RESEARCH DESIGN

MODULE II
 Important concepts relating to Research Design

 Variable : A concept which can take different

quantitative values is called a variable.
For example : weight ,height and income.

 In research, variable is a factor that can be

manipulated and measured.

 They take on different values which are

observed and measured to draw conclusion
about the subject under study.
Continuous variables : Phenomena which
can take on quantitatively different values
even in decimal points are called
‘continuous variables’.
Eg: Age

Discretevariables : Those variables which

can only be expressed in integer values
are called ‘discrete variables’ or non-
continuous variables.
Eg: Number of Children
 An independent variable is the variable
that is changed or controlled to determine
the value of the dependent variable.
 If one variable depends upon or is a
consequence of the other variable, it is
termed as a dependent variable.
 For example, height depends on age. Here,
age is the independent variable and height
is the dependent variable.
 We say, demand depends on price. In this
case, price is the independent variable and
demand is the dependent variable.
Extraneous variables : Independent
variables, that are not related to the
purpose of study, but may affect the
dependent variable are termed as
‘extraneous variables’.
For example, to test plant growth at
different temperature levels. The
dependent variable is plant growth,
independent variable is temperature
level and the soil type and amount of
water are the extraneous variables.
 Another example, to study the impact of
chocolates on the increase in weight of the
children.
 A researcher can study this by feeding
different quantities of chocolates to different
children and observing the increase in weight.
 Here, the quantity of chocolate in the diet of a
child is the independent variable, the weight
of the child a dependent variable.
 Other factors such as the original weight, the
living conditions of the child are termed as
extraneous variables.
Control : When the study is designed by
minimizing the extraneous independent
variables, the technical term ‘control’ is
used.

Research hypothesis : The research

hypothesis is a predictive statement that
relates an independent variable to a
dependent variable. When a prediction
is to be tested by scientific methods, it is
termed as research hypothesis.
 Hypothesis–testing research : When the
purpose of the research is to test a research
hypothesis, it is termed as hypothesis-
testing research.
 It can be of the experimental design or of
the non-experimental design.
 Research in which the independent variable
is manipulated is called the experimental
hypothesis-testing research and a research
in which the independent variable is not
manipulated is called the non-experimental
hypothesis-testing research.
 The independent variable varies between
conditions and the dependent variable is the
“outcome” variable ,it is the outcome that is being
measured.
 For example, Imagine you believe that high
communication skills enhances individual
confidence.
 Here, the independent variable is communication
skills and dependent variable is individual
confidence.
 For experimental testing, a sample of people are
selected and are divided into two groups: and one
group is given communication training and the
other group does not receive any.
 At the end of the training period, you can
see if there is a difference in individual
confidence between the group that
received the communication training and
those who didn't.
 In a non-experimental research, we just
separate them into two groups: high
communication skills and low
communication skills, and then you see if
there is a difference in the two groups in
terms of individual confidence.
Experimental and control groups: In an
experimental hypothesis-testing research
when a group is exposed to usual
conditions, it is termed a ‘control group’,
but when the group is exposed to some
special condition it is termed an
‘experimental group’.
Treatment : The different conditions
under which the experimental and
control groups are put are usually
referred to as ‘treatments’.
 Experiment : The process of examining the
truth of a statistical hypothesis related to
some research problem, is known as an
experiment. Experiments can be of two
types : absolute experiment and comparative
experiment.
 For example, to determine the impact of a
fertilizer on the yield of a crop ,it is absolute
experiment . If we want to determine the
impact of one fertilizer compared to some
other fertilizer, then it is a comparative
experiment.
 Types of Research Designs
 Based on the fundamental objectives or
purposes, research design can be
classified into:

1) Research design for Exploratory research

studies
2) Research design in case of Descriptive
and Diagnostic research studies
3) Research design in case of Experimental
Studies (hypothesis-testing research
studies
Research design in case of Exploratory
or Formulative research studies

 Applied when the researcher is not

acquainted with the problem.
 Aims at gaining familiarity with a
new phenomenon.
 A specific problem or hypothesis is
formulated from an operational
observation.
 Thus also called formulative studies.
The following methods are used
for exploratory research:

a) Survey of literature
b) Experienced survey
c) Analysis of insight stimulating
experiences
Research design in case of Descriptive
and Diagnostic research studies
 Aim is to obtain complete and
accurate information.
 Less biased and more reliable
information, with due concern for the
economical completion of the
research study.
 Firstly, the objective of study is
formulated.
 Selecting the sample.
 Designing the methods of data
collection.
 Processing and analysing the data.
 Reporting the findings.
Research design in case of experimental
or hypothesis-testing research studies

 Experiment is the procedure for gaining

knowledge by collecting observation
under controlled conditions.
 The researcher tests a causal
hypothesis.
 A causal hypothesis is one that states
the cause and effect relationship
between two or more variables.
 Explains the structure of an
experiment.
 Involve plans for testing of the causal
hypothesis.
 Decides the number of observations to
be taken and also the order in which
experiments are to be carried out.
 Which mathematical model to be used
for explaining the experiment.
 Basic Principles of Experimental
Research Design

Principle of Replication
Principle of Randomization
Principle of Locus Control
 Principle of Replication – The experiment is
repeated, treatment is applied in many
experimental units. It is done in order to
increase the accuracy of the experiment.
 Principle of Randomization – It considers
the extraneous variables that may occur in
the experiment and can control the
systematic error.
 Principle of Locus Control – The extraneous
variable is made to vary deliberately
and ,such that it can be calculated and
eliminated from the systematic error.
 Procedure of Experimental Research
design
Experimental Research design involves
the following steps:
1) Construction of experimental and
control groups
2) Pre-test
3) Exposure of the experimental groups to
the experimental variable
4) Post-test
 Classification of Experimental Design

The Experimental Design can be classified

into broad categories namely:
a) Informal Experiment designs
b) Formal Experiment designs
Types of Formal Experimental Design
The formal experimental design can be
further classified into:
i. Completely Randomized design
ii. Randomized block design
iii. Latin square design
iv. Factorial design
i. Completely Randomized design (C.R.
design)

 Used when a researcher is investigating the

effect of one independent variable on the
dependent variable.
 Considering different categories of the
independent variable, each category forms
the treatment.
 But, the extraneous variables that could
influence the outcome are not considered
in this design.
 Example, sales (dependent variable)of a
product is influenced by the price level
(independent variable).
 Considering three levels of price high,
medium and low, the most effective price
levels to be determined.
 The test units being the stores, which are
randomly assigned to the three treatment
levels, the average sales for each price level
is computed and examined to calculate the
significance difference in sales at different
price-levels.
The influence of extraneous variables is
not considered, here being the size of
the store, competitor’s pricing and so
on.
This design is very simple and
inexpensive to conduct.
ii. Randomized block design

 In this design, the influence of

extraneous variable is also
considered.
 In the example discussed, considering
three levels of pricing(high, medium
and low) and dividing the stores into
low, medium and high, the extraneous
factor is limited.
 The size of the store is treated as blocks and
each treatment appears in each block atleast
once.
 In this experiment ,one might randomly assign
12 small-sized stores to three price levels in
such a way that there are four stores to three
price-levels. Similarly applied to 12 medium
and 12 large-sized stores as well.
 Then the technique of analysis of variance
could be employed to analyse the effect of
treatment on the dependent variable and
separate out the influence of the extraneous
variable from the experiment.
iii. Latin square design

 To separate out the influence of two

extraneous variables.
 In the example besides the store-size
another extraneous variable can be type of
packaging.
 Each treatment occurs only once to both the
extraneous variables whose effect is to be
controlled and measured.
 Complex setup and quite expensive.
iv. Factorial design

 To measure the effect of two or more

independent variables at various
levels.(eg,price levels and store-size).
 Different combinations of the
independent variables are formed on
random basis and the interaction
effect is studied.
Types of Informal Experimental Design

 The informal experimental design is

classified into:
i. After only design
ii. After only with control design
iii. Before and after without control design
iv. Before and after with control design
v. Ex post facto design
i. After only design
 This design consists of applying the
experimental variable to an experimental
group and measuring the dependent
variable after the application of the
experimental variable.
 If a manufacturer wants to know the impact
of advertisement on sales, he can compare
the sales of the product after and before
the advertisement.
 Not fully reliable, because sales can
improve due to other factors as well.
ii After-only with control design
 In this design two groups (test group and
control group) are selected and the
treatment is introduced in the test area only.
 The dependent variable is measured in both
the areas at the same time.
 Treatment impact is assessed by subtracting
the value of dependent variable in the
control area from its value in the test area.
 The assumption is that the two groups are
exactly similar in all respects.
iii Before and after without control design
 A single test group is selected and the
dependent variable is measured before the
introduction of the treatment.
 The treatment is then introduced and
dependent variable is measured after the
treatment has been introduced.
iv. Before and after with control design
 In this design, two areas are selected and
dependent variable is measured in both the
areas for an identical time-period before the
treatment.
 The treatment is then introduced into the test
area only, and the dependent variable is
measured in both for an identical time period
after the introduction of the treatment.
 The treatment effect is determined by
subtracting the change in the dependent
variable in the control area(Z-A) from the
change in the dependent variable in the test
area(Y-X) i.e., (Y-X) – (Z-A).
v. Expost Facto Design
 In this design, the researcher is required to
select two or more groups for his study.
 The event has already taken place in one
group and the other group is free from the
happening of the event.
 Example, to study the impact of metro-rail
in a city.
 For this study, the researcher has to
depend on the past experience.
 SAMPLING DESIGN
Population
Element
Census
Sample
Sampling Unit
Sampling design
Law of statistical regularity
Sampling in real life
Characteristics of a good sample design

True representation of the

population
Small sampling error
Economical
Less time and effort
 Steps in a Sample design
i. Type of Universe : First step is to clearly
define the set of objects to be studied. They
can be finite (eg, population of a city) and
infinite (eg,number of stars in the sky).
ii. Sampling unit : It may be a geographical
area (state, district, village), social unit
(family, club, school) or an individual.
iii. Source list : Also called sampling frame,
from which sample is to be drawn. It
contains the names of all items of a finite
universe and such a list has to be correct,
reliable and appropriate.
Size of sample: This refers to the number of
items to be selected from the universe to
constitute a sample. It should not be too
large nor too small, but optimum.
Parameters of interest : The parameters of
interest in a research study has to be taken
care of while deciding the size of the sample.
Budgetary constraints : Cost considerations
from the practical point of view, have a
major impact upon the size as well as the
type of sample.
Sampling procedure : Finally, the
researcher must decide about the
technique to be used in selecting the
items for the sample. Obviously, a
design that results in a smaller
sampling error is always chosen by the
researcher.
 TYPES OF SAMPLING
Probability sampling
Non-probability sampling
PROBABILITY SAMPLING

Probability sampling is
a sampling technique wherein the
samples are gathered in a process
that gives all the individuals in the
population equal chances of being
selected.
1) Simple Random Sampling

 Simple Random Sampling is a probability

sampling design where every element of
the population has both known and equal
chance of being selected in the sample.

 They can be of two types:

 Simple Random Sampling with Replacement

 Simple Random Sampling without
Replacement
2) Complex Random Sampling Designs

Systematic sampling
Stratified sampling
Cluster sampling
Area sampling
SYSTEMATIC SAMPLING

Selects the ‘k’ th item in a list, which is

selected randomly.
Overcomes the limitation of the simple
random sampling that the sample may not be
a representative one.
Used in those cases where list of the
population from which the sample is to be
drawn is available.

Firstly, the sampling interval (K) is calculated
using the formula , K = N/n, where N= the size
of the population and n = size of the sample.
Then a random number (C) is selected from 1
to K.
Thus the first element to be selected from the
ordered population would be C, the next
element would be C+K, then C+2K and so on
till a sample size ‘n’ is selected.
Easier, less costly and more convenient to use.
STRATIFIED SAMPLING
When the population is not homogeneous
and contains variety of items, this type of
sampling is selected.
The entire population is divided into
strata(groups),which are mutually exclusive
and collectively exhaustive.
Mutually exclusive means one element can
belong to only one strata and collectively
exhaustive in the sense if all the elements of
various strata put together completely cover
all the elements of the population.
How to form strata?
How should items be selected from each
stratum?
How to determine the sample size of each
stratum?
CLUSTER SAMPLING
The total population is divided into a number
of small units called clusters.
The elements within the clusters are
heterogeneous but there exists homogeneity
among the clusters.
Some of these clusters are then randomly
selected for sample formation.
Can be adopted when there would either be
no reliable list of the units of the population
or it would be very expensive to collect such a
list.
AREA SAMPLING
If cluster happens to be some geographic
subdivisions, cluster sampling is referred to as
Area sampling.
Non-Probability Sampling Designs
Also called deliberate sampling or
purposive sampling.
Rather than probability techniques, the
personal choice of sample selection of
the researcher takes place.
Convenience Sampling

When population elements are

selected for inclusion in the sample
based on the ease of access, it can be
called convenience sampling.
Used to obtain information quickly and
inexpensively.
At times this procedure may give very
biased results particularly when the
population is not homogeneous.
Judgemental Sampling
The researcher’s judgement is used for
selecting items which he considers as
representative of the population.

Efforts are taken to locate and gain access

to the individuals who have the required
information.
Snowball Sampling
Used when it is difficult to identify the
members of the desired population
e.g., deep-sea divers, families with
triplets and so on.
The challenge is making the initial
contact, each respondent is then asked
to identify one or more in the field.
Difficult to identify a representative
sample.
Quota Sampling
The sample includes a minimum
number from each specified subgroup.
The sample is selected on the basis of
certain demographic characteristics
such as age, gender, occupation, income
and so on.
The researcher is asked to choose a
sample that conforms to these
parameters.
Sampling Error :
o This error arises when a sample is not
representative of the population.

o Sampling error reduces with the

increase in the sample size.
Non-Sampling Error :
o Error that occurs at stages ither than the
sampling stage.
o Wrong information
o Error that can occur while transferring
data from the questionnaire to the
system.
o During analysis stage.
o Non-availability of the respondent.
 Indian industries in the post-globalization era, have
undergone a sea-change in their operation, selection,
transfer and acquisition of technology have become
the talk of the business today. Global
Competitiveness Report (GCR) 1998 has revealed
that India ranks 41 out of 56 countries against United
States as no.1. So it is worthwhile to measure the
strength and weakness of technological
competitiveness of Indian industries in this
liberalized, privatized and globalised environment.
The competitiveness of Indian industries is to be
compared with global players. In view of the above
facts, it is required to carry a study of technological
competitiveness of Indian industries.
 You, as a researcher are asked to
answer the following:-
1. What could be the objective of your
study?
2. What is sampling frame and how do
you make the sampling plan to make
the above study?
3. Mention few hypotheses that can be
tested out from the data of the study.