Marketing Research Chap 6 - Experimentation and Causal Research

Factorial design

The most common statistical designs, used to measure the effects of two or more independent variables at various levels and to allow for interactions between variables, formatted as a table.

Static Group Design

there are 2 groups: experimental group (EG), which is exposed to the treatment, and the control group (CG). Measurement on both groups are made only after the treatment, and test units are NOT assigned at random. Lack of randomization leaves the experiment open to some extraneous effects, two groups might differ before the treatment. USUALLY the control group receives the current level of marketing activity rather than no treatment at all, it is impossible to reduce marketing input (ex. price) to zero. Represented by Experimental Group: X O(1) Control Group: O(2).

advantages of statistical designs

1) Effects of more than one independent variable can be measured. 2) Specific extraneous variables can be statistically controlled.

experimentation limits

1) Time: More time = more accurate field experiments (ex. examining long term effect of promotional campaign) 2) Cost: New product research can be extremely expensive. 3) Administration: Controlling the effects of extraneous variables is an essential aspect of experimental research.

Conditions for causality inferences

1) concomitant variation 2) time order of occurrence of variables 3) elimination of other possible causal factors. These conditions are NECESSARY, but not sufficient to demonstrate causality.

goals for experiments

1) draw valid conclusions about the effects of the independent variables on the test units (concerns internal validity) 2) make valid generalizations to a larger population of interest (concerns external validity)

experimental design

1) the test units and sampling procedures, 2) independent variables 3) dependent variables 4) how to control the extraneous variables

External validity

A determination of whether the cause-and-effect relationships found in the experiment can be generalized. Can the results be generalized beyond the experimental situation and in what settings? Threats to external validity arise when the set of experimental conditions do not realistically take into account other relevant variables in the real world.

trade off

It is desirable to have both internal and external validity, but in applied marketing research, we often must trade one type for the other. Ex) To control for extraneous variables, a researcher may conduct an experiment in an artificial environment. Factors that threaten internal validity may also threaten external validity, most serious of these being extraneous variables

Experimentation

Research technique used in causal research, primary method for establishing cause and effect relationships in marketing.

Disadvantage of Factorial Design

Statistical procedures are used to analyze the treatment effects, main disadvantage of factorial design is that the number of treatment combinations increases multiplicatively.

selecting an experimental design

You can attain high internal and external validity by using different designs at different stages of the project. Lab experiments offer tight internal validity should be used in the beginning, and pre-experimental designs may suffice. Field experiments could be used during later stages of the project to enable generalization of results. True experimental and factorial designs are more appropriate during the later stages.

One shot case study

a single group of test units is exposed to a treatment X, and then a single measurement on the dependent variable is taken. Test units are NOT assigned at random. Lacks a control group and randomization, so WEAK internal validity. Represented by X O(1).

test marketing

application of a controlled experiment done in limited but carefully selected test markets. It involves a replication of the planned national marketing program for a product in the test markets. Two main objectives: 1) determine market acceptance of the product 2) test alternative levels of marketing mix variables.

concomitant

condition for INFERRING causality, requires that cause and effect occur together and vary together as predicted by the hypothesis

pre-experimental designs

designs that don't control for extraneous factors by randomization.

Posttest-Only Control Group Design:

does not involve any premeasurement. Subjects are randomly assigned to either experimental or control group. After administering the treatment to the experimental group, both groups are measured: Experimental group: R X O(1) Control group: R O(2) The treatment effect is the difference between the experimental and control group measurements: TE = O(1) - O(2) Advantages: time, cost, and sample-size. MOST POPULAR experimental design in marketing research. BUT since you didn't randomize to equalize groups, without pretest, there is no way to verify group similarity.

statistical design

experiment design that allows for the statistical control and analysis of extraneous variables.

One-Group Pretest-Posttest Design

group of test units is measured twice, before and after exposure to the treatment. Test units are NOT assigned at random. Also, lacks a control group for comparison, but better than one-shot. BUT weak validity because extraneous variables are uncontrolled. Represented by O(1) X O(2).

role of evidence

if evidence is strong and consistent, it may be reasonable to conclude that there is a causal relationship. Accumulated evidence from several investigations increases our confidence that a causal relationship exists.

internal validity

measure of accuracy of an experiment. It measures if the manipulation of the independent variables actually caused the effects on the dependent variables. It is the basic minimum that must be present in an experiment before any conclusion about treatment effects can be made.

dependent variables

measure the effect of the independent variables on the test units, ex. Sales, profits, and market shares.

Causality

occurrence of X increases the probability of the occurrence of Y. Experimentation is used to infer causal relationships. Marketing effects are caused by MULTIPLE variables, relationship between cause and effect is PROBABILISTIC. You can NEVER prove causality; only infer a cause-and-effect relationship.

true experimental designs

researcher can randomly assign subjects/ test units to experimental groups and randomly assign treatments to experimental groups.

confounding variables

same as extraneous variables, used to illustrate that extraneous variables can confound the results by influencing the dependent variable. They can be controlled by RANDOMIZATION = randomly assigning test units and treatment conditions to experimental groups by using random numbers.

Pretest-Posttest Control Group Design

subjects are randomly assigned to either experimental or control group. Pretreatment measure is taken on each group. After administering treatment to experimental group, both groups are measured again. Experimental Group: R O(1) X O(2) Control Group: R O(3) O(4) Treatment effect is measured as: (O2-O1) - (O4-O3)

Interaction

when the simultaneous effect of two or more variables is different from the sum of their separate effects. Ex) someone likes coffee and their favorite temperature level is cold, but they might not like cold coffee.