The Research Design

DETAILS OF STUDY MEASUREMENT

Purpose of the Types of Extent of researcher Study setting Measurement DATA

and
study investigation interference ANALYSIS
measures
Exploration Establishing: Minimal: Studying 1. Feel
Description – Causal relationships events as they normally Contrived for data
Operational
Hypothesis testing – Correlations occur Noncontrived
definition
PROBLEM STATEMENT

– Group differences, Manipulation and/or

Scaling
ranks, etc. control and/or
simulation

2. Goodness
of data

Data-
Unit of analysis Sampling Time
collection
(population to be Design horizon
method
studied) 3. Hypothesis
Individuals testing
Observation
Dyads Probability/ One-shot
Interview
Groups nonprobability (crosssectional)
Questionnaire
Organizations Sample size (n) Longitudinal
Physical
Machines
measurement
etc.
 Sampling
 It is a process of obtaining information about an entire population
by examining only part of it
 If a banker is interested in investigating the saving habits of BPO
workers in India
Essentials of a Good Sample

 Representativeness
 sample should possess the same characteristics as the original
population from which it has been drawn
 Independence
 Items should be independent of each other. In other words,
selection of any one item should not affect the selection of
any other item of the population for the same sample.
 Homogeneity
 There should not be any basic difference in the nature of units
of the population and that of the sample. If two samples from
the same population are taken they should be similar
 Adequacy
 Ie. The sample should be fairly large. The no. of units in the
sample should be adequate so as to make the results more
accurate and reliable.
Sampling frame

 List of elements from which the sample is to be

drawn
 It is a correct list of population members
 5 questions have to be answered in developing
a sample
 What is the target population?
 What are the parameters of interest?
 What is the sampling frame?
 What is the appropriate sampling method?
 What is the sample size needed?
Sampling design
 A definite plan for obtaining a sample from the sample frame
 Steps of sampling design
 Finalize parameters of interest
 Define type of the universe
 Decide the sampling unit or unit of analysis (house, person,
object)
 Prepare the sampling frame (List)
 Decide sample size
 Finalise budget
 Finalise sampling procedure (type & method)
Sampling error

 Sampling errors are those errors which arise on account

of sampling

Sampling error = frame error + chance error + response

error
Types of sampling
 Non- probability sampling (probability of choosing a particular
member of population is unknown)
- Convenience sampling
- Purposive sampling (quota & judgement sampling,
snowball sampling, volunteer)
 Probability sampling (every element in the population has a
known probability of selection)
- Simple random sampling
- Complex (restricted) random sampling
 cluster (heterogeneity within group),
 systematic
 stratified sampling (homogeneity within group)
 Area and double sampling
Probability Sampling
 Random Samples
 Selected using chance method or random methods
 Simple or Unrestricted Random Sampling
 A sample is considered a random sample if its members
are drawn in such a way that each observation of the
population has an equal chance of being included in the
sample, and every possible combination of observations
in the population has the same chance of being
included.
 Lottery method (a slip is made for every item in the
population and someone who is neutral is asked to
select the required no. of slips)
 Tippet’s method (Standard tables of random nos. like
Tippets random no. tables, Fisher and Yates tables and
Random Corporation tables are used and the population
units are numbered from 1 to 100 and selected using
these tables)
 Probability Sampling
 Random Sampling
 Restricted Random Sampling
 Special type of random sampling
 Mainly used
 when the data is not homogenous; in this case stratified sampling
or cluster sampling is used.
 When a short-cut method of obtaining a random sample is
required; in this case a systematic sampling is used.

 Stratified Sampling
 Used when population is heterogeneous and consists of different
homogenous groups with different characteristics, hence popultn.
is divided into different strata or groups and are selected in such a
way that the variance of the characteristic under study is smaller
within the group and larger between the groups. Units are
sampled at random from each of these strata.
 Homogeneity within the groups and heterogeneity between the
groups
 Eg. Two income group people of low income and high income
Probability Sampling

 Random Sampling
 Restricted Random Sampling
 Cluster Sampling
 Population is divided into groups called blocks
or clusters which are representative of the
population as a whole.
 There is considerable variation within each
group but no variation between the different
groups. (heterogeneity within the groups &
homogeneity between the groups)
 Eg. A market research team wants to determine
the average no. of TV sets per household in a
large town; they can divide the town into a no.
of blocks and then choose a certain no. of
blocks for interviewing each household.
Probability Sampling

 Random Sampling
 Restricted Random Sampling
 Systematic Sampling
 Each sample element in a population has an
equal probability of being selected
 Eg. If in a firm there are 100 employees,
numbered 0 to 99 and we choose a sample of
10 by picking up every 10th employee beginning
with 1, then those numbered 2,3,4 and 5 have
no chance of being selected altogether.
 This type of sampling is used when the
complete and upto date list of sampling units is
available and the units of the population to be
sampled are ordered in a specific manner.
 Eg. The arrangement of names in a telephone
directory in an alphabetical order can be used
for selecting a sample of subscribers by the
systematic sampling method.
 Non-Probability Sampling
 Non-Random Sampling
 Can be drawn by 3 methods
 Judgement / Purposive Sampling
 Individual items are selected by the investigator by
his own judgement
 Quota Sampling
 Quotas are fixed for each group and units are chosen
depending on the prescribed quotas
 Convenience Sampling
 Convenient slice of the population is selected
according to the convenience of the investigator
 Determining the Sample size
 Nature of universe
 Nature of study
 Type of sampling
 Standard of accuracy and acceptable
confidence level
 Availability of finance
 Higher the sample size greater is the
generalizability
Determining the Sample size Sample
Size Formula

2
 zs 
n 
E
Question
 A survey researcher, studying expenditures on soft
drinks, wishes to have a 95 percent confidence level (Z
standardized value corresponding to the conf. level) and
a range of error (E) of less than Rs. 2. The estimate of
standard deviation is Rs. 29.00. Determine sample size.
Sample Size Formula -
Example

 1.96  29.00  
2 2
 zs 
n   
E  2.00 
2
 56.84 
    28 . 42  2
 808
 2.00 
Sample Size Formula -
Example

Suppose, in the same example as the

one before, the range of error (E) is
acceptable at Rs.4.00, sample size is
reduced.
Sample Size Formula -
Example

 1.96  29.00  
2 2
 zs 
n   
E  4.00 
2
 56.84 
    14 . 21 2
 202
 4.00 
 Calculating Sample Size
99% Confidence

2 2
(2.57)(29)  (2.57)(29) 
n  n 
 2   4 
2 2
74.53  74.53 
   
 2   4 
 37.265  18.6325
2 2

1389  347
Types of Data

 Primary and Secondary

 Qualitative (Categorical)
 Quantitative (Numerical)
 Primary
 Data originally collected in the process of
investigations
 Secondary
 Data already collected by some other
persons/for some other studies but can be
used for this particular study.
 Primary data collection
methods
 Direct Personal Interview (investigation)
 Face to face
 Indirect Personal Interview (investigation)
 Telephone
 Computer assisted tel. interviews (CATI)
 Computer assisted per. Interviews (CAPI)
 Information from correspondents
 Mailed Questionnaires
 Questionnaires filled by enumerators
 Observational studies
 Projective techniques (indirect projection of beliefs
and feelings)
Primary Data
 Focus Groups
 An unstructured free flowing interview with a small
group of people (one-time)
 A moderator plays the role in steering the
discussions which help in obtaining the respondents’
interpretations, opinions etc.
 Videoconferencing helps several focus groups at
different locations
 Prim. Advantages: relatively brief, easy to execute,
quickly analyzed and less expensive, provides fairly
dependable data within a short time
 Disadvantages: rarely representative, cannot replace
quantitative studies
Primary Data

 Panels
 Groups that meet for more than once for
discussions
 Certain effects or changes have to be studied
over a period of time
 Static (same members serve for extended
periods of time)
 Dynamic (members change from time to time
as various phases of the study are in progress)
 Adv: measure changes from time to time
 Disadv: bias can develop due to the
repetitiveness of discussions
Primary Data

 Unobtrusive measures
 Originate from a primary source that does not directly
involve people
 Eg. no. of different soft drink cans found in trash bags
provide indication about their consumption levels
 Collection of data on the births, deaths etc.
 Company records provide a lot of personal information
about their employees
Comparison

 Primary Data
 Merits
 Original in nature
 Reliable, Authentic and Accurate
 Free from bias
 Exactly matches the need of the project
 Demerits
 Expensive
 Time-consuming
 Questionnaires and interview forms have to be developed
Sources of Secondary Data

 Published sources
 Government publications
 Publications of international organizations
 Semi-official publications
 Reports of committees and commissions
 Private publications
 Journals & newspapers
 Research institutions
 Professional trade bodies
 Annual reports of companies
 Articles, market reviews and reports

 Unpublished sources
 Unpublished data collected by research institutions, trade
associations, universities etc.
Comparison

 Secondary Data
 Merits
 Readily available
 Less expensive
 Less time consuming
 Demerits
 Possibility that proper procedure might not have been
followed in their collection
 May not be free from personal bias and prejudices
 May not have the needed accuracy or reliability
 May be outdated

Exercise 3
In an article in the Wall Street Journal titled “Kellogg to
study work of salaried staff, setting stage for possible job
cutbacks”, it was stated that Kellogg's earnings remained
under heavy competitive pressure and its cereal market
continued to slip. It was also stated that Kellogg was seeking
to regain its lost momentum through the first three
strategies listed below, to which the last two are added:
 Increasing production efficiencies
 Developing new products
 Increasing product promotion through advertising
effectiveness
 Tapping creative ideas from organisational members at
different levels
 Assessing perceptions of organisational health and vitality
Discuss the sampling design for each of the five strategies
above. Give reasons.
Exercise 2

 A company having chain of restaurants in Mumbai, wants to start a

new one in Navi Mumbai. Before starting, it wants to conduct a
survey of employees of various institutions in that area to know
about their eating habits, preferences, etc. Suggest a suitable
sampling scheme for the same.