Deck 8: Quasi-Experimental Designs for Public Health

Question 1

The need for quasi-experimental design studies is proof that randomized experimental studies are not effective in interventions at the community level or higher.

Answer 1

False

Question 2

The nonequivalent groups pretest-posttest design does not require that the comparison community:

A) be demographically similar to the intervention community.
B) be geographically similar to the intervention community.
C) agree to regularly scheduled waves of data collection.
D) have pre-existing public health practices and policies similar to those of the intervention community.

Answer 2

B

Question 3

NEGDs are especially susceptible to the threat of differential selection. Which of these is the most unlikely cause?

A) There were errors in the baseline data collection for the communities.
B) A historical effect took place that affected the intervention.
C) The chosen communities were not similar enough.
D) The investigator had adjusted for differences between the two communities.

Answer 3

D

Question 4

Why would you use a posttest-only design study instead of a pretest-posttest design?

A) It would require less data collection.
B) The preintervention differences would not affect your assessment.
C) You want to evaluate the effectiveness of a program after it has begun.
D) You are certain of the outcome from the intervention.

Question 5

What is it called when a historical event affects one group but not the other?

A) Differential selection
B) Differential attrition
C) Selection history
D) Extraneous variable

Question 6

Nonequivalent groups pretest-posttest design studies can be best improved by:

A) ensuring the intervention and control communities are similar.
B) maintaining a rigorous data collection process, before and after the intervention.
C) utilizing a switch replication method in the design.

Question 7

Switching replication can help control for maturation, instrumentation, and differential selection.

Question 8

Nonrandomized stepped-wedge design studies:

A) allow each cluster to receive the intervention simultaneously to remove the "randomness" found in other experimental designs.
B) have equally long control and intervention phases.
C) have drastically varying experimental timelines, depending on the intervention.

Question 9

Interrupted time-series design (ITSD) studies are known for the strong evaluation of:

A) cross-sectional studies.
B) quasi-experimental studies.
C) cohort studies.
D) longitudinal studies.

Question 10

A research study is determining whether providing discount codes for Lyft and Uber on the weekends has an effect on the number of DUIs given to people between the ages of 25 and 35. The intervention lasted for 12 months, and researchers collected data on a weekly basis, even 3 months after the intervention ended. This ITSD:

A) requires no changes.
B) is missing data prior to the intervention.
C) could be considered an ARIMA model.
D) has accounted for maturational change.
E) has accounted for seasonal trend.