Conceptualizations of test bias and adverse impact : implications of recent policy proposals

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Anita S. Tesh, Associate Dean for Undergraduate Programs (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site:
Richard M. Jaeger

Abstract: The purpose of this study was to investigate selected implications of recent proposals for the imposition of a set of test assembly procedures called the "Golden Rule procedures" in a wide variety of testing situations. These test assembly procedures stipulate that items selected for test inclusion should exhibit differences in performance between groups below a specified level, and should not be more difficult than specified, for either majority or minority groups. This study used data from the performances of 1807 examinees on four standardized tests completed during their eighth and tenth school grades. Synthetic tests composed of items from the original tests which conformed to the stipulations of the Golden Rule procedure were created. The effects of applying the procedures on the adverse impact and racial bias of test use were examined through comparisons between and among the properties of and results of using original tests and corresponding synthetic tests.

Additional Information

Language: English
Date: 1990
Educational tests and measurements
Test bias
Students $x Rating of

Email this document to