Chapter 6
Causal Designs
common sense notion of causality is true?
a. The scientific notion suggests there is a single cause of an event while the
common sense notion suggests there are a number of causes of an event.
b. The common sense notion implies a completely deterministic relationship while
the scientific notion implies a probabilistic relationship.
c. Both the common sense notion and the scientific notion recognize that we can
never prove that a relationship exists.
d. There is no difference between the scientific notion of causality and the common
sense notion.
e. The empirical evidence is most consistent with the common sense notion.
a. one of the notions holds that a researcher can discover a single cause of an event.
b. one of the notions implies a completely deterministic relationship between a cause
and an effect.
c. one of the notions implies that a researcher can always infer, but never prove that
a relationship exists.
d. a and c.
e. all of the above are differences.
a. The scientific notion of causality differs from the common sense notion in that the
common sense notion suggests that X (the cause) always leads to Y (the effect).
b. The common sense notion of causality implies a completely deterministic
relationship while the scientific notion implies a probabilistic relationship.
c. The scientific notion of causality rests upon the belief that we can only prove that
X (the cause) leads to Y (the effect) in an experiment.
d. b and c.
e. They are all true.
descriptive research design is
a. elimination of other possible causal factors.
b. time order of occurrence of variables.
c. concomitant variation.
d. a and c.
e. none of the above.
a. If association is found, it means that one variable clearly causes the other.
b. Absence of association implies the absence of causality.
c. Causation can be clearly established by experimental as well as descriptive
studies.
d. Causation requires association.
e. All the statements are false.
between two variables
a. the technique can only be applied with quantitative variables.
b. we are looking for a one-to-one correspondence between the two variables and
this is usually found.
c. when evidence of concomitant variation is found, the hypothesis is proved.
d. the absence of initial evidence of concomitant variation implies that there is no
causation between the two variables.
e. there may be no causal relationship between variables even though there is initial
evidence of concomitant variation.
a. Concomitant variation is a measure of the extent to which two variables vary
together.
b. The “pure” case of concomitant variation is found in 20-30 percent of all causal
research studies.
c. If perfect concomitant variation is found, we can say that our hypothesis is
proven.
d. If there is no evidence of concomitant variation in a two-way cross-classification
table, we can be sure that there is no causal relationship.
e. They are all false.
a. Cross-classification tables are useful in proving cause and effect relationships.
b. The presence of association in a cross-classification table demonstrates
conclusively that X and Y are related in the real world.
c. A causal factor must occur before the effect or event it causes.
d. When considering the time order of occurrence of variables as evidence, each
term in the relationship must be either a cause or an effect. They cannot be both.
e. They are all false.
relationship between two variables
a. it is possible for each term in the relationship to be both a “cause” and an effect of
the other term.
b. the occurrence of a causal factor must follow the occurrence of the event.
c. it is fairly easy for the researcher to understand the time sequence governing a
phenomenon.
d. an effect can be produced by an event that occurs after the effect has taken place.
e. a relationship between the two variables can be proved.
a. knowledge of the product’s manufacturer
b. the use of brand names on the product
c. the use of the letters A,M, and Z on three products
d. a and b
e. All of the above will affect consumer responses.
a. all other explanations for the effect have been eliminated.
b. we can conclusively prove relationships.
c. only replication will increase our confidence in the resulting conclusions.
d. we substitute method knowledge for conceptual knowledge.
e. a and b are true.
a. An experiment has greater ability to supply evidence of causality because it takes
longer to complete than a descriptive study.
b. In marketing, experimental treatments are generally elements of the marketing
mix whereas test units are the entities that serve as objects for a firm’s marketing
efforts.
c. An experimental design limits one to supplying evidence of concomitant
variation.
d. An experimental design is a research design in which the investigator has direct
control over at least one independent variable and manipulates at least one
independent variable.
e. b and d are true.
a. is one in which the investigator has direct control over at least one independent
variable.
b. is less capable of providing evidence of causal relationships than is an exploratory
design.
c. is less internally valid than a descriptive study.
d. is an example of ex-post-facto research.
e. limits one to supplying evidence of concomitant variation.
a. manipulates the independent variable under carefully controlled conditions.
b. executes a turnover analysis.
c. attempts to find one or more causal variables.
d. tries to supply evidence of the time order of occurrence of variables.
e. observes the criterion variable.
a. laboratory and survey.
b. laboratory and field.
c. field and survey.
d. survey and ex post facto.
e. ex post facto and cross sectional.
a. their manipulation of variables.
b. their cost.
c. their validity.
d. their control.
e. their environments.
a. A laboratory experiment is conducted under artificial conditions while
manipulating some marketing variables and controlling others.
b. A field experiment is conducted in a natural or realistic setting in which one or
more independent variables are manipulated by the experimenter under as
carefully controlled conditions as the situation will permit.
c. Laboratory experiments are distinguished from field experiments primarily in
terms of the degree of control over the variables, e.g. price, being manipulated.
d. The specially designed conditions of a laboratory experiment provide more
control than a field experiment.
e. They are all true.
a. a descriptive design.
b. an exploratory design.
c. a field experiment.
d. a sample survey.
e. a laboratory experiment.
a. is more internally valid than a field experiment.
b. affords the researcher greater control.
c. better eliminates confounding factors.
d. b and c only.
e. all of the above.
difference in the specific instance under investigation, one is referring to
a. external validity.
b. internal validity.
c. concomitant variation.
d. cross-sectional analysis.
e. interaction errors.
a. External validity is concerned with the issue of whether or not the specific
treatment in which the researchers are interested has actually brought about the
effect they have observed.
b. A consumer good purchase experiment conducted in a mock-up store is more
externally valid than the same experiment in a real store setting.
c. External validity deals with the problem of projecting one’s results from the
experimental setting to the real world.
d. a and c.
e. a, b, and c.
reproducible in other situations, the researcher is attempting to enhance the
experiment’s
a. internal validity.
b. convergent validity.
c. external validity.
d. construct validity.
e. none of the above are correct.
a. you can’t have one without the other.
b. procedures used to establish internal validity will also establish external validity.
c. they are matters of degree rather than all-or-nothing proportions.
e. internal validity is a much more important issue in marketing than external
validity.
d. all of the above are true.
a. A study lacking external validity may nevertheless have high internal validity.
b. Internal validity is a matter of degree, whereas external validity either exists or
doesn’t exist.
c. In most experiments we can conclude that the observed response is due to our
experimental manipulations.
d. a and b.
e. a, b, and c.
a. Do adults in the western United States consume more 100% fruit drinks than
those in the Northeast?
b. Will one brand name produce higher level-of-satisfaction reports than another
brand name?
c. Do consumers purchase more chicken per capita on a retail weight basis now than
they did two years ago?
d. Does the average two-person household purchase more than $65.00 worth of
groceries a week?
e. a and c.
a. O refers to the observation of the test unit(s).
b. movement through time is represented by the vertical arrangement of X’s and O’s.
c. O, X, O2 indicates that one or more test units were exposed to an experimental
variable and that their response was measured at two different points in time.
d. a and c.
e. a, b, and c.
a. Its effects on a research conclusion are often subtle.
b. Its effects are eliminated in a laboratory experiment.
c. It refers to events external to the experiment which occurred before or during the
experiment.
d. a and c.
e. all of the above are true of the history effect.
a. The one-group pretest-posttest design is an example of a true experimental design.
b. In an experiment with a large sample, it is safe to assume that the people who did
not respond were equal to the people who did.
c. History primarily jeopardizes the internal validity of marketing experiments.
d. The use of randomization to eliminate selection bias in experimental designs
works best when the groups being compared are small.
e. None of the above is true.
a. refers to the changes occurring within the test units that are due to the impact of
the experimental variable.
b. refers to events external to the test units.
c. is limited to test units composed of people.
d. does not depend on the timing of the specific experiment.
e. refers to the processes operating within the test units as a function of time per se.
a. Students take a test and score 84; two weeks later they take a different test and
score 90.
b. A person is asked about her attitudes toward Ford automobiles; she then watches
Ford commercials more closely than before.
c. After Dan is asked to be on a consumer panel that reports purchasing behavior, he
changes his purchasing behavior.
d. While answering a questionnaire, Dawn tries to answer the last questions
consistent with how she answered the first questions.
e. a through d are all examples of the testing effect.
a. refer to specific events that occur prior to the experiment that may affect
respondent’s attitudes.
b. on a research conclusion are usually obvious.
c. primarily jeopardize the internal validity of a marketing experiment.
d. refer to changes occurring within the test units that are not due to the impact of
the experimental variable, but result from the passage of time.
e. c and d.
a. The main testing effect is the effect of a prior observation on a later observation.
b. The interactive testing effect is exhibited when subjects who are asked about their
attitudes on national security become much less sensitive to defense spending
promotional campaigns.
c. The main testing effect is exhibited when subjects who are asked about their
smoking habits on two separate occasions respond so as to be consistent.
d. The main testing effect usually exerts its greatest impact on the internal validity of
an experiment, while the interactive testing effect most typically affects the
external validity of a conclusion.
e. As long as there is a before measurement, there can be no testing effect.
a. When it is not possible to go back after an experiment and determine if the
experimental and control groups were equal before the fact, selection bias may be
present in the experiment.
b. Matching is the preferred procedure to ensure equality of comparison groups.
c. To avoid selection bias, the general principle is: select subjects at random and
then arbitrarily assign subjects to groups.
d. The use of randomization to eliminate selection bias in experimental designs
works best when the groups being compared are small.
e. They are all false.
taking the test for the first time. This is an example of the
a. history effect.
b. maturation effect.
c. interactive testing effect.
d. main testing effect.
e. selection bias effect.
behavior, may change his or her REPORTED purchasing behavior but not his or her
ACTUAL behavior is an example of the
a. main testing effect.
b. interactive testing effect.
c. maturation effect.
d. experimental mortality effect.
e. history effect.
a. Mortality effects refer to the deaths of subjects during hazardous experiments
using electronic devices.
b. Maturation effects can only be measured in test units composed of people.
c. Experiments lasting more than two hours are likely to suffer from history effects.
d. a and b.
e. They are all false.
the experimental variable is the
a. main testing effect.
b. maturation effect.
c. interactive testing effect.
d. concomitant variation effect.
e. history effect.
a. does not result from variation in administration of the interview.
b. can be acute when there are multiple observers or interviewers.
c. is not attributed to the way with which the interviewer handles the assignment.
d. does not occur if a minor modification is made, such as a slight change in the
wording of a specific question so as to make it more understandable without
changing its meaning.
e. does not occur with interviews conducted by the same interviewer.
position during the course of an experiment refers to
a. statistical regression.
b. turnover analysis.
c. instrument variation.
d. the main testing effect.
e. selection bias.
toward Brand X compared the attitudes of a sample of respondents who did not watch
the ABC Sunday Night Movie (one of the sponsors of which was Brand X) against
those who did. This is an example of
a. statistical regression.
b. history.
c. the interactive testing effect.
d. selection bias.
e. the main testing effect.
time, this is said to be evidence of
a. statistical regression.
b. concomitant variation.
c. selection bias.
d. the interactive testing effect.
e. instrument variation.
established?
a. randomization and statistical regression
b. matching and turnover analysis
c. turnover analysis and cross-sectional analysis
d. matching and randomization
e. longitudinal analysis and cross-sectional analysis
experimenter considers two test units as a pair by means of some external criterion
before assigning one unit randomly from each pair to each group, he is said to be using
a. cross-sectional analysis.
b. turnover analysis.
c. time order of occurrence of variables.
d. statistical regression.
e. matching.
comparison groups in an experiment because
a. it is easy to match test units on almost any characteristic.
b. matched test units are equal on most important variables.
c. matching is cheaper.
d. matching takes less time.
e. matching is not preferred.
project has experienced
a. selection bias.
b. experimental mortality.
c. maturation.
d. subject variation.
e. b and c.
a. they are next to impossible to eliminate or adjust the results for.
b. they do not allow the experimenter to conclude that the observed response was
due to experimental manipulations.
c. we do not know whether or not they have operated.
d. they always work in a negative direction.
e. they can only be eliminated by randomization and matching
a. pre-experimental; true experimental; and post-experimental.
b. true-experimental; post-experimental; quasi-experimental.
c. true-experimental; quasi-experimental; regular experimental.
d. quasi-experimental; true-experimental; pre-experimental.
e. quasi-experimental; pre-experimental; regular experimental.
a. In an experiment it is safe to assume that the people who did not respond were
equal to the people who did.
b. The ability to control who will be exposed to the treatment variable is a feature
that distinguishes true experiments from quasi-experiments.
c. True experimental designs eliminate all doubts in interpreting the results of the
research.
d. b and c.
e. They are all false.
once for their response. This is what type of study?
a. a one group pretest-posttest study
b. a post-experimental study
c. a static group comparison
d. a one-shot case study
e. a longitudinal study
a. can be diagrammed: X 01 02.
b. results from the addition of a pretest to the one-shot case study design.
c. is a type of quasi-experimental mortality.
d. controls for history.
e. can be used to measure the effectiveness of the experimental variable through the
calculation d = 01 – 02.
a. E + U.
b. E + U + I.
c. U + I.
d. only I.
e. none of the above.
a. the groups have not been created by randomization.
b. the effectiveness can be measured by the equation, d=02 – 01.
c. extraneous error is eliminated.
d. its data is more valid than data from any other quasi-experimental design.
e. its results can be generalized to the larger population.
a. describes a design in which there are two groups created by randomization and
where one group has experienced X and the other group has not.
b. designs contain groups that have not been created by randomization.
c. involves measuring the difference in the attitudes of the two groups by the same
equation as the difference in a one-group pretest-posttest design, that is O2 – O1.
d. a and b.
e. a, b, and c.
a. history and selection bias.
b. main testing effect.
c. selection bias and experimental mortality.
d. experimental mortality and history.
e. none of the above.
a. the experimenter can arbitrarily assign test units to the experimental and control
groups once they are randomly selected from the population.
b. the error due to extraneous factors can be measured by O4 – O3.
c. the main testing effect operates, making the design a less than ideal experimental
procedure.
d. it is possible to estimate the size of the interactive testing effect.
e. They are all false.
a. Random assignment of the test units to the experimental and control groups is the
feature that distinguishes the after-only with control group design from the static
group comparison.
b. One of the advantages of the after-only with control group is that there is no
chance for an interactive testing effect.
c. The after-only with control group design is particularly vulnerable to selection
bias, and where random assignment is not possible, it remains suspect.
d. The after-only with control group design is particularly susceptible to problems of
experimental mortality.
e. They are all true.
from any pre-experimental designs is
a. matching.
b. randomization.
c. the absence of extraneous sources of error.
d. the fact that the results can always be generalized to the larger population.
e. the fact that each of the test units is measured before and after the introduction of
the experimental stimulus.
didn’t control for
a. the main testing effect.
b. the interactive testing effect.
c. any extraneous factors.
d. the effect of the experimental treatment.
e. none of the above.
control group is the requirement that
a. the experimenter can arbitrarily assign test units to the experimental and control
groups.
b. he will have to estimate the impact of history by means external to the
experiment.
c. it is up to the test unit to self-select whether it will be a member of the control or
experimental group.
d. each of the test units in both groups is measured before and after the introduction
of the experimental stimulus to the experimental group.
e. the experimenter has to make an estimate of what the prior measurements should
be.
(R) 01 X 02
(R) 03 04
The difference represented by (02 – 01) – (04 – 03) estimates the EFFECT of
a. history.
b. maturation.
c. history, maturation and all other relevant sources jeopardizing internal validity.
d. experimental variable.
e. statistical regression.
a. history effect.
b. maturation effect.
c. main testing effect.
d. interactive testing effect.
e. selection bias.
a. E + U + I.
b. E + U.
c. E.
d. nothing (it is irrelevant).
e. none of the above.
a. the effects of the extraneous influences.
b. the difference between the control group and the experimental group.
c. the change in attitude of the subject.
d. the interactive effect of testing.
e. the effect of the experimental stimulus.
a. the main testing effect.
b. the interactive testing effect.
c. maturation.
d. statistical regression.
e. history.
design should be used?
a. the before-after with control group design
b. the one-shot case study design
c. the four-group-six-study design
d. a time series design
e. the static group comparison design
a. one control group and three experimental groups.
b. three control groups and one experimental group.
c. four experimental groups.
d. two control groups and two experimental groups.
e. none of the above.
the introduction of the experimental stimulus are
a. the two experimental groups.
b. the first experimental group and the second control group.
c. the first control group and the second experimental group.
d. all four groups.
e. the first experimental and control groups.
a. E.
b. E + U.
c. E + U + I.
d. E + I.
e. none of the above.
four-group-six-study design when there is an interactive testing effect by comparing
a. the second control group and the first experimental group.
b. the two experimental groups.
c. the second experimental and control groups.
d. the first control group and the second experimental group.
e. the first experimental and control groups.
is provided by comparing
a. the second experimental and control groups.
b. the first control group and the second experimental group.
c. the two experimental groups.
d. the first experimental and second control groups.
e. the first experimental and control groups.
in marketing?
a. The design is expensive in terms of time.
b. Marketing samples are not always large enough to divide the samples of test units
into four groups.
c. If the group samples are small, it is unlikely they will be equal even if assigned
randomly.
d. The design is expensive.
e. All of the above.
of change?
a. the four-group-six-study design
b. the before-after design with control group
c. the after-only with control group design
d. the time series design
e. the one-group pretest-posttest design
a. pre-experimental design.
b. true-experimental design.
c. post-experimental design.
d. quasi-experimental design.
e. b and d.
a. samples of cheese.
b. people receiving the free samples.
c. discount coupon use.
d. coupon redemption rate.
e. purchases of cheddar cheese measured by the number of coupons redeemed.
a. history and maturation.
b. statistical regression and selection bias.
c. maturation and experimental mortality.
d. selection bias and experimental mortality.
e. the main testing effect and history.
A local cheese producer feels their biggest problem is to get consumers acquainted
with their product. An experiment is run in which 50 free samples of cheddar cheese
are distributed to a randomly selected sample of families. This group and another
group of 1,000 randomly selected people to which free samples had not been given are
both sent coupons for the purchase of cheddar cheese at a discount at local grocery
stores. The keyed coupons show that 387 coupons are returned from the group
receiving the free sample and 130 are received from the other group.
Answer the following three questions on the basis of this information.
a. R 01 X 02
R 03 04.
b. R X 01
R 02.
c. R X 0102
R 0304.
d. R 01 X 02
R 03.
e. none of the above.
a. allows analysis of individual cases of change.
b. cost relative to other true experimental designs.
c. time relative to other true experimental designs.
d. b and c.
e. all of the above.
of change?
a. to study the effect of coupons on product trial
b. to study the effect of product acquaintance on the purchase of cheddar cheese
c. to study the effect of free samples on the use of discount coupons
d. to study the demographic characteristics of cheddar cheese purchasers
e. to study the effect of free samples on product trial
advantages it possesses with respect to sample size, cost, and time is
a. the four-group-six-study design.
b. the after-only with control group design.
c. the time series design.
d. the before-after design with control group.
e. the static-group comparison design.
a. the before-after with control group design.
b. the one group pretest-posttest design.
c. the after-only with control group design.
d. the one-shot case study.
e. the static group comparison.
a. are the only form of quasi-experimental designs that exist.
b. allow the researcher repeated access to the same test units.
c. are fundamentally weak because of the failure to control maturation.
d. a and b.
e. a, b, and c.
for
a. maturation.
b. history.
c. selection bias.
d. the main testing effect.
e. experimental mortality.
A retail hardware chain is considering a self-service system for its 326 stores. Ten
stores have been selected for a trial period of three months. Ten other stores, matching
the experimental stores as closely as possible, have been selected as a control group in
which the traditional clerk system will operate. Dollar sales for each of the twenty
stores will be recorded weekly for the three-month period. Prior to the study, weekly
sales will be recorded for a month, as a “before” measure. The difference in sales
between the experimental and control groups is expected to help management in its
decision.
Answer the following three questions on the basis of the above information.
a. the stores.
b. the trial period.
c. sales.
d. self-service.
e. none of the above.
a. the attitudes of consumers towards self-service.
b. dollar sales.
c. the stores.
d. the method of operation.
e. the product category (hardware).
a. method of operation (self-service vs. clerk)
b. differences in items offered in self-service vs. clerk operated stores
c. any special promotion by any store or stores
d. time period selected, e.g., near Christmas time
e. differences in physical attractiveness of the stores
a. equivalency of the control and experiment groups before observation cannot be
substantiated.
b. concomitant variation cannot be observed.
c. the effects of the extraneous variables cannot be controlled.
d. they are post facto research studies.
e. a, c, and d are true.
a. The administrative problems are typically greater in experimental research than in
descriptive research.
b. In descriptive research the investigator can establish the equality of the
experimental and treatment groups.
c. Descriptive studies are likely to cost more than well-designed experiments.
d. a and b.
e. a and c.
a. is done in a limited but carefully selected part of the marketplace.
b. aims to predict the sales or profit consequences of a proposed marketing action.
c. is not an experiment.
d. is a and b.
e. is b and c.
marketing in particular EXCEPT
a. control.
b. cost.
c. time.
d. data-collection form design.
e. a and d.
a. production costs.
b. cost of designing the data-collection instrument and sample.
c. cost of the marketing strategy support programs.
d. wages paid to field staff.
e. all of the above are cost incurred in test marketing.
a. One year is the recommended minimum amount of time to test market a product.
b. A controlled test market is also known as the forced-distribution test market.
c. A market test is an example of a controlled experiment.
d. Test markets can only be used to determine the sales potential of a new product or
an improved version of an established product.
e. Simulated test marketing relies heavily on computer models.
a. less than 3 months.
b. 3-6 months.
c. 7-9 months.
d. 10-12 months.
e. 13-18 months.
a. lack of cooperation of retailers.
b. over-attention to in-store conditions, such as always-stocked shelves.
c. attempts by competitors to sabotage test market results.
d. a and c.
e. a, b, and c.
a. the product is distributed through normal distribution channels.
b. the test is conducted by a service agency external to the firm.
c. the test market service pays retailers for shelf space.
d. the test market service coordinates the trade promotion program.
e. the product is guaranteed distribution.
a. Simulated test markets are not useful for eliminating weak products before they
go to a standard test market.
b. Electronic test markets, popular in the 1970’s, are now a declining segment of the
test market industry.
c. A prime advantage of simulated test markets is the protection from competitors
that they provide.
d. In a standardized test market the entire test program is conducted by an outside
service.
e. They are all false.
a. the providers of the service have a great deal of demographic data on their
recruits.
b. the test market service monitors each household’s television viewing behavior.
c. each household has an identification card for supermarket purchases.
d. Both b and c are correct.
e. Any of the above are correct.
a. consumers are shown commercials for a new product and competitive products.
b. consumers are given a chance to purchase products in the product category of
interest.
c. consumers are interviewed at malls or in their home and are exposed to a new
product and asked to rate its features.
d. the information gathered is used as input for a computer model which estimates
future sales.
e. All of the above are true.
a. provide protection from competitors.
b. are good for assessing trial and repeat purchase behavior.
c. are faster and cheaper than standard market tests.
d. are good for estimating likely competitive reaction.
e. a, b and c.
a. when it is important for the firm to test its ability to secure support from the trade
and get distribution.
b. when the firm desires control over the whom of measurement.
c. when the capital investment is significant and the firm needs a prolonged test
market to assess its technical ability to produce the product.
d. when the firm needs to build an experience base, but only on a limited scale.
e. when the test requires a more natural environment.
a) Do children from upper classes drink more carbonated beverages than those from lower classes?
b) Will gasoline with XYZ additive provide better mileage than gas without the additive?
c) Does the average male use cosmetics?
d) Is the standard of living higher today than ten years ago?
e) None of the questions can be answered by experimentation.