Research Design
Chapter 3
Chapter Overview
1. Understand the research characteristics? 2. What are the sources of marketing information? 3. What types of errors affect research designs? 4. What methods are there for dealing with potential errors? 5. How do we choose a research design?
What is a Research Design?
• A way of arranging the environment…
– People, Places, Activities, Objects to be studied.
What is a Research Design?
• A way of arranging the environment…
– People, Places, Activities, Objects to be studied.
• There are three broad classes of research designs. These reflect the investigation purpose: – Exploratory – Descriptive – Causal
Exploratory Studies
• Purpose: identification and precise formulation of problems; formulation of new, alternative courses of action. • Often used in an introductory phase of a larger study in order to focus the study’s scope. • It has three stages:
– A search of secondary information sources
– information that has already been published
– Obtaining information from knowledgeable persons
– through interviews and focus groups
– The examination of analogous situations
– case histories and simulations; suggestive rather than conclusive
Exploratory Studies
• Focus Groups:
– 8-12 participants with a common background. – Used to identify problems and provide background rather than to provide solutions. – If the sole purpose is to create ideas, individual interviews are more effective than focus groups. – More specific uses: determine perceptions, detect the “language” of the customers, interpret new questions raised by quantitative research. – Can be conducted on-line through chat sessions and thus target hard-to-get segments. (this segment is better educated, younger, affluent)
Focus Groups
http://child.cornell.edu/army/focus.html Focus groups guide the design and conducting of subsequent large-scale quantitative survey. In this context, group interviews are used to accomplish several goals: 1. Identify and understand consumer language relating to the product category in question. What terms do they use? What do they mean? 2. Identify the range of consumer concerns. How much variability is there among consumers’ perception of the product, and in the considerations leading them to accept or reject the product?
3. Identify the complexity of consumer concerns. Do a few simple attitudes govern consumer reaction toward the product, or is the structure complex, involving many contingencies?
4. Identify specific methodological or logistical problems that are likely to affect either the cost of the subsequent research, or one’s ability to generate meaningful, actionable findings.
Volvo Focus Group
• Explored the differences between Volvo buyers and Volvo considerers (people who thought about buying a Volvo,but in the end bought some other car) What were the reasons for buying and resistances to buying a Volvo. •
Descriptive Studies
• Purpose: to describe market characteristics or functions • Areas:
– Product – Promotion – Distribution – Pricing – Existing groups and phenomena
Descriptive Studies
• Attempt to determine the association between variables AND to infer, not to establish, a causal relationship. • The investigator already knows about the research problem and follows a preplanned and structured design. • Critical aspects: proposed data analysis and project output.
Hockey Hot Dog Study
• Specific Questions number(s) and variable name(s):
Questions 5.0 and 5.1 ask the number of packages of hot dogs consumed per month by household usual number of hot dogs per package number of loose links consumed per month. Variable name is USAGE1, USAGE2, USAGE3.
•
Compute Household Usage of Hot Dogs per Month: Multiply number of packages per month by dogs per package and add loose links to obtain the total household usage of hot dogs per month. Reasons for including question(s): We want to segment into submarkets based on usage level. Usage volume predicts future purchase patterns, and may relate to interest in proposed new hot dog products and may support media selection that will reach high-usage households with the least cost.
•
Hockey Hot Dog Study
• Additional Analysis and Information Value: Relation to importance of product features:
User segments will be compared in terms of their mean responses to the ten-point importance-attribute scales, which include: package brand name quality contents taste texture smell shape color of the hot dogs.
Causal Studies
• Purpose: find the causes of what it is being predicted. • Terminology:
– Causal relationships: “Did the price by itself cause the increase in sales?” – Deterministic causation: “Did the price cut cause the sales increase?” – Probabilistic causation: “How likely is price reduction of X to produce (probabilistic cause) the sales increase of Y?”
Causal Relationships
A manufacturer of color television sets reduces the wholesale price of his sets by 10 percent, and that this reduction is passed on to the consumer. Further assume that sales to the consumer rose by 15 percent during the succeeding three months, compared to a similar period prior to the price reduction.
Did the price cut by itself cause the increase in sales (Figure 3.1A), Did it only contribute, along with other marketing actions (Figure 3.1B), Did it have no effect whatsoever (Figure 3.1C)?
Causal Relationships
Deterministic Causation
Rodeo Advertising Increases Sales: Increase in purchase rate for spectators x Market Size x Profit Margin Source: Wrangler.com
Bases for Inferring Causal Relationships:
1. Associative variation (or “concomitant variation”): measure of the extent to which occurrences of two variables are associated.
• Two types: (a) between two variables, (b) between the changes of two variables.
•
A high degree of association ≠ proof of a causal relationship
2. Sequence of events: the causal factor occurs first.
3. Absence of other causal factors.
Causal Studies
Issues in Determining Causation: • What is the source of causality? (X Y or Y X?) • What is the direction of causality? Is the influence positive or negative? • Is X a “necessary and sufficient” or a “necessary but not sufficient” cause of Y? Is X’s causation deterministic or probabilistic? • Which value of the believed cause exerts a causal influence – its presence or absence? • Are the causes and effects the states themselves or changes in the states? Is the relationship static or dynamic?
Causal Studies
• Causal inference studies:
1. Natural experiments:
• • Involve investigator intervention only to the extent of measurement Provide associative variation and sequence of events
Causal Studies
• Causal inference studies:
2. Controlled experiments:
• • • Manipulate at least one causal variable Random assignment of subjects to experimental and control groups Provide associative variation, sequence of events and evidence of the absence of other possible causes.
Sources of Marketing Information
1. Secondary sources 2. Respondents
• • From communication with respondents From observation of respondents
3. Natural experiments 4. Controlled experiments 5. Simulation
• “Experimenting” with a model of a real-life operational situation
Types of Errors Effecting Research Designs
Total error = Sampling error + Non-sampling error
Measured as total error variance:
(Total error)2 = (Sampling error)2 + (Non-sampling error)2 • Sampling error: introduces variability into the precision with which a sample statistic is calculated • Non-sampling errors: random (example: misrecording) or systematic (example: loaded question) • Increasing the sample size reduces the sampling error and can increase the non-sampling error
How errors add up
Types of errors:
• • • • • • • •
Sources of errors:
Population specification • Researcher Sampling • Sample Selection • Interviewer Frame • Instrument Non-response • Respondent Surrogate information Measurement Methods for dealing with potential errors: Experimental • Minimize errors through research design • Measure or estimate the error or its impact
Summary Questions
• Company A is doing concept tests for a series of advertisements to determine which is best received. Is this an exploratory, descriptive or causal study?
• How could you make this study causal predictive?