A Comparison of the Nedelsky and Modified Angoff Standard-Setting Procedure On Evaluation Outcome

In criterion-referenced evaluation, several procedures for setting absolute standards of performance are available. However, it is not known if a given standard-setting procedure will yield consistent evaluation outcomes. To determine this, one of three parallel certifying examinations was administered to three classes of second year medical students. The examinations contained two item types and the standards were set by two procedures. The student responses to both item types were analyzed by the Rasch latent trait model. The results indicated that the standard-setting procedure is a significant factor in the determination of the evaluation outcome.