Market Research Wednesday Morning

1
Market ResearchWednesday Morning

• Nigel Bradley
• University of Westminster

2
The Week

• Wednesday Morning
• Sampling/Questionnaires
• Wednesday Afternoon
• Analysis/Segmentation
• Thursday Friday
• Promotion Product/ Price/Place/You Present

3
Learning Outcomes

• You will be able to select research techniques
  for specific tasks
• You will be able to describe their uses
• You will be able to describe their limits

4
Sampling (Echantillonage)

• Definitions
• Random
• Non-random
• Nigel Bradley
• February 2002
• CESEM/UoW

5
Census(recensement)
Universe examined in its entirety

• Disadvantages
• Expensive
• Slow
• Rare

• Advantages
• Complete picture
• True facts

6
UK Census

• Every ten yearsto inform social policy
• and to plan health and local policy
• - First census
• - Move to computers
• 1991 - Cost 135 million
• - Asked ethnic questions
• 2001 - Asked religion questions
• - More ticks for optical readers
• 2011 - Perhaps no census? (enough info)

7
Sample(echantillon)
part of the universe is examined

• Advantages
• Cheap
• Quick
• Easy to do

• Disadvantages
• Can be biased
• Cannot be perfect
• Assumptions made

8
Surveys

• Systematic collection, analysis and
  interpretation of information.
• Can be a census, sample or desk research.
• MR commonly uses sample surveys.
• (cf. enquete par sondage)

9
Sampling Frame(plan de sondage)

• A list or set of directions for identifying the
  target population.
• (Malhotra 1999).
• Examples.
• Register of electors, phone directory,
  membership lists.

10
Sample Size(taille, effectif de lechantillon)

• Depends on
• The population
• Type of information needed
• Cost

11
Sample Types

• Random sample
• Echantillon au hasard
• Echantillon probabiliste

Non- Random sample Echantillon
non-probabiliste. Les sondages non aleatoires ou
empiriques
12
Sample Design

• Probability sample
• A member has a positive, calculable probability
  of being chosen
• Response rate important
• Is expensive

• Non-probability sample
• Uses human judgement
• Subject to errors
• Also known as purposive

13
Sample Design

• Probability sample
• Simple random Systematic
• Stratified random
• Cluster
• Other types

• Non-probability sample
• Quota
• Judgement
• Convenience
• Other types

14
Quota Sampling
Aim of survey divides population into strata
(usually age, sex, social grade) researcher sets
numbers, interviewer finds target.

• Advantages
• Cheap
• Easy to do
• Adequate for many purposes

• Disadvantages
• Unrepresentative ?
• Interviewer bias
• Grade hard to identify

15
Types of Quota

• Interlocking
• Also called interrelated
• Strata are linked
• Greater control over interviewer
• Difficult task
• More representative

• Non-interlocking
• Also called independent
• Strata are unconnected
• Simple task
• Likely to skew to most co-operative people
• Less used

16
Judgement Sampling

• Consult secondary data and choose sample that
  best fits aims of study
• Common in industrial applications
• E.G. Census of main firms sample of rest

17
Convenience Sampling

• Picked on basis of convenience
• Cheap, easy to do, but is a self selecting sample
  
• Examples
• Staff or dealers evaluate new products, visitors
  at an exhibition are interviewed

18
Simple Random Sampling (Echantillonnage aleatoire
simple)
All members of population have equal probability
of selection - Lottery method, random numbers
or computer-generated

• Disadvantages
• Wide-dispersal
• Uneconomical

• Advantages
• No human judgement

19
Systematic Sampling
(or quasi-random sampling) sample interval
calculated by finding the ratio of the population
to the sample. (i.e.) N/n. Widely used.
Practical.
20
Stratified Random Sampling (Sondage aleatoire
stratifie)

• Pop divided into groups with similar attributes.
• within these strata random samples are selected

• Disadvantages
• Tempting to add irrelevant strata
• Many strata can complicate survey

• Advantages
• Accurate
• Less sampling errors than simple random sample

21
Cluster Sampling
Small groups are selected at random. Then every
unit is sampled. Interviews concentrated in
small groups. e.g. identify sales areas, randomly
select a few, interview all salesmen in
these.

• Advantages
• Cheap, fast.
• Good for personal interviews.
• Good for heterogeneous pop.

• Disadvantages
• Biased selection.
• Can miss whole partsof population.
• Sampling error increases.

22
Other Types Of Probability Sample

• Multi-stage sampling
• Replicated sampling(AKA interpenetrating
  sampling)
• Multi-phase sampling(not multi-stage sampling)

23
Random Route Sampling(or RANDOM WALK)
Interviewer starts at specified point (randomly
chosen) and calls on household at set intervals.
Advantages - good face validity Disadvanta
ges - more errors than quota? - more
expensive than random?
24
Random Location Sampling

• Like random-walk, but interviewers work in
  highly specific homogeneous areas and call at
  addresses until well-defined quotas are achieved.

25
Sampling is powerful

• Election surveys (and results) show how powerful
  sampling can be
• See the different outcomes with a sample of
• 2 thousand
• 15 thousand
• 31 million

26
UK Election 1997

• ELECTION EXIT POLL
  SURVEY
• 1 MAY 1 MAY
  29 APRIL
• 
  
  Labour 45 46 48
• Conservative 31 30 28
• Lib-Dem 17 17 16
• Others 7 7 8
• Election n 31,372,549 Turnout 71, weather
  good Exit Poll n 15,000 (MORI/ITN )
• Survey n 2,304 at 253 sampling
  points, face to face(Times/MORI)
• (Source Times 3/5/1997)

27
Panels

• Consumer Panels
• Households or individuals
• Occupational Groups
• eg Sofres, IPSOS, Nielsen

28
Panels

• Advantages
• Rich data source
• Increased precision
• Cost savings

Disadvantages Representative? Non-response
bias Changes in population
29
The Omnibus

• Representative sample
• Can buy one or more questions
• General public v specialist (businesses,
  agriculture etc.)
• See ESOMAR directory for info

30
Tasks

• Keywords echantillonage boule de neige
  samples, census, quotas, PAF, sampling frames.
• Reading Evrard Chapter 5
• 3. What social grade is a student, lecturer, OAP?

31
Questionnaires

• Agenda
• Purpose
• Types
• Scales

32
Questionnaire

• Definition 
• two or more questions containing carefully
  chosen vocabulary.
• Its purpose is to assist in the investigation
  of a specific subject.
• (Bradley 2001)

33
Questionnaire or Form

• Questionnaires are often used to interview a
  respondent or informant.
• Asking is one way of finding out about what a
  person feels, thinks or does. Alternatively one
  can observe, experiment or deduce.
• These forms are also used to capture data in
  observational research.

34
5 Types Of Information

• 1. Facts
• 2. Opinions
• 3. Motives
• 4. Past Behaviour
• 5. Future Behaviour

35
Definitions

• Attitude
• A learned predisposition to respond in a
  consistently favourable or unfavourable manner
  with respect to a given object
• Opinion
• The expression of an underlying attitude

36
Golden Rules

• Respondent must ...
• 1. Understand the question
• 2. Be willing to answer
• 3. Be able to answer
• Researcher must ...
• 1. Address objectives
• 2. Be able to analyse
• 3. Consider respondent (fatigue, ethics,
  knowledge)

37
Steps In Questionnaire Writing

• 1. Formulate hypotheses
• 2. Choose collection method
• 3. List topics of interest
• 4. Plan tables
• 5. Order topics/flowchart
• 6. Layout questionnaire
• 7. Pilot questionnaire
• 8. Fine-tune questionnaire

38
Steps In Questionnaire Writing

• 2. Choose collection method

39
Personal
Telephone
Questioning Mode
others
Self completion
internet
post
on-site
40
Types Of Questions

• Closed
• Open-ended
• Direct questions
• Indirect questions
• Unstructured questionnaires
• Structured questionnaires

41
Closed Questions
2 Answers - Yes/No More Answers -
Multiple - Single choice

• Advantages
• Speed
• Analysis
• Planning can avoid asking

• Disadvantages
• Needs pilot
• Error possible

42
Open Questions
- Interviewer/respondent writes answer

• Advantages
• Full answer
• Exploratory

• Disadvantages
• Misinterpretation
• Speed
• Coding/analysis

43
Timing Guide

• Closed 20 sec (3 Q per min)
• Open-ended 30 sec (2 Q per min)
• 5 Ratings 1 minute
• but ... pilot the questionnaire to estimate
  time taken

44
Direct V Indirect

• Eg. Do you, would you, Does your friend,
  Would your friend
• use of third person projection
• PROBLEMS
• people lie, avoid etc.

45
Unstructured ?

• Structure imposed by researcher may not be best
• An unstructured interview administered by mail
  gives analysis problems
• Large numbers best interviewed in structured way

46
Internet Questionnaires

• Interactive or static
• colour or not
• With/out edit checks
• With/out filters
• Closed/open
• Single scroll/multi-page
• E-mail or web

47
Data
48
Scales

• 1. Summated Rating Scale (Likert)
• 2. Semantic differential (Osgood)
• 3. Kellys personal constructs

49
Summated Rating Scale(Likert)

• How much do you agree/disagree

50
Semantic Differential Scale (Osgood)

• Rating of attitude dimensions for specific
  products

51
Semantic Differential

• Developed by Osgood et al to measure the
  "semantic space" of interpersonal experience.
• A series of 7-point bi-polar rating scales.
• Bi-polar adjectives anchor the beginning and end
  of the scale (e.g. good/bad, new/old)

52
Uses

• To compare the profile of one product against
  competing products
• One brand v. others
• Store v. others
• Firm v. others

53
Variations

• Originally was 7 points, can be 5, 9 or more
• Bi-polar adjectives replaced with descriptive
  phrases
• Bi-polar opposites replaced
  (often respondents didn't use the negative part
  of scales)

54
Original

• x
• Good - - - - - - - Bad

• 7 points
• Bi-polar adjectives

55
Variation (1)

• Prompt x
  Bad
• Service - - - - -
  Service
• by by
• Personnel
  Personnel

• 5 points
• Descriptive phrases

56
Variation (2)

• Extremely x
  Extremely
• poor - - - - -
  good
• performance
  performance
• Satisfactory
  Extremely
• performance - - - - -
  good
• 
  performance
• (negative side of scale eliminated)

57
Analysis

• Assign a weight to each position on the scale
• Traditionally is 7, 6, 5, 4, 3, 2, 1
• Alternatively 3 2 1 0 -1 -2 -3
• The mean or median is used to compare a product
  profile with competing products.

58
Notes

• Widely seen to provide interval data, but critics
  say it is only ordinal because the weights are
  arbitrary
• Osgood C, Suci G Tannenbaum (1957) The
  Measurement of Meaning. Urbana, Illinois,
  University of Illinois Press.

59
Personal Constructs(Kelly)

• The answer is the respondents personal
  construct and is used to form a semantic scale
• Use of cards to create different scales

60
Hostage Survey

• 16 Dec 1996 - Tupac Amaru (MRTA) take
  hostages at the Japanese Embassy, in Lima,
  Peru.
• 18 Dec 1996 - Alfredo Torres, Director of
  Apoyo, a market research agency -
  also a hostage, conducted a survey

61
Questions

• Q1 In general, how would you describe the
  treatment you are receiving from the MRTA?
• VERY GOOD FAIR BAD VERY GOOD BAD
• Q2 What is upsetting you most about the current
  situation?
• Q3 Why were the MRTA able to pull off the
  invasion?
• Q4 Do you think the government should negotiate
  or storm the embassy?

62
Sample (n80)

• Q1 Treatment
• Very good -
• Good 78
• Fair 21
• Bad -
• Very Bad -
• Q4 Negotiate or Storm?
• Negotiate 95 Storm - N/A 5

63
Q2 Most Upsetting Me...

• uncertainty 83
• no sleeping space 49
• no communication with family 29
• state of bathrooms 13
• food 4

64
Q3 Why Were MRTA Successful?

• Faulty state intelligence 52
• Embassy security 35
• Excellent MRTA organisation 13

65
Design Checklist ...

• 1. Ambiguity?
• 2. Jargon?
• 3. Hypothetical questions?
• 4. Double negatives?
• 5. Long-winded?
• 6. Can be read out easily?
• 7. Uses showcards well?
• 8. Flows

66
Tasks

• 1. Keywords scale, closed question
• Evrard Ch 7 and 8
• Write a questionnaire (telephone) to ask the
  public which cars they can name
  (spontaneous/prompted).

67
Learning Outcomes

• You are able to select research techniques for
  specific tasks
• You are able to describe their uses
• You are able to describe their limits

68
The Week

• Wednesday Morning
• Sampling/Questionnaires
• Wednesday Afternoon
• Analysis/Segmentation
• Thursday Friday
• Promotion Product/ Price/Place/You Present

69
Market Research(Etudes Marketing)

• Prepared by Nigel Bradley
• Email bradlen_at_wmin.ac.uk
• Harrow Business School,
• University of Westminster
• For lectures 18-22 February 2002
• at Cesem Mediterranee, Marseille
• 3602 Etudes Marketing