MR Test 2 Flashcards
What are the 11 steps for Determining the Research Design?
1) Establish the need for MR
2) Define the problem
3) Establish research objectives
4) Determine research design
5) Identify Information
6) Determine methods of accessing data
7) Design data collection forms
8) Determine sample plan and sizes
9) Collect data
10) Analyze data
11) Prepare and present the final research
Why do you have to determine research design?
Because every research problem is unique. Therefore, no one research will satisfy all types of research objectives
3 types of research designs
Exploratory
Descriptive
Causal
research has 3 objectives
To develop hypotheses
To measure a variable of interest
To test hypotheses that specify relationships between variables
Exploratory Research
Unstructured, informal research undertaken to gain background information
“Unstructured” – no formalized set of objectives, sample plan, or questionnaire
“Informal” – aimed at gaining additional information about a topic and generating possible hypotheses
Often conducted at the outset of research projects
Systematic, but very flexible
Fast (especially online research)
Relatively inexpensive
Some exploratory research is needed in any research project
Uses of Exploratory Research
To gain background information
To help define terms and concepts (e.g., what is “bank image?”)
To clarify (help define) problems, and hypotheses
“Hypotheses” – statements describing the speculated relationships among variables
To prioritize research projects
Types of Exploratory Research: Secondary data analysis
Using existing information relevant to the research objectives (“secondary” = data that have been collected for some other purpose)
Library, Internet – books, journals, reports, magazines, bulletins, news letters, etc.,
Often the “core” of exploratory research
Benefits, minimal costs
Types of Exploratory Research: Experience Surveys
Gathering information from knowledgeable people (experts) on the issues relevant to the research problem
Note that there is no formal attempt to ensure that results are representative of any defined group of subjects, unlike with surveys conducted as part of descriptive research
Types of Exploratory Research: Case Analysis
A review of available information about a former situation that has some similarities to the present research problem
Some caveats; determine relevance of case:
How similar are the phenomena?
What situational factors have changed?
Types of Exploratory Research: Focus Groups
Small groups of people brought together and guided by a moderator through an unstructured, spontaneous (but focused) discussion
Useful technique for gathering information from a limited sample of respondents
Can be used to generate ideas, to gain insights into basic needs and attitudes, and to learn consumers’ “vocabulary” when they relate to a product
Types of Exploratory Research: Projective Techniques
Seek to explore hidden consumer motives for buying goods and services by asking participants to project themselves into a situation and then to respond to specific questions regarding the situation
Example: “Andrea Livingston never buys frozen dinners for her family because…”
Descriptive Research
Undertaken to answer who, what, where, when, and how questions. Examples:
Who are our customers?
What brands do they buy?
Where do they buy these brands?
When do they shop?
How did they find out about our products?
Answers to these are typically found in secondary data or by conducting surveys
Marketing managers need answers to these questions before they can formulate effective marketing strategies
Important: Note that we cannot answer “WHY?” questions with descriptive research
Types of Descriptive Research: Cross-sectional Studies
Measure units from a sample of the population at only one point in time “snapshots”
Very common in MR
Example: Questions asked: age, occupation, income, educational level, etc.,
Employ fairly large sample sizes, therefore referred to as “sample surveys.”
Types of Descriptive Research: Sample Surveys
Cross-sectional studies whose samples are representative of a specific population
Require that their samples be drawn according to a prescribed plan and to a predetermined number
Online survey research becoming increasingly common
Types of Descriptive Research: Longitudinal Studies
Repeatedly measure the same sample units of a population over a period of time
“Movies” of the population
“Panel” – Sample units who have agreed to answer questions at periodic intervals.
Often, the demographics of the panel are proportionate to the demographics in the total population (Census Bureau statistics)
ACNielsen, NFO Worldwide
Not limited to consumers; can have panels of building contractors, supermarkets, physicians, lawyers, etc.
Longitudinal Studies: Continuous Panels
Same questions are asked on each measurement
Example: firms wishing to track changes in consumers’ purchases, attitudes, etc., Brand-switching studies
Longitudinal Studies: Discontinuous Panels
Questions vary from measurement to measurement
Example: a marketer wanting to know how consumers feel about two different product concepts
Uses of Longitudinal Studies
Brand-switching studies
Studies examining how many consumers switched brands from one time period to the next
Market tracking studies
Measure some variable of interest over time (e.g., market share, unit sales)
Causal Research
“Causality” – Change in x brings about change in y
If x, then y
“What will happen if….?” “What makes consumers behave the way they do?”
In marketing, what will cause a change in consumer satisfaction, gain in market share, increase in sales?
Establishing causality is very difficult
Causal relationships can ONLY be determined by the use of experiments
Experiments
“Experiment” – manipulating an independent variable (IV) to see how it affects a dependent variable (DV), while also controlling the effects of additional extraneous variables (EV).
“Independent variable” – a variable over which the researcher has control and wishes to manipulate
“Dependent variable” – a variable that we have little or no direct control over, yet we have a strong interest in
“Extraneous variable” – one that may have some effect on a DV but yet is not an IV.
Experimental Design
A procedure for devising an experimental setting such that a change in a DV may be attributed solely to the change in an IV.
Symbols:
O = the measurement of a DV
X = the manipulation, or change, of an IV
R = random assignment of subjects to experimental and control groups
E = experimental effect; change in the DV due to the IV
O1 or O2 = measurement of DV at Time 1 and Time 2
Pretest – Measurement of the DV taken before changing the IV.
Posttest – Measurement of the DV taken after changing the IV.
“True” experimental design – one that truly isolates the effects of the IV on the DV while controlling for the effects of any extraneous variables.
Three designs:
After-only
One-group, before-after
Before-after with control group
After-Only Design
Achieved by changing the IV and, after some time, measuring the DV:
X O1
Note that this design does not measure up to requirement for a “true” experimental design: “Quasi-experimental” design
No measure of E, the experimental effect
One-Group, Before-After Design
First measure the DV, then change IV, and finally, take a second measure of the DV:
O1 X O2
Change in DV from Time 1 to Time 2
But, we cannot attribute the change in DV solely to the change in the IV.
No measurement of E (no control for the effect of extraneous variables).
Quasi-experimental design
Before-After with Control Group Design
“Control group” = a group whose subjects have not been exposed to the change in the IV
“Experimental group” = a group whose subjects have been exposed to the change in the IV
This design achieved by randomly dividing the subjects into two groups, the Control and the Experimental group. Pretest measurement of the DV is taken on both groups. IV is changed only in the Experimental group. Posttest measurements taken after some time of the DV in both groups:
Experimental group (R) O1 X O2
Control group (R) O3 O4
where E = (O2 – O1) – (O4 – O3)
True experimental design - can make conclusions about causality
Design assumes that the two groups are equivalent in all aspects
An experiment is “valid” if
The observed change in the DV is, in fact, due to the IV, and
If the results of the experiment apply to the “real world”
Two forms of validity:
Internal validity = the extent to which the change in the DV was solely due to the IV
External validity = the extent to which the relationship observed between the IV and the DV during the experiment is generalizable to the “real world.”
Types of Experiments: Laboratory
The IV is manipulated and measures of the DV are taken in a contrived, artificial, setting (so as to control as many extraneous variables as possible)
Advantages:
Allow the researcher to control for the effects of extraneous variables
Quick, less expensive than field experiments
Disadvantage:
Lack of a natural setting; generalizability is questionable
Types of Experiments: Field
The IV is manipulated and the measurements of the DV are made in their natural setting
Supermarkets, malls, retail stores, homes
Advantage: Manager can have more confidence that the results of the study will hold up in the “real world.”
Disadvantage:
Expensive, Time consuming
Be alert to the effect of extraneous variables
In marketing, field experiments are known as “test marketing”
Test Marketing
An experiment, study, or test that is conducted in a field setting
Test market cities – geographical areas selected in which to conduct the test.
Uses:
To test the sales potential for a new product
To gauge consumer or dealer reactions to variations in the marketing mix for a product
Expensive, time consuming, but justified if they can improve a product’s chances of success
Types of Test Markets
Standard test market:
Firm tests product through the company’s normal distribution channels
Time consuming, very expensive, not confidential, but are the best indicators of how the product will actually fare in the market
Controlled test market:
Conducted by outside research firms through prespecified types and numbers of distributors
Advantage: fast access to a distribution system set up for the test market.
Disadvantage: may or may not properly represent the firm’s actual distribution system
Electronic test market
Panel of consumers carry ID cards that each consumer presents when buying goods
Demographic information is automatically gathered
Advantage: Speed, greater confidentiality, less cost
Disadvantage: not the real market; consumers may be atypical
Simulated test market (STM)
A limited amount of consumer response information is fed into a mathematical model, which generates likely product sales volume
Advantage: fast, less cost, flexible, can be accurate predictors
Disadvantage: are not as accurate as full-scale test markets; very dependent on the assumptions built into the models.
Criteria for test markets:
Representativeness – it must be representative of the marketing territory in which the product will ultimately be distributed
Degree of isolation – isolated geographically, so that there are minimum spillover effects
Ability to control distribution and promotion
Pros and cons of test marketing
Allows for the most accurate method of forecasting future sales
Allows firm to pretest marketing mix variables
Do not yield infallible results
Competitors may intentionally sabotage test markets
Cost
Why measurement?
MR needs Information
Information is gathered via measurement
Three types of question-response formats
Closed-Ended: provides response options that can be answered quickly and easily.
Dichotomous (only 2 response options are provided)
Multiple Category (more than 2 response options)
Scaled-Response: uses a scale to measure a construct.
Unlabeled (a purely numerical scale, or only endpoints are identified)
Labeled (all of the scale positions are identified with a descriptor)
Which option of measurement should you use? Some considerations
Certain data collection methods are better suited for certain question formats – telephone interviews, mail or self-administered questionnaires.
Match the question format with the abilities of the respondents.
If you wish to use the data for further statistical analyses, you need to follow the assumptions of those statistical methods. Examples: simple percentages vs average
Objective properties
concrete, tangible, verifiable (age, income, no. of cans purchased, store last visited), is simple.
Subjective properties
cannot be directly observed because they are mental constructs (attitudes, intentions, opinions, emotions), is difficult.
Scale development
designing questions to measure the subjective properties of an object.
Qualitative and Quantitative Variables
“Categorical” Scales measure Qualitative Variables (e.g., M/F, Yes/ No). This is a scale that is typically composed of distinct categories.
“Metric” scales: used for measuring Quantitative Variables (age, income, how many children, etc.)
Four Levels of Measurement
Nominal Scales
Those that use only labels (type of dwelling, occupation, gender, Yes/No)
Least amount of information conveyed. Therefore considered the “crudest” or least sophisticated scale. Cannot perform advanced or sophisticated statistics with this data
Ordinal Scales
Permit the researcher to rank order the respondents or their responses (Indicate your first, second, and third choice of brands)
Indicate relative size differences among objects. They possess description and order, but we cannot tell how far apart the descriptors are on the scale.
Do NOT possess distance or origin
Interval Scales
Those in which the distance between each descriptor is known or assumed to be equal (distance from 3 to 4 is the same as that from 4 to 5)
The labels connote a continuum and the check lines are equal distances apart.
Higher level of measurement than nominal or ordinal scales; can subject data to more sophisticated statistical analyses.
Ratio Scales
A true zero origin exists ($ spent, time taken, no. of brands purchased, no. of children in family, years of college education)
Allows us to construct “ratios” when comparing results: “twice as heavy,” “one-half the income of,” “three times as costly”
Contain the greatest amount of information, the most sophisticated scale. Allow advanced or sophisticated statistical analyses.
The Modified Likert Scale
Respondents are asked to indicate their degree of agreement or disagreement on a symmetric agree-disagree scale for a series of statements.
Scale captures the intensity of their feelings.
The Lifestyle Inventory
A special application of the modified Likert question form
“Psychographics” – Activities, Interests, Opinions (AIO). Based on the premise that more than just demographic information can be used to aid marketing decisions.
Users/ nonusers, sports enthusiast, outdoorsy, price-conscious, child oriented, fashion-conscious, etc.,
A large number of lifestyle statements are presented, ranging from very general to very specific ones.
Ideally, more than one question for each dimension.
Widely used as a basis for market segmentation.
Tend to become very lengthy; require a great number of respondents.
The Semantic Differential Scale
Contains a series of bipolar adjectives for the various properties of the object under study. Respondents indicate their preferred location along the continuum.
“Our website is:
Very unattractive_ _ _ _ _ _ _ Very attractive
Friendly-unfriendly, high quality-low quality, good-bad, for me-not for me, etc.,
Questions should be randomly flipped to avoid “halo effect” (this is a general feeling toward an object that can bias responses to specific questions).
Researcher can compute averages and plot a “profile” of the object, e.g., brand image
Summated scale
If respondents reply with checkmarks from a list of options (“check all that apply”), the researcher can sum the checks to get a measure of the construct.
Anchored scale
one in which the endpoints are identified with the beginning and ending numbers of the scale (e.g., 1= very dissatisfied; 5 = very satisfied)
Reliable Measurements
one in which the same respondent responds in the same way to an identical or nearly identical question.
Valid Measurements
the accuracy of the measurement; a valid measure is one that is truthful
Face validity – the degree to which a measurement “looks like” it measures what it is supposed to measure.
Basic Data Analysis: Descriptive Statistics
Data entry can be done by:
Manual keyboard entry
Computer scanning of entire sets of questionnaires
Data entry requires a step called data coding – assigning codes to the possible responses for each question
Typically, codes are numerical
In large projects, a data code book may be used.
*SPSS - Know how to use the File – Display Data File Info – Working File or Utilities - Variables commands to produce/ print out variable info.
Types of Statistical Analyses used in MR
Marketing researchers work with data matrices. A data matrix is the coded raw data from a survey.
Columns – represent answers to the questions.
Rows – represent each respondent or case
When confronting a data matrix, a marketing researcher faces the problem of data reduction – how do I condense the data matrix into a few representative measures which can help me summarize the data and convey their salient characteristics?
Data reduction
What does data reduction accomplish?
Summarizes the data
E.g., “The average respondent’s age is…”
Applies understandable conceptualizations
E.g., “Few respondents are younger than 30”
Communicates underlying patterns
E.g., “Most customers seem to be satisfied with our price”
Generalizes sample findings to the population
E.g., “This means that 70-80% of the target market is…”
Five types of statistical analysis
Descriptive analysis: used to describe the data set
Inferential analysis: used to generate conclusions about the population’s characteristics based on the sample data
Differences analysis: used to compare the mean of the responses of one group to that of another group
Associative analysis: determines the strength and direction of relationships between two or more variables
Predictive analysis: allows one to make forecasts for future events
Descriptive Analysis
Mean, median, mode, standard deviation (“SD”), range
Used to describe the sample data matrix in such a way as to portray the “typical” respondent and to reveal the general pattern of responses.
Typically used early in the analysis; become foundations for subsequent analysis.
Inferential Analysis
“Inferential analysis” – generalize the results of the sample to the target population that it represents.
Meaning, the researcher makes conclusions about the population based on the sample results.
Include hypothesis testing; estimating true population values based on sample information.
Difference Analysis
Used to test the difference between or among groups – are the groups really different on this variable?
For example: One group exposed to ad for Brand A; one group has not seen the ad. Variable of importance: purchase of brand. Are there real and significant differences between the two groups?
t-test for significant differences; analysis of variance (ANOVA)
Associative Analysis
Used to determine systematic relationships among variables. “Associative analysis” investigates if and how two variables are related.
Analysis may indicate how strong the association and /or the direction (positive? negative?)
Techniques include cross-tabulations (“cross-tabs”) and correlations.
Predictive Analysis
Used to help the marketing researcher make predictions or forecasts about future events.
Regression analysis, time series analysis are examples of techniques.
Measures of Central Tendency
Goal: to report a single piece of information that describes the most typical response to a given question.
“Central tendency” – what response is most typical? Which is the most frequent response?
Mode- The value in a string of numbers that occurs MOST often
median- Expresses the value that lies in the middle of an ORDERED set of values. It is the value such that half the other values are greater than the median, and half are less than it.
mean- The average value characterizing a set of numbers. It is a measure that indicates the central tendency of the set of numbers.
Range
The range identifies the distance between the minimum and the maximum value in an ordered set of values.
Standard Deviation
The SD indicates the degree of variation in the values translated into a normal distribution.
SD indicates the limits within which 95% of the area under the curve lies
± 1.96
indicates the limits within which 99% of the area under the curve lies
± 2.58
Sample statistics
Values that are computed from a sample
Parameters
values that are computed from a complete census of the population
Inference
a form of logic in which you make a generalization about an entire class based on what you have observed about a small set of members of that class
Statistical inference
a set of procedures in which the sample size and sample statistics are used to make estimates of population parameters.
3 Types of Statistical Inference
Parameter Estimation
Hypothesis Testing
Tests of Significant Differences
Parameter Estimation
This is the process of using sample information to compute an interval that describes the range of a parameter such as the population mean (µ) or the population percentage (π).
Involves the use of 3 values:
Sample Statistic
The mean or percentage is derived from a sample, so it is the sample statistic.
2. Standard Error (“SE”)
There is usually some variability (about the mean, for example) in the sample.
Standard Error is a measure of the variability in the sampling distribution based on what is theoretically believed to occur if we were to take many, many independent samples from the same population.
Confidence Intervals
Are the degree of accuracy desired by the researcher and stipulated as a “level of confidence,” expressed as a %
Why? Because there is always some sampling error when a sample is taken, we need to estimate the population parameter with a range.
Researcher has to decide first how confident he/she wants to be
Typically, 90% (SE = ±1.64), 95% (SE = ±1.96) and 99% (SE = ±2.58) are used. Most common: 95%