Question

True or False:

1) ____ In the context of a 2SLS estimation, it is usually risky to use models that are non-linear in the parameters to produce first-stage estimates, since in this case we can no longer guarantee that the first-stage residuals will be uncorrelated with the regressors included in the second-stage.

2) ____ Using first differencing or deviations from means (time-demeaning) to eliminate the "unobserved effect" generally yields identical coefficient estimates and inferences for the variables of interest.

3) ____ In a typical difference-in-differences (DD) analysis, the causal interpretation of the estimates of interest crucially relies on the "common trends" (or "parallel trends") assumption.

4) ____ The "clustering" procedure allows the errors to be arbitrarily correlated, but only within each cluster.

Answer 1

1. In the context of a 2SLS estimation, it is usually risky to use models that are non-linear in the parameters to produce first-stage estimates, since in this case we can no longer guarantee that the first-stage residuals will be uncorrelated with the regressors included in the second-stage.

False

2. Using first differencing or deviations from means (time-demeaning) to eliminate the "unobserved effect" generally yields identical coefficient estimates and inferences for the variables of interest.

True

3. In a typical difference-in-differences (DD) analysis, the causal interpretation of the estimates of interest crucially relies on the "common trends" (or "parallel trends") assumption.

False

4. The "clustering" procedure allows the errors to be arbitrarily correlated, but only within each cluster.

True