0% found this document useful (0 votes)
249 views134 pages

Notes From Introduction To Market Research - Naresh Malhotra

Market research involves systematically collecting, analyzing, and reporting information to assist with marketing decisions. It helps identify opportunities and problems, evaluate marketing actions, and improve understanding of the marketing process. The key steps in market research are: 1) Defining the problem through discussions with decision-makers to understand their needs. 2) Developing an approach such as qualitative research or a literature review to learn more about the problem. 3) Designing the research which details how information will be collected and analyzed. 4) Collecting data through methods like surveys, experiments, or observations. 5) Analyzing the results and communicating findings to decision-makers. Properly defining the problem is important
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
249 views134 pages

Notes From Introduction To Market Research - Naresh Malhotra

Market research involves systematically collecting, analyzing, and reporting information to assist with marketing decisions. It helps identify opportunities and problems, evaluate marketing actions, and improve understanding of the marketing process. The key steps in market research are: 1) Defining the problem through discussions with decision-makers to understand their needs. 2) Developing an approach such as qualitative research or a literature review to learn more about the problem. 3) Designing the research which details how information will be collected and analyzed. 4) Collecting data through methods like surveys, experiments, or observations. 5) Analyzing the results and communicating findings to decision-makers. Properly defining the problem is important
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 134

Introduction to Market Research

Notes for chapter 1 to 12.


Market Research

Why? How?
Market research is useful in learning more about your By specifying information required to address the
business, your audience, and your industry so as to… specific issue
• identify and define marketing opportunities, Design methods for collecting information
problems;
Manage and implement the data collection process
• Generate, refine and evaluate marketing actions
• Monitor marketing actions Analyze result and communicate the finding & its
• Improve understanding of the marketing process. implication
Market Research definition
• It is a process of systematic • R cognize info needs
and objective • E ffective decision making
identification, collection,
• S ystematic and Objective
analysis, dissemination
and use of information for • E xude/ disseminate info
the purpose of assisting • A nalysis of info
management in decision • R ecommendations for action
making related to
identification and • C ollect information
solutions of problems and • H elpful for Decision makers
opportunity in marketing.
Types of
market
research
Types of
market
research
Market research services and suppliers
Role of
market
research in
Marketing
Questions:

What decisions are made


Describe the tasks of by marketing managers? Define marketing
market research. How does research help research.
make these decisions?

Describe any one How should the decision


Describe the steps in
classification of to conduct market
marketing research.
marketing research. research be taken

What are the limitation


What is a syndicated of using social media for
research? conducting marketing
research?
Step 1 Problem definition

The 6 steps Step 2 Development of an approach to Problem

of Step 3 Research Design Formulation

marketing Step 4 Fieldwork or Data collection

research Step 5 Data Preparation and Analysis

Step 6 Report Preparation and Presentation


Ethical issues
in Marketing
research
Problem
definition and
development
of an approach
• Why is problem definition important?
• Gives a general statement of the problem and helps
in identifying the components of the market
research problem.
Problem • Ensures that the time and resources spent on MR is
definition and better utilized.
development of
an approach
• Discussion with decision makers:
• Problem Audit:
Problem • Events that led to the decision that action is
definition and needed.
• Alternative course of action
development • Criteria that will be useful to evaluate alternate
course of action
of an • Potential expected actions likely to emerge from
the research
approach • How much information to conclude on the
potential actions from research or alternate action
is available?
• How are the decision makers likely to use the
research information
• Decision making process in the organization.
• Prerequisites: Communication, Cooperation,
Confidence, Candor, Closeness and Creativity.
As a research buyer….

• One of the marketing team comes to


you and says

• You need to evaluate the request

14
Requests like this should ring alarm bells – why?
• What questions would you ask to help you evaluate the request?

What decisions have How confident are you about the How did you get to this
already been made? decisions you’ve already made point?

What decisions do you When do you need to make


need to make? these decisions?

What actions are you What would you do if you What’s the risk if you ‘got it wrong’
considering ? couldn’t do any research? at this stage?

15
Problem definition and development of an
approach

• Interview with Industry experts/ Key informant technique/ Lead user research:
• Particularly useful when:
• Little is known about the problem or issues
• Difficult to locate and obtain help from experts
• Secondary research
• Using Data collected for some other purpose than the problem at hand.
• Economical and is a quick source of background information.
• Qualitative survey
• Pilot survey
• Case studies
Problem definition and development of an
approach

• Environmental context: Set of factors that are impacted by the problem under
consideration for Market Research.
• Factors to be considered for understanding environmental context
• P ast information and forecast
• R esource and constrain
• O bjectives: Organizational, Departmental, Transactional, Individual
• B uyer behavior:
• demographic, usage and purchase behavior, preferences, media habits, price
sensitivity, distribution,
• L egal information: Regulatory norms
• E conomic Environment: Economic indicators
• M arketing and Technology skills:
Allow researcher to obtain all the info needed to
Characteristics address the management decision problem.
of Market
Research
Problem Guide the researcher in proceeding with the
project.

Watch outs?
Biases due to ones own Limited focus on problem
cultural values and beliefs. definition.
Developing an approach to MR problem

1. Component of the Approach:


1. Objective/ Theoretical framework: e.g. trilobal modal, AIDA, STDC,
etc.
2. Analytical model:
1. Verbal Model: AIDA, Trilobal, STDC etc.
2. Mathematical model: Y=a+bx
3. Graphical model
These models are complementary and must guide researchers on
identifying relevant research questions or forming hypothesis.
Research Question & Hypothesis

Research question: Hypothesis:


Refined statement of specific components of the An unproven statement or proposition about the
problem? factor or phenomenon that is of interest to the
Reflecting upon the components of Marketing researcher,
research problem using Objectives and Reflecting upon the Research question using
Theoretical framework can help form the Analytical model can help develop hypothesis.
research question.
Questions:
What are some reasons
What are the steps in Why is it important to What is the role of the
why the management is
Marketing research define MR problem researcher in problem
often not clear about
project? appropriately? definition process?
the problem?

What is the key What are the key


difference between difference between What are the types of
What is problem audit? Symptom and Problem Management Decision error in determining the
in the context of this Problem and Marketing Research problem?
chapter? Research Problem

What are the key


How are research Is it necessary for every
difference between What are the different
questions related to the research to have an
Management Decision types of analytical
components of the hypothesis? Why and
Problem and Market models
problem? why not?
Research problem
Research Design

• The research design is like an


complete SOP of the research
task. It details all elements of
the research.
The high level classification of the research
• General guidelines for choosing research designs:
• 1) If little is known about the problem situation – start
with exploratory research (general initial step)
• Exploratory research is appropriate for the following:
• a) When nature of topic cannot be measured in
structured, quantifiable manner
• b) When problem needs to be defined more precisely
• c) When alternative courses of action need to be
identified
• d) When research questions or hypotheses need to
be developed
• e) When key variables need to be isolated and
classified as dependent or independent

• 2) Exploratory research followed by descriptive or causal


research
• 3) Research design can also begin with descriptive or
causal research
• 4) Exploratory research may follow descriptive or causal
research
Exploratory Conclusive
To provide insights and
TO test specific hypotheses and
understanding of the nature of
Objectives examine relationships
marketing phenomena
Comparing To understand
Information needed may be loosely
To measure
Information needed is clearly

Exploratory defined.
Research process is flexible,
Characteristics unstructured and may evolve.
defined.
Research process is formal and
structured.
& Conclusive Samples are small.
Data analysis can be qualitative or
Sample is large and aims to be
representative.

research quantitative
Can be used in their own right.
Data analysis is quantitative.
Can be used in their own right.
May feed into exploratory
May feed into conclusive research.
Findings/results research.
May illuminate specific conclusive
May set a context to exploratory
findings.
findings.
Expert surveys
Surveys
Pilot surveys
Secondary data
Secondary data
Databases
Methods Qualitative interviews
Panels
Unstructured observations
Structured observations
Quantitative exploratory
Experiments
multivariate methods
Potential sources of error in
Research
Let’s step back a little How good
is my survey

…and they answered our


Well, we talked to some people… questions…

....but they were only a sample of


people….
…so how sure can we be about
generalising from this sample to the
real world?
Does the
survey
Sampling plan have high
external
validity?

5
How good
is my survey

…and they answered our


Well, we talked to some people… questions…

…but can we really believe what they


say?...
... are we really measuring what we think
we are measuring?

Does the survey Major design


have high internal decisions
validity?

6
Major Design Decision

Secondary Research Primary Research

Qualitative Quantitative

29
Basic differences

Qualitative research Quantitative research


Involves flexible Involves a structured questionnaire
discussions or measurement tool

What: Focuses on exploring Focuses on measuring

Why: Priority is depth or detail of Priority is robust representation


understanding of the ‘population of interest’
Smaller numbers of respondents Larger numbers of respondents
Who: sampled to reflect key characteristics sampled to be representative
Analysis of transcripts/videos Analysis by adding answers
How: – themes, comments – data tables, %’s
Greater internal validity Greater external validity

30
Most ‘issues’ would benefit from qualitative’ AND ‘quantitative’ inputs
• …as they provide different ‘types’ of understanding
Design decision: Which first?
First stage Second stage OR First stage Second stage
Qualitative Quantitative Quantitative Qualitative

e.g. To explore e.g. To measure how e.g. To assess the e.g. To provide
what the issues many people in the overall appeal of a detailed input from
are and the population hold new idea – and to the types of people
language people these attitudes identify the types identified as most
use to help design or behaviours – using of people who are interested to guide
a quantitative relevant language most interested further development
questionnaire
Flexible early stage Establishing which Broad perspective Focus and depth to
to avoid incorrect issues are most to identify biggest provide detailed
assumptions common/important opportunities recommendations 31
But often the budget won’t stretch to doing both ‘qual’ and ‘quant’ so a
choice must be made
• …accepting compromise in some areas
Qualitative OR Quantitative
Especially if: Especially if:
▪ investigating habitual behaviour or ▪ greater confidence is needed in the validity
subconscious attitudes where people need of results i.e. in generalisation to the
help to articulate why they do what they do ‘population’
▪ generating ideas e.g. for new products or e.g. high risk decisions
improvements as people need ▪ tracking key measures over time where
encouragement and time to be creative greater objectivity and consistency are
▪ the target respondents are hard to find/only needed
a small population e.g. finance directors or ▪ identifying the most important groups from
doctors a broad population is more important than
depth of understanding

32
A biscuit manufacturer has developed a range of
biscuit snacks designed as a nutritional and tasty
“energy boost for people on the go”. The snacks have
been produced in 7 different flavours. They now want
to commission research to help refine the range
further, and to gather information to inform a
Major Design marketing campaign.
Decision:
Qualitative If there was only budget for one stage of research
or what are the advantages of each approach?
Quantitative?

A qualitative approach

A quantitative approach

33
Based on MRS Advanced Certificate exam question
Exam revision:
Generic Summary of Qualitative
pros and cons

• Flexible – good for exploring (so good as early ▪ Moderator has a big influence on the
stage) outcome – increases the subjective
• Helps with quant design – q’aire and sample element – less easy to replicate
(?Reliability)
• Depth – subconscious/irrational/habitual – i.e.
if need help to articulate ‘why’ (internal
validity) ▪ The interview changes the respondent –
• Depth – detail/complex (esp. if 1:1 depth possibly now thinking in a way they don’t
interview) & richer context normally think - more sensitive to small
things/more considered than in real world
• Creative (especially if groups) - good for idea (?realism ?external validity)
generation
▪ Small samples – respondents selected to
reflect likely range of opinions rather than
• Cost effective (small scale)
• Practical – (esp. if hard-to-find people) to be numerically representative (?
generalise to population ?external validity)

• Easy for clients to watch first hand ▪ Clients ‘meddle’/don’t see the full
picture/more subjective interpretation
34
Exam revision:
Generic Summary of
pros and cons
Quantitative

• Structured – all interviews conducted in the ▪ Need to know more upfront to be able to create a
same way – fewer subjective elements. sensible questionnaire – risk of using wrong
Easier to replicate (greater reliability) language/missing important aspects
▪ Limited depth – hard to get at
• Good for tapping into top of mind thinking – subconscious/irrational and for subjects people
so can be more realistic of everyday find hard to articulate (?internal validity)
decision-making processes ▪ Limited depth – detail/complex/context – long
questionnaires would be needed and motivation is
likely to be low ( poor quality answers (?internal
• Bigger samples – objectively selected to be
representative so greater ability to validity)
generalise to the target population (better ▪ Not a good creative environment – low motivation
external validity)
▪ Often higher overall costs
▪ Impractical/expensive for hard-to-find/low
incidence/unwilling/busy targets 35
Exam revision:
Applying points to the Major Design Decision: Qualitative or
context
Quantitative?
A biscuit manufacturer has developed a range of biscuit snacks designed as a nutritional and tasty “energy boost for people on the go”.
The snacks have been produced in 7 different flavours. They now want to commission research to help refine the range further and to
gather information to inform a marketing campaign.

If there was only budget for one stage of research what are the advantages of each approach?

Qualitative Quantitative
A very competitive market with a wide range of products This is a major fmcg market so its likely there is quite
‘biscuit available – qual may be better at really understanding the a lot of secondary info on the biscuit market to help
manufacturer level of enthusiasm behind any interest and probing guide development of a questionnaire – i.e. it might
whether they would really be prepared to give up something not be too risky going for quant.
‘ from their existing repertoire to make way for this product
(nutritional ) What evidence have they used to get to this stage? If this concept What evidence have they used to get to this stage?
idea is not based on particularly strong evidence - and if it If they don’t have any profile information on
and tasty ‘people on the go’ it will be hard to decide what
possible to change elements of the concept (there are several
(“energy boost) ideas contained within it) – then it could be risky doing quant as types of people to do the qual amongst .
for (people on the reaction to the products themselves will be influenced by Quantitative is better at getting to a broad sample
the go)” expectations set up by the concept. In qual they would have more to help identify the profile of the core target
flexibility to explore the concept idea first – and to understand
and separate reaction to concept v reaction to product.

Based on MRS Advanced Certificate exam question 36


Continued…
Exam revision:
Applying points to the Major Design Decision: Qualitative or
context
Quantitative?

Qualitative Quantitative
(nutritional )and If elements of the concept can be changed e.g. If they are fairly confident in the decisions made to get to
tasty (“energy including what other flavours might be appealing - this point then quant will provide a more robust assessment
they might need more creative input – qual is better of appeal that can be generalised to the population
boost) for (people
at idea generation than quant
on the go)”
If refine the range means eliminating any clear losers However it they are trying to find the best one or two flavours
‘7 different among the flavours then qual would be a cost effective e.g. because they would be unlikely to get shelf space in retailers
way of doing this as even though it is small scale it is for more than one or two – then it would be more important to
flavours’
usually possible to identify any with no merit get a more robust assessment with a broader sample.
‘refine the range
further’ If they can make changes to the product formulations If there has been lots of internal development of the products
. and flavours then more specific detailed feedback on already and there is a feeling that these represent the best version
likes and dislikes would be useful. Often people are of each flavour – and the priority is to ‘refine’ them by identifying
not great at describing why they like/dislike which are ‘good enough to launch ’ – rather than to input to
something and need help to articulate this – qual is further development – then the greater external validity of quant
good at probing and exploring to help understanding would be more suited
to give greater internal validity.

Based on MRS Advanced Certificate exam question 37


Continued…
Exam revision:
Applying points to the Major Design Decision: Qualitative or
context
Quantitative?

Qualitative Quantitative
(nutritional )and If elements of the concept can be changed
If they are fairly confident in the decisions made to get
tasty (“energy e.g. including what other flavours might be
to this point then quant will provide a more robust
appealing - they might need more creative
boost) for (people input – qual is better at idea generation than
assessment of appeal that can be generalised to the
on the go)” population
quant

inform a If they are looking to fine tune materials If the priority is to understand which
qual would be better at providing specific groups to target media at then the
marketing broad reach of quant would be
detailed recommendations on which bits
campaign. work well/need improving, tone, language betterr at providing this
etc

Based on MRS Advanced Certificate exam question 38


Continued…
Reality check – quantitative research

People People are much


forget better at ‘what’ than
‘why’

And we can’t control And about ‘past’


what they forget rather than ‘future

39
One-off studies only give ‘a snapshot’ of performance

*
June

*
*
June Sep

40
Comparing results
• More valid if the same…. • Need to consider external factors….

• Sample definition • Market situation


• Sample profile • Marketing activity
• Sampling method
• Economic climate
• Mode of interview • Topical events
• (interviewer quality)
• Date/time of interview • Weather

• Length of interview
• Surrounding questionnaire
content
• Precise questionnaire wording
41
‘Repeat’ survey’s help to understand change
better….
‘Tracking studies’

* *
*
*
*
*
* *
June *
Jun Jul Aug Sep Oct Nov Dec Jan

* The frequency of tracking will


* vary according to the market
– e.g. speed of change, amount
of activity etc. (...and budget)
June Sep

42
An alternative to a tracking study is to set up a longitudinal survey
• – the same people are re-contacted again and again and are asked the same questions

• Longitudinal surveys • But…. Longitudinal surveys are fairly


can monitor more rare and are best used when the
▪ The questions we ask might ‘information gathering task’ is
directly how people influence them e.g. make them
change as they automated in some way.
more aware of issues so they
grow and as the change because of the research For example kantar worldpanel
market changes.
▪ They might decide to drop out
so
the profile of the sample starts Is a longitudinal panel of
to change unless they are respondents who ‘scan’ all their
replaced shopping when they get home
and download the data to us – we
▪ They are expensive to manage
aggregate it to provide market
and maintain
share data

43
Longitudinal panels play a key part in the marketing information chain

The Client knows how much And can keep track of how much it
product it has made has sold to retailers
But it doesn’t know….

1. How much is sold to customers? 1. Retail Audit panels

2. Who is buying it? 2. Purchasing panels

3. Who is using it? 3. Usage panels

4. Where to advertise it? 4. Audience panels


And they’ll have a lot of other specific
Ad Hoc (Qual/Quant)
questions to help them decide how to sell
more
Note: Nowadays the term ‘panel’ can be ambiguous

Longitudinal An ‘access’
PANEL PANEL

Large database of people

Different people selected for cross sectional


studies
Same people i.e.
Same information collected Different people
Different information collected
-again and again
- each time
45
A local council has introduced a new scheme for
recycling waste.
It wishes to monitor the ongoing success of the
scheme using quantitative methods.
Three potential approaches have been
suggested – there is a fixed budget.
Fundamental
design What would be reasons for
recommending/advantages of each of these
approaches?
decisions i. An annual study
ii. Monthly study
iii. Longitudinal panel providing continuous,
diary-based data collection

(scenario is based on a question used in the MRS Advanced Certificate exam)


46
Exam revision: Basic/generic points

• Annual • Monthly tracking • Longitudinal panel

• Cheapest • Easier to interpret – look at • Better understanding of how


trends, link data changes changes affect individuals as
more closely to activities the same people throughout
• Better memory for shorter • More (accurate) detail can be
period collected closer to point of it
happening

◼ Harder to interpret as a lot could ◼ More expensive ◼ Even more expensive


have happened within the year
◼ More reliance on memory
◼ Point in time chosen for annual
study could be critical if issue
affected by seasonality
47
Exam revision:
Points showing a deeper
The budget is likely to be set ...which brings in other
appreciation of trade-offs considerations

• Annual • Monthly tracking • Longitudinal panel


Less project management costs so can afford: More project management costs and regular • Even more project management
analysis costs costs. More complex analysis
Bigger sample
costs
• More robust Small monthly samples
• Allows more detailed subgroup analysis • Data may bounce around due to sample • Smallest sample
(e.g. different area within the council variance each month • Limits subgroup analysis
geography to see if recycling services are • Limits subgroup analysis
consistent for all
Poorer quality sample
Better quality sample • May be limited to more cost effective
• Potentially better quality sample if face to interviewing methods e.g. street or online
face in home affordable • Harder to match profiles each month
• Easier to match profile year on year than
month to month Shorter questionnaire
• Ok if only looking at a few key measures but less
Longer questionnaire detail/context possible
• Slightly longer questionnaire possible–
which would be beneficial if more context Less input/servicing from agency
about the individual is required • Field & tab only (but Ok if simple?)
More input/servicing from agency
• Get them to present/write report
Exam revision: Applying points to the context

• A local council has introduced a new scheme for recycling waste. It wishes to monitor the
ongoing success of the scheme using quantitative methods
• local council has introduced a new scheme
• Councils tend to move slowly in their decision making – and perhaps the scheme involves contracts for a set period?: if
they can’t take quick actions to amend the scheme then there’s less benefit in monthly or panel – annual might be
better use of resources
• Councils have to be accountable to their residents: if the data is to be put in the public domain/form the basis for
spending then there is likely to be greater emphasis on robustness – pointing to larger sample sizes and more inclusive
data collection methods (e.g. face to face rather than on line). Annual studies are more likely to be able to achieve this
within budget.
• Success …for a recycling waste scheme
• How success is defined might point to different approaches:
• Increased awareness? – e.g. effectiveness of comms to raise awareness. How is this building? Can rises be attributed to
specific activities? Monthly tracking would offer an advantage over annual (and longitudinal no good as can’t keep
repeating an awareness Q – it will prompt them to become more aware!)
• More positive attitudes? – e.g. shift in opinion of council’s approach. Attitudes change more slowly – unlikely to see
reliable month to month changes so annual would be more sensible in this instance (and longitudinal not suitable for
frequent measurement of attitudes as they will start to think differently the more you ask).
• Increased usage – seasonality is likely in recycling - so advantage for monthly. Longitudinal panel with a diary could put a
large load on the respondent – unlikely to get full cooperation from those not passionate about recycling i.e. bias sample
and poor measure on overall recycling level. However – if success is measured in terms of committed recyclers being able
to recycle a wide range of items then longitudinal panel could be the better approach 49
Omnibus
• A syndicated fieldwork vehicle – different clients can ‘buy space’ for their questions*

• Cost effective if you only have a few questions to ask (typically £250-800 per question)
• Runs regularly – typically at least weekly
• Fast turnaround – project management processes are in place and don’t need to be set up
• Larger sample size/more robust sampling than a client might be able to afford on their own (typically
1000 – 2000)

Usually nationally representative - can use part samples to hit a particular group e.g. men in London (Can
run over several waves to build sample sizes of low incidence/subgroups)

• Can be face to face, on line or telephone

*data from their questions is confidential to that client 50


When would you use an omnibus?
• If only a few key metrics are required but these need to be measured regularly e.g. advertising or brand awareness
tracking
• For evidence to support claims in advertising/marketing – access to a better sample than the client could afford on their
own e.g. a large face to face omnibus which uses random sampling would hold up to critical scrutiny better than an online
panel
• When other cost effective methods (like online panels) are likely to have coverage bias that affect the results in the subject
area - e.g. an online panel is likely to overread measures linked to online familiarity/behaviour whereas a face to face
omnibus would provide a more representative read of online and offline behaviours. An omnibus could be used in
conjunction with an online approach e.g. using an online panel as a cost effective vehicle for accessing a particular user
group and asking detailed questions – and then getting a more realistic read of actual incidence in the population via a
face to face omnibus to help with grossing up to market size.
• To get results quickly – e.g. to get a quick read on the impact of topical events such as negative publicity in the media (e.g.
reports of food contamination scare etc.) to help determine whether a full scale PR/action plan is needed
• To provide input and guidance to other research stages
• e.g. if estimates of incidence and profile of users are needed to help cost/assess feasibility, or to help design sample
structure/quotas/weighting
• e.g. to create a prompted list for a larger study but you don’t know the range of likely answers – an open ended question
on the Omnibus could be coded up to create a list for use elsewhere
• e.g. to pilot a question or alternative versions e.g. to decide how many points to use on a scale
Limitations of an omnibus
There will be a mix of topics in the survey and some may influence later topic areas
• e.g. questions about attitudes to healthy eating may influence later thought processes for a set of
questions about a snack bran.
Quality of data tends to deteriorate through the interview due to boredom and order
effects
• Questions sets that come nearer the start are likely to be better ‘protected’ from these effects

If there is an especially high demand the length of the questionnaire may become long
• i.e. exacerbating the above issues – and also impacting on sample quality as the type of people willing
to take part in a long survey may be different to others

The above effects could be especially problematic for tracking surveys


• i.e. they rely on comparability across waves to be able to link any changes to market effects rather than
changes in the survey method
*data from their questions is confidential to that client 52
• 2.
• Experimental designs
• e.g. when testing new concepts, products, ads, comms materials

53
Scenario: Our client has developed an ad – it is due to go on air May-June.
They want to measure the effectiveness of the ad campaign e.g. to aid decisions about the level of media spend for a
subsequent burst later in the year

They could do a survey after the first ad burst.- a key decision will be how soon after the ad burst?
- Straight away? good for getting the ‘maximum’ read e.g. of awareness
- After a pause? giving a better feel for longer term impact


* *
*
March May - June July August Sept-Oct
Ad on air Ad on air
Post campaign
evaluation

54
If they only take a ‘post’ campaign read there will be a lack of certainty about how much measures have improved

i.e. they will have no direct comparison and will have to rely on respondents’ claimed data

A ‘pre wave’ would provide a more objective point of comparison


- but it will be more costly to conduct two waves/or if the budget is fixed it will mean smaller sample sizes for
each wave


* *
* *
March May - June July August Sept-Oct
Ad on air Ad on air
Pre campaign Post campaign
benchmark evaluation

55
An additional read mid-campaign would improve understanding of how awareness has built

More regular reads (‘tracking’) would improve understanding further and might be especially useful if:

- immediate changes to media buying are possible


- there are different ads in the campaign: so any movements could be more directly attributed to a particular ad

*

* * *
* *
March May - June July August Sept-Oct
Ad on air Ad on air
Pre campaign Mid campaign Post campaign
benchmark read of build evaluation

56
Often media activity is accompanied by other instore activity e.g. promotional offers, instore merchandising. To help
separate the effect of this (and reduce reliance on the respondents’ ability to identify the source of their awareness or
perceptions a ‘control’ element could be included

i.e. a pre and post survey in regions where there was no TV advertising.

*

* *
* *
* *
March May - June July August Sept-Oct
Ad on air Ad on air
Pre campaign Mid campaign Post campaign
benchmark read of build evaluation

Pre Campaign Post Campaign


Control benchmark Control evaluation
– TV regions not – TV regions not
showing ad showing ad

57
Is this a good product?
% Is that
“good enough”?
Excellent 5
Very good 18 76% rate it positively
Good 25
Fairly good 28 Median
Neither good nor poor 19
Slightly poor 3
5% rate it negatively
Poor 2

58
We must always be thinking ahead to how we will interpret
and use the results
Often a point of comparison is needed
e.g. scores from other studies conducted in a similar way.

A key selling point of many agencies unique ‘solutions’ is their database of


previous results that provide ‘norms’.

59
If a respondents
3 sees more than one concept (ad/product) their
frame of reference will change during the interview

• See first concept • See second concept • See third concept

Evaluate v ‘memory’ Evaluate v memory + Evaluate v memory +


recent experience 1 recent experiences 1 and 2

Frame of Reference Frame of Reference Frame of Reference


Previous experience New factors taken into More new factors taken
account into account

Respondents tend to become more critical as the test progresses

27
Design options
Monadic design Sequential monadic design Proto monadic design Comparative design

Cell 1 Cell 2 Cell 1 Cell 2 Cell 1 Cell 2 Cell 1 Cell 2

See A See B See A See B See A See B See A See B

See B See A
Ratings Ratings Ratings Ratings Ratings Ratings
Preference Preference
See B See A See B See A
Preference Preference
Ratings Ratings only only

Preference Preference

The choice involves weighing the benefits of


- quality of information v fuller information
- and greater sensitivity v greater realism

61
Monadic Testing
Advantages…
• Recreates the ‘real’ world best – i.e. people tend to use one When used…
product at a time. Major (rather than subtle) differences in products
• Results are not influenced by other the products/concepts etc or concepts
• Can explore more depth behind reaction to products within a Performance needs to be assessed in absolute
sensible q’aire length sense
When need to be closest to realism
Disadvantages… e.g – go/no go before launch
• Not very ‘sensitive’ as no direct comparison e.g. – is it worth it? Will it really be
• Hence only major differences emerge noticed in the real world?
• Bigger sample sizes needed as sample variation within each When need to compare results with a
• Need ‘matched samples’ database of norms
• Cost implications

62
Pure Comparative Design

Advantages…
• Smaller samples than monadic as same person tries both so any sample ‘wobble’
affects both the same
• Cost savings
• Probably the most sensitive to differences
• Respondent preference is not influenced by any questions (i.e. we’re not directing
them to consider any specific elements/characteristics before giving their preference)

Disadvantages...
• Only identifies the best out of 2 concepts. Doesn’t tell us how good (or bad) the best
one is
• In reality they may be unlikely to ever try products side by side so it could lack realism
Sequential Monadic Design
• The data collected in the second test is not truly monadic as it is not being evaluated in isolation –
and will be compared with the previous product/concept.
• Questionnaire becomes long – and respondent starts thinking as a ‘tester’ rather than a
‘consumer’
• Hence, nowadays a protomonadic approach is often favoured…..
Protomonadic Design

Analysis…
Monadics for most realistic evaluation
Comparative data for preferences - to tease things apart if monadics are close

Advantages…
‘Best’ of both worlds? Kept shorter by not repeating specific ratings for second product/concept

Disadvantages?
Not quite as cost effective as fully comparative designs (not as short). Preference is not so pure – as
the questions asked for the first product/concept may have an affect on how the second is perceived
64
‘A concept test’ – what decisions have already been made?

• . • ‘Natural spring water from Britain’

We all want a healthier and more


natural lifestyle.
This is an untouched spring water that
• ‘Natural spring comes from Britain’s national parks,
water from from natural, protected, rural sources.
Britain’ Drinking this water everyday makes me
feel at one with nature and replenishes
me and my family with rural goodness

65
In product tests we also need to decide how to introduce the
product as this will affect the results
Product Led
(‘Try this tomato soup’) Gives pure product performance

Concept Led
(‘Try this refreshingly tangy Gives performance of product in light
low calorie tomato soup’) expectations set up by the concept

Total Mix
(Show ad, price, pack first) Gives performance in the most realistic
context
Q: Thinking just about the product itself how would you rate it ?

Blind –
plain packaging, no brand
name

Brand Leader Store Own


Brand

67
Respondents can’t ‘forget’ about the influence of the brand
Blind – Branded/Identified
plain packaging, no brand
name

Brand Store Own Brand Store Own


Leader Brand Leader Brand

68
Test location can be important for product testing

In Home?
respondent can use the product in the way they normally would, at a time they normally would

more ‘realistic’ reaction

Central location (‘hall’ test)?


we have more control over how the product is used

more ‘sensitive’ to differences as we have limited the number of possible varying factors
A key design consideration…

Do you want greater realism


or greater sensitivity?
Scenario 1: (dairy spreads)
• Brand A is the client’s major brand – it is the market leader in dairy spreads but its nearest
competitor is close behind

• The client has developed a new recipe ‘B’ for the brand that will deliver better profit
margins than current recipe A

High risk so need sensitivity –will


? Design? only introduce if confident it
won’t be worse
? In home or In hall? - Comparative
- In hall
? Blind or Branded?
- Blind
? Regular brand users or category users? - Regular brand users

71
Scenario 2: (Sausage casserole cook in sauce)
• Previous work indicated a weakness in Current Brand X’s product recipe v the main
competitor who is the clear market leader
• Many new recipes have been screened via small scale internal taste testing
• and New Recipe Y has been identified as the best option. If progressed it will involve
investment in changes in the production process that will impact on profits.
• New Recipe Z also represented an improvement over X. It would not involve changes in
the production process but in the internal taste tests it did not perform as well as New
Recipe Y
New Y will need investment
– is it really worth it? Will it close the ga
? Design? on the competition in real life?
Monadic
? In home or In hall? New Y, New Z & competitor
Branded – Concept led
? Blind or Branded?
‘new improved’
? Regular brand users or category users?
72
Introduction to Market Research

6 Steps in Market Research

• Step 1: Problem definition


• Step 2: Development of an approach to
So far… Problem
• Step 3: Research Design Formulation
• Step 4: Fieldwork or Data collection
• Step 5: Data Preparation and Analysis
• Step 6: Report Preparation and
Presentation
Measurement & Scaling
Levels of measurement of scales:

• Measurement: Assignment | pre-specified


rule
• Scaling: Continuum on which the measured
objects are located.
MEASUREMENT
& SCALING Scale characteristics

CONCEPTS • Descriptors
• Order: Relative size or position of the
descriptor
• Distance: The absolute difference between
scale descriptors
• Origin: fixed starting or true zero.
• Primary Scales of measurement :
• Nominal:
• Classification | Descriptor
• Ordinal:
MEASUREMENT • Relative position | Descriptor, Order
& SCALING • Interval:
CONCEPTS • Difference between position are comparable
but arbitrary zero point | Descriptor, Order,
Distance
• Ratio:
• has property of all above scale plus has
absolute zero point. Thus, can compute
ration of the scale value. | Descriptor, Order,
Distance, Origin
MEASUREMENT & SCALING CONCEPTS
MEASUREMENT & SCALING CONCEPTS

Chose preferred packaging option.


MEASUREMENT & SCALING CONCEPTS
Questionnaire design
What is a questionnaire?

A tool for collecting information to describe, compare, or explain an event or situation, as well
as, knowledge, attitudes, behaviours, and/or socio-demographic characteristics on a particular
target group.

The two main methods of collecting primary data are asking questions and observing.

Survey research ( in its various forms) relies on the use of a questionnaire.

A questionnaire is a document consisting of a set of questions and scales to gather primary


data.
81
Questionnaires 82

• The primary purpose of a questionnaire is measurement.


• Questionnaires can be used to measure customer’ s behaviour, attitudes,
awareness and characteristics.
• A questionnaire provides standardisation and uniformity in the data
gathering process.
• Standardisation of wording and sequencing of the ?ns allows field workers
to achieve speed and accuracy in recording the data.
Objectives of a questionnaire 83

• It must translate the information needed into


questions that the respondent can and will answer.
• It must motivate and encourage the respondent to co-
operate and complete the interview.
• It must minimise error.
• Questionnaire design is a major source of non-
sampling error.
• Respondent error:
• Non-response errors
• Response errors
Questionnaire Pre work Specify the Information Needed
Design Process
Specify the Type of Interviewing Method

Determine the Content of Individual Questions

Guidance/ check Design the Question to Overcome the Respondent’s


Inability and Unwillingness to Answer

Que type Decide the Question Structure

Creation content Determine the Question Wording

Flow Arrange the Questions in Proper Order

Identify the Form and Layout

Production Reproduce the Questionnaire

Test Eliminate Bugs by Pre-testing


Effect of Interviewing Method on Questionnaire
Design
Department Store Project
Mail Questionnaire
• Please rank order the following department stores by picking out the store that you like most and
assigning it a number 1. The second most preferred department store gets a number 2. Continue this
procedure. The least preferred store should be assigned a rank of 10. No two stores should receive
the same rank number.
Store Rank Order

1. Parisian ____________
2. Macy's ____________
.
.
10. Wal-Mart ____________
Effect of Interviewing Method on Questionnaire
Design
Telephone Questionnaire
• I will read to you the names of some department stores. Please rate them in terms of your
preference. Use a ten-point scale, where 1 denotes not so preferred and 10 denotes greatly
preferred. Again, please remember that the higher the number, the greater the degree of
preference. Now, please tell me your preference to shop at .......(READ ONE STORE AT A TIME)
Store Not So Greatly
Preferred Preferred
1. Parisian 1 2 3 4 5 6 7 8 9 10
2. Macy's 1 2 3 4 5 6 7 8 9 10
.
.
.
10.Wal-Mart 1 2 3 4 5 6 7 8 9 10
Effect of Interviewing Method on Questionnaire
Design
Personal Questionnaire
• (HAND DEPARTMENT STORE CARDS TO THE RESPONDENT). Each store name written on a separate
card. Please examine these cards carefully. (GIVE RESPONDENT TIME). Now, pull out that card
which has the name of the store you like the most, (RECORD THE STORE NAME AND KEEP THIS CARD
WITH YOU). Examine the remaining nine cards. What is your most preferred store? (REPEAT UNTIL
RESPONDENT HAS ONLY ONE CARD LEFT)

Store Rank Name of the Store


1. 1 __________________
2. 2 __________________
.
10. 10 __________________
Effect of Interviewing Method
on Questionnaire Design

Electronic Questionnaire
• This question for e-mail and Internet questionnaires will be very similar
to that for the mail questionnaire.
• In all these methods, the questionnaire is self-administered by the
respondent.
Individual Question Content ─ Are Several
Questions Needed ?

• Avoid double-barreled questions, because two or more questions are


combined into one. Instead, two or more distinct questions should be
asked:
Overcoming Inability To Answer – Is the
Respondent Informed?
• When not all respondents are informed about the topic, filter questions measuring
familiarity and past experience should be asked before questions about the topics
themselves.

• A “don't know” option appears to reduce uninformed responses without reducing


the response rate.

• Write the question in a way that does not tax the respondent’s memory too much
• The following question will be difficult for people to answer:
• Q: In the past six months, have you experienced any nausea or dizziness?
Reasonable time frames

• Use the right time frame for your question.

Event Reasonable time frame


Auto accident w/ injury Several years
Hospital stay, car purchase One year
Appliance purchase Three months
Grocery purchase Two weeks
Food consumption One day
Overcoming Inability To Answer – Can the
Respondent Articulate?

• Respondents may be unable to articulate certain types of responses,


e.g., describe the atmosphere of a department store.

• Respondents should be given aids, such as pictures, maps, and


descriptions to help them articulate their responses.
Overcoming Unwillingness To Answer
Effort
• If too much effort is required, people will not answer

Context
• Respondents are unwilling to respond to questions which they consider to be inappropriate for the given context.
• The researcher should manipulate the context so that the request for information seems appropriate.

Legitimate Purpose
• Explaining why the data are needed legitimizes request and increase the respondents' willingness to answer.

Sensitive Information
• Respondents are unwilling to disclose, at least accurately, sensitive information because of embarrassment, loss
of prestige or self-image.
Overcoming Unwillingness To Answer –
Increasing the Willingness of Respondents

• Place sensitive topics at the end of the questionnaire.


• Preface the question with a statement that the behavior of interest is
common.
• Ask the question using the third-person technique: phrase the question
as if it referred to other people.
• Hide the question in a group of other questions which respondents are
willing to answer.
• Provide response categories rather than asking for specific figures.
• Use randomized techniques.
Choosing Question Wording

• Define the issue


-Define the issue in terms of who, what, when, where, why, and
way (the six Ws). Who, what, when, and where are particularly important.
• Use ordinary words
-Avoid the use of technical terms
• Use unambiguous words
-Avoid words that are likely to have different meanings for different
people
• Avoid leading or biasing questions
Choosing Question Wording

• Avoid implicit alternatives


- An alternative that is not explicitly expressed in the options is an implicit alternative.

• Avoid implicit assumptions


- Questions should not be worded so that the answer is dependent upon implicit
assumptions about what will happen as a consequence.

• Avoid Generalizations and Estimates

• Dual Statements: Positive and Negative


- Questions that are in the form of statements should be worded both positively and
negatively
Choosing Question Structure – Unstructured/
Structured Questions

• Unstructured questions are open-ended questions that respondents answer in


their own words.
• What is your occupation?
• Who is your favorite actor?
• What do you think about people who shop at high-end department
stores?

• Structured questions specify the set of response alternatives and the response
format. A structured question may be multiple-choice, dichotomous, or a
scale.
Choosing Question Structure –
Multiple-Choice Questions

• Example of multiple-choice question

Do you intend to buy a new car within the next six months?
____ Definitely will not buy
____ Probably will not buy
____ Undecided
____ Probably will buy
____ Definitely will buy
____ Other (please specify)
Choosing Question Structure –
Dichotomous Questions

• A dichotomous question has only two response alternatives: yes or no,


agree or disagree, and so on.
• Often, the two alternatives of interest are supplemented by a neutral
alternative, such as “no opinion,” “don't know,” “both,” or “none.”

Do you intend to buy a new car within the next six months?
_____ Yes
_____ No
_____ Don't know
Choosing Question Structure –
Scales
• Example of a Scale:

Do you intend to buy a new car within the next six months?

Definitely Probably Undecided Probably Definitely


will not buy will not buy will buy will buy
1 2 3 4 5
Consideration for type of questions in research
instrument
Layers or response from
Access to respondent respondents Examples of techniques

• Spontaneous
Highly structured questionnaires (e.g. telephone
Public • Reasoned
surveys)
Communicab • Conventional
le
Questionnaire with a proportion of open-ended
Aware • Concealed
questions to allow probing (e.g. face-to-face
• Personal
surveys)
Semi-structured to unstructured interviews on an
• Intuitive
Private individual or group basis (e.g. focus group
• Imaginative
Non- discussion)
communicabl
e • Unconscious
Unaware Unstructured interviews and observations (e.g.
• Repressed
naturalistic approaches on a one-to-one basis
Determining the Order of Questions

Opening Questions
• The opening questions should be interesting, simple, and non-threatening.

Type of Information
• As a general guideline, basic information should be obtained first, followed by
classification, and, finally, identification information.

Difficult Questions
• Difficult questions or questions which are sensitive, embarrassing, complex, or
dull, should be placed late in the sequence.
Determining the Order of Questions

Effect on Subsequent Questions


• General questions should precede the specific questions (funnel
approach).

Q1: “What considerations are important to you in selecting a


department store?”

Q2: “In selecting a department store, how important is convenience of


location?”
(Correct)
Determining the Order of Questions

Logical Order

The following guidelines should be followed for branching questions:

• The question being branched (the one to which the respondent is being
directed) should be placed as close as possible to the question causing
the branching.

• The branching questions should be ordered so that the respondents


cannot anticipate what additional information will be required.
Flow Chart for Questionnaire Design
Fig. 10.2
Introduction

Ownership of Store, Bank, and


Other Charge Cards

Purchased Products in a Specific Department Store


during the Last Two Months

Yes No

How was Payment made? Ever Purchased in a


Department Store?

Credit Cash
Yes
Other

No
Store Bank Charge Other
Charge Card Charge
Card Card

Intentions to Use Store, Bank, and


other Charge Cards
Form and Layout

• Divide a questionnaire into several parts.

• The questions in each part should be numbered, particularly when


branching questions are used.

• The questionnaires should preferably be precoded.

• The questionnaires themselves should be numbered serially.


Example of a Precoded Questionnaire

The American Lawyer


A Confidential Survey of Our Subscribers

(Please ignore the numbers alongside the answers. They are only to help us in data
processing.)

1. Considering all the times you pick it up, about how much time, in total, do you spend
reading or looking through a typical issue of THE AMERICAN LAWYER?

Less than 30 minutes.....................-1 11/2 hours to 1 hour 59 minutes.........-4

30 to 59 minutes............................-2 2 hours to 2 hours 59 minutes...........-5

1 hour to 1 hour 29 minutes..........-3 3 hours or more.................................-6


Reproduction of the Questionnaire
• The questionnaire should be reproduced on good-quality paper and have a professional
appearance.
• Questionnaires should take the form of a booklet rather than a number of sheets of paper clipped
or stapled together.

• Vertical response columns should be used for individual questions.


• Grids are useful when there are a number of related questions they use the same set of
response categories.

• The tendency to crowd questions together to make the questionnaire look shorter should be
avoided.

• Directions or instructions for individual questions should be placed as close to the questions as
possible.
Pretesting
Pretesting refers to the testing of the questionnaire on a small sample of respondents to identify
and eliminate potential problems.

• A questionnaire should not be used in the field survey without adequate pretesting.
• All aspects of the questionnaire should be tested, including question content, wording, sequence,
form and layout, question difficulty, and instructions.
• The respondents for the pretest and for the actual survey should be drawn from the same
population.
• Pretests are best done by personal interviews, even if the actual survey is to be conducted by mail,
telephone, or electronic means, because interviewers can observe respondents' reactions and
attitudes.
Pretesting
• After the necessary changes have been made, another pretest could be conducted by mail,
telephone, or electronic means if those methods are to be used in the actual survey.
• A variety of interviewers should be used for pretests.
• The pretest sample size varies from 15 to 30 respondents for each wave.
• Protocol analysis and debriefing are two commonly used procedures in pretesting.
• Finally, the responses obtained from the pretest should be coded and analyzed.
Questionnaire Design Checklist
Table 10.1

Step 1. Specify The Information Needed

Step 2. Type of Interviewing Method

Step 3. Individual Question Content

Step 4. Overcome Inability and Unwillingness to Answer

Step 5. Choose Question Structure

Step 6. Choose Question Wording

Step 7. Determine the Order of Questions

Step 8. Form and Layout

Step 9. Reproduce the Questionnaire

Step 10. Pretest


Questionnaire Design Checklist
Table 10.1, cont.

Step 1. Specify the Information Needed


1. Ensure that the information obtained fully addresses all
the components of the problem. Review components of
the problem and the approach, particularly the research
questions, hypotheses, and specification of information
needed.
2. Prepare a set of dummy tables.
3. Have a clear idea of the target population.

Step 2. Type of Interviewing Method


1. Review the type of interviewing method determined based
on considerations discussed in Chapter 6.
Questionnaire Design Checklist
Table 10.1, cont.
Step 3. Individual Question Content
1. Is the question necessary?
2. Are several questions needed instead of one to obtain
the required information in an unambiguous manner?
3. Do not use double-barreled questions.
Questionnaire Design Checklist
Table 10.1, cont.

Step 4. Overcoming Inability and Unwillingness to Answer


1. Is the respondent informed?
2. If respondents are not likely to be informed, filter questions
that measure familiarity, product use, and past experience
should be asked before questions about the topics
themselves.
3. Can the respondent remember?
4. Avoid errors of omission, telescoping, and creation.
5. Questions which do not provide the respondent with cues can
underestimate the actual occurrence of an event.
6. Can the respondent articulate?
Questionnaire Design Checklist
Table 10.1, cont.

Step 4. Overcoming Inability and Unwillingness to Answer


7. Minimize the effort required of the respondents.
8. Is the context in which the questions are asked appropriate?
9. Make the request for information seem legitimate.
10. If the information is sensitive:
a. Place sensitive topics at the end of the questionnaire.
b. Preface the question with a statement that the behavior of
interest is common.
c. Ask the question using the third-person technique.
d. Hide the question in a group of other questions which
respondents are willing to answer.
e. Provide response categories rather than asking for specific
figures.
f. Use randomized techniques, if appropriate.
Questionnaire Design Checklist
Table 10.1, cont.

Step 5. Choosing Question Structure


1. Open-ended questions are useful in exploratory research and
as opening questions.
2. Use structured questions whenever possible.
3. In multiple-choice questions, the response alternatives should
include the set of all possible choices and should be mutually
exclusive.
4. In a dichotomous question, if a substantial proportion of the
respondents can be expected to be neutral, include a neutral
alternative.
5. Consider the use of the split ballot technique to reduce order
bias in dichotomous and multiple-choice questions.
6. If the response alternatives are numerous, consider using
more than one question to reduce the information processing
demands on the respondents.
Questionnaire Design Checklist
Table 10.1, cont.

Step 6. Choosing Question Wording


1. Define the issue in terms of who, what, when, where, why, and way
(the six Ws).
2. Use ordinary words. Words should match the vocabulary level of the
respondents.
3. Avoid ambiguous words: usually, normally, frequently, often,
regularly, occasionally, sometimes, etc.
4. Avoid leading questions that clue the respondent to what the answer
should be.
5. Avoid implicit alternatives that are not explicitly expressed in the
options.
6. Avoid implicit assumptions.
7. Respondent should not have to make generalizations or compute
estimates.
8. Use positive and negative statements.
Questionnaire Design Checklist
Table 10.1, cont.

Step 7. Determine the Order of Questions


1. The opening questions should be interesting, simple, and non-
threatening.
2. Qualifying questions should serve as the opening questions.
3. Basic information should be obtained first, followed by classification,
and, finally, identification information.
4. Difficult, sensitive, or complex questions should be placed late in the
sequence.
5. General questions should precede the specific questions.
6. Questions should be asked in a logical order.
7. Branching questions should be designed carefully to cover all
possible contingencies.
8. The question being branched should be placed as close as possible to
the question causing the branching, and (2) the branching questions
should be ordered so that the respondents cannot anticipate what
additional information will be required.
Questionnaire Design Checklist
Table 10.1, cont.

Step 8. Form and Layout


1. Divide a questionnaire into several
parts.
2. Questions in each part should be
numbered.
3. The questionnaire should be pre-coded.
4. The questionnaires themselves should
be numbered serially.
Questionnaire Design Checklist
Table 10.1, cont.

Step 9. Reproduction of the Questionnaire


1. The questionnaire should have a professional appearance.
2. Booklet format should be used for long questionnaires.
3. Each question should be reproduced on a single page (or
double-page spread).
4. Vertical response columns should be used.
5. Grids are useful when there are a number of related
questions which use the same set of response categories.
6. The tendency to crowd questions to make the questionnaire
look shorter should be avoided.
7. Directions or instructions for individual questions should
be placed as close to the questions as possible.
Questionnaire Design Checklist
Table 10.1, cont.

Step 10. Pretesting


1. Pretesting should be done always.
2. All aspects of the questionnaire should be tested, including question content,
wording, sequence, form and layout, question difficulty, and instructions.
3. The respondents in the pretest should be similar to those who will be
included in the actual survey.
4. Begin the pretest by using personal interviews.
5. Pretest should also be conducted by mail or telephone if those methods are
to be used in the actual survey.
6. A variety of interviewers should be used for pretests.
7. The pretest sample size is small, varying from 15 to 30 respondents for the
initial testing.
8. Use protocol analysis and debriefing to identify problems.
9. After each significant revision of the questionnaire, another pretest should be
conducted, using a different sample of respondents.
10. The responses obtained from the pretest should be coded and analyzed.
Sample Design
Population and Sample
• The population is the entire
group that you want to draw
conclusions about.

• The sample is the specific


group of individuals that you will
collect data from.
Population and Sample
• Non-probability
sample: individuals are
selected based on non-
• Probability random criteria. Not
sampling means every individual has a
that every chance of being
member of the included.
population has a • Weaker inference
chance of being about population.
selected. • Easier, cheaper with
• If you want to higher risk
of sampling bias.
produce results
that are • Often used in
representative of exploratory & qualitative
the whole research, to develop an
population, initial understanding
of a small or under-
• most valid choice. researched population
or generate
hypotheses
Sampling techniques: Probability v/s Non-
Probability Sampling
Probability Sampling Methods Non-Probability Sampling Methods
Samples from a larger population are chosen using a Samples chosen are based on the researcher’s subjective
Definition
method based on the theory of probability. judgment rather than random selection.
Alternatively
Random sampling method. Non-random sampling method
Known as
Population
The population is selected randomly. The population is selected arbitrarily.
selection
Nature The research is conclusive. The research is exploratory.
Sample Conclusive represents population demographics. The population demographics representation is skewed.
Takes longer to conduct since the research design defines
This type of sampling method is quick since neither the
Time Taken the selection parameters before the market research
sample or selection criteria of the sample are undefined.
study begins.
This type of sampling is entirely unbiased and hence the This type of sampling is entirely biased and hence the
Results
results are unbiased too and conclusive. results are biased too, rendering the research speculative.
There is an underlying hypothesis before the study
The hypothesis is derived after conducting the research
Hypothesis begins and the hypothesis is tested using this sampling
study.
method.
Random sampling techniques • Systematic sampling:
• Simple random sampling: Every member of the
Every member of the population is listed with a
population has an equal number, but instead of
chance of being selected. randomly generating
• Your sampling frame should numbers, individuals are
include the whole chosen at regular
population. intervals.
• Random number generator • No hidden bias in in list
is used

• Cluster sampling also


• Stratified random sampling: divides the population
Divides the population into into subgroups. Each
subgroups (stratas) e.g. subgroup should have
based on Gender, Age range, similar characteristics as
income group, etc. whole sample; Randomly
• More precise representation select entire subgroups.
of subgroups. • (preferred) include all
individuals in the
• Sample size of strata= (size sampled
of strata)/ (population). cluster. (alternatively) use any random sampling
• Random or systematic methods (Multi-Stage sampling) .
sampling used to select • Used for large & dispersed populations. Higher risk of
samples in each subgroup. error in case of inter cluster differences.
Non-Random sampling techniques • Purposive/ Judgement sampling:
The researcher uses their expertise
to select a sample that is most
• Convenience useful to the purposes of the
sampling: Simply research.
includes the
individuals most • Used in qual research, for obtaining
accessible to the detailed knowledge rather than
researcher. statistical inferences, or for very
small and specific niche population
• Easy & inexpensive, (TG).
No way to tell if the
sample is • Must have a clear criteria and
representative of the rationale for inclusion.
population

• Quota sampling also divides


the population into subgroups.
Select a pre-specified number
of sample for each sub-group.
• The sub-group sample may or
• Snowball sampling: may not be PPS.
Hard to access sample.
It relies on reference for • Sub group sampling may not
samples recruited follow Random sampling
process
Sampling techniques: Cluster v/s Stratified Sampling
Sampling techniques: Cluster v/s Stratified
Sampling
Factors for
Cluster Sampling Stratified Sampling
Comparison
The population is first divided into clusters. Each clusters will be
Members of this sample are randomly chosen
Definition heterogenous within itself but will be homogenous in
from non-overlapping, homogeneous strata.
comparison to other cluster.

Purpose Cost reduction and increased efficiency. Enhanced precision and population depiction.

Sample First a cluster is sampled and then members within the cluster is Selection of the sample is done by randomly
selection sampled and including all the members from chosen clusters. selecting members from various formed strata.

Selection of
elements that Conjointly (collectively) Distinctively (individually)
form a Sample
Division type Naturally formed Depends on the researcher
Heterogeneity Internally, with the clusters Externally, between various strata
Homogeneity Externally, between various clusters Internally, with the strata
Goals and
Precision level
objectives

Confidence Population
Determine level variability
Sample Size:
General Factors Response rate
Consider Your
audience

Focus on your
survey’s
objectives
Population size

Determine Confidence level


• 90% – Z Score = 1.645
Sample Size: • 95% – Z Score = 1.96

Statistical • 99% – Z Score = 2.576


The margin of error (confidence interval)
computation • 10%, 5%, 1%

Standard deviation
Simple formula for difference in means
Represents the
desired power
Sample size in each (typically .84 for 80%
group (assumes equal power).
sized groups)

2 ( Z  + Z/2 )
2 2

n= 2
difference Represents the desired
Standard deviation of level of statistical
the outcome variable Effect Size (the
difference in significance (typically
means) 1.96).
Simple formula for difference in proportions
Represents the
desired power
Sample size in each (typically .84 for 80%
group (assumes equal power).
sized groups)

2( p )(1 − p )( Z  + Z/2 ) 2

n=
(p1 − p2 ) 2

Represents the
A measure of Effect Size (the desired level of
variability (similar to difference in statistical
standard deviation) proportions) significance
(typically 1.96).
End of chapter 1 to 12

You might also like