Uncertain Responses on Multiple-Choice Examinations

Abstract

100 students were asked to label item responses on two multiple-choice tests as sure, uncertain, and guess. The uncertain category was significantly different from the sure and guessed response categories, and from the theoretical chance expectancy level. This demonstrates that the basic assumption of the correction-for-errors scoring formula (that all wrong responses are pure guesses) penalizes the student by utilizing only those responses of which he feels certain.