Get access

Not Seeing Clearly With Cleary: What Test Bias Analyses Do and Do Not Tell Us

Authors


  • Adam W. Meade, Department of Psychology, North Carolina State University; Scott Tonidandel, Department of Psychology, Davidson College.

Adam W. Meade. E-mail: awmeade@ncsu.edu
Address: Department of Psychology, North Carolina State University, Campus Box 7650, Raleigh, NC 27695-7650

Abstract

In recent decades, the Cleary (1968) approach for testing for differences in regression lines among demographic groups has been codified as a central approach to evaluate a test for bias. However, this approach is fraught with numerous shortcomings, a preponderance of implicit assumptions, and outcomes that are not sufficient to conclude that there is a problem with a test. We believe these shortcomings are poorly understood by many industrial–organizational (I–O) psychologists, that this method for evaluating test bias is overrelied on by our profession, and that it is interpreted improperly by those wishing to evaluate tests for bias in applied settings. Moreover, eliminating differential prediction may be impossible in some cases, undesirable in others, and places an undue burden on organizational researchers.

Ancillary