DCI Flashcards
Define market research
+
Basic Process
The systematic and objective process of generating information that aids in making marketing decisions involving:
- Specifying the information required to address market issues
- Designing the method for collecting information
- Managing and implementing the data collection process
- Analysing the results
- Communicating the findings and their implications
Importance of market research / value
- Business Component
(marketing management needs research, manager’s need for knowledge & business decisions, reduce uncertainty of marketing strategies and tactics) - Social / welfare component
(Social marketing, broader societal issues)
Importance of market research / why
- Identifying and evaluating opportunities (market envionment - opportunities and threats)
- Analysing and selecting target markets
(segment characteristics, purchase motivations) - Planning and implementing marketing mix
(product, price, place, promotion) - Analysing marketing research
(monitor performance)
- Increasing access to big data
Market research process
- Research problem & Scope (research questions & objectives)
- Method: qual, quant or both - consider secondary data
- Collect data from right group
(e.g sample, size, probability/non-probability) - Analyse data, conduct test (SPSS, NVIVO)
- Communicate results
Difference between Qualitative and Quantitative research: Qualitative Research
Qualitative
- Initial discovery / explore
- Understand
- Words, visuals, open-ended, probing questions
- Interviews, focus groups, observations
- Semi-structured questions usually, order doesn’t matter as much
- Small samples
- Limited generalisability
Difference between Qualitative and Quantitative research: Quantitative Research
- Certainty
- Confirmation/assessment
- Explain: descriptive or causal
- Numbers, measures, scales
- Surveys
- Structured, ordered questions
- Larger sample
- Samples representative of population / inferences
Focus of qual VS focus of quant
Qual
- Insights into preferences, experiences, motivations
- Understand customer journey
- Test new ideas / products
- Develop questions for surveys
Quant
- Assess attitudes, product usage, recall, likelihood of purchase, market potential
- Consumer segmentation
- Test hypothesis and theories
- predict behaviour
Advantages and disadvantages of qualitative research
Advantages
- Rich data
- Preliminary insights into behaviours and attitudes
- Preliminary frameworks / how variables are related
Disadvantages
- Lack of generalisability
- Inability to quantify, measure differences
- Insights depend on researcher’s skills and training
Tools of qualitative research - focus group adv and dis
Small group w/ moderator
Adv
- can be quick
- gain multiple perspectives
- Flexibility
- Inexpensive
- contrast opinions
- can demonstrate a concept / conduct exercise
- Synergy
- Snowballing (1 comment -> responses)
- Serendipity (group idea generation)
- Security
- Spontaneity
- Structure
-Scientific scrutiny through observers and recordings
Dis
- Results do not generalise
- difficult for sensitive topics
- difficult for people with busy schedules or want a specific person
- Require sensitive and effective moderators
- Participants may dominate conversation
- social bias
- sampling issues
Tools of qualitative research - depth interviews
One-on-one interview, probing questions
Adv
- Considerable insight
- Understand unusual behaviours
Dis
- Results do not generalise
- Very expensive per interview
Tools of qualitative research
- semi-structured interviews
Open-ended questions
Adv
- Can address more specific issues
- Results can be easily interpreted
- More cost-effective than focus groups and depth
Dis
- Lack flexibility that is likely to produce truly creative or novel explanations
Applications of interviews and focus groups
- Test advertising / integrated marketing communications
- Generate new ideas about a product or delivery method
Advantages of interviews
Adv
- Gain deeper insight from each individual
- Good for understanding private or unusual behaviours
- Can cover sensitive topics
- Respondents are not influenced by others
- More flexibility on time and location
Writing appropriate interview questions
- avoid yes/no questions
- minimise researcher bias / leading questions
- based off research objectives
Question order
- Intro -> topic A -> topic B -> conclusion
- introductions and warm up
- Grand tour questions (general to specific)
- Floating prompts (encouragement deliberate, concrete examples)
- demographic questions
- information sheet, consent
Role of the researcher
- know what will be said
- be prepared to ask follow up questions
- Listen more than talk
- Active listening
- probing and clarifying questions
What is qualitative analysis
- Nonnumerical explanation and interpretation of observations
- seek to understand phenomena, behaviour, thoughts, emotions
- equally art as science
Themes vs codes
Code
- Identifying themes in accounts and attaching labels (codes)
- process to get to theme
- can also be sub-themes
Theme
- Features of participants’ accounts
NVIVO: Node groups, and can be both
Conducting qualitative analysis
- Organise data
- Develop coding framework
- Allocate data into framework
- Interpret the data (core data and theory)
Organising and collating qualitative data
- Transcribe interviews
- Clean interviews
- Records of interviews
- De-identify data
- data to form it can be manipulated on NVIVO
Discovering patterns in in qualitative data
- frequencies
- magnitude
- structures
- processes
- causes
- consequences
Process of qualitative analysis
- Coding
- Annotating (draw attention to most important)
- Labelling (grouping data)
- Selection (choose important items)
- Summarising (one or more examples to illustrate findings)
3 coding framework
- Open coding
(initial classification and labelling) - Axial coding
(reanalysis, relationships between codes, general concepts) - Selective coding
(central concepts)
NVivo process
- Clean transcript
- Import transcript
- Create nodes -> open coding
– highlight and drag text to the nodes and label
– or right-click, new node - Axial coding
(go through codes, common concepts, start grouping)
– drag node onto another - Selective coding
(refine into core categories) - Begin visualising
NVivo tools for visualising
Word frequency (Mind maps, word maps, word trees)
Hierarchy chart
Bar chart
Concept map
Interpreting qualitative research
- Capture and share information
- Find themes
- Making sense of findings
(links between themes, dig deeper?) - Define insights
(what was surprising, connect with business problem)
- insight statement -> rephrase theme into statement - Evaluating themes / insights
- useful, innovative, explain, resonate, address problem
Visualising qualitative research
– canva, NVivo, worditout
- Infographics
- Quotes
(voice of customer, support findings) - Word Clouds
(importance of certain words) - Word Trees
(How one word is connected to others) - Conceptual framework (e.g processes - think funnel / customer journey)
Qual vs Quant research design
- Interview guide (qual) vs questionnaire / survey (quant)
- Semi-structured, open-ended (qual) vs closed, ratings
- small sample size - 30 or saturation (qual) vs large - 100 or more (quant)
- Words, sentences, text, images (qual) vs variables, constructs (quant)
Advantages and disadvantages of surveys
Advantages
- Large samples = generalisability
- Estimates differences
- Easy to administer and record answers to structured questions
- Can assess abstract concepts and relationships
Disadvantages
- Questions to accurately measure attitudes and behaviour can be hard to develop
- Data from open-ended (?) questions may not be enough to record all concepts -> can’t probe
- Response rate
- Over-surveying
Building survey questions
Option 1: Adapt items from secondary research
- change wording
- Item removal / addition
Option 2: Use qualitative research
- usually to justify variables, adapt scales
Scale vs item
Scale = set of items that measures something
Item = component of a scale, or singular statement that measures something
Variable vs construct
Variable = Observable attribute of an object, measurable, one item
Construct = Abstract, set of related questions
Survey questions wording (what to avoid and aim for)
Avoid:
- Jargon / slang / abbreviations
- Ambiguity / confusion
- Emotional language
- Double-barrelled questions
- Leading questions
Aim for
- understandable
- useful
- explanations of concepts
Different levels of measurement
Nominal, ordinal, interval, ratio
Nominal data
- basic level
- values or categories with no quantitative value
- assigned number codes but still no order
Ordinal
- meaningful order
- distance between answers may not be equal
- include likert scales
Interval data
- Distance is meaningful
- differences between answers can be measured not just classified
- scemantic scales 1,2,3,4,5 etc.
Ratio data
- Distance is meaningful
- Measurable
- There is a true zero value
Probability vs non-probability sampling
Probability
- Every member of population has known and equal chance of being chosen.
- Selection is based on a chance
- no error related to researcher
judgment in selecting respondents
- standard error and confidence intervals can be calculated
- generalisations can be made
Non-probability
- no population list
- every member of pop. does not have known and equal chance of being chosen
- sampling error not known
- limited generalisations
- units selected from personal judgement or convenience
- can be bias
Types of probability sampling
- Simple random sampling
- Systematic sampling
- Stratified sampling
- Cluster sampling
Types of non-probability sampling
- Convenience sampling
- Judgement (purposive) sampling
- Quota sampling
- Snowball sampling
Simple random sampling
Probability
RNG, all in a hat
Systematic sampling
Probability
Select members at a regular interval from a sampling frame (e.g every 4)
Stratified sampling
Probability
Random sampling from sub-groups (e.g gender)
Cluster sampling
Probability
Segment into geographic areas
Random cluster chosen
Convenience sampling
Non-probability
Units most conveniently available
try to have people from different households
Judgement (purposive) sampling
Non-probability
Based on researcher’s judgement on who would be best
Quota sampling
Non-probability
Ensures representation of certain subgroups
e.g Males 50%, females 50%
- bias -> may not be representative of actual population
Snowball sampling
Non-probability
Initial respondents selected with random, but additional respondents obtained from these
e.g referrals
Independent samples t-test
Test significant differences in rating between two groups
Nominal (2 categories) by metric (ordinal, interval, ratio)
Independent samples t-test in SPSS
Analyse, compare means, independent samples t-test, categorical grouping variable, metric test variable, assign codes to grouping variables (e.g female = 1)
Independent samples t-test interpretation of output
Independent samples t-test
1. Levenes >0.05, equal variances assumes - top row
< 0.05, not assumed, bottom row
- sig (2-tailed): >0.05, not sig
< 0.05, sig
Group statistics
3. Look at means
Paired samples t-test
Tests differences between two variables for the same group
Paired samples t-test in SPSS
Analyse, compare means, paired samples t-test, select variables of interest
Paired samples t-test interpretation of output
Paired samples test
1. sig (2-tailed)
Paired samples statistics
2. Compare means (which is higher, lower)
- can also look at mean in paired samples test
Visualising tests of difference
bar graphs
pie charts
infographics
Association between variables
Presence of association - level of significance
Direction of association
Strength of association - correlation coefficients
Strength of associations
Non-existent - r=.00
Weak (small) - r = 0.01-0.4
Moderate (medium) - r = 0.41-0.6
Strong (large) = r = 0.61-1
Pearson’s correlation
Interval / ratio variables
Spearman’s correlation
Ordinal by ordinal/interval/ratio
Correlation SPSS process
Analyse, correlate, bivariate, tick either spearman or pearson
Correlation SPSS output interpreation
Correlations
1. sig (2-tailed)
> 0.05 = no correlation
< 0.05 = correlation
- Correlation coefficient
r = __ (weak, moderate, strong)
Tests of association visualisation
Balls, arrows, numbers
scattergraph can be difficult for managers to interpret
