Sex-Based Salary Disparity and the Uses of Multiple Regression for Definition and Remediation

Abstract

Although sex differences in salaries are widely condemned, some organizations tolerate these disparities. When organizations are found to discriminate against women in salaries for the same jobs as men, a number of remedies can be pursued. This study illustrates how a state university responded to demonstrated sex-based salary disparities identified in a class-action lawsuit. In particular, a regression analysis was conducted for male and female faculty members that also included seven other predictors. Statistical differences between male and female faculty members were found. As part of the remediation, the salaries of 39 women faculty members were increased. Subsequent regression analysis indicated that the effect for faculty member sex no longer reached traditional levels of statistical significance. The court-order remediation was deemed to have removed the sex-based salary disparity. Yet, it is not clear that the post-remediation salaries of the women faculty were perceived as fair. This case demonstrates how concepts from regression can be (mis)used in legal cases of salary fairness and disparity.