Much research-and much litigation-has been conducted on the disparity between the salary levels of men and women. Research reported in Work and Occupations (Nov. 1992) analyzes the salaries for a sample of 191 Illinois managers using a regression analysis with the following independent variables:

xz = Race of manager = 0 if not x3 = Education level (in years)
x4 = Tenure with firm (in years)
x5 = Number of hours worked per week The regression results are shown in the table below as they were reported in the article.
Variable 3 p-Value XI 12.774

< .05 X2 .713 > .lo x3 1.519 < .05 X4 .320 < .05 Xs .205 < .05 Constant 15.491

-

a. Write the hypothesized model that was used, and interpret each of the p parameters in the model.

b. Write the least squares equation that estimates the model in part

a, and interpret each of the j3 estimates

c. Interpret the value of R2. Test to determine whether the model is useful for predicting annual salary.Test using a = .05.

d. Test to determine whether the gender variable indicates that male managers are paid more than female managers, even after adjusting for and holding constant the other four factors in the model.Test using a = .05.

[Note: The p-values given in the table are two-tailed.]

e. Why would one want to adjust for these other factors before conducting a test for salary discrimination?