Sie sind auf Seite 1von 51

Research Design

Session 2
Dr. Plavini Punyatoya
Xavier Institute of Management,
Xavier University, Bhubaneswar
Research Design: Definition
A research design is a framework or blueprint for conducting the
marketing research project. It details the procedures necessary for
obtaining the information needed to structure or solve marketing
research problems.
Research design

Exploratory research Conclusive research


design design

Descriptive Causal
research research

Cross-sectional Longitudinal
design design

Single Multiple
cross-sectional cross-sectional
design design
Example
Amway is one of the leading international brands in the nutrition
product category. It had an assortment of good nutritional products for
adults (Nutrilite tablets). Then it decided to develop some new and
different nutrition product for entire family.
They consulted a MR agency IMRB and asked for help. IMRB started the
research by meeting Amways marketing professional to list what
customers want. Then they discussed with Food technologists
(engineers) to understand present technology, competency and its
applicability.
Considering the input, IMRB conducted a group discussion with 20
people (consumer, external designers, marketers and specialist) & came
up with 140 new product concepts.
It is Exploratory research.
Provide insights & understanding of the problem confronting the
researcher

Front end of the total research design

Tentative, normally followed by conclusive research

Small sample, mostly qualitative data results

Include: expert interview, discussion, secondary data analysis

Although it is initial step, it may not be done at times. It researcher


defined problem well & certain about the approach.
Example
Amway refined and validated the 140 concepts. Selected a few. Then
go for a descriptive survey among consumers.
Consumers were questioned: which product they desire or prefer to
buy. The questionnaire contained the product description and
features and benefits. The purchase intention was measured among
the respondents.
Based on their input Amway came up with Nutrition tablets for
children and old-age people. The tablet will help children to grow
faster (taller). The other tablet developed for older people will help
them fight the joint pain.
It is Descriptive Research.
Describe something

Include: characteristics of consumer, market share study, sales analysis


study, product usage study, perception of product characteristics, media
consumption habit, perception of firm

You have prior information. It is planned & structured.

Clear: who, what, when, where, why & way (in what way)
(6 Ws of research)

Quantitative data results


It is Descriptive Research.
Research on department store patronage

Who is a patron?
What information will be obtained?
When to collect data?
Where to collect? (in store, outside, parking, home)
Why require the information? (improve store image, make promotion strategy)
In what way we obtain information? (email, telephone)
It is Descriptive Research.
Conducted through survey or observation

Response rate : % of total attempted interviews/surveys that are


completed

Structured: researcher specified what is to be observed


Observation
Unstructured: observer monitor all aspects

Personal
Mechanical
Observation
Audit: examine physical objects, records, inventory

Content analysis: analyse words,themes,characters,topics

Trace analysis: evidence, past behavior, physical trace


It is Descriptive Research.
Mystery shopping: trained observers pose as consumers and shop at
company or competitor-owned stores to collect data about
customer-employee interaction, marketing variables and consumer
insight etc
-question store employees, mentally take note of answers, observe
Example
To check store loyalty:
Any person who did a purchase of more than Rs 1000/- on December
20th at BigBazar (one BigBazar in Mumbai, one in Delhi, one in Kolkata,
one in Chennai) =population
Departmental stores were randomly selected in cities.
Customers surveyed outside the store.
Sample: Every 3rd person doing purchase > 1000/-
Respondents surveyed on that day only and result was analysed
It is Cross-sectional design.
Collect information from any given sample of population only once

Single cross-sectional: only one sample drawn from target population &
information obtained only once

Multiple cross-sectional: 2 or more samples of respondents &


information obtained only once
-different samples surveyed at different time intervals
(December 20th , January 20th, February 20th )
Example
Jawed Habib Hair and Beauty Salon decided to introduce 2 different
schemes of customer promotion in the year 2013. To test the
effectiveness of these two schemes it decided to conduct MR. The best
scheme will be repeated in the next year 2014.
From last many years documentation, it found the regular
men/customers list (visiting 6 times a year for last 2 years). Mumbai is
selected as the city to conduct research. The Salon randomly contacted
some regular customers and asked if they are interested in the research.
130 of them agreed and theyll get an incentive to participate in the
research.
February: A survey conducted to know perception about the salon
April: head massage free with hair cut
May: survey about salon perception among same men sample
July: Rs 5000/- for haircut, hair spa and hair coloring
August: survey about salon perception among same men sample
It is Longitudinal design.
A fixed sample of population elements is measure repeatedly on the
same variables

Change over time

Panel: a sample of respondents who have agreed to provide


information at specified intervals over an extended period

Eg: Election poll


Favor congress govt. in Delhi?
Then, Food inflation
Now, Favor congress govt. in Delhi?
Example
Microsoft regularly develops and enhances its product portfolio by
innovations and product improvements. Before launching a new
product to the market it checks its usability.
Before launching MS Office 2007 to the market it decided to conduct a
usability research. Three computer user groups are formed and asked
to use Office 2007, Office 2003 and Office XP. Then they were asked
about ease of use, experience and product capability. It is compared
and Office 2007 was found to be better than previous versions and
launched to the market.
IV= Office type
DV= ease of use, experience and product capability
It is Causal research.
To get the evidence of cause-and-effect relationship

Descriptive research talks about degree of association between


variables & causal research talks about causal relationship

Control environment

Substantial knowledge, skill & capability required from researchers

Exploratory and Descriptive more use and common in market


research
Sampling Plan
Population and Sample
Population: (universe) is the collection of all members of a group
Eg: All students of XIMB, All employee of a bank, All milk packets

Sample: a portion of the population selected for analysis


Objective of sampling: to obtain maximum information about the
population with the minimum effort
Eg: 20 students of XIMB, 100 milk packets (MG road, Kanpur)

Census: gathering data from the entire population.


A complete enumeration

Sampling Frame: A set from which sample would come


Sample Size
Number of elements to be included in the study
Types of Sampling

2 methods of selecting samples from population

Nonrandom/judgment sampling Random/probability sampling

Personal knowledge, expertise and All items in population have a chance


opinion are used to identify the items of being chosen in the sample
from population

Sometimes-used as pilot or trial Rigorous statistical analysis possible


sample to decide how to choose
random sample
Convenient valid

Opens the selection bias Eliminates bias in the selection


process
Sampling Techniques

Nonprobability sampling Probability sampling

Convenience Judgmental Quota Snowball


sampling sampling sampling sampling

Simple random Systemic Stratified Cluster Other


sampling sampling sampling sampling

Proportionate Disproportiona
te
Probability sampling can be :
With replacement
Without replacement
1.Nonprobability Sampling
Convenience Sampling
Convenience is the main criteria to obtain the sample
Selection of sampling units primarily left to interviewer
No elaborative criteria for selection

Eg: students, mall-intercept interview without qualifying respondents,


people on street, member of a club or office

High selection bias

Least expensive & least time-consuming


Judgmental Sampling
A form of convenience sampling in which: Selection is based on the
judgment of researcher
Eg: Selection of test market for new product, purchase engineer
survey on industrial market research

Low cost, convenient, quick, but affected by experts judgment


Quota Sampling
A two-stage restricted Judgmental sampling
1. Develop control category/quotas (can be based on sex, age, race etc)
2. Sample elements are selected based on convenience or judgment

Eg: Categories/Quota based on age. Population =10,000


Sample to be taken =100
Quota: 18-30, 31-45, 46-60, over 60
From the age group of18-30, we will select 25 people
Snowball Sampling
Here through reference we select the respondents
Initial gp selected at random (probability) & they are asked for
referrals (nonprobability)

Eg: Select a group who purchased HP laptop recently.


We ask: which friend or relative or person in your area purchased HP
laptop in the last 12 months?

More use in industrial research (buyer-seller pairs)


2.Probability Sampling
(a) Simple Random Sampling
Every individual or item from the population has an equal chance of
being selected in the sample

Each possible sample have an equal probability of being picked

Simple to use, but may not be a good representation of the


populations underlying characteristics

How to do random sampling

-Assign a random digit & select


-Write names on a slip, put in a box & select
(b) Systematic Sampling
Elements are selected at a uniform interval that is measured in time, order
or space.

Each element has an equal chance of being selected but each sample does
not have an equal chance of being selected

Have error

Less time & cost than simple R S

Eg: Interview every 20th student in the institute


Solid paper waste in 100 houses every Monday
(b) Systematic Sampling
Purchase orders for the previous fiscal year are serialized 1 to 10,000 (N =
10,000).

A sample of fifty (n = 50) purchases orders is needed for an audit.

k = 10,000/50 = 200 = (population size/sample size)

First sample element randomly selected from the first 200 purchase
orders. Assume the 45th purchase order was selected.

Subsequent sample elements: 245, 445, 645, . . .


(C) Stratified Sampling
Divide population into two or more subgroups (called strata) according to
some common characteristic

A simple random sample is selected from each subgroup, with sample sizes
proportional to strata sizes
or draw equal size from each strata and give wt. based on stratas
proportion to total population
Samples from subgroups are combined into one

Strata are relatively homogeneous (within it) & Strata are mutually
exclusive and collectively exhaustive (among them)

Eg: divide based on age group or SECs

However sample elements selection is probabilistic as compared to quota


sampling where it is Convenience or Judgment based
(D) Cluster Sampling
Population is divided into several clusters, each representative of the
population

A simple random sample of clusters is selected


All items in the selected clusters can be used, or items can be chosen from
a cluster using another probability sampling technique

Eg: To find avg. no. of TV sets/house in a city


Divide city into blocks and perform sampling
(D) Cluster Sampling

Stratified :use Cluster :use

When each group has small variation Opposite case-


within itself but there is a wide variation When there is a considerable variation
between the groups within each group but the groups are
essentially similar to each other
Variation
Within=less Within=more
Among=more Among=less
Eg: divide consumers based on age group Eg: divide city by blocks
Other Probability Sampling Technique
Sequential (sampling some elements, then analyzing and deciding upon
whether to sample additional population elements or not)

Double (certain elements sampled twice): 1st a sample is analyzed,


then a subsample from it again analyzed
Eg: households in Bhubaneswar is sampled; then who used vacuum
cleaner among them sampled
Measurement & Scaling
Scales of measurement
Nominal Scale
Numbers identify & classify objects
Objects/Classes are mutually exclusive & collectively exhaustive
Eg: Sex, No. assigned to runners, store types
Number dont reflect amount of characteristics

Ordinal Scale
Ranking scale where numbers show relative position of objects, but
not the magnitude of difference between them
Eg: Rank of runners after finishing the run, quality ranking, preference
ranking, market position, social class
x> or <y
1-2 # 2-3
Scales of measurement
Interval Scale
Numbers are used to rate objects
Equal distance on scale represent equal distance in characteristics
being measured
1-2 = 2-3; Zero point is arbitrary
Eg: Temperature (0C), attitude, opinion

Ratio Scale
Absolute zero point; have all characteristics of other scales
You can classify, rank, compare
Eg: height, weight, money, age, costs, sales, no. of customers
Scaling Techniques
Scaling Techniques

Comparative scales Noncomparative scales

Continuous Itemized
Paired Rank Constant
Q-Sort rating rating
comparison order sum
scales scales

Semantic
Likert Stapel
Differential
Comparative scales
Direct comparison of stimulus objects, Eg: you like Coke or Pepsi
Ordinal data results

Paired comparison: presented with 2 objects & asked to select one based on
some criteria, Eg: I prefer Colgate more than Babool

Rank order: presented with several objects simultaneously and rank based on
some criteria, Eg: Rank based on your preference-colgate, pepsodent, babool,
sach, sensodyne, dabur-red

Constant sum: allocate a sum of units (points, dollars,chips) among set of objects
Eg: allocate 100 points to attributes of a toothpaste based on importance (mildness,
foam, price, packaging, taste, germ-fight)

Q-Sort : use rank order to sort objects (piles) based on similarity based on
some criteria, Eg: 100 brands put in 11 piles
Noncomparative scales
Each object is scaled independently
It is a metric or monadic scale, which results in interval or ratio data
Eg: rate Coke in 1-6 preference scale
Most widely used in MR

Continuous rating scales: graphical rating


-rate objects by placing a mark at appropriate position
Eg: How you rate Pantaloon departmental stores?
Probably the worstI..Probably the best
0 25 50 75 100

Likert scales: to show degree of agreement with each statement about a object
1=strongly disagree
2=disagree
3=neither agree nor disagree
4=agree
5=strongly agree
Noncomparative scales
Semantic Differential scales: end points associated with bipolar labels of
semantic meanings
Modern 1 2 3 4 5 6 7 Old-fashioned

Staple scales: unipolar rating scale with 10 categories number from -5 to +5


without a neutral point (zero)
-5-1 High quality +1 +5
-5.-1 Poor Service +1+5
Questionnaire and
Form design
Questionnaires
A formalized set of questions for obtaining information from
respondents

Survey method: Telephone, Personal, Electronic, Mail


Questionnaires
Open-ended questions
Completely unstructured:
What is your opinion about Bingo mad angle?

Word association:
Which is the first word that comes to your mind when you hear ITC?

Sentence completion:
I like Bingo, because.
Story Completion

Picture: fill the empty balloon

Thematic apperception test (TPT): shown a picture & asked what


comes to their mind.
Close-ended questions
Structured questions with set of response alternatives
Dichotomous: yes/no
Multiple choice
Likert scale: agreement
Semantic differential
..
Questionnaire Wording
Define the Issue
Use Ordinary Words & Avoid use of Unambiguous Words: usually, often,
sometimes
Avoid leading or biasing questions
-Leading Question: question is phrased in such a manner that it appears that
researcher expect certain answer

Avoid Implicit Alternatives


-Alternative that is not explicitly expressed
Eg: Do you go to office by your own car?
Correct: Do you go to office by your own car or use public transport?

Avoid Implicit Assumptions


Eg: Do you like investment?
Correct: Do you like to invest in Reliances stock?

Avoid Generalization and estimates


Eg: What is your annual per capita grocery expenditure?
Correct: What is the monthly/weekly expenditure on grocery in your household?
Close-ended questions-requirements
All will understand & Same meaning to all respondent

Response set: Respondent answers all question in 1 direction


Solution: change format, sequence, avoid lumping questions of
same topic, reverse code

Social desirability

Emotionally loaded questions

Threatening questions: embarrassing & difficult to answer

Double-barreled questions
Other facts: questionnaire
Filter question: asked before the questions themselves?
-To check whether the respondent meets the requirement criteria

Pretest

Scale evaluation
-Reliability
-Validity
Thank You
References
Malhotra, N.K. and Dash, S. (2013). Marketing Research: An Applied
Orientation, Pearson Education.

R.I. Levin and D.S. Rubin (2011). Statistics for Management (SFM), Pearson
Education, New Delhi.

Das könnte Ihnen auch gefallen