MKT Research Final

Beta Coeff.

NOT a measure of central tendancy

Secondary Databases

Comes in 4 flavors:
Bibliographic
Offer possible sources for further query
Numeric
For example, categorical data, meaningful reports based on census data
Directory
A listing of relevant information
Full text
Online databases of professional publications

Secondary Data (Ch. 13)

data that already exists and can be used to uncover facts and build business related models
(often historical & in a form suitable for your research purposes)

Secondary Data USES:

Fact-finding
- Identify consumption patterns
- Track trends
- Scan environment
Regression-type model building
- Predictor variables are used to explain an outcome variable
e.g., estimate market potential, forecast sales (moving average), select new trade areas (retail saturation index)

Internal Secondary Data

Sources:
Accounting information; Sales information; Backorders; Customer complaints
(+) Advantages:
Suitable geographic and product breakdowns
Minimal time lags
(-) Disadvantages:
Hard-to-handle volume of information
Inputs tied to compensation
Data in accounting format

External Secondary Data

Sources:
Government agencies; Syndicated research services
Trade and professional associations; Custom research firms
Books, newspapers, & periodicals; Internet
(+) Advantages:
Inexpensive relative to primary data
Can be gathered rapidly and are readily available
Provide information that otherwise may not be accessible
Aid in the design of primary research
Enhance existing primary data
(-) Disadvantages:
They may not be consistent with your needs
They may be dated
Difficult to assess their credibility

Evaluating External secondary data

Must evaluate:
- The provider's purpose
- The data collector
- How data were collected
- What data were collected
- When data were collected
- The consistency of these data to data from other sources
Assessing Web sites:
- Purpose
- Authority
- Scope
- Audience
- Format

In-Depth Interviews and Focus Groups
(CH. 14)

...

Qualitative Methods

#NAME?

In-depth Interviews

Non-directed (conversations)
- Interviewer must be able to adapt
Semi-structured
- Rely on a more detailed and structured set of questions
- Provide a comfortable environment
- Offer summary statements
- When appropriate, ask probing questions
- Avoid comments and behaviors that inhibit interviewees

Focus Groups

In-person or online (videoconferencing, text-based)
Recruitment screener
Definitive research problem and question
Importance of a script
Effective moderation is essential
Intra-group similarity
Capitalize on group synergy

...

...

Projective Techniques (Ch. 15)

Provide safer avenue for people to reveal their opinions because those opinions (seemingly) aren't reflective of themselves.
Ex) projecting/assigning one's attitudes, preferences, and behaviors onto a safe person or object.

Projective Techniques (Book)

require people to respond to more ambiguous "multiple interpretation", unstructured stimuli
- no RIGHT or WRONG way to respond
- useful for controversial topics

Projective encourages....

honest responses
people reveal their true motives, opinions, and subconscious since they do not feel personally threatened.

Thematic Apperception Test (TAT)

focus = on picture interpretation
Respondents are asked to concoct a STORY based on a picture
That story should include:
- important precursors of the depicted event
- a description of the depicted event
- the cognitions and emotions of the depicted person
- A conclusion
(+'s): Pictures are fun and easy to understand; pictures are adaptable
(-'s): Pictures are fictional; coding and interpreting data may be cumbersome

Pro's (+) of TAT

Pictures = fun and easy to understand; adaptable

Cons (-) of TAT

Pictures are fictional = causing "who cares" mentality bc depicted situation are imaginary;
coding and interpreting data may be cumbersome = TIME CONSUMING / EXPENSIVE

TAT stories

Psychologists assume that these stories reveal people's conscious and subconscious needs, motives, emotions and conflicts, which in turn influence people's behaviors

as if" answers

reveal a respondents true beliefs

TAT Example #1

by answering "as if" they were the IT person
respondents project their beliefs about upgrades onto her
= safer than admitting to those opinions

TAT Example #2

Porsche 911's have a certain image, so their drivers maintain that certain image as well
Asking respondents to GUESS what "huge smirking Joe" is going in his porsche = unveils their true impressions
a response like "he's going for a ride in the country to enjoy the fall colors" is very different from "tooling up and down the main street picking up hot chicks

TAT Example #3

Positioning MILK
respondents choosing that Ms. Smith (old & out of shape) consumes milk over Ms. Jones (young & active)
shows researchers that they should advertise milk consumption through young and active people

Word Association

asks respondents to reply to a series of preplanned words or statements with the first words that come to their minds
actual word answers = IRRELEVANT
Time needed to answer = OF INTEREST

Word Association focus

TIME needed to respond
- the longer the response time, the more complex the respondents thoughts are about the named entity
- greater response time = being more conflicted

Pros (+) of Word Association

Requires short answers
(encourages meaningful responses)

Cons (-) of Word Association

May be onerous (burdensome) for respondents
struggle to believe what is a relevant response; response quality and willingness to complete the task will diminish

Sentence Completion

Sentence completion (focus is on word combinations)
No hints are provided about appropriate responses

Third-person role playing

(focus is on projective answer to a third person)
Either the third person is someone else or the respondent
(+'s) encourage more elaborate projections, providing richer insights into consumers' attitudes, preferences and behaviors
(-'s): Responses aren't seriously considered (lazy and answer superficially)
;social desirability bias may surface = respondents can be asked to answer as if a close friend --> respondents may prefer to depict their bff in a socially favorable light

Observation Methods

Data are captured in real time with or without respondent knowledge
Scrutinized bc "YOU SEE, BUT DON'T OBSERVE"
Types of Observation Research:
Human (mystery shopper) vs. mechanical
Visible vs. hidden
Direct (natural environment)

Human vs Mechanical

mystery shopper
vs.
store scanners; eye tracking; pupilometer; psychogalvanometer; voice pitch; people meter
Traffic counters; Web site monitors

Visible vs Hidden

observers may be visible = the people being observed know they are being observed
OR
observers = hidden (ethical implications)

Direct vs Contrived

natural environment (ethnography - inductive: starts with data not theory)
vs.
contrived (artificial environment)
direct observation = ppl are observed in their natural environment
contrived observation = researches create artificial environments that encourage people to behave in certain ways

Contrived Observation

researchers create an artificial environment
Ex) "Soft Drink Experiment"
people first shown several ads for new sodas, then allowed to pick one to drink. Most frequently chosen drink assumed that its relative popularity was inspired by its ad
Ex) "Mystery Shopper"
hired shopper follows carefully scripted activity --> ordering a SPECIFIC fast food meal at a SPECIFIC franchise location at at SPECIFIC day and time
then provide detailed reports about their experience

Observation Methods can observe:

- Physical action; Verbal behavior; Expressive behavior
- Spatial relations; Temporal patterns; Physical objects

Pros (+) of Observational Methods

#NAME?

Cons (-) of Observational Methods

Cognitive phenomena aren't observable
Data interpretation may be problematic
Causes of behavior may not be recordable
Observer bias is possible

Good Human Observers must be:

sensitive to consumer behaviors that are relevant to the marketing problem of interest

(Ch. 16) Conducting Experiments and Test Marketing

Experiments = studies conducted under controlled conditions

Casual Relationships

#NAME?

Establishing Causal Relationships

3 criteria needed to PROVE:
Temporal ordering or sequence
- Independent (treatment) & Dependent variables
Concomitant variation
Control for spurious variables (internal validity)
- Hold conditions constant for each participant (test unit)
Control & Experimental group
- Randomly assign subjects to experimental treatments
- Laboratory experiments & Field tests or experiments (test markets)

Temporal Ordering

means that a change in one entity always precedes a change in the other entity

Concomitant Variation

refers to 2 things that can change in tandem (one behind the other)

Experiments can establish:

a CAUSAL RELATIONSHIP
requiring
-concomitant variation of variables
- temporal ordering of variables
- control over other possible causes that may influence the relationship btwn the variables

Internal Validity Definition

the ability of the experiment to determine whether the treatment was the sole cause of changes in a dependent variable
(did the researchers manipulation do what it was suppose to do?)

Internal Validity Threats

History effects
Maturation effects
Testing effects
Instrumentation
Selection bias
Mortality effects
Hawthorne effects
Placebo effects
Demand effects

History Effects
(Threat)

changes in the environment that are irrelevant to the effects of interest but may modify scores on a dependent variable

Maturation
(threat)

changes in study participants that occur during an experiment
responses about later-viewed ads DIFFER from early-viewed ads bc they are tired (rather than some characteristic of the ad)

Testing
(Threat)

occurs before the participants are exposed to an experimental treatment
"Pre-Exposure Testing": may sensitize your participants to the nature of the experiment --> they may respond differently to post-exposure testing

Instrumentation
(threat)

changes in the instrument you use;
certain questions may --> diff responses
change in observers --> diff judgment of behaviors

Selection Bias
(threat)

occurs when participants arent randomly assigned or there experiments lacks a control group

Mortality
(threat)

People who drop out of an experiment over time for any reason

Hawthorne
(threat)

...

Placebo
(threat)

...

Demand
(threat)

...

Test Market

(type of field experiment)
used to refine marketing strategies for new and established brands and to decide whether to discontinue newly introduced products

Traditional (test mkt)

Sales are compared across multiple cities
(-'s)
They're costly and allow competitors time to respond
Criteria for screening possible test markets:
- Sufficient population
- Proper representation

Simulated (test mkt)

(+'s)
Time and cost efficient; they don't tip off competitors
(-'s)
Artificial environment; they induce unnatural behavior

Controlled

Rely on universal product codes, checkout scanning equipment, and marketing information systems
Require small-city grocers and consumers
(+'s)
They provide complete store data; can accurately track coupon use; they're more accurate than store audits or purchase diaries
(-'s)
They're limited to smaller markets and certain types of retailers

Virtual

(+'s)
They provide efficient store manipulation; they foster respondent cooperation; they're cheaper than traditional test markets
(-'s)
Virtual environment; consumers are more involved than normal; some influencing factors can't be accounted for; target market representativeness may be lacking

Collecting and Preparing Your Data
CH. 17

...

Collecting Data

In-house or outsource fieldwork service
Initial contact strategies
Ask questions properly
Use proper probing tactics
Using silence effectively
Summarizing respondents' comments
Repeating the question
Asking a neutral question
Terminate the interview appropriately

Data Preparation and Entry

Pre-entry
- Validate and Edit
For personal interviews, check for:
- Omissions; Ambiguities; Inconsistencies
- Improper skip patterns; Not properly recording answers
For self-administered questionnaires, check for:
- All key questions were answered
- Respondents took questionnaire seriously
Coding data
- Subjective vs. Objective
Creating and cleaning data files
Controlling for missing responses
- Leave responses blank
- Delete records
- Substitute mean response
- Substitute an imputed response

Tools for Analyzing Your Data
CH. 18

...

Descriptive Analysis

Frequency tables
Measures of central tendency
Measures of dispersion (spread in the data)
- Range, variance, standard deviation
Cross-tabulation (joint frequency distribution)
- Chi-square (?2)
- For nominal data, used to determine if observed counts significantly differ from counts that are expected by chance

Difference Tests

Independent samples t-test
Analysis of variance (ANOVA)

Correlation Tests

Pearson (bivariate) correlation (r)
Multiple regression (R2, F, beta)
- Mediator vs. Moderator variables

Creating Effective Research Reports
CH. 19

...

Goals of Research Report

Explain why the research was conducted
- Value of the study
State the specific research objectives
- Indicate research questions
Explain how the research was conducted
- Describe methodology and sampling procedure
Present the research findings
Provide both conclusions and recommendations
- Relevance of the findings to practitioners

Components of Research Report

Prefatory parts:
- Title page
- Letter of transmittal
- Letter of authorization
- Table of contents
- Executive summary
Main body:
- Introduction
- Methodology
- Research design (type of study, data sources, means of data collection)
- Sample design
- Data collection and fieldwork
- Analysis
- Results
- Limitations
- Conclusions and Recommendations
Appendices
- Data-collection forms
- Detailed or unique calculations
- Visuals (e.g., tables, graphs, figures)
- Primary materials (e.g., interview transcripts)
- Secondary materials (e.g., journal articles)
- Bibliography

Writing the Report

Steps to a winning report
- Organize your report components before you begin writing
- Write a first draft
- Revise your first draft (multiple times)
- Have several sets of eyes review your final report
Do's
- Say what you mean in as few words as possible
- Use present tense verbs and active voice
Don'ts
- Don't provide insufficient explanation
- Don't fail to connect with your objectives
- Don't indiscriminately use quantitative methods
- Don't provide a sense of false accuracy
- Don't offer single-number research

Preparing your Presentation

Components
A presentation outline
Visuals
Charts and graphs
Include a number, a descriptive title, explanatory legend, and a list of secondary sources used for development
Pie charts
Depict what percent of all respondents is comprised by a response category
Bar charts
Indicate changes in a Dv (vertical axis) at specific intervals (e.g., time, percent) of the Iv (horizontal axis)
For plotting multiple Dvs simultaneously, use clustered column charts or stacked bar charts
Area graphs
A continuous version of a bar chart (depict trends)
Place least-varying items at the bottom and most-varying items at the top
Multi-line graphs
Indicate relationships between multiple Dvs (e.g., sales by region) and time (e.g., first through fourth quarter)
Scatterplots
Indicate joint values of people or objects on two measures
Box and whisker plots
Depict data generated by descriptive data (e.g., mean, median, inter-quartile range, standard deviation)
An executive summary
Copies of the final report

Ten Useful Research Tips for Business Operators
CH. 20

...

Research Tips

Look to university help first
Take a statistics or research class
- Show initiative...peruse the literature
View research as an ongoing process
Avoid research method myopia
- Don't rely on one type of research method
Prior to collecting data, clearly define and carefully formulate your research questions
Don't ignore opportunity costs
- Missed opportunities can prove damaging
Pretest your method and questionnaire prior to actual data collection
Study your consumers thoroughly
- Trilogy approach to decision making
Make incentives a part of your research program
Share research results with employees
- Internal marketing strategies are essential to building a brand

Ten Useful Statistical Methods
CH. 21

...

Independent samples t-test

To examine mean differences between two independent groups
- Iv=Grouping variable (categorical)
- Dv=Ratio or interval-scaled variable

Paired samples t-test

To examine mean differences between two non-independent groups
- Iv=Grouping variable (categorical)
- Dv=Ratio or interval-scaled variable

One-Way Analysis of Variance (ANOVA)

To examine mean differences among three or more independent groups
- Iv=Grouping variable
- Dv=Ratio or interval-scaled variable

Linear Multiple Regression

To explain variance in a Dv with multiple Ivs
Ivs & Dv=Ratio or interval-scaled variable
- Importance of R2 and F-value or score

Conjoint analysis

To determine how consumers value the different components of a good or service
- Respondents indicate their relative preferences for various combinations of product attributes
- A utility score is used to express the value of each attribute level

Exploratory factor analysis (EFA)

To identify a reduced number of factors from a larger pool of measures

Multidimensional scaling (MDS)

To map objects (like products or retailers) based on respondent perceptions of specific attributes or object similarities
- Helps in identifying characteristics that consumers use to differentiate products or retailers
- Niche markets can be identified

Cluster analysis

To assign people or objects to groups so that intra-group similarity and inter-group dissimilarity are maximized
- Identifies naturally occurring groups (rather than predefined...no Iv and Dv distinction)
- Because algorithms (not formulae) are used, groupings can be based on non-metric or metric data

Discriminant analysis

To determine which characteristics discriminate between two or more naturally occurring groups
- Seek to predict group membership

Logistic regression

To predict probability of an event
Dv=binary response (dichotomous)
1 (probability of success 'x')
0 (probability of failure '1 - x')
Iv=Continuous or dichotomous
No data distribution assumptions for predictor variables