ASSESSMENT OF DIFFERENTIAL ITEM FUNCTIONING FOR PERFORMANCE-TASKS

Although the belief has been expressed that performance assessments ar e intrinsically more fair than multiple-choice measures, some forms of performance assessment may in fact be more likely than conventional t ests to tap construct-irrelevant factors. The assessment of differenti al item functioning (DIF) can be helpful in investigating the effect o n subpopulations of the introduction of performance tasks. In this stu dy, two extensions of the Manter-Haenszel (MH; 1959) procedure that ma y be useful in assessing DIF in performance measures were explored. Th e test of conditional association proposed by Mantel (1963) seems prom ising as a test of DIF for polytomous items when the primary interest is in the between-group difference in item means, conditional on some measure of ability. The generalized statistic proposed by Mantel and H aenszel may be more useful than Mantel's test when the entire response distributions of the groups are of interest Simulation results showed that, for both inferential procedures, the studied item should be inc luded in the matching variable, as in the dichotomous case. Descriptiv e statistics that index the magnitude of DIF were also investigated.