An Analytical Method of Identifying Biased Test Items
Barbara Sterrett Plake, Hiram D. Hoover · The Journal of Experimental Education · 1979
When using an ANOVA procedure to examine a test for biased items, a follow-up technique is needed to identify the items contributing to the items-by-groups interaction. The method described in this paper includes distribution theory for assessing level of significance and is equally sensitive to items at all difficulty levels. In addition, any number of groups can be considered and the methodology is generalizable to any ANOVA design. The technique is powerful yet controls experiment-wise error rate to a.