CHAPTER 5 TEST BANK
Experimental and Quasi-Experimental Designs
TRUE/FALSE
1. Experimentation is best suited for descriptive studies.
2. The defining feature of an experiment lies in the control of the independent variable by
the experimenter.
3. In a classical experiment, the independent variable must be a ratio level variable.
4. In a classical experiment, subjects are measured on the independent variable before the
experiment begins and again after the dependent variable has been manipulated by the
researcher.
5. Construct validity refers to generalizing from our experimental observations to causal
processes in the real world.
6. Experiments in criminal justice typically require only one experimental and one control
group for each study.
7. The threat of statistical regression is a concern any time the researcher begins with
subjects who exhibit extreme values on the dependent variable.
8. Construct validity is concerned with the ability to generalize from the results of the
experimental group to the control group.
60
9. Random assignment cannot be used in some criminal justice research for legal and ethical
reasons.
10. In case-oriented research, a great number of cases are examined in order to understand a
small number of variables.
MULTIPLE CHOICE
1. Which statement is most accurate?
a.
Experiments are only useful in the physical sciences such as chemistry and
physics.
b.
Experiments involve observing phenomena but do not try to produce them.
c.
Experiments can be used in scientific as well as nonscientific human inquiry.
d.
Experiments in the classical form require at least three groups of subjects..
2. Which statement is most accurate?
a.
Experimentation are rarely appropriate for hypothesis testing.
b.
Experiments are well suited to research involving poorly-defined concepts.
c.
There is no need to generate a hypothesis for a classical experiment.
d.
Experiments are appropriate for evaluation research.
3. An experiment examines the effects of a(n) ____ variable.
a.
dependent
b.
control
c.
independent
d.
masking
4. Which statement is true?
a.
A variable can be an independent in one experiment and dependent in another.
b.
The dependent variable is manipulated by the experimenter.
c.
A classical experiment examines the effects of the dependent variable.
d.
In a classical experiment, the independent variable is viewed as the outcome.
5. Which statement is most accurate?
a.
Neither the independent nor the dependent variable need be operationally defined.
b.
Only the dependent variable needs to be operationally defined.
c.
Only the independent variable needs to be operationally defined.
d.
Both the independent and dependent variables must be operationally defined.
6. In most cases, the methods used to select subjects must meet the scientific norm of ____.
a.
generalizability
b.
informed consent
c.
equivalence
d.
neutrality
7. Which procedure is suitable for random assignment to groups?
a.
Assigning subjects in the order they arrive, so that one group is filled before the
next is started
b.
Letting subjects express a preference and then assigning them to a non-preferred
group
c.
Flipping a coin assigning subjects to the control group and to the experimental
group.
d.
Asking each subject to secretly decide which group they want to be in
62
8. Which statement best fits randomization as it is used in experiments?
a.
“all other things being equal”
b.
“use your best guess”
c.
“it is close enough for government work”
d.
“you are never 100% sure, so don’t worry about it”
9. Which statement is a potential threat to internal validity in an experimental design?
a.
generalization
b.
instrumentation
c.
randomization
d.
correlation
10. Which situation presents a threat to internal validity?
a.
subjects dropping out of an experiment
b.
selecting a random sample
c.
randomization of group assignments
d.
very high reliability of outcome measures
11. Shortening the time between pretest and posttest or perhaps even offering cash payments
to participants in an experiment are techniques that may be used to ____.
a.
limit diffusion
b.
decrease experimental mortality
c.
decrease the effects of instrumentation
d.
control threats to testing
12. _______________ validity is the correspondence between the empirical test of a
hypothesis and the underlying causal process that the experiment is intended to represent.
a.
Construct
b.
Compensatory
c.
Correlational
d.
Criterion
63
13. Which statement is accurate?
a.
Threats to external validity are decreased by carefully controlling conditions.
b.
Threats to internal validity are decreased by carefully controlling conditions.
c.
Threats to generalizability are inherent in research and cannot be changed.
d.
Threats to internal and external validity covary, and if one is high, the other will
also be high.
14. When using a nonequivalent-groups design, the researcher will handle subject assignment
to groups by ____.
a.
random assignment to experimental and control groups
b.
allowing subjects to pick which group they want to be in
c.
matching subjects in the experimental and comparison groups
d.
using waves of cohorts
15. In Widom’s study of the effects of child abuse and its impact upon later criminal
offenses, she selected a sample of children with court records of being abuse victims and
gathered a comparison group that were matched on gender, race, age and socioeconomic
status. The technique used in this case was ____.
a.
individual matching
b.
aggregate matching
c.
randomization
d.
probabilistic aggregation
16. Graduates majoring in criminal justice and graduates majoring in psychology from the
State University were part of a study to assess employment success of those receiving
their degrees in May 2002. Six months after graduation (November 2002) a questionnaire
was sent out asking about their job seeking success. What type of design is this?
a.
a double-blind experiment
b.
a non-equivalent groups design
64
c.
a time-series design
d.
a cohort design
17. The history threat refers to the possibility that ____.
a.
because of their personal experiences, some subjects will be uncooperative
b.
most important research in the social sciences has already been done
c.
events external to the experiment that will impact the results
d.
the experiment will be judged to be unimportant
18. The maturation threat refers to _____.
a.
older subjects being more likely to take research participation seriously
b.
the impact of natural, developmental processes on experimental outcomes
c.
the natural development of researcher expertise as more research is completed
d.
younger subjects generally do better because they are easier to manipulate
19. Which action would be most likely to result in selection bias?
a.
soliciting volunteers from a college class
b.
randomly assigning experimental and control groups
c.
matching experimental and comparison groups
d.
running a double blind experiment
20. Experimental mortality is especially likely to be a problem when ____.
a.
subjects are paid an excessive amount to participate
b.
participation requires only one session, but it is an hour long
c.
students must participate as a course requirement
d.
participation in the experiment requires substantial effort by subjects
21. You have just completed the LSAT and scored 85. Your roommate, who is no smarter
than you, took the LSAT with you and scored 165. You were so upset with your score
that you decide to take the test again and you convince your roommate to take it with you
for moral support. You both receive scores on the second exam of 120. What accounts for
the differences in scores?
a.
statistical regression
b.
testing
c.
instrumentation
d.
experiment mortality
22. Neither the subjects nor the researcher can identify which group is the control group and
which is the experimental group. Which design is being used?
a.
cross-sectional
b.
double-blind
c.
a blind experiment
d.
of no value
23. External validity in an experimental design refers to _____.
a.
the accuracy of measurements
b.
the lack of concern with issues internal to the study
c.
the ability to apply findings to other populations
d.
whether appropriate outcome measures have been used
24. Threats to internal validity in an experimental design can be minimized by _____.
a.
proper selection and assignment of subjects to groups
b.
large sample size
c.
small sample size
d.
nothing can rule out threats to internal validity
66
25. Which technique can be used to help reduce the threat to experimental mortality?
a.
increasing the sample size
b.
decreasing the sample size
c.
shortening the time between the pretest and the posttest
d.
selecting only subjects who are female
26. Random assignment to experimental and control groups should reduce the threats to
_____.
a.
instrumentation
b.
maturation
c.
reliability
d.
small sample
27. If randomization is not possible, the researcher would be well-advised to ____.
a.
do nothing, but mention the issue in the research report
b.
switch to a classical design
c.
switch to a quasi-experimental design
d.
use pseudorandomization
28. An interrupted time-series design can be very useful in ____ research.
a.
applied
b.
basic
c.
primary
d.
double-blind
29. A questionnaire is sent to 1,000 police departments in Florida and gathers information on
the number of sick days officers use after being involved in a shooting incident. This is
an example of ____ research.
a.
variable-oriented
b.
case-oriented
c.
quasi-experimental
d.
cohort
67
30. Experimental mortality is also known as ____.
a.
selection
b.
regression
c.
attrition
d.
relevance
COMPLETION
1. In a classical experiment, the researcher uses a(n) _______________ to offset the effects
of the experiment itself.
2. A(n) _______________ experiment guards against experimenters’ tendency to prejudge
results by eliminating the possibility of either the researcher or the subjects knowing
which group is the control and which group is the experimental.
3. The goal of _______________ is to ensure that the experimental and groups will be
statistically equivalent.
4. _______________ relates to whether results from experiments conducted in one setting
would be obtained in other settings.
5. When random assignment to experimental and control groups is not possible, the
researcher may be forced to use a(n) _______________ design.
6. Instead of making one pre-test and one post-test measure, the _______________ makes a
longer series of observations before and after introducing an experimental treatment.
7. In _______________ research, a large number of variables that are associated with a
small number of cases or subjects are studied.
8. A(n) _______________ is a type of variable-oriented research.
9. Increasing experimental control will result in decreased _______________ validity.
10. _______________ refers to the extent to which findings from an experiment can be
applied to the real world.
ESSAY
1. Explain why experimentation is not well suited for the purposes of description and
exploration.
2. Explain the classical experiment in terms of the following:
Independent and dependent variables
Pre-testing and Post-testing
Experimental and Control groups
3. Explain the role of randomization in classical experiments.
69
4. Identify and discuss at least five (5) threats to internal validity in experimental designs.
Include in your answer ways of controlling for each of the threats you identify.
5. Give two (2) specific examples of research where a quasi-experimental design would be
more appropriate than a classical experiment. Why would each of your examples not fit
the classical experiment model?
6. Describe an experimental design to test the causal hypothesis that D.A.R.E. reduces drug
use. Is your experimental design feasible? Why or why not?
7. Experiments are often conducted in public health research where a distinction is made
between an efficacy experiment and an effectiveness experiment. Efficacy experiments
focus on whether a new health program works under ideal conditions; effectiveness
experiments test the program under typical conditions that health professionals encounter
in their day-to-day work. Discuss how efficacy experiments and effectiveness
experiments reflect concerns about internal validity threats on the one hand and
generalizability on the other.
8. Crime hot spots are areas where crime reports, calls for police service, or other measures
of crime are especially common. Police in departments with a good analytic capability
routinely identify hot spots and launch special tactics to reduce crime in these areas.
What kinds of validity threats should researchers be especially attentive to in studying the
effects of police interventions on hot spots?