SEMESTER II

BUSINESS RESEARCH
METHODS
Session
Session 11-13
11-13 Sessions
Sessions14-17
14-17 Session
Session 18
18
Scales
Scales & Measurement––Attitude
& Measurement Attitude Survey
Survey Research––Data
Research DataCollection
Collection Data Preparation
Data Preparation
Measurement-
Measurement- ComparativeVVNon-
Comparative Non- Methods Questionnaire Design
Methods Questionnaire Design Coding,
Coding,Outliers,
Outliers,
comparative scales
comparative scales Qualitative
QualitativeVVQuantitative
QuantitativeMethods
Methods Missing Values
Missing Values
Reliability
Reliability&&Validity
ValidityTests
Tests Observation, Indepth Interviews,
Observation, Indepth Interviews,
Projective
ProjectiveTechniques
Techniques(Guest
(GuestSessions
Sessions
Sessions
Sessions10
10 from industry experts) Mid Sem
from industry experts) Mid Sem
Case
Case Analysis––CEC
Analysis CEC22
SPSS Session
Session 19-23
19-23
SPSS Multivariate
MultivariateTechniques
Techniques––
Introduction
IntroductionMultiple
Multiple
Sessions
Sessions 99SPSS
SPSSLabLab Regression
Analysis Regression -Testing
- Testingofof
Analysis & Interpretationofof
& Interpretation goodness
goodnessofoffitfitDummy
Dummy
output
output- -ANOVA
ANOVA STRUCTURE
STRUCTURE
Variables,
Variables, SPSSLab
SPSS LabAnalysis
Analysis
&&Interpretation
Interpretation
Business
BusinessResearch
Research Numerical
Numerical Case
CaseAnalysis
Analysis
Methods
Methods CEC
CEC 4 4
Semester
SemesterIIII
Session
Session 24-28
24-28
Multivariate
MultivariateTechniques
Techniques––
Interdependence
InterdependenceTechniques
Techniques
Factor
Factor Analysis - SPSSLab
Analysis - SPSS Lab
Analysis & Interpretation
Analysis & Interpretation
Numerical
NumericalCase
CaseAnalysis
Analysis

Session
Session 29-30
29-30
Session
Session11-4-4 CEC
CEC 5 Presentations- -
5 Presentations
Introduction
IntroductiontotoResearch-
Research-Meaning
Meaning Projects
Projects
&&Definition, Significance, Overview
Definition, Significance, Overview
ofofMethodology Session
Session 31-32
31-32
Methodology. .Categories
Categoriesofof Session Other
Research,
Research,Research
ResearchTypes
Types, , Session33
33 Other
Process Report Writing MultivariateTechniques
MultivariateTechniques––
Process- -Identification
IdentificationofofProblem
Problem Report Writing
Discriminant, Cluster
Stages of the Research Process- The Fourth Stage
Management
Dilemma
Defining Research
Problem
Research
Formulating Report
Research
Hypothesis

Data analysis
Developing Research & interpretation
Proposal

Research Data Refining

framework -Design And Preparation

IBS
Data Collection
 Session Plan
• The fourth stage of Research Process
• Experimental Research Design
• Causality
• Conditions for Causality
• Concepts
• Validity
• Factors affecting Internal Validity
• Factors affecting External Validity
• Classification of Experimental Designs
• Pre Experimental
• Quasi Experimental
• True Experimental
• Statistical (ANOVA & SPSS for all the four designs)
07/08/20 IBS 4
 Research Designs
• Exploratory (Already covered)
• Descriptive (Already Covered)
• Causal

07/08/20 5
IBS
 Experimental Research Designs

• Experiment : An experiment is generally used to infer

a causality.
• Researcher actively manipulates one or more causal
variables and measures their effects on the
dependent variable of interest.
• The relationship between cause and effect often
tends to be probabilistic in nature
• One can infer cause-effect relationship.

07/08/20 6
IBS
 Causality Contd….
• Training given to sales personnel causes an increase
in sales in the territory by 20%
• Cause and effect?
• Can it be concluded that because of the new training
program sales has increased?
• But the increase in sales is due to new training
program alone?
• It could occur because of host of factors like:
– Reduction in the price of product
– A strike at competitor’s plant
– Price of competitor’s product
– Weather conditions

07/08/20 IBS 7
 Conditions for making causal
Inferences
1. Concomitant Variation – Extent to which cause X and effect Y occur
together or vary together
Strong association between training program and increased sales
Strong association does not imply causality .
Change may be due to presence of extraneous variable
affecting both the variables or as a result of random variations.
(May be)
2. Time Order of Occurrence of Variables- Causal variable must occur
prior to or simultaneously with effect variable.
Mere prior occurrence would not mean causality again
3. Absence of other possible causal factors- A researcher to enable
smooth study of variations need to block/control extraneous
variables . In case of sales training program all other variables
have to be kept constant to study the influence of training
program on sales.

07/08/20 IBS 8
 Experiments – A tool for Causal
Inferences
Experiments are used to seek help in identifying a
cause-effect relationship. The objective of
experiment is to measure the influence of the
independent variable while keeping the effect of
other extraneous variables constant.
Experiments are used to arrive at conclusive
answers in some of the following situations.
Can a change in the package design of a product
enhance its sales?
Does mentoring help in acclimatizing a person to
the organizational culture?

07/08/20 IBS 9
 Concepts used in Experiments
• Independent Variables : Explanatory variables or treatments. The levels
of these variables are manipulated/changed by researchers to measure
their effect on the dependent variable. In sales training example training
is independent variable
• Test Units- Test units are those entities on which treatments are applied.
Effect of treatment on test units are measured. The test units are
individuals, organizations etc. In sales training example, sales personnel
sent for training program are the test units.
• Dependent Variables- These variables measures the effect of treatments
(independent variables) on the test units. Sales in the example is the
dependent variable.
• Experiment – An experiment is executed when the researcher
manipulates one or more independent variables and measures their
effect on the dependent variables while controlling the effect of
extraneous variable
• Extraneous Variables- These are the variables other than the
independent variables which influence the response of test units to
treatments. Examples could be weather condition, a strike at
competitor’s plant etc..
IBS 10
07/08/20
 Validity in Experiments – Internal &
External
• Specify the following:
– Treatments to be manipulated
– Test Units to be used
– Dependent variables to be measured
– Procedure for dealing with extraneous variables
– The researcher has two goals while conducting an
experiment:
•To draw valid conclusions about the effect of
treatments on the dependent variables ( Internal
Validity)
•To make generalizations about the results of a
wider population. (External Validity)

IBS 11
07/08/20
 Internal Validity & External
Validity
• Internal validity tries to examine whether
observed effect on a dependent variable is
actually caused by the treatment
(independent variable) in question .
• Without internal validity the experiment gets
confounded.
• External validity refers to generalisation of the
results of an experiment. (for specific
population)

07/08/20 IBS 12
 Experimental Groups &
Control Groups
• Experimental group is a group of test units exposed to change
in independent variable
• Control group is a group of test units not exposed to change in
independent variable
To study the impact of package design on sales of a product
…..
A group of supermarkets (Experimental group) is selected &
each package design is displayed for a month
Another group of supermarkets (Control group) continue to
carry regular package design for that period
Sales of product are measured for both groups & difference in
sales is analysed to find out impact of design on sales

07/08/20 IBS 13
 Notations
• X: Treatment or manipulation of independent variable
• O: Observation or measurement of dependent
variable
• R: Random assignment of test units to experimental
groups
• EG : Experimental group exposed to experimental
treatment
• CG :Control group not exposed to experimental
treatment
• Notations in the diagram are read from left to right

07/08/20 IBS 14
 Arrangements
• Movement from left to right indicates
movement through time
• Horizontal alignment of symbols implies that
all those symbols refer to specific treatment
group
• Vertical alignment of symbol implies that
those symbols refer to activities or events that
occur simultaneously

07/08/20 IBS 15
 Arrangements….
• X O1 O2 means that a given group of test units
was exposed to treatment variable X and the
response was measured at two different points in
time ,O1 and O2
• O1 X O2 – Explain?
• R X O
means that a group of test units were randomly
assigned to a treatment group

07/08/20 IBS 16
 Factors affecting internal validity of the
experiment- Extraneous/lurking

• History
• Maturation
• Testing
• Instrumentation
• Statistical Regression
• Selection
• Mortality

07/08/20 IBS 17
 Extraneous/lurking Factors
• History - Specific events in the environment between
the Before and After measurement that are historic in
nature. Happens due to time difference
• O1 X O2
• O2 – O1 = Treatment Effect (Because of IV)
• But the effect could be due to the some happening
during the time gap. (History) Higher the time
difference between O2 and O1 , higher are chances
of History confounding the experiment.
• O1- Sales before training program
• O2- Sales after training program
• Between the two there is a possibility of economic
condition improving due to which sales may have
registered an increase
07/08/20 18
IBS
 Extraneous/lurking Factors
• Maturation – Similar to history except that it is concerned with
changes in a test unit due to maturity. Subjects change during
the course of the experiment due to passage of time. People
becoming older, more experienced, tired or uninterested .
• O1 X O2
• O2 – O1 = Treatment Effect
• But the effect could also be due to experience because of the
time gap. Longer the time difference between O2 and O1 ,
higher are chances of Maturation confounding the experiment.
• In sales training example, sales people might have gained
maturity with passage of time.

07/08/20 19
IBS
 Extraneous/lurking Factors
• Testing- It is concerned with possible effect on experiment of
taking a measurement on the dependent variable before
presentation of treatment. The Before measure alerts or
sensitizes subject to nature of experiment for second measure.
The first observation influences the second one.
O1 X O2
In sales training example if sales people know that they are
being sent for training to know its effectiveness, they would
become sensitized and behave differently.
For example, an experiment is conducted to study the effect of
advertisement on attitude towards a certain brand. A prior
measure was taken before the commercial was aired embedded
in an appropriate programme. Also post measure was taken.
The prior measure sensitized and may influence the post
measure thus confounding the actual effect of ad.

07/08/20 20
IBS
 Extraneous/lurking Factors
.
•Instrumentation – Refers to changes in
measuring instrument used for taking an
observation.
Difference in rupee sales before and after the
training program is used to measure the
effectiveness of the training program, a price
difference during the time interval could make a
substantial difference in the inference. Another
example could be presenting the pre and post test
questionnaire in a different fashion

07/08/20 21
IBS
 Extraneous/lurking Factors
• Statistical Regression – Occurs when the test
units with extreme scores are chosen for
exposure to the treatment (Extreme
favourable/unfavourable)
• In sales training example, extremely poor
performers are sent to training , then change in
sales will be more , variation would be more.
• Random occurrences like weather etc may also
influence the dependent variable.

07/08/20 22
IBS
 Extraneous/lurking Factors
• Selection Bias– test units selected for
experimental group differ from those assigned
to control group. Improper assignments of test
units to treatments result in selection bias.
This happens when it is done on non random
basis.
• If it is not random, subjective bias creeps in.
• The test units chosen for studying the effect of
a commercial during a sports tournament
telecast were different, the pre test units were
youth and post test units were of older age
groups.
07/08/20 23
IBS
 Extraneous/lurking Factors
• Test Unit Mortality -. Some of the test
units might drop out from the
experiment while it is in progress-
Sample attrition
• In case of sales training program some
sales people may quit job before
completing the training .

07/08/20 IBS 24
 Examples …
• Extra classes for non mathematics students
...Experiment/IV/Treatment
• Test units- Sample taken from class
• DV-Performance in test (pre /post)
• O1 = 24/50 O2= 35/50 (Average)
• Difference is 11
• Is this difference due to classes or there are EV ?
• What are those which disturbs the internal validity ?
• History: During the pre and post measure, festivals were there ,
students after the festival were better prepared but before they
were planning to go home town etc.
• Maturation: Students became matured, got acclimatized to
evaluation patterns.

07/08/20 IBS 25
 Some Examples…contd
• Instrumentation: The difference in pre test and post test (pattern
of exam changed from MCQ to subjective)
• Testing Effect: Students (Test units) because of pre test
measure became aware of the experiment, perform poorly post
test, they may not want to sit for extra classes. Or else perform
very well if they want to.
• Statistical regression: There could be a possibility, there are
extreme good performers or extremely bad ones in the sample
units chosen for study.
• Selection Bias: Sample selection is such that a researcher who
is in for supporting the extra classes may choose such a sample
so that the effect may seem attractive. He may choose very
poor performers , then the effect would be much more.
• Test Unit Mortality: Test units chosen for study, some of them
may leave the college
07/08/20 IBS 26
 Factors affecting external validity
• Experiments conducted in natural settings offer more
external validity as compared to experiments in
controlled conditions.
• Cinema Hall viewing V TV for the same advertisement
shown to prospective customers may have different
responses.
• Environmental factors influence results
• Populations used for experimentation of the test may
not be similar to the population where results of the
experiments are to be applied. Suppose a company
wants to launch an ice cream in Delhi NCR . The
results conducted in May-June cannot be generalised
for December January.
07/08/20 IBS 27
 Methods to Control Extraneous
Variables
• Randomization – Random assignment of test units to
experimental groups. (Which of the extraneous factor is
successfully blocked?)
• Matching- Match various groups by confounding variables . 120
people to be distributed in three groups, 45 of them are
females , each group should have 15 females. Effect of gender
can be distributed across the groups.
• Use of Experimental Designs- Some of the experimental
designs are very useful in eliminating the influence of
extraneous variables.
• Statistical Control- If all the above discussed methods of
eliminating effect of extraneous variables among the treatment
group fail, then some statistical methods (ANCOVA) can be
applied to control the effect.
07/08/20 IBS 28
 Experimental Environment
• Laboratory Environment: Experiments conducted in controlled
conditions
---Showing ads or products to select consumers in controlled
conditions
----blind taste tests
• Advantages-
Effects of extraneous variables can be minimised
History effects eliminated
Provides more internal validity
Time & cost effective
Reduces the risk of information about products & ideas
being passed on to competitors
• Disadvantages
Experiments conducted in artificial conditions & results may not
hold well in actual conditions
Less external validity
Results influenced by testing effects ,Test units aware of
being tested & may not respond naturally
07/08/20 IBS 29
 Field Environment

• Experiments conducted in natural settings

--Launching products in select regions
--Observing consumer behaviour regarding a POP
display in supermarkets
--Analysing customer response to trial offers
• Advantages
-high degree of external validity
• Disadvantages
-Low degree of internal validity
-Require greater time & efforts
07/08/20 IBS 30
 Laboratory Experiment Field Experiment

Artificial-Low Realism Natural-High Realism

Few Extraneous Many Extraneous

Variables Variables
High control Low control
Low Cost High Cost

Short Duration Long Duration

Subjects Aware of Subjects Unaware of
Participation Participation

07/08/20 IBS 31
 Experimental Design

Pre- Quasi- True Statistical

Experimental Experimental Experimental

One-Shot Case Time Series Pre-Test Post Completely

Study Test Control Gr Randomized

One Gr. Pre- Multi Time Post-Test Only Randomized

test-Post Test Series Control Group Block

Solomon Four Latin Square

Static Group
Group

Factorial

07/08/20 IBS 32
 Pre-Experimental Designs
They lack proper control mechanism to deal
with effects of extraneous variables
--- Randomization is absent
One Shot Design (After Only)
One Group Pretest-Posttest
Static Group Design

07/08/20 IBS 33
 One Shot Design (After Only)
• It involves exposing experimental group to treatment X after
which the measurement (O) of dependent variable is taken .
• : X O
• Only one test group is subjected to treatment X .
• To measure effectiveness of a test commercial
TV programme containing test Ad (X) is selected
Telephone interviews are conducted of respondents who
report having watched the TV programme previous night
Respondents are asked whether they recall having seen
the Ad. Measurements of Dependent variable are recalls
(O)
Recalls are compared with norm scores for interpretation
In sales training example, researcher selects sales personnel as
per his liking or convenience and examine the results

07/08/20 IBS 34
 One Group Pre test –Post test
• Also called before -after without control group design
• Exposing an experimental group to experimental treatment (X) and
taking measurements of dependent variable before(O1) and after(O2)
the treatment
• : O1 X O2
• Difference between O1 and O2 measures the impact of treatment on
dependent variable
• To measure effectiveness of a test commercial
Respondents are recruited on the basis of convenience
Administered a personal interview to measure attitude towards the
product
Then they watch a TV programme containing Commercial
After watching TV programme they are again administered a personal
interview to measure attitude towards product
Effectiveness of test commercial is measured as O2-O1
Influenced by maturation ,testing ,mortality, selection bias and
regression
07/08/20 35
IBS
 Static Group Design
• Two groups of test units are involved EG & CG
• EG is exposed to treatment CG is not exposed
• Measurements are taken for both groups after experiment.
• EG: X O1
• CG: O2
• Difference between O1& O2 is effect of treatment
• Exposure and experimental treatment not under the control
of the researcher.
• Segregation not made by the researcher
• To test effectiveness of TV commercial
Two groups of respondents are recruited
EG is exposed to TV programme containing Ad
Attitude towards product of both EG &CG are measured
Effectiveness of test Ad is measured as O1-O2
Presence of mortality
But testing effect is successfully blocked
07/08/20 IBS 36
 Quasi Experimental Designs
• Are used when randomisation is not possible & timing of
test presentation lacks control of researcher
• Not as effective as true designs but better than pre-design
• Prominent is Time Series Design
A series of measurements are taken on dependent variable
for a group of test units before & after the experimental
treatment
• O1,O2,O3 X O4, O5, O6
Trends before treatment are compared with trends after
treatment to determine effectiveness of treatment
• Aids in identifying permanent change & temporary change
• Because of multiple observations, effect of maturation,
main testing effect, statistical regression can be ruled out
• Threats on account of history

07/08/20 IBS 37
 Some examples…
• To find out the effect of promotional campaign
on the product sale ,purchases of panel
members before ,during and after the
campaign are examined to find out whether
there is a short term effect ,long term effect or
no effect

• To find out impact of price change on sales of

a product, a series of observations are taken
before the price change & after the price
change. Trends before treatment are
compared with trends after treatment to
determine change
07/08/20 IBS 38
 Possible Results of Time Series
Design under Three Different
Situations

07/08/20 IBS 39
 Multiple time-series design
• Similar to time series except that another group of
test units is added to serve as control group
• EG:O1,O2,O3 X O4,O5,O6
• CG:O`1,O`2,O`3 O`4,O`5,O`6
• The measurement on the sales for both the groups is
taken after the training programme. The treatment
effect (sales training in our example) is found by
comparing the average sales of the two groups
before and after the training programme
• Possibility of interactive effect in EG

07/08/20 IBS 40
 True Experimental Designs

In this type of design randomisation is used to

one or more control groups to reduce effect of
extraneous variables. Three variants :
Pre test – Post test Control Group Design
Post –test only Control Group Design
Solomon Four Group design

07/08/20 IBS 41
 Pre test-Post test Control Group Design
• Two groups of test units EG & CG are considered
• Test units are assigned to these 2 groups randomly
• Pre test measurements of dependent variable are
taken for both groups
• EG is exposed to treatment
• Post test measurements of dependent variable are
taken for both groups
• EG: R O1 X O2 (Presence of treatment &
Extraneous V)
• CG: R O3 O4 (Presence of only Extraneous V)
• Treatment Effect (TE) is (O2-O1)-(O4-O3)

07/08/20 IBS 42
 An Illustration….
To measure effectiveness of a TV advertisement:
Sample of respondents is selected at random
Half of these would be randomly assigned to EG & other half
would form CG
Both groups would be administered a questionnaire to obtain
pre-test measurement on attitude towards the product
Only EG would be exposed to TV programme showing
advertisement
Then respondents in both groups would be administered a
questionnaire to obtain post test measurements on attitude
towards product
Effectiveness measure is (O2-O1)-(O4-O3)
This design provides accurate results & addresses most of
extraneous variables except the testing effect.

07/08/20 IBS 43
 Post test only Control Group Design
Also referred to as after-only with one control group
Just as static group comparison except this design has
randomization.
Widely used in marketing research
Both EG & CG are formed by random assignment of test
units
EG is exposed to experimental treatment CG is not
Post test measurement of dependent variable is taken for
both the groups
EG: R X O1 (treatment & Extraneous V)
CG: R O2 (Extraneous V)
Treatment Effect (TE) equals O1- O2
Effect of instrumentation and testing effect is blocked.
07/08/20 IBS 44
 Examples….
To measure effectiveness of a TV advertisement-
Sample of respondents is selected at random
Half of these would be randomly assigned to EG & other half
would form CG
Only EG would be exposed to TV programme containing
commercial
Then respondents in both groups would be administered a
questionnaire to obtain post test measurements on attitude
towards product
Difference in attitudes of EG & CG is used as effectiveness
measure

07/08/20 IBS 45
 Solomon Four Group Design

• It involves 4 groups ,2 EGs & 2 CGs

• Six measurements are taken 2 pre test & 4 post test
• Also known as 4 group 6 study design
• EG: R O1 X O2
• CG: R O3 O4
• EG: R X O5
• CG: R O6
• Design provides various observations which can be
analysed
• O2-O1; O4-O3; O2-O4; O5-O6
• Addresses all extraneous variables but is expensive
& consumes time & efforts

07/08/20 IBS 46
 Solomon Four Group Design
• O5 – O6 , it has been proved to be the
treatment effect.
• This test gives the maximum internal
validity

07/08/20 IBS 47
 Statistical Designs
• Allow for statistical control and analysis of
external variables.
• The effect of more than one level of
independent variable on the dependent
variable can be manipulated.
• The effect of more than one independent
variable can be examined.
• The effect of specific extraneous variable can
be controlled.

07/08/20 IBS 48
 Completely Randomized Design
• Allows a researcher to investigate the effect of one independent
variable on the dependent variable. Normally independent
variable is required to be measured in nominal scale. (no of
categories) in ANOVA
• The basic assumption of this design is the similarity of the test
groups. Some illustration:
– Sale of a product is influenced by its price level , supposing
there are three price levels, low, medium and high.
– To find out at which price level the sale is highest some test
units are randomly chosen (stores)to the three treatment
level (price)
– The average sale for each price level is computed to see
whether there is any significant difference in the sale .
– The statistical technique to test such a difference is called
ANOVA.

07/08/20 IBS 49
 Working Illustration
• A company has three versions of
promotional campaign (AD), wanted to
understand the responsiveness of
these. It randomly selected 18
respondents, six of each shown the
first, second and the third version
respectively. The ratings (interval scale)
were obtained. The next slide shows the
data collected.

07/08/20 IBS 50
 The Ad version Illustration
S NO AD Version Rating
1 1 6
2 1 7
3 1 5
4 1 8
5 1 8
6 1 8
7 2 4
8 2 4
9 2 5
10 2 7
11 2 7
12 2 6
13 3 5
14 3 5
15 3 4
16 3 7
17 3 8
07/08/20 IBS 51
18 3 7
 Illustration continued…

• The code in the Ad represents different versions (the

three versions)
• The next column represents rating given by the
respondents after seeing the ad.
• Six respondents have rated each ad
• 18 respondents were randomly assigned each of the
three ad versions. Such assigning is called
completely randomized design

07/08/20 IBS 52
 ANOVA – One Way
Sources of Degrees SS MS F Sig/ p
Variation of value
Freedo
m
Treatments k – 1 TrSS TrSS/(k- TrSS/MS
1) E
Error k (n-1) SSE SSE/(k
(n-1))
Total kn -1 TSS

07/08/20 IBS 53
 SPSS Output

07/08/20 IBS 54
 Sig/p value
• The Sig/p-value is the level of marginal significance
within a statistical hypothesis test representing the
probability of the occurrence of a given event. The
Sig/p-value is the probability of obtaining the
observed results of a test, assuming that the null
hypothesis is correct.
• The Sig/p value is the probability that if null
hypothesis were true, sampling variation would
produce an estimate that is further away from the
hypothesised value than our data estimate.
• The Sig/p-value is used as an alternative to rejection
points to provide the smallest level of significance at
which the null hypothesis would be rejected is correct
07/08/20 IBS 55
 An example for understanding p value or sig
value
To analyse and study the performance of Div F in Sem I
end sem
Ho- Mean = 70
Ha- Mean > 70
S D = 6 -> S. E = 6/6 = 1
sample – 36 and mean of sample = 72
z = (72- 70) /S.E = 2
Taking right tailed test at 5 % LOS , table value of z is
1.65
Comparing it with calculated value and from the sketch,
it is clear that the value is falling on the rejection region,
thus H0 is rejected. It can be concluded that Div F has
outperformed .
07/08/20 IBS 56
 An example for understanding p value or
sig value

As the statistical software

and tools will be used,
there is a better way of
assessing , understanding
and analysing and
interpreting the results of
study.
Against the calculated
value of Z , find the area to
the right (in this case)
Against z = 2, the area =
0.5 - .4772 = .0228
0.0228 is called the p
value.
If p value is less than 5 %
which is the default sig
level, Ho is rejected.

07/08/20 IBS 57
 Illustration contd…Completely
Randomized Design - ANOVA
• Where the significance of the F test is given. It is .203
in this case, for the factor Ad_version. This indicates
that at the confidence level of 95 percent,
(corresponding to significance level of 0.05), the F-
test proves the model is not significant.
• In other words, the Ratings given to the three ad copy
versions are not significantly different from each
other.

07/08/20 IBS 58
 Illustration
•For example, the ratings for the ad copy version 1 are 6,7,5,8,8,8 and
the mean rating is (6+7+5+8+8+8) / 6, or 42/6 = 7. Similarly, the mean
rating of ad copy version 2 is (4+4+5+7+7+6) / 6, or 33/6 = 5.5. The
mean rating for ad copy version 3 is (5+5+4+7+8+7) / 6, or 36/6 = 6.
•At a glance, the three mean ratings appear to be different – 7, 5.5 and
6. But the ANOVA tells us that this difference is not statistically
significant at the 95 percent confidence level.
•It does this by performing an F-test. The null hypothesis for this F-test
is that there is no significant difference in the mean ratings for the
three ad copy versions. (H0: where 1, 2 and 3 are the
mean ratings for the three versions of ad copy). Thus, in this case, we
have accepted the null hypothesis (or failed to reject the null
hypothesis), at the 95 percent confidence level. Calculated value of F
is 1.78 while table value is 3.68.
• If the significance of F in the last column had been less than 0.05, we
would have rejected the null hypothesis. In that case, we would have
concluded that significant differences exist between mean ratings
given to the three ad copy versions.

07/08/20 IBS 59
 Class Exercise
• Cholesterol contents of the four competing diet foods on the basis of
the following data (in mgs per package) which were obtained for three
randomly taken 6 ounce packages of each of the diet foods,
measurement performed in three different labs. The first value of each
sample came from one laboratory, the second value came form another
laboratory and the third value from a third laboratory.

A 3.6 4.1 4
B 3.1 3.2 3.9
C 3.2 3.5 3.5
D 3.5 3.8 3.8

Test whether the difference among sample means can be

attributed to chance at 5 % level of significance?

07/08/20 IBS 60
 SPSS Lab
S NO CC Diet Food
1 3.6 1
2 4.1 1
3 4 1
4 3.1 2
5 3.2 2
6 3.9 2
7 3.2 3
8 3.5 3
9 3.5 3
10 3.5 4
11 3.8 4
12 3.8 4

07/08/20 IBS 61
Generate Output and interpret

• Class Exercise-

07/08/20 IBS 62
 Randomized Block Design
•An extension of the completely randomized design in which a
single extraneous variable that might affect test units’ response to
the treatment has been identified and the effects of this variable
are isolated by blocking out its effects.
•One extraneous variable can be blocked to provide a clear picture
of impact of treatment on test units
•In the price level example all the test units (stores) were assumed
to be more or less equal. However , all stores need not be of same
size .
•We can classify the stores as small, medium, large.
•In this design extraneous variable like stores size can be treated
as blocks.
•The treatments are randomly assigned to the blocks in such a
way, such that each treatment appears in each block at least once

07/08/20 IBS 63
 Randomised Block Design – The store
size is the blocking variable

Price Level Small Medium Large

Store Store Store
Low
Medium
High
It is assumed that blocking variable is correlated with dependent
variable (Sales) Blocking is done prior to the application of treatment

07/08/20 IBS 64
 ANOVA Table (Randomized
Block Design)
Sourc Degrees of SS MS F Sig
es of Freedom
Variat
ion
Treatm k -1 TrSS MSTr =
ents TrSS/(k-1) MSTr/MSE

Blocks n-1 SSB MSB = MSB/MSE

SSB/(n-1)
Error/ Bal Fig SSE MSE =
Residu SSE/Df
al Error
Total N-1 TSS

07/08/20 IBS 65
 Randomised Block Design
:

Fig. 3
sr. adcopy rating magazine
no.
1 1 6.00 1
2 1 7.00 2
3 1 5.00 3
4 1 8.00 4
5 1 8.00 5
6 1 8.00 6
7 2 4.00 1
8 2 4.00 2
9 2 5.00 3
10 2 7.00 4
11 2 7.00 5
12 2 6.00 6
13 3 5.00 1
14 3 5.00 2
15 3 4.00 3
16 3 7.00 4
17 3 8.00 5
18 3 7.00 6

07/08/20 IBS 66
 Illustration contd…
• Three versions of ad were used in 6 different
magazines. These 6 different magazines are
coded from 1 to 6 and are recorded in column
titled magazine. Out of the people who saw
these ads, 18 randomly chosen respondents
are picked , one from each magazine. Thus
we have one respondent who have seen a
given version of ad in a given magazine. We
have one respondent for every combination
of ad and magazine.

07/08/20 IBS 67
 SPSS Output: Completely Randomized Design &
Randomized Block
Dependent Variable: Ratings
Source Type III Sum df Mean F Sig.
of Squares Square
a

Ad_Version 7.000 2 3.500 9.545 .005

Magazine 25.833 5 5.167 14.091 .000
Error 3.667 10 .367   
Total 36.5 17     
a. R Squared = .995 (Adjusted R Squared = .991)

07/08/20 IBS 68
 SPSS Output: Analysis & Interpretation
• The computer output for this problem using a randomised block
design is shown in Tests of significance for RATING using UNIQUE
sums of squares.
• This table is similar to the output table of the one-way ANOVA we
got earlier in completely randomized design, except that there is an
additional source of variation called ‘Magazine’ in the first column.
This is the ‘block’ we have used, to test the null hypotheses –
• The first null hypothesis is that mean rating of the Ad_Version is
the same for all 3 versions. This is the same as the null hypothesis
we had used earlier for the one-way ANOVA.
• What should be H1?
• .The second null hypothesis is that the ‘block’ used (Magazine in
this case) has no effect on mean ratings given to Ad_Version
versions by respondents. Write H1

07/08/20 IBS 69
 SPSS Output: Analysis & Interpretation
•1. To test if the null hypotheses are rejected or not, we turn to the last
column of the output of ANOVA, which gives the result of an F-test for any
assumed confidence level. 95% is the default level or 5% is level of
significance.The same has been considered in this example as well.
•2. The significance level of F or the p value in the last column should be
less than 0.05 for the null hypothesis to be rejected. For both the rows
labelled Ad_version and MAGAZINE, the significance of F is less than .05.
It is .005 for Ad_version and .000 for MAGAZINE. This means that both the
null hypotheses are rejected. The same can be checked through the table
value of F as well with given df for numerator and denominator.
• 3. We conclude that the mean ratings given to the 3 versions of ADCOPY
are significantly different, and also that the MAGAZINE in which the
ADCOPY appears has an impact on its rating.
• 4. Please note that the Blocking Factor being considered separately has
now led us to a different conclusion from that in a completely randomized
test of the same basic data. This makes the randomized block test a
better test when we suspect that a blocking factor affects the relationship
between the independent variable and the dependent variable.

07/08/20 IBS 70
 Class Exercise
• Cholesterol contents of the four competing diet foods on the
basis of the following data (in mgs per package) which were
obtained for three randomly taken 6 ounce packages of each of
the diet foods, measurement performed in three different labs.
The first value of each sample came from one laboratory, the
second value came form another laboratory and the third value
from a third laboratory.
Diet Food Lab 1 Lab 2 Lab 3
A 3.6 4.1 4
B 3.1 3.2 3.9
C 3.2 3.5 3.5
D 3.5 3.8 3.8

Perform an ANOVA using 5 % Level of

significance. Write the hypotheses. Analyse and
interpret the output generated from SPSS
07/08/20 IBS 71
 ANOVA Table – Please
complete
Source SS DF MS F Sig/p
Value
Diet Food ? ? ? 4.909 .047
Lab ? ? ? 5.727 .041
Error .220 ? ?
Total ? ?

07/08/20 IBS 72
 ANOVA Table

Source SS DF MS F Sig/p Value

Diet Food ? (4 -1) ? 4.909 .047

Lab ? (3-1) ? 5.727 .041

Error .220 Balancing ?

fig
Total 4*3-1

07/08/20 IBS 73
 Latin Square Design
• The design is employed when the researcher
is interested in seperating the influence of two
extraneous variables.
• Suppose in the price level sales example, the
sales could be influenced by store size and
type of packaging.
• For application of Latin Square design, the
number of categories of two extraneous
variables should be equal to the number of
levels of treatments.

07/08/20 IBS 74
 Latin Square Design – The store and
Packaging are two variables to be
blocked
Packaging

Store Size I II III

Small X1 X2 X3
Medium X2 X3 X1
Large X3 X1 X2

07/08/20 IBS 75
 Latin Square Design –
Contd…
• Rows and columns represent those
extraneous variables whose effect is to be
controlled and measured. 3 X 3 Latin Square.
• The treatment should be randomly assigned
to cells in such a way that each treatment
occurs once and only once in each row and in
each column.
• This design requires a complex set up and
quite expensive to execute

07/08/20 IBS 76
 Latin Square Design contd.

• The Latin Square Design is an extension of

the Randomized Block Design. It consists of
one independent variable (FACTOR) and two
Blocks, instead of one which we saw in the
Randomized Block Design.
• It has no special significance in marketing
research.

07/08/20 IBS 77
 An Illustration
• The production manager of a firm is keen on
analysing the hourly production volume of a
product. He feels that operators (A B C D E,
Machines (M1,M2.M3,M4,M5) and suppliers
of raw materials (S1,S2,S3, S4, S5) will affect
the production volume of the product.
• Fill in the blanks . State null and alternative
hypotheses .
• State the design used

07/08/20 IBS 78
 Latin Square Design
ANOVA TABLE

Degrees Mean
Source of of Sum of sum of
Variation Freedom Squares Squares F Ratio Sig
Between
Operators ? ? ? 2.14 
Between
Machines ? ? ? 1.57 
Between
Suppliers ? ? ? 0.15 
Error ? ? 10    
Total
07/08/20 IBS 79
 Solution
Source of d.f SS MS F
Variation
Rows m-1 RSS MSR = RSS/(m- MSR/MSE
1)
Columns m-1 CSS MSC = CSS/(m- MSC/MSE
1)

Treatment m-1 TrSS MST = TrSS/(m F= MST/MSE

– 1)

Error Balancing fig ESS MSE = ESS/(df F Ratio d f for

of error) Numerator m-1
and d f for
Denominator is
df of error or
residual
Total m2 - 1
07/08/20 IBS 80
 Class Exercise
• Complete the table……

07/08/20 IBS 81
 Factorial Design
•A factorial design may be employed to measure the
effect of two or more independent variables at various
levels. The factorial designs allow for interaction
between the variables.
•An interaction is said to have taken place when the
simultaneous effect of two or more variables is different
from the sum of their individual effects. (Mango ice
creams)
The sales price example, if the store size is categorized
into two , then we have a two-factor design .A table with
three rows and two columns could be formed. Six
different level of treatment combinations would be
produced each with a specific price level and store size.
07/08/20 IBS 82
 3 X 2 Factorial Design

Price/Store Small (B1) Big (B2)

Low Level (A1) A1B1 A1B2

Medium Level (A2) A2B1 A2B2

High Level (A3) A3B1 A3B2

Respondents in each cell receive a specified treatment

combination

07/08/20 IBS 83
 Factorial Design Contd..
• Advantages:
It is possible to measure the main effects and
interaction effect of two or more independent
variables at various levels.
It allows saving of time and effort because all
observations are employed to study the
effects of each factor.
The conclusion reached using factorial design
has broader applications as each factor is
studied with different combinations of other
factors
07/08/20 IBS 84
 Format of Output
Source of SS DF MS F
Variation
Row RSS r-1 RSS/(r-1) MSR/MSE
Column CSS c-1 CSS/(c-1) MSC/MSE
Interaction ISS (r-1) (c-1) ISS/((r-1) (c- MIS/MSE
1))
Error ESS Balancing fig ESS/(df of
error)
Total TSS N -1

07/08/20 IBS 85
 Factorial Design :Explanation
•If cosmetic soap company is interested to understand the effect of two
Factors (independent variables) – Pack Design and Price - on Sales
(dependent variable). One is the company would like to know (1) if each of
the Factors independently affects Sales (called the Main Effects), and (2) if
there is a combined effect of Pack Design and Price (called the 2 way
Interaction Effect) on Sales.

•Incidentally, if there are 3 factors in a study, then we could test for all 2-
way interaction effects and the 3-way interaction effect, in addition to the
Main Effects of the individual factors.

•Here, the experiment is conducted in a simulated environment on 18

randomly selected respondents. There are 3 levels of price – Rs. 8, Rs. 11
and Rs. 14, and 3 levels of Pack Design – designated by the main colours
used – Blue, Red and Green.
• The coding of these variables is 1, 2, 3 respectively for Rs. 8, 11 and 14
and 1, 2, 3 for Blue, Red and Green in the case of Pack Design.

07/08/20 IBS 86
 Input Data
The
sr. inputsales
no. dataset ispackdesn
shown price
1 500 1 1
2 440 2 1
3 360 3 1
4 300 1 2
5 280 2 2
6 250 3 2
7 200 1 3
8 150 2 3
9 250 3 3
10 600 1 1
11 450 2 1
12 510 3 1
13 400 1 2
14 350 2 2
15 300 3 2
16 250 1 3
17 275 2 3
18 220 3 3

Column 1 is Sales, column 2 is Pack Design and Column 3 is Price. Please note that
even though Price is a continuous metric variable, for the purpose of ANOVA, being an
independent variable, it has to be treated as a categorical variable. Hence the coding
(1, 2, 3) for Price.
07/08/20 IBS 88
 Analysis & Interpretation
•Also note from that each combination of Price and Pack Design
appears twice in the dataset. For example, Packdesign = 1 and
Price = 1 appears in Row 1 and also Row 10. This is known as a
replication in design of experiments. This is similar to having a
higher sample size in a survey.

• Depending on the number of Factors and the number of levels of

each Factor, the minimum sample size required for ANOVA may
go up. In such cases, multiple observations or replications become
necessary. In general, replications reduce chances of random error
affecting the results of ANOVA experiments, similar to the effects
of increasing sample size in surveys.
•Output:
•The output data for our factorial experiment are presented in
previous slide

07/08/20 IBS 89
 Analysis & Interpretation
• A look at Sources of Variation listed in the first column reveals
main effects and interaction effect. The last source of variation
listed is the Residual or error term.
• In this case, we are testing three hypotheses –
 The mean level of Sales remains the same for all 3
levels of Pack Design (Main Effect 1).
 The mean level of Sales remains the same for all 3
levels of Price (Main Effect 2).
 The mean level of Sales remains the same for all
combinations of Pack Design and Price (Interaction
Effect).
• Please write the alternative hypotheses for each.
• Assuming 0.05 level of significance, we check whether for each
of the rows corresponding to the above hypotheses, the
significance of F or the p value is below 0.05 in the last column
of output
07/08/20 IBS 90
 Analysis & Interpretation
• The significance or p value of F are
• Pack Design - .248 (Main Effect 1)
• Price - .000 (Main Effect 2)
• Pack Design by Price - .646 (Interaction Effect)
• Therefore, only the Price effect, one of the two main
effects, is significant statistically, at 95 percent confidence
level. This means that hypothesis no. 2 is rejected.
• Hypothesis 1 and 3 cannot be rejected, as the significance
of F values are greater than .05 in both cases .248 and .
646 respectively).
• Thus, we conclude that Price alone has an impact on
Sales. Neither Pack Design alone nor the combination of
Pack Design with Price have any significant impact on
Sales of the toilet soap.
07/08/20 IBS 91
 Exercise
• The following data refers to the salary
package ( in Lakhs of rupees) offered to
MBA graduates with different
specializations and having studied at
four different B- Schools. For sake of
simplifications only two students are
taken for each interaction between the
institute and field of specialization.

07/08/20 IBS 92
 Exercise Contd…..
Specialization BS I BS II BSIII BSIV
Marketing 6 4 8 6
5 5 6 4
Finance 7 6 6 9
6 7 7 8
Operations 8 5 10 9
7 5 9 10

Test the hypothesis at a significance level of 5 %

i) whether the difference between the average pay packages
across all B Schools are equal.
ii) Average pay packages by all specializations are equal
iii) the average pay package for 12 interactions are equal
Can we attribute it the difference is due to chance?
07/08/20 IBS 93
 Please write all the
hypotheses

07/08/20 IBS 94
 SPSS Lab
S package SPL BS
6 1 1
5 1 1
7 2 1
6 2 1
8 3 1
7 3 1
4 1 2
5 1 2
6 2 2
7 2 2
5 3 2
5 3 2
8 1 3
6 1 3
07/08/20 IBS 95
 SPSS Lab
S package SPL BS
6 2 3
7 2 3
10 3 3
9 3 3
6 1 4
4 1 4
9 2 4
8 2 4
9 3 4
10 3 4

07/08/20 IBS 96
Generate Output and Interpret

07/08/20 IBS 97
 SPSS output Factorial Design
Source SS DF MS F Sig/p value
Spl 23.083 2 11.542 16.294 .000
BS 22.458 3 7.486 10.569 .001
Spl*BS 17.917 6 2.986 4.216 .016
Error 8.5 12 .708
Total 71.958 23

07/08/20 IBS 98
 Output from SPSS

07/08/20 IBS 99
 Experiments & Post hoc Analysis
• ANOVA design helps to understand the existence of difference
between groups, but to study which of the groups are
significantly different from others, a post hoc study especially
using Tukey Test is used. The same helps us to comprehend
the relevant variables which actually are significantly different or
popularly called Honestly significant different.
• Post hoc analysis analyses tests that were not specified before
the data was seen. While an Ad Hoc analysis consists of
analysis designed for a specific purpose rather than for general
usage. Post Hoc analysis as the phrase indicates “ after this”,
refers to reasoning, discussion, or explanation that takes place
after something has already transpired.
• Experimental study is undertaken using some pre thought while
post hoc study is taken in a scientific experiment in case the
results of the experiment need to be analysed so.
07/08/20 IBS 100
 Applications of Statistical Designs

• Experiments are today widely used in many ways in Business

Research.
• An experimental procedure that provides an opportunity to
test a new product or a new marketing plan under realistic
market conditions to measure sales or profit potential.
• For example, test marketing of new concepts, products or
prototypes is usually done through procedures explained
above, or similar to these.
• STM or simulated Test Marketing procedures are extensions
of the basic ANOVA type experiments, with the added tools of
forecasting based on the results of experiments conducts.

07/08/20 IBS 101

 Recap…
• The fourth stage of Research Process
• Experimental Research Design
• Causality
• Conditions for Causality
• Concepts
• Validity
• Factors affecting Internal Validity
• Factors affecting External Validity
• Classification of Experimental Designs
• Pre Experimental
• Quasi Experimental
• True Experimental
• Statistical ( ANOVA & SPSS) for the four Designs

07/08/20 IBS 102