Regression analysis is frequently used to test for wage discrimination at the firm level. The Employee Data set (on Moodle) contains the employee data from Firm ABC. Specifically, the data set contains the salary, years of experience (at Firm ABC), last year’s annual performance rating*, years of education past high school, a dummy that takes the value of 1 if the employee is in a management role, a dummy that takes the value of 1 if the employee is female, and a dummy that takes the value of 1 if the employee is non-white. Estimate the following regression (where y is salary) and answer the questions that follow:

y= β_0+β_1 Exp+ β_2 Perf+ β_3 Ed+ D_1 MGT+ D_2 Gender+ D_3 Race+ε

What variables significantly affect an employee’s salary at Firm ABC?

Is there any evidence of gender or racial discrimination at Firm ABC?

What about the model assumptions? Are there any concerns?

Finally, there is a concern that female managers at Firm ABC earn less than male managers, all else equal (which means assuming the same years of experience, performance rating, and education level). Use regression analysis to test if there is statistical evidence of this kind of discrimination. Hint: You will need to create a new dummy variable which equals to 1 if the person is both a female and in management position, and zero if not a woman or not in management position. An easy way of doing it is by multiplying MGT variable with the Gender variable.

*The annual performance rating is based on a possible score of 1 (worst possible employee performance) to 10 (best possible employee performance). The scores are given each year with the employees’ performance reviews.