Lecture 1: Introduction
Introduction
Main goal: to acquire the knowledge and skills on how to approach marketing-related
problems from a rigorous, analytical, data-based perspective
- Identify and understand common marketing research problems faced by managers
- Develop proficiency in working with diAerent types of data
- Master a range of data analysis techniques commonly used in marketing practice
and academia
- Apply these techniques in SPSS/R/Excel and interpret these results
Classifying Marketing Research / Analysis
Type of data:
- Quantitative research: seek to provide depth, insight and understanding
o Good for: mapping the customer’s overall range of behavior and attitudes,
pinpointing the motivations behind people’s behavior, stimulating new and
creative ideas
- Qualitative research: seek to quantify data through measurement and statistical
analysis
o Good for: profiling detailed usage and behavior, highlighting variations between
diAerent sub-groups, precisely measuring consumer preferences for diAerent
products and services
Research design:
- Exploratory research: research in which the major emphasis is on gaining ideas and
insights
o Purposes: increase familiarity with problem, clarify concepts, develop specific
hypotheses
o Approaches: interviews, case studies, focus groups, text analysis
- Descriptive & predictive research: often guided by an initial hypothesis
o Purposes: describe the characteristics of certain groups, estimate the proportion
of people in a specified population who behave in a certain way, examine
associations between two or more variables, make specific predictions
- Causal research: research in which the major emphasis is on determining a cause-
and-eAect relationships
o Make use of experiments: laboratory experiments, field experiments
Data source:
- Secondary data: data previously collected for purposes other than the research at
hand
o Internal sources: accounting records (e.g., sales invoices, marketing
expenditures) customer transaction databases, clickstream data, operating
records (e.g., warranty cards, customer complaint services), previous market
research studies
, o External sources: market and industry research publishers (e.g., Datamonitor),
trade associations, government agencies
o Syndicated research: large-scale, regularly updated marketing research that is
undertaken by a research firm to be sold, often on a subscription basis, to a
number of clients (e.g., consumer panel data / scanned data )
§ E.g., Kantar NIPObase, Nielsen TV Ratings, Gfk Consumer Panel
§ Consumer panel: panel of households or individuals whose purchases are
monitored on a continuous or ongoing basis
- Primary data: data collected specifically to answer the question(s) posed by the
current research objectives
o Types of primary data: demographic/socioeconomic/lifestyle characteristics,
attitudes/opinions, awareness/knowledge, motivations, intentions and behaviors
o Collecting primary data:
§ Communication: questioning respondents to secure the desired information
(via surveys, focus groups, etc.)
§ Observation: the situation of interest is watched, and the relevant facts,
actions or behaviors recorded
Lecture 2: Measurement and Scaling / Reliability, Validity, Dimensionality
Measurement and Scaling
Observable variables: e.g., gender, age, income, education level, number of
complaints, type of contract, customer tenure, premium amount paid, digital app usage
frequency
Unobservable constructs: e.g., involvement, need for uniqueness, customer loyalty,
impulsive buying tendency, price sensitivity, brand attitude, satisfaction, tendency to
give socially desirable answers, materialism, susceptibility to social comparisons
Relational Net Promotor Score (NPS): would you recommend (brand)?
- 0 = not at all likely – 10 = extremely likely
Measurement: the assignment of number to characteristics (attributes)
of objects of people according to a pre-specified rule
Scaling: generation of continuum upon which measured objects are located
Levels of measurement: every scale you design
sits at one of these four levels
- Nominal scale: just labels, no order
o E.g., gender
- Ordinal scale: ranked, but unequal gaps
between points
o E.g., Likert scale
- Interval scale: equal intervals, no true zero
o E.g., IQ-score, temperature
- Ratio scale: a true zero point
o E.g., income
,Why the level of measurement matters?
- Identification of the measurement scale depends on the basic nature of the attribute
- Higher levels of measurements allow for more in-depth statistical analyses
- Critical in determining which data analysis techniques are the most applicable to
the data
o Chi square: nominal
o Pearson correlation coeQicient: interval & ratio
o Linear regression: DV is interval & ratio
o Logistic regression: DV is nominal
Types of Scales in Marketing Research
Comparative scaling techniques: comparison of stimulus objects on a scale
- Paired comparison scaling: respondent needs to select one object (out of 2)
according to some criterion
o E.g., Heineken or Grolsch, Heineken or Bavaria
- Rank order: respondent need to order several objects according to some criterion
o E.g., 1. Grolsch, 2. Heineken, 3. Bavaria
- Contant sum scaling: respondent allocate a fixed amount of points among several
objects according to some criterion
o E.g., Grolsch 50, Heineken 30, Bavaria 20
Non-comparative scaling: each stimulus is scaled independently of the other object
- Likert scale: respondents need to indicate degree of agreement to a series of
statements
o E.g., How much do you like the taste of Heineken? à 0 = not at all likely – 10 =
extremely likely
- Semantic diQerential scale: rating scale with endpoints
associated with bipolar labels
o E.g., Please rate your perception of Heineken among the
following dimensions à see picture
- Continuous rating scale: respondents
rate the objects by placing a mark at the
appropriate position on a continuous line
o E.g., How satisfied are you with the
following? à see picture
o Advantages: fine-grained data
collection, intuitive and user-friendly, interactive
, - Item decisions:
o The number of scale categories: 5 to 9-point scales, depends on respondents’
knowledge and involvement
o Odd vs even number of categories: neutral response allowed or not
o Nature and degree of verbal descriptions: words, numbers, pictures
o Label all vs. only extremes
Multiple-Item Measurement Theory
Single-item measurements: popular in industry, simple, single item, easy to
benchmark
- What exactly is it measuring, and what is it missing?
- Are rarely capture a complex construct like customer loyalty full
Multiple-item measurement:
- Why do personality quizzes, exams and
intelligence test consist of several
questions?
- Why do we often use multi-item (questions)
scales instead of single-item scales?
o Reliability: to minimize random error
o Validity: to capture the
complex/heterogeneous nature of the
construct/characteristic
- Sometimes single-item scales are suAicient
o Objective, concrete constructs (age, income, purchase intention)
- Error
o Xo: observed score (e.g., your measures loyalty score)
o Xt: true score (e.g., your real, underlying, customer loyalty)
o Xs: systematic, non-random, error (e.g., timing/context bias)
o Xe: random error (e.g., being distracted for a moment, mood on the day)
Scale Evaluation
Reliability: the extent to which the measurement instrument gives the same results
when the measurement is repeated à is it stable?
- Consistency in measurement
- Xe = 0
Validity: the extent to which the measurement instrument
indeed measures the construct that we intend to measure
à does it measure what it’s supposed to?
- With a valid measurement, diAerences in scores among
respondents are only due to real diAerences in the
construct (characteristics) among respondents
- Xo = Xt; Xs = 0; Xe = 0
Introduction
Main goal: to acquire the knowledge and skills on how to approach marketing-related
problems from a rigorous, analytical, data-based perspective
- Identify and understand common marketing research problems faced by managers
- Develop proficiency in working with diAerent types of data
- Master a range of data analysis techniques commonly used in marketing practice
and academia
- Apply these techniques in SPSS/R/Excel and interpret these results
Classifying Marketing Research / Analysis
Type of data:
- Quantitative research: seek to provide depth, insight and understanding
o Good for: mapping the customer’s overall range of behavior and attitudes,
pinpointing the motivations behind people’s behavior, stimulating new and
creative ideas
- Qualitative research: seek to quantify data through measurement and statistical
analysis
o Good for: profiling detailed usage and behavior, highlighting variations between
diAerent sub-groups, precisely measuring consumer preferences for diAerent
products and services
Research design:
- Exploratory research: research in which the major emphasis is on gaining ideas and
insights
o Purposes: increase familiarity with problem, clarify concepts, develop specific
hypotheses
o Approaches: interviews, case studies, focus groups, text analysis
- Descriptive & predictive research: often guided by an initial hypothesis
o Purposes: describe the characteristics of certain groups, estimate the proportion
of people in a specified population who behave in a certain way, examine
associations between two or more variables, make specific predictions
- Causal research: research in which the major emphasis is on determining a cause-
and-eAect relationships
o Make use of experiments: laboratory experiments, field experiments
Data source:
- Secondary data: data previously collected for purposes other than the research at
hand
o Internal sources: accounting records (e.g., sales invoices, marketing
expenditures) customer transaction databases, clickstream data, operating
records (e.g., warranty cards, customer complaint services), previous market
research studies
, o External sources: market and industry research publishers (e.g., Datamonitor),
trade associations, government agencies
o Syndicated research: large-scale, regularly updated marketing research that is
undertaken by a research firm to be sold, often on a subscription basis, to a
number of clients (e.g., consumer panel data / scanned data )
§ E.g., Kantar NIPObase, Nielsen TV Ratings, Gfk Consumer Panel
§ Consumer panel: panel of households or individuals whose purchases are
monitored on a continuous or ongoing basis
- Primary data: data collected specifically to answer the question(s) posed by the
current research objectives
o Types of primary data: demographic/socioeconomic/lifestyle characteristics,
attitudes/opinions, awareness/knowledge, motivations, intentions and behaviors
o Collecting primary data:
§ Communication: questioning respondents to secure the desired information
(via surveys, focus groups, etc.)
§ Observation: the situation of interest is watched, and the relevant facts,
actions or behaviors recorded
Lecture 2: Measurement and Scaling / Reliability, Validity, Dimensionality
Measurement and Scaling
Observable variables: e.g., gender, age, income, education level, number of
complaints, type of contract, customer tenure, premium amount paid, digital app usage
frequency
Unobservable constructs: e.g., involvement, need for uniqueness, customer loyalty,
impulsive buying tendency, price sensitivity, brand attitude, satisfaction, tendency to
give socially desirable answers, materialism, susceptibility to social comparisons
Relational Net Promotor Score (NPS): would you recommend (brand)?
- 0 = not at all likely – 10 = extremely likely
Measurement: the assignment of number to characteristics (attributes)
of objects of people according to a pre-specified rule
Scaling: generation of continuum upon which measured objects are located
Levels of measurement: every scale you design
sits at one of these four levels
- Nominal scale: just labels, no order
o E.g., gender
- Ordinal scale: ranked, but unequal gaps
between points
o E.g., Likert scale
- Interval scale: equal intervals, no true zero
o E.g., IQ-score, temperature
- Ratio scale: a true zero point
o E.g., income
,Why the level of measurement matters?
- Identification of the measurement scale depends on the basic nature of the attribute
- Higher levels of measurements allow for more in-depth statistical analyses
- Critical in determining which data analysis techniques are the most applicable to
the data
o Chi square: nominal
o Pearson correlation coeQicient: interval & ratio
o Linear regression: DV is interval & ratio
o Logistic regression: DV is nominal
Types of Scales in Marketing Research
Comparative scaling techniques: comparison of stimulus objects on a scale
- Paired comparison scaling: respondent needs to select one object (out of 2)
according to some criterion
o E.g., Heineken or Grolsch, Heineken or Bavaria
- Rank order: respondent need to order several objects according to some criterion
o E.g., 1. Grolsch, 2. Heineken, 3. Bavaria
- Contant sum scaling: respondent allocate a fixed amount of points among several
objects according to some criterion
o E.g., Grolsch 50, Heineken 30, Bavaria 20
Non-comparative scaling: each stimulus is scaled independently of the other object
- Likert scale: respondents need to indicate degree of agreement to a series of
statements
o E.g., How much do you like the taste of Heineken? à 0 = not at all likely – 10 =
extremely likely
- Semantic diQerential scale: rating scale with endpoints
associated with bipolar labels
o E.g., Please rate your perception of Heineken among the
following dimensions à see picture
- Continuous rating scale: respondents
rate the objects by placing a mark at the
appropriate position on a continuous line
o E.g., How satisfied are you with the
following? à see picture
o Advantages: fine-grained data
collection, intuitive and user-friendly, interactive
, - Item decisions:
o The number of scale categories: 5 to 9-point scales, depends on respondents’
knowledge and involvement
o Odd vs even number of categories: neutral response allowed or not
o Nature and degree of verbal descriptions: words, numbers, pictures
o Label all vs. only extremes
Multiple-Item Measurement Theory
Single-item measurements: popular in industry, simple, single item, easy to
benchmark
- What exactly is it measuring, and what is it missing?
- Are rarely capture a complex construct like customer loyalty full
Multiple-item measurement:
- Why do personality quizzes, exams and
intelligence test consist of several
questions?
- Why do we often use multi-item (questions)
scales instead of single-item scales?
o Reliability: to minimize random error
o Validity: to capture the
complex/heterogeneous nature of the
construct/characteristic
- Sometimes single-item scales are suAicient
o Objective, concrete constructs (age, income, purchase intention)
- Error
o Xo: observed score (e.g., your measures loyalty score)
o Xt: true score (e.g., your real, underlying, customer loyalty)
o Xs: systematic, non-random, error (e.g., timing/context bias)
o Xe: random error (e.g., being distracted for a moment, mood on the day)
Scale Evaluation
Reliability: the extent to which the measurement instrument gives the same results
when the measurement is repeated à is it stable?
- Consistency in measurement
- Xe = 0
Validity: the extent to which the measurement instrument
indeed measures the construct that we intend to measure
à does it measure what it’s supposed to?
- With a valid measurement, diAerences in scores among
respondents are only due to real diAerences in the
construct (characteristics) among respondents
- Xo = Xt; Xs = 0; Xe = 0
