Example Exam questions week 3
Impact Evaluation (UC-17-ECB304)
Erasmus Universiteit Rotterdam
Aanbevolen voor jou
Reacties
Preview tekst
Example Exam questions week 3 – Fixed Effects and Difference-in -Difference
- Consider the following figure
What would be the expected outcome for the treatment group at time 1 in absence of treatment (i. “the counterfactual”) in a Difference-in-Difference (DiD) design? a. 6 b. 4 c. 2 d. 14
What could be a problem with randomization that leads to bias in the estimation of the Average Treatment Effect on the Treated (ATT)? a. Non-parallel trends b. Small treatment effects c. Possible spillover effects to non-treated d. Limited common support
Consider the following figure, in which treatment takes place between t = 0 and t=
What is a problem when conducting a Differences-in-Differences design in this case?
a. There may be observed time-invariant factors that influence the outcome b. The observed outcomes at time 0 (before treatment) are different, such that the estimates are confounded by this initial difference c. The parallel trends assumptions is unlikely to hold, since pre-treatment trends are different d. The outcomes for the treated are initially higher than the ones for the control group
What is the parallel trends assumption? a. Treated and non-treated would otherwise have had the same outcomes b. In an arbitrarily small interval around the relevant threshold, treatment assignment can be considered random c. Outcomes of treated and non-treated would have changed in the same way in absence of treatment d. There are no other determinants of outcomes changing between pre and post intervention
Twin fixed effects is a popular method to establish the causal effect of some treatment
variable on an outcome. Say that we have two twins 1 and 2, Y is the outcome, T is treatment, X are family specific variables, and G are genes specific to the twin pair p.
l fp f p
l l fp
l fp f p
l l fp
fp
fp Y T X G
Y T X G
2
1 2 2
1 1
In order to obtain the causal effect of treatment T on the outcome Y, typically the difference between two twins, Y 1 – Y 2 is taken. Which statement is NOT true regarding Twin Fixed Effects estimation? a. Twins should differ in their treatment status T, otherwise we cannot use the twin pair in the estimation of the causal effect of treatment on the outcome b. Twin fixed effects studies are particularly vulnerable to measurement error because one has to take differences within twin pairs c. Within-twins pair treatment variation may be correlated with unobserved within-twins factors that directly affect the outcome d. Twin fixed effects studies can estimate the coefficient β of unobserved family characteristics Xf
Answers:
D
C
C
C
A
A
D
Example Exam questions week 3
Vak: Impact Evaluation (UC-17-ECB304)
Universiteit: Erasmus Universiteit Rotterdam
Dit is een preview
Toegang tot alle documenten
Onbeperkt downloaden
Hogere cijfers halen
