Q1. How many law schools are in the dataset?

Q2.

Which of the following is the correct interpretation of b1 in Model 1?

a. If you increase GPA by one unit, then log median salary will increase by b1 units

b. If School A has median college GPA that is one point higher than School B, then we predict School A will have log median salary b1 units higher than School B

c. If School A has median college GPA that is one point higher than School B, then we predict School A will have log median salary b1 units lower than School B

d. If you increase GPA by one unit, then log median salary will decrease by b1 units

Q3. Theoretically, would you expect b1 and b2 in Model 2 to be positive of negative?

a. b1 negative and b2 positive

b. b1 and b2 both positive

c. b1 and b2 both negative

d. b1 positive and b2 negative

Q4. Which of the following is the correct interpretation of b1 in Model 2?

a. If School A has median college GPA that is one point higher than School B and if both schools have the same median LSAT score, then we predict School A will have log median salary b1 units higher than School B

b. If we increase the median college GPA, then log median salary will increase by b1 units.

c. If we increase the median college GPA and hold median LSAT constant, then log median salary will increase by b1 units

d. If School A has median college GPA that is one point higher than School B, then we predict School A will have log median salary b1 units higher than School B

Q5. What is the value of b1 in Model 1?

Q6. What is the value of the R2 in Model 1?

Q7. What is the value of b1 in Model 2?

Q8. The value of b1 differs between Models 1 and 2 because

a. There is no way to know

b. Model 2 has a higher R-squared

c. Model 2 adds a new variable (LSAT) that is correlated with GPA and log salary

d. Model 2 adds a new variable (LSAT) that is uncorrelated with GPA and log salary

Q9. What is the value of b1 in Model 3?

Q10. Suppose School A and School B have the same values for all the variables on the right-hand side in Model 3, except School A is ranked 10 places higher than School B. What is the difference in predicted difference in log median salary between the two schools?

Q11. Suppose School A and School B have the same values for all the variables on the right-hand side in Model 3, except School A costs $10,000 and School B costs $15,000. What is the difference in predicted difference in log median salary between the two schools?

Q12. Suppose a school were to double its cost. Would you expect graduates from the program to earn higher salaries? Why?

a. Yes, because the coefficient on cost in the regression is positive, which means that salaries tend to be higher at higher cost schools, holding constant rank, LSAT, and GPA scores.

b. No, because the coefficient on cost in the regression is positive, which means that salaries tend to be lower at higher cost schools, holding constant rank, LSAT, and GPA scores.

c. Yes, because the R-squared in Model 3 is higher than in Models 1 and 2

d. The coefficient on cost in the regression is positive, which means that salaries tend to be higher at higher cost schools, holding constant rank, LSAT, and GPA scores. However, increasing the cost of the school may not increase salaries because correlation is not the same as causation.

Q13. Re-estimate Model 1, except add north, south, east, and west as additional right hand side variables. What is wrong with this regression?

a. This model suffers from multicollinearity because north+south+east+west=1

b. Its R-squared is lower than Model 3

c. It excludes the LSAT variable

d. The value of b1 is different from that in Model 1

Q14. Consider the following two regression equations:

Estimate b1. This is a two-step process. First, you need to estimate the first regression model and save the errors. Then, you regress log(salaryi) on the errors (vi). What is your value of b1?

IMPORTANT TIP: Before estimating the first regression model, type the command “drop if lsalary==.” This will delete the observations for which lsalary is missing. There are 148 observations on LSAT and GPA, but only 141 on lsalary. You need to use the 141 observations in both regressions.

Q15. Your value of b1 in the previous question is

a. larger than in Model 1

b. the same as in Model 2

c. larger than in Model 2