Use of Multiple Regression for Evaluating Pay Equity: Prospects and Pitfalls

Fri Jan 14 15:24:00 EST 2011
By Dr. Alex Grecu, et al.

The recent pace of change in enforcement of pay equity has left many heads spinning. Legislative activity, changes in agency enforcement and regulatory requirements, and key court decisions have spurred new interest among employers and their counsel in self-critical reviews for evaluation of racial, ethnic, or gender equity. Although the Office of Federal Contract Compliance Programs has rescinded its previous guidelines for using multiple regression to examine pay equity, it remains one of the most powerful and frequently used tools for undertaking such assessments. The Equal Employment Opportunity Commission and private Title VII litigants often rely on regression for evaluation of claimed pay discrimination. Although it is too early to know what approaches the Dodd-Frank Offices of Minority and Women Inclusion may use, their enforcement activities could spawn more extensive use of this approach as well.

In this primer on the use of multiple regression analysis, Vice President Dr. Alex Grecu and former Senior Vice President Betsy Becker, argue that simply using regression is not an assurance that a reliable evaluation of pay equity has been made, or that self-imposed remedies will be immune from legal challenges. An analysis has to be carefully specified and applied to capture the realities of the workforce that is being evaluated. Lacking advice of expert economists or statisticians, a study can succumb to many potential pitfalls. The authors note that inappropriate conception, execution, or interpretation of the results may not only render a study useless but may open the door to unintended legal risks. Moreover, practitioners need to be aware that the solutions to problems can be as varied as the employers and employees being evaluated.

This primer is intended to assist employers and their counsel as they embrace this tool, with all its complexities, to evaluate equity in their workforce.