Deck 12: Quasi-Experimental Designs

Question 1

Cross-sectional designs have a high degree of internal validity because they show how causal processes occur over time.

Answer 1

False

Question 2

Time-series designs with many measurement points control for maturation.

Answer 2

True

Question 3

To avoid or alleviate practical pitfalls when carrying out experiments in social work agencies,researchers should:

A) Include agency staff in the research planning phase.
B) Keep research staff away from program staff during the study.
C) Avoid developing or using treatment manuals for the experimental or control conditions.
D) Refrain from recruiting clients on an ongoing basis throughout the study.

Answer 3

A

Question 4

The internal validity of quasi-experiments using nonequivalent comparison groups without random assignment is weakened when switching replications are employed.

Question 5

When we switch replications in a non-equivalent comparison group design,we:

A) administer the treatment to the control group before posttesting.
B) can determine whether the difference between the groups at the first posttest may be attributed to the lack of comparability between the two groups.
C) decrease the chances of determining whether the difference between the groups at the first posttest may be attributed to a selection bias.
D)decrease the internal validity of the design.

Question 6

Recommendations to maximize compliance with the case assignment protocol in experiments or quasi-experiments include

A) having agency staff implement the case assignment decisions.
B) NOT explaining to agency staff the purpose and nature of the case assignment protocol.
C) pilot testing the randomization procedure in the agency.
D)assigning control of the case assignment protocol to multiple agency staff members.

Question 7

Quasi-experimental designs are often used instead of experimental designs because

A) agency constraints often make experiments infeasible.
B) they have more internal validity than experiments
C) quasi-experiments randomly assign participants to groups.
D) they use random assignment.

Question 8

Which of the following statements is/are true about case-control designs?

A) They are rarely feasible.
B) Data must be collected at multiple points in time.
C) Retrospective data are collected about past differences that might explain differences in outcome.
D)We can usually be highly confident about their internal and external validity.
E)We can usually be highly confident about the direction of causal influence.

Question 9

After the delivery of their baby,new parents in hospital A were given a poison-control test.While at the hospital they attended a training session on poison control.At the end of a week,they were given the poison-control test again.After the delivery of their baby,new parents in hospital B were given the poison-control test.One week later they were given the test again.What kind of design is being used?

A) pretest-posttest control group:
B) time series.
C) nonequivalent comparison group:
D) cross-sectional design.
E)multiple time-series.

Question 10

When selecting a comparison group in a quasi-experimental design,one should

A)select a group unlike the experimental group:
B)select a group as similar as possible to the experimental group:
C)use random assignment.
D)avoid the use of switching replications.

Question 11

Which of the following is/are NOT a characteristic of a cross-sectional study?

A) It is conducted at only one point in time.
B) It can be exploratory, descriptive, or explanatory.
C) It concentrates on the changes that take place within a specific sample over a period of time.
D) It provides a means for studying a large variation in the population at the same point in time.

Question 12

Which of the following statements is true about carrying out experiments and quasi-experiments in social agencies?

A) The intervention being tested is almost always implemented in the intended fashion.
B) The control condition almost never is contaminated by interaction between clients in the experimental and control groups.
C) Practitioners might violate the research protocol for case assignment to make sure that clients receive the intervention being evaluated.
D) Outside agencies can be relied upon to refer enough clients to ensure obtaining a sufficient sample size.

Question 13

If gun control legislation were passed to reduce the number of homicides,which of the following graphs would best indicate that the legislation might be producing the intended effect? [In all graphs the horizontal axis x axis is time and the vertical axis y axis is number of homicides.]

A)

Question 14

Nonequivalent comparison groups designs can be strengthened by

A)using multiple pretests or switching replications.
B)using smaller sample sizes .
C)using no pretests
D)using replication without switching.

Question 15

Case-control designs have a high degree of internal validity because they eliminate the problem of recall bias.

Question 16

Quasi-experiments using nonequivalent comparison groups without random assignment can be credible,assuming that the comparability of the experimental and comparison groups is plausible-especially if the researcher provides substantial evidence of that comparability.

Question 17

Time-series designs with many measurement points control for statistical regression.

Question 18

Which of the following is descriptive of simple time-series designs?

A) They guard against a process of maturation that could occur and would otherwise go undetected.
B) Regular fluctuations that occur aparrom the experimental stimulus cannot be identified.
C) They serve as a protection against the possible influence of some extraneous factor.
D) The fewer data point, the better.

Question 19

Which of the following statements is true about client recruitment and retention in experimental and quasi-experimental evaluations in social work agencies?

A) Referring agencies almost never dump clients that they don't want to serve.
B) Referring agencies are usually eager to refer clients.
C) Referring agencies might resent having their referrals assigned to a control condition.
D)Clients usually welcome the use of randomized procedures to determine which service they receive.

Question 20

Administering multiple pretests in a non-equivalent comparison group design:

A) enables us to detect whether participants are already engaged in a change process before treatment begins.
B) increases the likelihood that improvemenrom pretest to posttest is a function of statistical regression.
C) makes it more difficult to detect trends in the data..
D)decreases the internal validity of the design.

Question 21

The best way to avoid contaminating the control condition is to locate the experimental and control conditions in the same building

Question 22

It is safe to assume that agency staff members who initially approve of an experiment will remain in close compliance with its research protocol throughout the study.

Question 23

Describe and explain the logical arrangements used in nonequivalent comparison groups designs,simple time-series designs,and multiple time-series designs.

Question 24

Discuss the advantages and disadvantages of cross-sectional studies and the case control design.

Question 25

Videotaping treatment sessions should be used only as a last resort when assessing intervention fidelity.

Question 26

Discuss how qualitative methods can be used to try to avoid or alleviate the pitfalls identified above in essay number 2.

Question 27

Administering multiple pretests in quasi-experiments using nonequivalent comparison groups helps in detecting whether participants were already engaged in a change process before treatment began.

Question 28

Qualitative methods offer a number of techniques that on-site research staff members can use in attempting to observe research implementation pitfalls.

Question 29

Identify and explain four practical pitfalls commonly encountered when carrying out experiments and quasi-experiments in social agencies.

Question 30

Explain why the validity of quasi-experimental designs is strengthened by using multiple pretests,switching replications,and selecting a comparison group as similar as possible to the experimental group: