, and also proper way of
Page No. 47
entries as stipulated in the response form. The field research staff should refrain
themselves from making any entry or correct the omission themselves.
Central Editing:It should be done when all „response forms‟, as completed by the
respondents, have been collected by the central office.Here, editing is done by a
single editor or by the same editing team. The editor(s) may correct only the obvious
errors like entry made at wrong place, wrong units written when correct unit arevery
obvious. The editor may strike out an obviously wrong reply when it is not possible to
guess the correct reply that the respondent would have given himself.
6.1.2 Coding
6.1.3 Classification
Many times, the large volume of raw data has to be reduced into homogenous
groups, if we are to get meaningful relationships.For this purpose, the data is
arranged in terms of groups or classes on the basis of common properties. Data with
one common property is kept in one particular class. In this way, the total data is
placed into number of groups or classes. Two types of classifications are common.
(a) Classification as per the Common Attributes: Such common attribute can
be of two types, namely descriptive property (like gender, literacy etc) and
numerical property/characteristics (like weight, income, etc). By selecting one
property or characteristic, data can be divided in two groups: one group
having the selected characteristics and other group not having it.
(b) Classification as per Class Intervals:Data with numerical properties, as
mentioned above,can be classified using class intervals of statistics of
Page No. 48
variables namely age, height, income etc).For example, if the basis of
classification is in terms of annual income, say in thousands of pounds, then
class intervals may be 10-20, 20-30, 30-40, 40-50, 50-60 etc. Data can then
be arranged as per such class intervals, i.e.grouping the data for a class
interval together as one group. Each group has a lower limit and one higher
limit; and interval; between the two limits is called range of classification.
6.1.4 Tabulation
For ease of handling the data, it is generally arranged in some concise and logical
order. Tabulation is one method for systematic and logical summarisation of
research data. Tabulation is an orderly arrangement of data in columns and rows.
Earlier tabulation was done manually, but now with easy availability of computers,
the tabulation of even large amount of data can be done quickly with the help of
computers.
Page No. 49
(b) Inferentialanalysis is also called as statistical analysis. Being inferential in
nature, it involves study of two variables at a time, and also study of more
than two variables.Such analysis generally involves study of two or more
variables using statistical techniques. Common types of inferential analysis
include bivariate and multivariate analysisnamely causal analysis, correlation
analysis, multiple regression analysis, and multiple discriminant analysis.
Advance research on complex issues involve other types of analysis namely
analysis of variance, and canonical analysis etc. Common types of inferential
analysis are described below.
Bivariate Correlation Analysis: The relation between two variables
can be studied through data collection in the group and analysing the
responses to investigate correlation i.e. relation between the two
variables. This can be done using statistical software package. The
data obtained from the members of the group is stored as two different
„arrays‟ corresponding to the two variables. By identifying the two
variables on the statistical software, „command‟ is given for computing
bivariate correlation between these variable.
Causal Relationship or Multiple Correlation Analysis:Investigation
regarding how one or more variables affect changes in another variable
is called causal analysis. In statistical term, this study is called as
multivariate analysis.In the statistical software, the data for the group
regarding various variables is stored as different „arrays’, and
„command‟ is given as per prescribed format for computing multiple
correlation. The results are displayed as the „output‟ provided in the
tabular format i.e. various scores on the output table indicate
correlation between the set of two particular variables.
Advanced Statistical Analysis: When relationship between more
than two variables (generally larger number of variables) is studied,
variety of advanced types of analysis are done using statistical
software packages.
Page No. 50
6.3 Data Analysis and Use of Statistics
As stated in the above sections, data analysis is generally done through use of two
types of analysis: descriptive analysis and inferential analysis.
Page No. 51
Range: It is the simple way of describing the dispersion defined by the
difference between the values of the extreme items of a variable.
Mean Deviation: It is the average of the „deviation‟ of each item, from
the mean of all variable values. If we compute the difference between
each item value and the mean value, then the sum of all such
„differences‟ so computed divided by number of items is the mean
deviation, with negative sign being ignored.
Standard Deviation:It is defined as the square-root of the average of
„squares of deviations‟, when deviations are the difference between
item value from the arithmetic average.
d) Measure of Relationship:
Simple Correlation:This is a popular method for determining degree
of relationship (correlation) between two variables. It is based on the
assumption that there is linear relationship between the two variables.
Partial Correlation:It measures relationship between two variables in
such a manner that the effect of all other variables has been
eliminated.
Simple Regression:Regression provides the relationship between two
variables in terms of a regression equation as given under:
Y=a+bX
Here „X‟ is the independent variable value; „b‟ is the coefficient of
correlation, and „a‟ is the constant of regression. Then „Y’ is the
estimated value of dependent variable Y.
MultipleRegression: When a dependent variable is function of two
independent variables, the analysis concerning the relationship is
called multiple correlation and the equation describing the relation is
called multiple regression equation.
The above analysis can be easily carried out using the statistical
software packages. Here values of the variables measured through data
collection is stored in terms of „arrays‟ for each variable, then appropriate
„command‟ is provided to the software, which provides the desired values as
the output printout.
Page No. 52
6.4 Statistical Software Packages
Further, another packages popular with new researchers is the SPSS Package
which “is a comprehensive software system designed to handle all steps in an
analysis from data listing, tabulation, and descriptive statistics to complex statistical
analysis”. SPSS is a registered trademark of SPSS Inc, USA. Many other statistical
software packages are also available in the market for use by statistics experts. But
the above two software are most common among the new researchers.
Page No. 53
UNIT 7: INTERPRETATION & PRESENTATION OF RESULTS
Page No. 54
Generally, the descriptive statistics regarding independent and dependent variables
is presented in tabular format, indicating the findings regarding range,minimum
value, maximum value, mean, median, mode and standard deviation. These are
interesting findings and should be presented while pointing implication of such
values on the research subject. Variable variance can also be presented in terms of
percentages of total responses received. Some researchers use pie diagrams to
present such variances. „High‟ and „low‟ values taken by some variables may be an
important issue in interpretation of findings.
The relationships among variables brought out through the study need to be
analysed in depth and their impact on the subject of study also needs to be
presented accordingly. Causal relations are the important findings in any research
study. Many-time, one variable appears to be influencing the other variable. But at
the same time, other variables may also the influencing the value of the variable
being studied. Therefore, careful analysis of multi-variable relationship is necessary
which should be interpreted very carefully.Partial correlation is very useful in which
correlation between two variables is studied while the effect of other variables is
eliminated. Relationship prediction is done using regression analysis. Therefore, in
a research involving number of research variables, relationship study should be
carried out using variety of methods, both for correlation and regression.
Page No. 55
have been achieved. Number of times, the results may not be on expected lines,
but still this presents an important finding and provides platform for conducting
further research on the subject matter.
As the next step, the procedure and findings are elaboratedfurther for better and
complete communication to the reader who may be conversant with overall subject
matter, but may not be an expert or authority on the subject. However, the write-up
must not be found to be wanting in presentation even by the experts. Finally,
bibliography and references are written and made part of the research report.
Page No. 56
(a) Introductory Write-up: The purpose of this part of the report is to make the
reader understand the background for research, reasons for carrying out the
research, previous knowledge and research on the subject matter. It covers
the followings:-
Title of the report
List of contents
General background
Need for this study
Objectives of the study
Scope of the study
Brief outline of research methodology
Relevance of the study
(b) Present research study – planning & preparation: Before starting the
research study, in its physical form, considerable planning and preparatory
work is required to be done, which starts from extensive literature review and
covers identification of variables and ending with finalising methodology for
data collection & measurement. This part of the research report covers the
following specific steps:-
Literature review
Conceptual model for the research study
Variables for the study,
Formulation of research questions and the hypotheses to be tested
Research methodology
Design of the questionnaire
Sampling design
Scaling & other techniques used for measurement of data
Page No. 57
Research findings and the interpretation of results
Page No. 58
The ordering in case of publication in a journal may be as under:
Authors‟ names
Journal‟s name within sign of quotations (in italics)
Year and Issue Number
Page Numbers from where matter is being quoted
(g) Annexures to the report: The text of questionnaire or schedule used for data
collection is to be placed as appendices to the report. It may also present the
details in terms of tabular format, pie diagram etc.
Generally, research reports are presented to the professional peers and/or to the
gathering of experts. As a modern trend, oral presentation is made by the
researcher while taking support from the computers. Such presentation is called
computer based oral presentation. The “power-point presentation software” available
as part of MS OFFICE software package is used.
The text regarding the research is typed in the power-point software, where facility
exists for showing tables, graphs, histograms, and pie diagrams. It facilitates making
diagrammatic presentation of data and results. Pictures stored at other place in the
computer can also be imported in the power-point presentation.
Page No. 59
Objectives & Scope of the Study
Research Evidences collected through Literature Survey
Research Variables for the Research Study
Research Questions and / or Hypothesis for the Study
Research Methodology covering Sampling & Scaling adopted
Data Collection & Analysis (in brief giving only the highlights)
Research Findings & Results
Implication of Results
Recommendations Emerging from the Research Findings & Results
Scope for Future Research
Page No. 60