CHAPTER 8
CAUSAL RESEARCH DESIGN: EXPERIMENTATION
OPENING QUESTIONS
1.
2.
3.
4.
5.
6.
7.
8.
9.
How is the concept of causality defined in marketing research and how do we distinguish
between the ordinary meaning and the scientific meaning of causality?
What are the conditions for causality? Can a causal relationship be demonstrated
conclusively?
How do we define and differentiate the two types of validity: internal and external?
What are the various types of experimental designs, and what are the differences among
preexperimental, true experimental, quasi-experimental, and statistical designs?
How can we compare and contrast the use of laboratory versus field experimentation and
experimental versus nonexperimental designs in marketing research?
What is test marketing and how does it involve experimentation?
Why is the internal and external validity of field experiments that are conducted overseas
generally lower than in the United States?
How does technology facilitate causal research?
What ethical issues are involved in conducting causal research, and how can debriefing
address some of these issues?
AUTHOR'S NOTES: CHAPTER FOCUS
This chapter provides an overview of causal research and experimentation. The concept
of causality and the conditions for causality are described. Internal and external validity in
experimentation are discussed in detail. A classification of experimental designs is presented and
pre-experimental, true experimental, quasi-experimental, and statistical designs are described.
Comparisons between laboratory versus field experiments, and experimental versus
nonexperimental designs are made. Test marketing is presented as an application of
experimental designs. This chapter covers the material on causal research contained in
competing texts; yet, the treatment of experimentation is more extensive than that found in
competing texts.
This chapter could be taught by focusing on the opening questions sequentially. Greater
emphasis could be placed on the concept and conditions for causality (Opening Questions 1 and
2), internal and external validity (Opening Question3), and the various types of experimental
designs (Opening Question 4). However, quasi experimental designs and statistical designs may
not be covered in detail. Conducting experiments internationally (Opening Question 7), ethical
issues (Opening Question 9), and use of the Internet should be discussed.
CHAPTER OUTLINE
1.
2.
3.
Overview
Concept of Causality
Conditions for Causality
i. Concomitant Variation
ii. Time Order of Occurrence of Variables
103
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
iii. Absence of Other Possible Causal Factors
iv. Role of Evidence
What is Experimentation?
Definition of Symbols
Validity in Experimentation
i. Internal Validity
ii. External Validity
A Classification of Experimental Designs
Pre-experimental Designs
i. One-Shot Case Study
ii. One-Group Pretest-Posttest Design
iii. Static Group
True Experimental Designs
i. Pretest-Posttest Control Group Design
ii. Posttest-Only Control Group Design
Quasi Experimental Designs
i. Time Series Design
Statistical Designs
i. Factorial Design
Selecting an Experimental Design
Laboratory Versus Field Experiments
Limitations of Experimentation
i. Time
ii. Cost
iii. Administration
Application: Test Marketing
Summary Illustration Using the Opening Vignette
International Marketing Research
Technology and Marketing Research
Ethics in Marketing Research
Summary
Key Terms and Concepts
Acronyms
TEACHING SUGGESTIONS
Opening Question 1
* Discuss the concept of causality as used in marketing.
You may want to begin the class by asking a few students what the term ‘causality’ means
and writing components of their definitions on the board. Afterwards, be sure to stress to the
students that marketing effects are caused by multiple variables and the relationships tend to be
probabilistic. As such, it is not possible to conclusively prove causality. Thus, we can only infer
a cause-and-effect relationship between the variables.
* Via an example, explain the three necessary conditions to infer causality -
104
1.
2.
3.
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. Evidence
pertaining to concomitant variation can be obtained in a qualitative or quantitative manner.
Thus, we may hypothesize that increasing product price will lead to an increase in profits.
Time order of occurrence of variables: the causing event must occur either before or
simultaneously with the effect; it cannot occur afterwards. Thus, our profits cannot increase
until after we increase the price.
Absence of other possible causal factors: means that we can infer causality if all other
factors affecting the subject are absent, i.e., these factors are held constant. Thus, other
factors like store penetration, expanded consumer spending, or decreased costs must be held
constant.
* Review the key definitions related to experimentation -
1.
2.
3.
4.
5.
6.
Independent variable: variables or alternatives that are manipulated and whose effects are
measured and compared, such as price levels, different package designs, and advertising
themes.
Test units: individuals, organizations, or other entities whose response to the independent
variables or treatments is being examined, such as consumers, stores, and geographic areas.
Dependent variables: variables that measure the effect of the independent variables on the
test units, such as sales, profits, and market shares.
Extraneous variables: all the variables other than the independent variables or treatments
that affect the response of the test units to the treatments, such as store size, store location,
and competitive effort.
Experiment: the process of manipulating one or more independent variables to determine
their effect on the dependent variable, controlling for the effect of extraneous factors.
Experimental design: a set of procedures specifying (1) the test units and how these units
are to be divided into homogeneous sub-samples; (2) independent variables or treatments
that are to be manipulated; (3) dependent variables are to be measured; and (4) how the
extraneous variables are to be dealt with.
* Introduce the symbolic notation used in marketing research.
1.
X
=
2.
O
=
3.
R
=
the exposure of a group to an independent variable or
treatment or event, the effects of which are to be
determined
the process of observation or measurement of the
dependent variable on the test units or group of units
the random assignment of test units or groups to
separate treatments
The following conventions are associated with the symbols:
1.
Movement from left to right indicates movement through time.
2.
Horizontal alignment of symbols implies that all those symbols refer to a specific
treatment group.
105
3.
Vertical alignment of symbols implies that those symbols refer to activities or events that
occur simultaneously.
Thus, the first example in the text, X
O1 O2, indicates a treatment was given to a
sample and measurements at two different points in time were made.
Opening Question 2
*
Discuss the conditions for causality.
At least three conditions must be satisfied in order to justify the inference of a causal
relationship between two variables: (1) concomitant variation, (2) time order of occurrence of
variables, and (3) absence of other possible causal factors. These conditions are necessary but not
sufficient to demonstrate causality. By that we mean that they must be satisfied to justify
drawing a causal inference; however, their presence does not guarantee that we have isolated the
true variables responsible for the effects we are observing.
Describe and illustrate the three conditions to the students.
Opening Question 3
*
Define validity and describe the difference between internal and external validity.
Start by defining validity, in an experimental setting, as the ability to draw appropriate
conclusions about the effects of independent variables and to make reasonable generalizations to
a larger population of interest. With this definition, we can distinguish internal and external
validity. Internal validity examines whether the manipulation of the independent variables or
treatments actually caused the effects on the dependent variables, while external validity
examines whether the cause-and-effect relationships found in the experiment can be generalized.
Because of the conceptual difficulty in understanding validity, some examples that depict
a research design lacking one of the two types of validity would be instructive. For example, if
we hypothesize that selling through national chains will increase profits, we must hold all other
causes of increased profits constant in order to conclude internal validity. External validity can
be seen in a study that randomly selects stores from all markets in which a national chain
operates. Assuming that the study takes into account all realistic factors, any findings that hold
for all these markets will generalize to all stores in the national chain.
Opening Question 4
* Discuss the differences in the types of experimental designs.
1.
2.
3.
Pre-experimental Designs: do not apply randomization, thus, they can be influenced by
extraneous factors.
True Experimental Designs: randomization is included, thus, allowing researchers to
randomly assign test units and treatments to experimental groups.
Quasi Experimental Designs: lacks control over the scheduling of treatments and the
ability to randomly expose test units to treatments.
106
4.
Statistical Designs: a series of several basic experiments that are conducted
simultaneously. These designs suffer from the same shortcomings as the actual designs
of the basic experiments.
Figure 8.4 can be used here to organize your discussion on experimental designs.
* Describe pre-experimental designs and their shortcomings -
Note that pre-experimental designs include the one-shot case study, the one-group pretestposttest design, and the static group design. Stress that the salient feature of these designs is that
they do not employ randomization procedures to control for extraneous factors. For example, in
the one-group pretest-posttest design, an observation on attitudes toward the brand is taken
before and after watching a commercial, but randomization is not used in selecting observers;
thus, the results cannot be extended to the population as a whole.
*
List the significant features of a true experimental design and provide examples.
First, to introduce the ensuing discussion, distinguish true experimental designs by the
fact that the researcher can randomly assign test units to experimental groups and also randomly
assign treatments to experimental groups.
Examples of True Experimental Designs:
1. Pretest-Posttest Control Group Design: this design controls for most of the
extraneous variables. Selection bias is eliminated by randomization, and interactive
testing effect is not controlled due to the effect of the pretest measurement on the
reaction of units in the experimental group to the treatment.
2. Posttest-Only Control Group Design: compared to the Solomon four-group design,
this design is fairly simple to implement because there is no premeasurement, thus,
the testing effects are eliminated. However, this design is sensitive to selection bias
and mortality.
3. Solomon Four-Group Design: the Solomon four-group design overcomes the
limitations of the pretest-posttest control group and the posttest only control group
designs in that it explicitly controls for interactive testing effect, in addition to
controlling for all extraneous variables. This design allows the researcher to update
and measure the various effects.
It will prove instructive to use a running example in explaining these designs. A standard
example is to use advertising research in the measurement of consumer attitudes toward a
commercial or the brand being advertised.
* Explain the analysis of quasi experimental designs.
Mention that in a quasi experimental design, the researcher can control when
measurements are taken and on whom they are taken, but the researcher is lacking control over
the scheduling of the treatments and the ability to randomly expose test units to the treatments.
Quasi experimental designs are quicker and less expensive than true experimentation, but should
only be used when a true experimental design is not feasible.
107
* Explain the nature and uses of statistical designs.
Describe statistical designs as a series of basic experiments that enable statistical control
and analysis of external variables; thus several basic experiments are conducted simultaneously.
Advantages:
1.
The effects of more than one independent variable can be measured.
2.
Specific extraneous variables can be statistically controlled.
3.
Economical designs can be formulated when each test unit is measured more
than once.
Examples of a randomized block design, Latin Square design, and Factorial design can be
constructed for demonstration purposes. See Table 8.1
Opening Question 5
* Start a class discussion on whether laboratory or field experiments are more useful in
marketing research.
Begin this discussion by simply asking students which form of experimentation is better.
You may want to note that laboratory experiments make up the bulk of consumer research
because of their ability to control extraneous variables and their relative efficiency in gathering
data. However, for certain studies, field experiments are used. For example, Coca-Cola counts
the shelf space it and its competitors receive in local grocery stores when promotional variables
are manipulated. Thus, each type of experimentation has its role to play in marketing research,
but because causality cannot be inferred from field experiments, laboratory experiments
predominate.
See Table 8.2 for a summary of laboratory vs. field experimentation.
* Discuss why nonexperimental designs are limited for inferring cause-effect
relationships.
Mention that descriptive studies are sometimes used to infer causality, but this may not be
appropriate. In descriptive studies, prior equivalence of the groups with respect to both the
independent and dependent variables is difficult to establish, as is time order of occurrence of the
variables. Finally, it offers little control in eliminating other extraneous factors.
* Discuss the limitations of experimentation Start by asserting that theoretically, experimentation is a very powerful way to uncover
causal relationships. However, in practice, real world considerations prevent it from achieving
its optimal effectiveness. These considerations are:
108
1.
2.
3.
Time: experiments can be time consuming; however, the experiments should be
long enough in duration so that the post-treatment measurements include most, or all
of the effects of the variables.
Cost: the requirements of experimental group, control group, and multiple
measurements significantly add to the cost of the experiment.
Administration: experiments can be difficult to administer, and it is not always
feasible to control for the effects of the extraneous variables.
Opening Question 6
* Define and classify test marketing.
Define test marketing as an application of a controlled experiment, done in a limited but
carefully selected part of the marketplace called test markets, which involves a replication of the
planned national marketing program for a product. In a standard test market the product is sold
through regular distribution channels. The company's own salesforce is responsible for
distributing the product, stocking the shelves, and taking inventory at regular intervals. A
controlled test marketing program is conducted by an outside research company. The research
company is responsible for all activities associated with the test market. Simulated test markets
yield mathematical estimates of market share based on initial reaction of the consumers to the
new product.
Opening Question 7
* Examine the difficulties involved with international experiments.
It is difficult to control for the time order of occurrence of variables and it is difficult to
control for the absence of other possible causal factors, two of the necessary conditions for
causality in an experiment.
• Control of extraneous variables is particularly problematic.
• It may not be possible to address this problem by adopting the most appropriate
experimental design, as environmental constraints may make that design infeasible.
• The internal and external validity of field experiments conducted overseas is generally
lower than in the United States.
EXAMPLE: M&M/Mars has set up massive manufacturing facilities in Russia and
advertises its candy bars on television. Yet, sales have been low. Is Mars advertising too much,
too little, or just right? While the answer could be determined by conducting a field experiment
that manipulated the level of television advertising, such causal research is not feasible given the
tight control of television programming and advertising by the Russian government. TV stations
are owned and operated by the government with severe restrictions on television advertising.
This makes field experiments manipulating advertising levels extremely difficult.
109
Opening Question 8
* Describe how technology is facilitating experimentation.
Virtual Reality (VR) is a real-time, 3D environment made to represent either reality or an
environment out of someone's imagination. These environments are created by high-powered
computer systems. As with many of the technological innovations used in marketing research,
VR was not developed specifically with research in mind, but VR is finding its way into the field
of marketing research nonetheless and holds great promise for conducting causal research. Using
VR, the researcher can create an environment that represents the field (marketplace) and yet
exercise the degree of control possible only in a laboratory setting. Discuss Visionary Shopper
by MarketWare Simulation Services.
Opening Question 9
* Discuss the ethical responsibilities resulting from disguised experiments.
It is often necessary in experimentation to disguise the purpose of the research to produce
valid results. Disguising the purpose of the research should not lead to deception, however.
Although this seems like a paradox, one solution would be to disclose the possible existence of
deception before the start of the experiment and allow the participants the right to redress at the
conclusion of the experiment. The following four items should be conveyed: (1) inform
respondents that in an experiment of this nature a disguise of the purpose is often required for
valid results, (2) inform them of the general nature of the experiment and what they will be asked
to do, (3) make sure they know that they can leave the experiment at any time, and (4) inform
them that the study will be fully explained after the data have been gathered and at that time they
may request that their information be withdrawn. The procedure outlined in item (4) is called
debriefing.
One further ethical concern in experimentation involves using the appropriate experimental
design to control errors caused by extraneous variables. It is the responsibility of the researcher
to use the most applicable experimental design for the problem. Determining the most
appropriate experimental design for the problem requires not only an initial evaluation but also
continuous monitoring.
Internet Emphasis
* Indicate that the Internet and computers are helpful in the analysis of experimental
design data.
Different experimental treatments can be displayed at different Web sites. Respondents
can then be recruited to visit these sites and respond to a questionnaire that obtains information
on the dependent and extraneous variables. Thus, the Internet can provide a mechanism for
controlled experimentation, although in a laboratory type of environment.
To complement the Internet, microcomputers and mainframe software can be used in the
design and analysis of experiments. The comprehensive statistical-analysis software package
Minitab can be used to design experiments. Although similar in use to, SPSS or SAS, Minitab
110
includes functions and documentation specifically for industrial quality control work where
factorial designs are encountered.
BE AN MR! AND BE A DM!
It should be noted that a variety of answers are appropriate. The ones given here are merely illustrative.
Be a DM!: FTC:
To ensure the smooth operation of our free market system, the Federal Trade Commission (FTC)
enforces federal consumer protection laws that prevent fraud, deception and unfair business practices.
The Commission also enforces federal antitrust laws that prohibit anticompetitive mergers and other
business practices that restrict competition and harm consumers. Whether combating telemarketing
fraud, Internet scams or price-fixing schemes, the FTC’s primary mission is to protect consumers.
- Information overflow; difficult for customers to differentiate noise and useful information
- Information flow cannot be controlled or traced. This can lead to misrepresentation of
information, fraud etc.
- Dumping of products especially from foreign countries (web based selling of drugs by Canadian
and Indian companies)
- Increased possibility of collusion between companies
Be an MR!: FTC:
The conditions necessary to conclude such causality are:
1. Concomitant variation - the information available has increased with increased consumer use.
2. Time Order - increased information lags or concurs with increased information use.
3. Absence of other possible causes
Be an MR!: Fox News
Information on consumers’ preferences for network news channels can be obtained from sources such as
http://www.rateitall.com/t-8-tv_news_shows.aspx.
A preexperimental design would not be appropriate as there are three experimental groups. A statistical
design is called for.
Be a DM!: Fox News
The popularity ratings for channels can be used to assess the attributes that audiences look for in
Network News Channels.
Based on experimentation, causality can be assigned to factors that make the news more attractive.
Improve the attributes preferred by audience.
Convey the change in form and content to audience through appropriate PR and Advertising channels.
EXPERIENTIAL LEARNING
Canon
1.
Students should be encouraged to visit www.bestbuys.com and identify the relevant
information.
2.
Students should be encouraged to search the Internet and identify the relevant information
on consumers’ price sensitivity for digital cameras.
111
3.
4.
A factorial design is appropriate. The price of digital cameras could be varied and
consumers’ willingness to buy measured. It would be important to control for income
and digital camera usage.
It is likely that the market is segmented with respect to price sensitivity. Different prices
can be charged to the different segments with some differentiation in the cameras offered.
A Local Pizza Maker
1.
Students should be encouraged to search the Internet and identify the relevant information.
2.
It would be highly beneficial for the firm to conduct a factorial experimental sign because
this involves measuring the effects of two or more independent variables at various levels. In
addition, it is highly recommended that management test market in their largest store in the local
area.
Here is our proposed format:
a. Controlled Variables: While test marketing at a store location, the amount of pizza
sauce will remain unchanged as respondents determine if their preference is a 2:1 ratio of
provolone/mozzarella or a 1:2 ratio. Another ingredient, such as crust, will also remain
the same so that it would prevent undue influence upon each respondent’s perceptions of
the new pizza formula. Store location will also be non-factor so that respondents do not
disproportionately influence their perceptions on price and pizza formulation.
b. Manipulated Variables: One such manipulated independent variable could include
evaluating various promotional tactics, versus none at all, when evaluating price and the
new provolone/cheese mix concept. For example, the team could evaluate such measures
like “Buy One, Get One Free” promotions at the local pizza maker’s store as part of the
test market strategy. Management would also determine how this particular manipulated
independent variable affects ratings for the new pizza formulation. When evaluating the
new pizza cheese mix, it would be imperative to gauge the price level at three stages
(raised, lowered or unchanged) to one that is kept constant and its affect on ratings. It
would also be informative to manipulate the ratio of pizza cheese (2:1 or 1:2) and evaluate
the consumer’s response level.
c. Response Variables: One obvious dependent variable that would be measured is
consumer’s perceptions of overall brand quality ratings. Other measures that would be
useful are consumers overall preference, and willingness to buy.
d. Control Group: One experimental treatment will consist of the current price and the
current cheese mix; this can serve as a control group. A separate control group is not
required.
3.
Although costs and time allocation is of significant important, a field experiment would be
more useful because our client is seeking external validity in which a cause-effect relationship
found in the experiment could be generalized. It also provides a realistic forum for our
experimental design in which our results could benefit our client more effectively. Most
importantly, it enables the researchers to evaluate the behavior, attitudes and perceptions of pizza
customers in a natural environment. We suggest a test market for the new provolone/mozzarella
cheese mix in the largest store in the Metroplex or largest shelf-space at a local convenient store.
This would determine if the new pizza product should expand to all locations for full market
rollout.
112
4.
Incentives for this project could total $3600 dollars.
180 respondents x $20 (for participation)=$3600 (see Table 1 below)
Table 1. Factorial Design
Provolone/Mozzarella Raise Price
Constant Price
Lowered Price
2:1 ratio
30 respondents
30 respondents
30 respondents
1:2 ratio
30 respondents
30 respondents
30 respondents
Total Incentives
$3600
5.
I would probably switch to the less time consuming and fiscally conservative method of a
laboratory environment. Although the research team would have more control of the
environment, it is more subject to reactive error, perhaps low in external validity and is more
artificial when evaluating consumer behavior.
6.
Pizza Hut can increase its market share by using the experimental results to identify the
most desired combination of cheese formulation and price.
WHAT WOULD YOU DO?
It should be noted that a variety of answers are appropriate. The ones given here are merely illustrative.
The Marketing Research Decision
1. A, B & E
2. A lot of secondary data about apparel industry is available from online as well as traditional
third party databases. This can be used initially to gather information about the success of
private labels and in-house brands. The focus groups are helpful in gaining further
understanding of the conclusions drawn from the analysis of secondary data, understanding
consumer priorities, perceptions and preferences. The specific causal design that is
recommended is a factorial design. The factor, brand name can be varied at three levels:
Levi’s, Gap, and Arizona. Respondents can be randomly assigned to one of these brands and
asked to express preference for the brand rate its power. Current usage of these brands
should be controlled.
3. The causal design, as specified, will provide a direct relative measure of the power of the
three brands when the mean levels are compared.
113
The Marketing Management Decision
1. B and C
2. Comparing the prices, it appears to be one of the reasons why Levi’s is loosing the battle for
market share. It would be a good idea to launch a cheaper line of jeans to appeal to the cost
conscious market segment. To reach the new target segment, the company would want to
increase advertising spend in targeted marketing efforts. The new line of jeans should be
distributed through different channels than that of the existing high priced segment. This will
ensure the necessary reach.
REVIEW QUESTIONS
1.
2.
3.
4.
There are three requirements or conditions that must be satisfied before a causal
relationship can be inferred. The three conditions are:
1. Concomitant variation of the two variables
2. Time order of occurrence of variables
3. Elimination of other possible causal factors
The difference between internal and external validity is that internal validity is concerned
with the question of whether the observed effects on the dependent variable have been
caused by variables other than the treatment, and control of extraneous variables is
necessary for establishing internal validity. External validity, on the other hand, is
concerned with whether the cause-effect relationship can be generalized. More often than
not a tradeoff is required between internal and external validity. While internal validity
can be raised in an artificial environment, this reduces the generalizability of the results
and consequently, the external validity.
The distinguishing feature of the true experimental design, as compared to the preexperimental design, is randomization. In true experimental designs, the researcher can
randomly assign test units to experimental groups and also randomly assign treatments to
experimental groups.
The steps involved in implementing the posttest-only control group design are:
1. Select a sample of respondents at random.
2. Randomly assign the respondents to two groups (i.e. experimental group and control
group).
3. Respondents in the experimental group would then be exposed to a treatment.
4. Post-treatment measurements are obtained from both groups using a measuring
instrument like a questionnaire.
The design is described symbolically as follows:
EG: R X 01
CG: R
02
5.
Treatment effect is given by:
TE = 01 – 02
A time series experiment is a quasi experimental design. It involves periodic
measurement on the dependent variable for a group of test units, and then the treatment is
administered by the researcher or occurs naturally. After the treatment, periodic
114
measurements are continued in order to determine the treatment effect. A time series
experiment may be symbolically described as:
01 02 03 04 05 X 06 07 08 09 010
6.
7.
8.
9.
This design is used in cases where the researcher lacks control over the scheduling of the
treatment and the ability to randomly expose test units to the treatment, but can control
when measurements are taken and on whom they are taken.
A multiple time series design is different from the basic time series in that a multiple time
series design has an additional group of test units added to serve as a control group.
The advantages statistical designs have over basic designs are:
a. In the case of basic designs, the effects of only one independent variable or treatment
can be measured, while in case of a statistical design, the effects of more than one
independent variable can be measured.
b. In the case of statistical designs, specific extraneous variables can be statistically
controlled.
c. Economical designs can be formulated when each test unit is measured more than
once.
A laboratory experiment is one where the researcher creates an artificial environment for
the purpose of experiment; thus, he has a great degree of control over extraneous
variables. On the other hand, a field environment is one where the experiment is
conducted under actual conditions.
Laboratory experiments have some advantages and some disadvantages vis-à-vis field
experiments. The advantages are that there is a greater control over extraneous variables,
internal validity is high, and time required is short. It is also economical. The
disadvantages are that the environment is artificial and external validity is low.
Therefore, the results cannot be safely generalized. Also, the reactive error is high due to
artificiality of the environment, which may cause the respondent to react to the situation
itself, rather than to the independent variable.
Test marketing is an application of a controlled experiment done in a limited but carefully
selected part of the marketplace called test markets. It involves a replication of a planned
national marketing program for a product in the test markets. The three types of test
markets are:
1. Standard test market: The product is sold through regular distribution channels and,
typically, the company's own salesforce is responsible for distributing the product. It
involves a one-shot case study.
2. Controlled test market: The entire test marketing program is conducted by an outside
research company, which also handles distribution and field sales operations in the
test market.
3. Simulated test market: These are also called laboratory tests and yield mathematical
estimates of market share based on initial reaction of consumers to the new product.
115
APPLIED PROBLEMS
1.
2.
3.
4.
5.
6.
7.
8.
a.
Independent Variable - The commercial
Dependent Variable - Attitude towards abortion
b.
Posttest-only control group design
External validity may be affected because this was a lab experiment, not an actual field
experiment, thus, the results may not be generalizable to those obtained when the
commercial is actually viewed in home.
If the respondents had been selected by convenience, a static group design would result.
Any experimental effect could also be attributed to selection or mortality.
a.
This is an example of the one-shot case study. The internal validity is threatened
by history, maturation, selection, and mortality.
b.
This is an example of the posttest-only control group design. Internal validity is
not hampered by extraneous variables.
Many answers are possible for each design. One solution is provided as an example.
a. Measuring the attitude towards a new advertisement or new product
b. Evaluating the effectiveness of a change in a product attribute or the effectiveness of
a new marketing program when time and cost are not major considerations
c. Evaluating the attitude toward a new advertisement or product when time and cost are
not significant factors, forcing researchers to reduce sample size, quickly determine
results, and/or keep costs down
d. When great precision is needed and the effectiveness of an advertising program needs
to be monitored over time to show the process of change in the population
e. A new marketing program is being initiated which must be evaluated to determine its
effectiveness. The goal is to determine which of the components of the program
(price, promotion, channels, etc.) are the most important factors leading to sales and
to assess the interaction of these components on sales.
Several of the designs can be implemented. A two factor design, where factor one
denotes the value of coupon (none, low, medium, and high), and factor two denotes
product use (none, low, medium, and high), seems to be very appropriate. Students
should select a specific product and also define the levels (low, medium, and high) for
each factor.
The existing and the three new packages can be displayed, one at each specific location
on the Internet. Respondents can be recruited and randomly assigned to each package
location (URL), asked to view the package and answer several attitudinal and behavioral
intent questions. The package with the most favorable attitude and/or the highest
behavioral intent can then be identified.
The previous (existing) and the new version of EXCEL can be made available, one at
each specific location on the Internet. Respondents can be recruited and randomly
assigned to each EXCEL version, and then asked to do a specific task using that version.
Then they answer several questions relating to ease of use, performance, features, and
overall preference. The most favorable version can then be identified.
Respondents would be randomly recruited and randomly assigned to one of two Internet
locations (URLs). The experimental group of respondents would first see the print ad for
Toyota Camry and then respond to a battery of attitudinal and behavioral intent questions.
The control group respondents would not see the print ad but respond to the same battery
116
of attitudinal and behavioral intent questions. The effectiveness of the print ad can be
determined by comparing the responses of the two groups.
GROUP DISCUSSIONS
1.
Issues which can be discussed include the role of inference in drawing conclusions, the
difficulty of isolating all possible sources of variation in social science experimentation,
the historic value of experimentation in deriving theories and inferring relationships, and
the nature or definition of science, (i.e., is science only limited to proofs which account
for all sources of variation?).
117