Transcript Research Methodology

Research Methodology
Definition: Business Research
 Systematic and Scientific inquiry
aimed at providing information to
solve managerial problems
Basic (Fundamental) Research
 Generalization of natural
phenomenon or human behavior
 Conducted to solve perplexing
questions of a theoretical nature
E.g. Cars running on compressed air
Applied Research
 Solution to some practical problems
 Conducted to reveal answers to
specific questions related to action or
policy needs
E.g. Study of declining sales of
particular car model
Why business research?
 Manager wants improve decision
making skill by widening his
knowledge
 to reduce uncertainty
 to justify his decision
Business Research
 Corporate research:





Forecasting
Economic trends
Business & Industry trends
Technology trends
Global environment
 Financial Research:
 Portfolio analysis
 Business opportunities & threats
analysis
 Cost analysis
 Production:
 Total quality management
 Sales & Marketing:







Market potential
Market share
Market segmentation
Sales analysis
Sales forecasting
New product testing
Effectiveness of marketing strategies
 Human resources:




Morale & job satisfaction
Employee productivity
Organizational effectiveness
Training needs
 Information system:
 Data mining
 Technical support satisfaction
 Enterprise Resource Planning
(ERP)/Customer Relationship
Management (CRM) systems
Types Of Research
 Exploratory
 Descriptive
 Predictive
Exploratory Research
 Gather Information around the topic
at elementary level
 Leads to insights into the problem
E.g. a pharmaceutical company wants to
expand
Descriptive Research
 Fact finding inquiries
 Who, what, when, how, where etc
 No powerful inference is drawn
E.g. mining company databases for
describing the nature customer
complaints.
Predictive Research
 Cause –effect relationship
 Conducted for well understood
problems
E.g. sale of small cars depends on
price, fuel efficiency, engine
Research Process
 Formulating Research problem
 Select Research design
 Data collection and presentation
 Data analysis and interpretation
 Research reporting
Formulating a Research Problem
Problem:Dilemma faced by manager
E.g. Software consulting firm is
witnessing high attrition of employees
Management Question
 What are the reasons for high
attrition rate?
 What steps to be taken to control it?
Research Question
 What is the employee satisfaction
level in the organisation?
 What are the corrective measures
available?
Investigating Question
 Timings, travelling, facilities available,
work culture
 Annual increment, additional
incentives, permission to pursue
higher education
Measurement Question
 Questions to be asked to employees
regarding timings, travelling, facilities
available, work culture, additional
incentives, permission to pursue
higher education etc.
Formulated Research Problem
 What characteristics are to be
studied?
 What data is to be collected?
 What relations need to be explored?
Research Design
 Research design is a plan and
structure of investigation to obtain
answers to research questions
Exploratory Studies
Aim is to find the practicality to do
formal research in this area
The techniques used are
 Secondary data analysis
 Experience Survey
 Focus groups
 Depth interviews
Secondary Data Analysis
 Review of prior research studies
 Survey of concerned literature: Catalogs,
subject guides and electronic indexes in
libraries help to identify periodicals and
books
 Online service using computer
This method supplies excellent background,
identifies suitable methodology and
decision making patterns.
Experience Survey
 Seeking information from persons
experienced in the area of study by
interviewing
This technique gives practicality of doing
study, facilities available, factors need
to be controlled
Focus Groups
 Panel of people led by a trained
moderator
 Facilitator introduces the topic and
encourages the group to discuss it
among themselves
Quickly and inexpensively grasp the
core issues
Depth Interviews
 Carried out to obtain information on
sensitive issues to discover motives
and desires
 Respondent is taken in confidence
 Answers are obtained by probing
 Indirect questions provide information
on attitude towards the subject
Descriptive Studies
The research design in such studies should
be rigid ,no bias ,reliable and should aim
at
 Description of the characteristics
associated with the population
 Estimates of the proportion of the
population possessing the characteristics
 Studying association among these
characteristics
Causal Studies
 If Changes in one variable are
responsible for changes in another
variable then the former variable is
independent and later is dependent
 Drawing inferences about causal
relationship require procedures that
will reduce bias and increase
reliability
Design of Experiments
 Replication: repetition of observation
 Random assignment: each person
must have equal chance of exposure
 Control: All other factors must be
held constant
Informal Experimental Designs
Before and after without control
Design
Level before--treatment-- level after
X
Y
treatment effect : Y-X
After only control Design
Level control group after Z
Level test group after
Y
treatment effect : Y-Z
Before and after with control
Design
Control group
Level before
------------ level after
A
Z
Test group
Level before--treatment-- level after
X
Y
 treatment effect : (Y-X)-(Z-A)
Formal Experimental Designs




Completely randomised design
Randomised block design
Latin square design
Factorial design
Completely randomised Design
Det. A
Det. B
Det. C
Randomisation, Replication
Randomised block Design
Det/Mat Det. A
Det. B
Det. C
Cotton
Poly.
Ter.
Randomisation, Replication, Local control
Latin square Design
Mat/Cat worker
Cotton
Poly.
Terr.
A
B
C
Assistant manager
B
C
C
A
A
B
Randomisation, Replication, Local control
Factorial Design
 Treatments are combination of factors
at different levels.
 Effects of the factors rather than
treatments are studied.
Ex Post Facto Design
 Widely used in business research
 Not possible to assign treatments
 Study of subjects exposed to
treatments to seek causal explanation
Sampling Design
 Census versus sample
Census: small population
Sample: greater speed, less cost, greater
accuracy, inevitable
Sample must represent all characteristics of
the population.
(no bias, precise)
Types of Sampling
 Probability sampling
 Non probability sampling
Probability Sampling





Simple random sampling
Systematic sampling
Stratified sampling
Cluster sampling
Multi-stage sampling
Non probability Sampling
 Judgment sampling
 Quota sampling
Sample Size
Sample size should be proportional to
population size
Larger sample is required if
 Greater dispersion within population
 High confidence level in the estimates
 Narrow interval range
budget constraint may influence sample size
Measurement and Scaling
 Measurement: assigning numbers to
objects with a set of rules
 Data types:
Classification - nominal
Order - ordinal
Distance - interval
Origin - ratio
Tests of Measurement
 Validity: the extent to which the
measuring tool measures what we are
interested in measuring
 Reliability: Accuracy and precision of
measurement procedure
 Practicality: economy, convenience
and interpretability
Scaling
Procedure for assignment of numbers to
objects
Types of measurement scale
 Rating
 Ranking
 Categorization
Rating Scales
E.g.
Survey on mobiles
Simple category Scale
Which payment scheme are you
currently using for your mobile?
 Pre-paid
 Post-paid
Nominal data
Multiple choice Single response
Scale
Which mobile service according to you is the
best?
 Vodafone
 BPL
 Airtel
 Reliance
 Tata Indicom
 Dolphin/Trump
Nominal data
Multiple choice Multiple response
Scale
What factors influence you while choosing
mobile service?
 Network
 Brand name
 Price
 Customer service
nominal data
Likert scale Summated Rating
“Mobile users are spending on unnecessary calls “
Strongly
agree
agree
interval
Neither
agree nor
disagree
disagree
strongly
disagree
Semantic Differential Scale
Reliance service
Good __:__:__:__:__:__:__: Bad
Handset
Handset
High
__:__:__:__:__:__:__: Low
quality
quality
Network
Network
interval
Numerical Scale
Extremely
favorable
5
4
3
2
1 Extremely
unfavorable
Handset with camera:__
More than one handset per person:__
interval
Multiple Rating List Scale
“Please indicate how important or
unimportant each parameter is”
Important Unimportant
Network congestion
5 4 3 2 1
Billing errors
5 4 3 2 1
Clarity
5 4 3 2 1
Advertiser’s SMS
5 4 3 2 1
interval
Fixed sum Scale
Considering web services and all
services other than calling, what is their
relative importance to you?
 Web services
: ____
 All other services: ____
sum = 100
ratio
Staple Scale
Vodafone
Technology
Leader
ordinal
-3
-2
-1
0
1
2
3
Exciting
product
-3
-2
-1
0
1
2
3
Graphic Rating Scale
Rate the satisfaction level from the
current service. (Place an X at the
position along the line that reflects
your judgment)
Very good _______________ Very bad
ratio
Ranking Scale
 Paired comparison scale
 Forced ranking scale
 Comparative scale
For each pair of Mobile service provider,
put a tick besides the one you would most
prefer
-- Vodafone
-- BPL
-- Airtel
-- BPL
-- Vodafone
-- Airtel
Paired comparison Scale
Comparison between Vodafone, BPL, Airtel
Suggestion
Vodafone
BPL
Vodafone
--60
BPL
40
--Airtel
80
50
Total
120
110
Rank order
1
2
M=(C+0.5N)/nN
0.5667
0.5333
Z
0.17
0.08
R
0.42
0.33
ratio
Airtel
20
50
--70
3
0.4
-0.25
0
Forced Rank Scale
Rank the following mobile services.
Place the number 1 next to most
preferred, 2by the second choice etc.
Vodafone
BPL
Airtel
Reliance
Tata Indicom
Dolphin/Trump
ordinal
Comparative Scale
Compared to NOKIA 3310,how do you
rate the your handset?
Superior
1
2
ordinal
all the same
3
4
inferior
5
Measurement Scale Construction




Arbitrary scaling
Consensus scaling
Item analysis scaling
Cumulative scaling
Arbitrary Scaling
 Designed by collecting several items
appropriate to the given topic
 Each item is scored from 1 to 5
depending degree of favorableness
 Total represents respondent’s view
Consensus Scaling
 Panel of judges evaluate items on the
basis of relevance to topic, potential
for ambiguity and level of attitude
 Time consuming, costly ,rarely used
in management research
Item analysis Scaling
 Designed by collecting several items
appropriate to the given topic
 Each item is scored from 1 to 5
depending degree of favorableness
 On the basis of total score two groups
are formed (top 25% and bottom
25%)
 Comparative analysis of each
statement of members of these two
groups is carried out
Item analysis Scaling
Low High
Response
X
f
f
Strongly agree
5
5
20
Agree
4
10
18
Neither ag.nor dis.
3
22
25
Disagree
2
25
12
Strongly disagree
1
23
10
If calculated value of t>1.75 the statement is
said to be a good discriminator of the
measured attitude
Cumulative Scaling
From total score it is possible to
estimate which items were answered
positively and negatively
Scalogram
 i) Vodafone service is good: agree/disagree
 ii) I prefer Vodafone service over other
services: agree/disagree
 iii) Vodafone service suits me: agree/disagree
 iv) I strongly recommend Vodafone:
agree/disagree
Agree=1, disagree=0
Score: 4-aggreed to all,3 agreed to i),ii),iii)
2 agreed to i),iii) 1 agreed to iii)
Data collection and Presentation
 Primary sources (original work of
research): Government data (census,
economic, labour) laws, regulations
 Secondary sources (interpretation of
primary data): magazines, newspapers,
newscasts, annual reports
Government Sources
 National sample survey
 Economic survey
 Census survey
http://www.censusindia.net
 Central Statistical organisation:
Statistical Outlines of India, annual
survey of industries
http://mospi.nic.in
Other databases to Business
Researchers
 CRISIL INFAC www.crisil.com
 NSE publications
 National readership survey (cinema, TV,
radio, newspaper)
 A C Nielsen company ( consumer products)
www2.acnielsen.com
 ORG (pharmaceutical products)
Mining Internal sources
 TAM (Television Audience
measurement)
 Adex ( advertising exposures )
 IRS (Indian Readership survey)
 NRS (National Readership survey)
 indiacar.com
 autocarindia.com
Survey Methods





Observation method
Interview method
Questionnaire/schedules
Projective techniques
Case study method
Observation Method
 Information obtained by direct
observation without asking the
respondents
 Scientific tool when systematically
planned and recorded
 No non-response
Interview Method
 Personal interview: Information
directly obtained from person
concerned
 Use of set of predetermined
questions: structured interview
 Telephone interview
Questionnaire/schedules Method
 Questionnaire is mailed to the
respondents and replies collected
 Schedules are similar to questionnaire
but are being filled in by the
enumerators
Projective Techniques




Word association
Sentence completion
Thematic appreciation test
Role play
Case Study Method
 Qualitative analysis
 Observation on a social unit or a
person or an institution or a
community
 An exhaustive study of each and
every aspect of the concerned unit
and its relationship to other units
Questionnaire
 Question: short, simple, easy to
understand, arranged in logical sequence,
self explanatory, objective with possible
answers, no numerical calculation, provide
cross check
 Classification questions, target questions
 Covering letter
 Test on small group, remove deficiency,
final draft