, Question 1
A social work researcher wants to test whether a new housing intervention
reduces homelessness recidivism while accounting for unmeasured
time-invariant confounders. Which design best balances internal validity and
ethical feasibility?
A. Randomized controlled trial with waitlist control
B. Regression discontinuity design using a cutoff score
C. Interrupted time series with nonequivalent comparison group
D. Difference-in-differences with propensity score matching
Correct Answer: D - Difference-in-differences with propensity
score matching
RATIONALE
Difference-in-differences with propensity score matching controls for
both observed and unobserved time-invariant confounders, offering
strong internal validity when randomization is infeasible. RCTs are
ethical but often impractical for housing interventions; regression
discontinuity requires a sharp cutoff; interrupted time series lacks a
clear counterfactual.
Question 2
A scale measuring depression yields a Cronbach's alpha of 0.92, but item-total
correlations for two items are below 0.30. What is the most appropriate
psychometric interpretation?
A. The scale has excellent internal consistency and should be retained as
is.
B. The high alpha indicates redundancy; the two items should be removed
to improve validity.
C. The low item-total correlations suggest those items may measure a
different construct, warranting further factor analysis.
D. Alpha is inflated by the large number of items; the scale should be
Page 2
A social work researcher wants to test whether a new housing intervention
reduces homelessness recidivism while accounting for unmeasured
time-invariant confounders. Which design best balances internal validity and
ethical feasibility?
A. Randomized controlled trial with waitlist control
B. Regression discontinuity design using a cutoff score
C. Interrupted time series with nonequivalent comparison group
D. Difference-in-differences with propensity score matching
Correct Answer: D - Difference-in-differences with propensity
score matching
RATIONALE
Difference-in-differences with propensity score matching controls for
both observed and unobserved time-invariant confounders, offering
strong internal validity when randomization is infeasible. RCTs are
ethical but often impractical for housing interventions; regression
discontinuity requires a sharp cutoff; interrupted time series lacks a
clear counterfactual.
Question 2
A scale measuring depression yields a Cronbach's alpha of 0.92, but item-total
correlations for two items are below 0.30. What is the most appropriate
psychometric interpretation?
A. The scale has excellent internal consistency and should be retained as
is.
B. The high alpha indicates redundancy; the two items should be removed
to improve validity.
C. The low item-total correlations suggest those items may measure a
different construct, warranting further factor analysis.
D. Alpha is inflated by the large number of items; the scale should be
Page 2
