1. Introduction to Marketing Research
CHAPTER 8: CAUSAL RESEARCH DESIGN - EXPERIMENTATION
Idil Yaveroglu Lecture Notes

2. Experimentation as Conclusive Research
Descriptive
Causal
Experimentation
Field
Experiments
Laboratory

3. Concept of Causality
Causality is the relationship between an event (the cause) and a second event (the effect), where the second event is understood as a consequence of the first.

4. Conditions for Causality
Concomitant variation is the extent to which a cause, X, and an effect, Y, occur together or vary together in the way predicted by the hypothesis under consideration.
- X & Y are correlated
The time order of occurrence condition states that the causing event must occur either before or simultaneously with the effect; it cannot occur afterwards.
The absence of other possible causal factors means that the factor or variable being investigated should be the only possible causal explanation.

5. Conditions for causality (1)
Sales levels are higher at times of increased advertising spending.
Not sufficient for causality: the fact that advertising and sales move together does not prove casuality. Condition 1:
1. The two variables must covary with one another (concomitant variation)

6. Conditions for causality (2)
Example: A study finds that children with anti-social behavioral problems have worse relationships with their parents than well-behaved children.
Not sufficient for causality: direction of causality? Condition 2:
2. Cause must precede the effect in time (time order of occurrence)

7. Conditions for causality (3)
We let prisoners choose: some receive psychological council and some don’t. The first show more improvement in behavior.
Not sufficient for causality: Those prisoners who volunteer for council may be more motivated to get on the right track. Condition 3:
3. Covariation should not be caused by a third variable

8. Evidence of causality?
“School children with larger shoe size have higher math ability”
“In periods of increased ice cream sales, more people die due to drowning”
“Areas with more storks have higher birth rates”

9. Definitions and Concepts
Independent variables are variables or alternatives that are manipulated and whose effects are measured and compared, e.g., price levels.
Test units are individuals, organizations, or other entities whose response to the independent variables or treatments is being examined, e.g., consumers or stores.
Dependent variables are the variables which measure the effect of the independent variables on the test units, e.g., sales, profits, and market shares.
Extraneous variables are all variables other than the independent variables that affect the response of the test units, e.g., store size, store location, and competitive effort.

10. Experimental Design
An experimental design is a set of procedures specifying
the test units and how these units are to be divided into homogeneous subsamples;
what independent variables or treatments are to be manipulated;
what dependent variables are to be measured; and
how the extraneous variables are to be controlled.

11. Validity in Experimentation
Internal validity refers to whether the manipulation of the independent variables or treatments actually caused the observed effects on the dependent variables. Control of extraneous variables is a necessary condition for establishing internal validity.
External validity refers to whether the cause-and-effect relationships found in the experiment can be generalized. To what populations, settings, times, independent variables, and dependent variables can the results be projected?
External validity is a main concern in experimental research, particularly when conducted in a contrived laboratory setting.
Take for example an experiment on the effect of electricity. If you shock someone (experimental treatment) and he jumps (observed effect), and he jumps only because of the shock and for no other reason, then internal validity exists.
Let’s say a small bakery in Denver, Colorado, wants to know whether or not putting additional frosting on its new cakes would cause customers to like the cakes better. Using an experimental design, it tests the hypothesis that its customers like additional frosting on their cakes. As the “amount of frosting” (independent variable) is manipulated, the bakery discovers that additional frosting allows the cakes to stay moist longer. Consequently, it might be the moistness and not the frosting that causes customers’ positive reaction to the new cakes. In assessing internal validity, control groups are developed consisting of customers who were not exposed to the independent variable but all other conditions were kept the same. Adding the control group to the experimental design reduces the possibility that the observed effect of the new cakes is caused by some other factor other than the treatment. In this case, if it is the frosting, then the treatment group should like the new cake more than the control group. But if it were the moistness, then both groups would like the new cake equally.
External validity: Surely, a fear-of-heights scale administered on top of a mountain would have different results from one administered in a classroom.

12. Construct validity and reliability
Repeated observations give same result
Internal consistency of a measurement scale: different measures are used and they measure the same thing (e.g. multi-item scales)
When you code variables: inter-rater reliability
(Internal) construct validity
Do I measure what I intend to measure?
Absence of systematic error
let’s assume researchers want to use a construct called “motivation to succeed” (MTS) as an independent variable in predicting the likelihood of individuals’ life success. Subjects measuring the highest on the MTS construct would be those most likely to have succeeded in life. The fundamental question that must be addressed is “What are the observable, real-life indicators of such success?” Using a process referred to as specifying the domain of observable subcomponents related to the construct,7 let’s assume that peer respect, academic achievement, and personal financial security represent success in the society being investigated. The precise nature of this specification is necessary so that hypothesized relationships can be empirically tested with real data.

13. Example of multi-item scale
Social satisfaction
This supplier expresses criticism tactfully.
Interactions between my firm and this supplier are characterized by mutual respect.
This supplier leaves me in the dark about things I ought to know. (R)
7-point completely disagree-completely agree
(R) = reverse coded

14. Construct validity and reliability
Analogy of a good shot
15 shots with three guns I, II, and III
Question: What is the quality of these guns (in terms of internal validity and reliability)? I II
III
Hits scattered around
Unreliable & not valid
Hits close to each other, but not in the bull’s-eye
Reliable & not valid
All hits in the bull’s-eye
Reliable & valid

15. Controlling Extraneous Variables
Randomization refers to the random assignment of test units to experimental groups by using random numbers. Treatment conditions are also randomly assigned to experimental groups.
Matching involves comparing test units on a set of key background variables before assigning them to the treatment conditions.
Statistical control involves measuring the extraneous variables and adjusting for their effects through statistical analysis.
Design control involves the use of experiments designed to control specific extraneous variables.

16. A Classification of Experimental Designs
Pre-experimental designs do not employ randomization procedures to control for extraneous factors. Examples are: the one-shot case study, the one-group pretest- posttest design, and the static-group.
In true experimental designs, the researcher can randomly assign test units to experimental groups and treatments to experimental groups. Examples are: the pretest-posttest control group design, the posttest-only control group design, and the Solomon four-group design.

17. A Classification of Experimental Designs (Cont.)
Quasi-experimental designs result when the researcher is unable to control the scheduling of treatments or must assign respondents to treatments in a nonrandom fashion.
Used mainly in field setting.
Examples are: time series and multiple time series designs.
A statistical design is a series of basic experiments that allow for statistical control and analysis of external variables. Examples are: randomized block design, Latin square design, and factorial designs.

18. A Classification of Experimental Designs
Pre Experimental
One-Shot Case Study
One Group Pretest-
Posttest
Static Group
Statistical
Factorial Design
True Experimental
Pretest-Posttest Control Group
Posttest-Only Control Group
Quasi-Experimental
Time Series
Multiple Time Series

19. One-Shot Case Study
X 01
A single group of test units is exposed to a treatment X.
A single measurement on the dependent variable is taken (01).
There is no random assignment of test units.
The one-shot case study is more appropriate for exploratory than for conclusive research.
Extraneous variables not controlled, weak on internal validity

20. One-Group Pretest-Posttest Design
01 X 02
A group of test units is measured twice.
There is no control group.
The treatment effect is computed as
02 – 01.
The validity of this conclusion is questionable since extraneous variables are largely uncontrolled.
Reactive error: consumers react to the situation rather than the variable
Demand artifacts: guess the purpose
Interactive testing effect: consistency

21. Static Group Design EG: X 01 CG: 02 A two-group experimental design.
The experimental group (EG) is exposed to the treatment, and the control group (CG) is not.
Measurements on both groups are made only after the treatment.
Test units are not assigned at random.
The treatment effect would be measured as

22. True Experimental Designs: Pretest-Posttest Control Group Design
EG: R 01 X 02
CG: R
Test units are randomly assigned to either the experimental or the control group.
A pretreatment measure is taken on each group.
The treatment effect (TE) is measured as:( ) - ( ).
Interactive testing effect is not controlled.

23. Posttest-Only Control Group Design
EG : R X 01
CG : R
The treatment effect is obtained by
TE =
Except for pre-measurement, the implementation of this design is very similar to that of the pretest-posttest control group design.

24. Quasi-Experimental Designs: Time Series Design
There is no randomization of test units to treatments.
The timing of treatment presentation, as well as which test units are exposed to the treatment, may not be within the researcher's control.

25. Statistical Designs: Factorial Design
Statistical designs consist of a series of basic experiments that allow for statistical control and analysis of external variables and offer the following advantages:
The effects of more than one independent variable can be measured.
Specific extraneous variables can be statistically controlled.
Economical designs can be formulated when each test unit is measured more than once.

26. An Example of a Factorial Design
Amount of Brand Information
Amount of Humor
No Humor
Medium Humor
High Humor
Low
Medium
High

27. Table 8.2 Laboratory versus Field Experiments
Factor
Laboratory
Field
Environment
Artificial
Realistic
Control
High
Low
Reactive error
Demand artifacts
Internal validity
External validity
Time
Short
Long
Number of units
Small
Large
Ease of implementation
Cost

28. Limitations of Experimentation
Experiments can be time consuming, particularly if the researcher is interested in measuring the long-term effects.
Experiments are often expensive. The requirements of experimental group, control group, and multiple measurements significantly add to the cost of research.
Experiments can be difficult to administer. It may be impossible to control for the effects of the extraneous variables, particularly in a field environment.
Competitors may deliberately contaminate the results of a field experiment.

29. What type of experiment is described?
A healthcare company wants to study the impact of unemployment on health. They send one survey of unemployed people in a specific local area to assess their health status.
>> One shot design

30. What type of experiment is described?
A soft drink manufacturer puts the same brand of orange juice in two containers with different designs. Subjects are randomly assigned to three groups. Two groups are given a package with different designs and asked about the taste. A third group is given the orange drink in an unlabeled package and asked the same questions.
>> Posttest only control group design
EG1 : R X 01
EG2 : R X 02
CG : R

31. Example Background of the managerial decision problem of a retailer:
One of the main challenges for retailers is how to organize their shelf-space to make shopping an attractive experience for their customers. Lately, customers have been complaining that the assortment of the biscuits section is not sufficiently diverse. The retailer realizes that it is physically not possible to increase shelf-space. However, it is possible to increase variety (by allocating less shelf-space to some biscuits and adding new biscuits). Of course, it remains to be seen if customers will purchase more biscuits when the assortment becomes more diverse.
(1) Formulate a marketing research question.
(2) Design an experiment that enables you to address the marketing research question that you have formulated. Motivate your choices.