Question

2. In class, we described a regression as a “machine” that transformed certain inputs into certain outputs. Which of the following is NOT an “output” that comes out of running a regression?
(A) The regression equation
(B) The estimated coefficients
(C) Fitted values
(D) Residuals
(E) All of the above are outputs that come out of running a regression.

3. All of the following would be considered categorical data EXCEPT
A) One’s gender
B) Which U.S. state one lives in
C) One’s telephone number
D) One’s zip code
E) All of the above would be considered categorical data.

4. Which of the following econometric techniques would help address reverse causality in a regression?
A) Control variables
B) Instrumental variables
C) Controls and instruments would both work.
D) Neither controls nor instruments would work.

5. You want to regress standardized test scores for student i on the student-teacher ratio in student i’s classes. All of the following would make good controls EXCEPT
A) whether student i is an English learner.
B) whether student i attends a public school.
C) the time of day student i took their standardized test.
D) education spending in student i’s school district.

Answer 1

2. The correct option is A. The regression equation is a model constructed to study the relationship, not the output of the analysis.

3. The correct option is C. Each telephone number is unique and we cannot categorise them.

4. The correct option is B. Instrumental variables are exogenous variables weakly correlated to the explanatory variable and help in removing the problem of reverse causality.

5. The correct option is C. The scores may not depend on the time of the day as it is unlikely that his learning abilities change during the course of the day.