A Model to Address Compression Inequities in Faculty Salaries

Abstract

This paper presents the details of an effort by Indiana State University to apply multiple regression analysis to the problem of eliminating faculty salary inequities. The services of a consultant were used to develop a multiple regression model to predict an individual's salary while excluding any independent variable addressing race or gender. Residuals were then computed by gender, gender by rank, race, and race by rank. The results showed no evidence of a consistent pattern of salary bias due to gender or race. The consultant's analysis appeared somewhat contradictory when considering salary compression in that a variable representing the number of internal promotions (a previously undocumented measure of compression) was statistically significant in the model, even though the analysis stated that no evidence of significant compression existed. After much discussion, the internal promotions variable was dropped. The revised model was run and residuals analysis was used to verify the presence of significant salary compression. The residuals became an important consideration in the adjustment of salaries to address the compression problems. The use of the model demonstrates how quantitative models can aid in addressing salary inequities and improve faculty salary allocation decisions by combining a data-driven approach with more traditional peer evaluation methods.