Abstract
Multiple regression procedures are commonly used to investigate gender equity in faculty salary. However, a review of 19 case studies indicates that many of them fail to (a) adequately develop a regression model that examines the linear and nonlinear (i. e., interactive and curvilinear) relations between predictors and the criterion and (b) appropriately apply regression diagnostic statistics throughout salary model development. A seven-step process is presented as a comprehensive framework for testing allegations of gender discrimination in faculty salary. Steps include (a) identifying predictors of faculty salary, (b) identifying and establishing criteria for interpreting statistical tests and diagnostic procedures, (c) determining the criterion variable used in the salary model, (d) developing a salary model, (e) testing for gender discrimination in pay, (f) conducting diagnostic procedures to confirm the appropriateness of the final salary model, and (g) testing the assumptions of the regression model. An application of this model is presented using a case study (N = 725 faculty).