Market Research Exam

Front 1

Problem definition

Back 1

1. Management decision problem
2. Marketing research questions/objectives
3. Constraints

Front 2

Secondary data:

Back 2

Data that exists
Can be internal or external

Front 3

Primary data:

Back 3

Data does not exist that is needed to collect for a particular project
Can be qualitative or quantitative

Front 4

Advantages and disadvantages of secondary data

Back 4

Advantages:
1. Availability/efficiency
2. More objective data
Disadvantages:
1. Might not be accurate/relevant
2. Source credibility
3. Appropriateness of methodology
4. Source bias

Front 5

Exploratory Research aims:

Back 5

1. Generates qualitative data
2. To understand the purpose of the research
3. To identify key variables to measure
4. To establish research priorities

Front 6

Technique in in-depth interviews

Back 6

1. Design interview protocol
2. Provide comfortable environment
3. Give time to think (don't lead/interrupt)
4. Summarise participants' words to ensure understanding
5. Ask for clarification if not clear
6. Avoid evaluative comments or leading questions
7. Let subject do the talking

Front 7

Analysing data of in-depth interviews:

Back 7

1. Find most often talked themes
2. Organise/structure themes
3. Describe themes
4. Measure frequency of response in each theme

Front 8

Projected interviews:

Back 8

a process, which helps to uncover subjects' true feelings or opinions about a topic (by talking about what "other people" would do)

Front 9

Observation method

Back 9

Direct vs Indirect
Disguised vs Undisguised
Structure vs Unstructured
Human vs Mechanic

Front 10

Descriptive Research aims:

Back 10

1. Describe profile of target market
2. Describe status of key variables in target market
3. Explore the relations between key variables

Front 11

Type of descriptive research

Back 11

Cross-sectional
Longtitudinal

Front 12

Survey is...

Back 12

a method of collecting primary data from a representative sample of population

Front 13

Advantages of survey (DR):

Back 13

1. Efficiency (cheaper/less time consuming)
2. Standardisation (easier to interpret/code)
3. Identify subgroups difference

Front 14

Factors for selecting a survey method

Back 14

1. Budget
2. Quality of responses needed
3. Length of survey
4. Sensitivity of the topic
5. Incidence rate
6. Time constraints
7. Stimuli to be presented during survey
8. Completion rate
9. Diversity of subject

Front 15

Response format:

Back 15

1. Open-ended
2. Categorical
3. Scaled

Front 16

How to choose response format:

Back 16

1. Nature of the variable needed to be ask
2. Survey method
3. Level of statistical analysis needed
4. Ability of participants
5. Previous research studies

Front 17

Type of measurement scales

Back 17

1. Comparative
2. Non-comparative

Front 18

Comparative scales

Back 18

1. Paired comparison
2. Rank order
3. Constant sum

Front 19

Non-comparative scales

Back 19

1. Continuous rating scale
2. Likert scale
3. Semantic differential scale

Front 20

Construct:

Back 20

Concept that we are interested in measuring (i.e friendship closeness)

Front 21

Variable:

Back 21

Factor that we will measure the construct on (i.e time spent together)

Front 22

Single item scale:

Back 22

Measure on 1 dimension of a variable (i.e like --> don't like)

Front 23

Multiple item scale:

Back 23

Measure on multiple dimensions of a variable (i.e positive, favourable, tasty...)

Front 24

Guideline to develop a questionnaire

Back 24

1. What should be asked
2. How to phrase a question
3. How to organise question
4. What is the layout of the questionnaire
5. How to pre-test a questionnaire

Front 25

What should be asked?

Back 25

1. Start with research objectives
2. Consider variables that need to be measured
3. Determine scale to be used

Front 26

Principles in phrasing questions

Back 26

1. Avoid extreme words (all, always, best, ever...)
2. Avoid loaded questions
3. Avoid ambiguous words
4. Avoid double barrelled questions
5. Avoid broad questions
6. Avoid abbreviations and jargons
7. Avoid leading questions
8. Avoid questions that are hard to answer
9. Avoid wrong/ambiguous scale
10. Avoid questions that may cause social desirability

Front 27

How to organise questions:

Back 27

1. Start with screening/filter questions
2. Warm-up questions
3. Use transition statements (include "skip" questions)
4. Divide questions into sections for main research objectives
5. Leave personal (demographic except screening) and possibly offensive questions until last

Front 28

How to pre-test:

Back 28

1. Test reliability of measurement items
2. Test readability of questionnaire
3. Measure the time required to complete questionnaire

Front 29

Steps in defining sampling method and size

Back 29

1. Define target population
2. Select sampling frame
3. Select sampling method
4. Determine sampling size
5. Determine number of contacts
6. Conduct field work

Front 30

How to define target population?

Back 30

Look at research objectives

Front 31

What is sampling frame?

Back 31

List of all eligible objects for your research

Front 32

Sampling methods:

Back 32

1. Simple sampling method
2. Systematic sampling method
3. Cluster sampling
4. Stratify sampling
5. Convenience sampling
6. Judgmental sampling
7. Referral sampling
8. Quota sampling

Front 33

What is simple random sampling? Adv. and disadv.?

Back 33

Assign number to each of member in target market, use program to randomly select a required number of participants for sample population.
Advantage: Not biased and most representative method
Disadvantage: Need to know everyone in the target population and assigned each a number

Front 34

What is systematic random sampling? Adv. and disadv.?

Back 34

Systematic sampling method is like SRS but objects are sorted in a particular way (i.e alphabetical) --> Choose a random starting point and skip a constant interval to select a member for sampling population.
Advantage: Easier and faster than SRS
Disadvantage: May have "hidden" pattern --> biased

Front 35

What is cluster sampling? Adv. and disadv.?

Back 35

Dividing target population into similar subgroups and choose to test on one of them.
Advantage: Convenient and quick
Disadvantage: Groups may not be similar to each other. One group may not be representative enough of the whole population.

Front 36

What is stratified sampling? Adv. and disadv.?

Back 36

Diving target population into homogenous group within and heterogeneous group between. Select number of representatives from each group (proportionate to the size of subgroups) to form a sampling group.
Advantage: Ensure sample representativeness and non-bias
Disadvantage: Hard to define basis for stratifying

Front 37

What is convenience sampling?

Back 37

Choose sample at the convenience of researchers

Front 38

What is judgmental sampling?

Back 38

Choose sample based on the judgement of the fit of a particular person to be sample participant

Front 39

What is referral sampling?

Back 39

Select number of representatives and ask them to invite more representative people (base on their judgment) to participate in the research.

Front 40

What is quota sampling?

Back 40

Like stratifying sampling method, however participants from each subgroups are not chosen randomly, but on a per-specified selection objectives to form a representative group. --> bias

Front 41

Ways to determine sampling size

Back 41

1. Rule of thumb (usually 5% of target population)
2. Past research
3. Cost-benefit approach
4. Statistical analysis approach
5. Confidence interval approach

Front 42

Data preparation process:

Back 42

1. Data Validation
2. Data Editing
3. Data Coding
4. Data Entry
5. Data Clean

Front 43

Data validation:

Back 43

To make sure that the data collection process was correct, unbiased and valid.
How? Re-test by contacting a number of participants in the sample --> may be impossible if there are no information about participants --> choose a trustable survey agent

Front 44

Data editing:

Back 44

A process of picking up data mistakes: strange patterns (non-correlation between answers, speeders), make sure respondents understand instructions.

Front 45

Data coding:

Back 45

A process of numbering responses in the questionnaire

Front 46

Data entry:

Back 46

Use SPSS to enter data

Front 47

Data clean:

Back 47

Missing values: if a lot people don't answer the question --> there may be problem with question
Outliers: respondents with strange patterns
Normality: check on kurtosis and skewness

Front 48

What is type I error?

Back 48

When we reject the null hypothesis when it is true.
The probability of making such mistake is alpha (level of significance)

Front 49

Causal Research:

Back 49

To find out the cause of effect and validate that it is the only cause

Front 50

Assumptions of causal research:

Back 50

1. Co-variation: there must be a relationship between IV and DV --> test for association of the data must be done prior
2. Time sequence: IV must occur prior or at the same time as DV
3. Systematic elimination: all other possible causal variables (extraneous) must be eliminated from the research

Front 51

Types of experimental design:

Back 51

1. After-Only
2. One group, before-after
3. After-Only with control group
4. Before-After with control group

Front 52

After-Only experimental design

Back 52

Measure the result after the treatment effect without prior measurement of that group. Assume prior knowledge as a baseline for comparison.
--> very subjective, but cheapest and quickest

Front 53

One Group, Before-After experimental design

Back 53

Measure the result before and after treatment on 1 group.
Problem:
If the time gap is too big --> there may be a lot of extraneous variables (mood, tiredness...)
If time gap is too small --> social desirability bias (participants may guess the aim of the research and response in a desirable way)

Front 54

After-Only with control group

Back 54

Have 2 similar groups. One group (G1) with treatment, other group (G2) without. Measure the effect on both group. Treatment effect = result of G1 - result of G2.
Problem: hard to control extraneous variables because there are 2 different groups --> choose similar demographic, geographic groups

With psychographic and behavioral differences --> randomization to minimize the difference between 2 groups

Front 55

Before-After with control group

Back 55

Have 2 groups chosen on random basis (minimize difference). Measure BEFORE effect of 2 groups, then measure AFTER effect of group 1 (with treatment) and group 2 (without treatment).
Treatment Effect = (G2A-G2B) - (G1A-G1B)
--> expensive, difficult but most effective

Front 56

Types of experiments validity

Back 56

Internal: X is the only cause of Y (extraneous variables are controlled)
External: This relationship can be generalized in the population (under other conditions like time, weather, country...)

Front 57

Threats to external validity:

Back 57

1. Sample does not represent the population
2. Artificial setting in lab experiments

Front 58

Main reasons for field experiment:

Back 58

1. Determine sales potential for new products
2. Test the effectiveness of marketing mix

Front 59

Types of relationship between variables:

Back 59

1. Linear: strength and direction of the variables are constant
2. Curvilinear: strength and direction of the variables changes over the course of the variable

Front 60

What to do when cell validity in Chi-Square is disrupted?

Back 60

1. Use another test
2. Merge or remove category with too few people in it

Front 61

When Chi-Square test is valid?

Back 61

When there are no cells with count less/equal one or more than 20% of cells with counts less than 5

Front 62

Steps in doing multiple regression analysis:

Back 62

1. Adjusted R-Square
2. F-value of ANOVA: statistical significance of overall model
3. p-value in coefficient table: the likelihood that relationship exist between IV and DV
4. Standardised beta-coefficient: importance of IV to DV
5. Collinearity statistics: tolerance >10; VIF<5 (no prob)
6. If multicollinearity presents --> check correlation between 2 IVs, remove one of IVs if correlation >0.70 and do test again

Front 63

Point of estimate:

Back 63

A statistic that estimate the value of population parameter

Front 64

Confidence Interval =

Back 64

Point of estimate +/- Margin of Error

Front 65

Confidence interval is:

Back 65

A range of value that is likely to capture the true population parameter

Front 66

Confidence level:

Back 66

The probability that a confidence interval will contain the population parameter

Front 67

Margin of error:

Back 67

The difference between sample statistic and population parameter

Front 68

Adv. & Disadv. of open-ended response format?

Back 68

Adv: allows respondents to use their own words
Disadv.: difficult to code and interpret

Front 69

Adv. & Disadv. of categorical response format?

Back 69

Adv: simple to administer and code, may alert respondents to options of which they were unaware/forgotten
Disadv: Oversimplify response options (give "other, please specify" option); Assumes that researcher knows the options

Front 70

Adv. & Disadv. of scaled response format?

Back 70

Adv: improve accuracy in response; allow the degree of attitude/feelings to be expressed
Disadv: assume that respondents understand the scale; different respondents may interpret scale differently

Front 71

How to decide on sampling method?

Back 71

Choose probability sampling if:
- Do descriptive or causal research
- Variability in population is high
- Budget is sufficient
Choose non-probability sampling if:
- Do exploratory or causal research
- Variability in population is low
- Budget is limited

Front 72

Factors to choose between single and multiple item scales:

Back 72

1. Characteristic of the variable
2. Survey method
3. Analysis method

Front 73

How to test reliability of single item scale?

Back 73

Use Test-retest technique: measure responses in T1 and T2 (or of 2 groups at the same time), if correlation between responses is high --> scale is reliable!

Front 74

How to test reliability of multiple item scale?

Back 74

Use SPSS to measure Cronbach's alpha, if it's >0.6 --> measure is reliable

Front 75

What is content validity?

Back 75

Subjective assessment of how well the variable represents the construct to be measured

Front 76

Factors determining adv. & disadv. of survey methods?

Back 76

1. Length of questionnaire
2. Quality control
3. Flexibility (demo product)
4. Data input (easy/hard)
5. Cost
6. Social desirability bias
7. Follow up opportunity (easy/hard)

Front 77

How to segment the market?

Back 77

1. Choose behaviour that you want to associate with the product
2. Find and describe consumer segment that would possess those behaviours