Comparison of Teacher–Written and Empirically Derived Distractors to Multiple–Choice Test Questions

Abstract

A comparison was made of the discrimination power of multiple–choice geometry test question alternatives supplied by (a) students taking an experimental test in completion format, (b) experienced teachers, and (c) student teachers. Neither experienced teachers nor student teachers were successful in supplying multiple–choice item distractors that best discriminated between high–and low–scoring students on the completion test. It was concluded that the most discriminating multiple–choice item distractors are produced by administering items in completion format, item analyzing the results, and selecting the most discriminating student errors.