Abstract

This study considers multiple-choice exams’ ability to measure student knowledge. Multiple-choice exams were adopted in the US early in the last century. They were believed to be more objective measures of knowledge than essay exams. Multiple-choice exams are useful in lower level business courses because the learning required tends towards the knowledge level and, multiple-choice questions are able to cover many topics in a short time. In addition, multiple-choice exams show economies to scale in preparation and use, and have a low cost of grading. This is especially important with lower level classes, which are generally large with students that display the whole range of knowledge and ability. The whole range of opinions about multiple-choice exams, test taking strategies, and grading schemes exists. Much of the literature considers guessing and finding ways to correct multiple-choice exam scores for guessing. At the University of Wisconsin-Whitewater, we teach the first course in financial accounting as a common course. During the fall semesters, we teach about 550 students and during the spring semesters, we teach up to 300 students. Multiple-choice exams determine 80% of the semester grade. Processing and determining what the scores actually mean has lead to some questions and suggestions about the scores and what they mean among the instructors. The prior research considers the relationship between scores, students’ reports of their knowledge and guessing. Generally, guessing is modeled using the binomial distribution and the assumption that guessing is random. This study uses prior research on the guessing component and considers this relationship by investigating associations between knowledge and scores. The analysis to date (i.e., one semester and 213 observations) suggests that multiple-choice exam scores are representative of knowledge.