Multiple choice assessments are one of the most common methods instructors use to predict the ability level of their students. However, despite the widespread prevalence of this kind of testing, the optimal way of analyzing the results to predict each student’s ability remains unanswered. This work compares the most commonly used dichotomous scoring, with various polytomous methods and answer-until-correct procedures. Each of these scoring methods were then modeled using item response theory (IRT) and student ability estimates were extracted out of each model. This then creates a 4 (scoring methods) x 3 (extraction techniques) array of various scores for each student. Each of the cells in this array were compared to determine which score comes closest to predicting the students true underlying ability. The quantitative results from this analysis were further validated with a qualitative approach to student scoring. Both approaches demonstrate the added benefit of assigning partial credit and the utility of IRT.

Search for the optimal measurement of student chemistry ability using multiple-choice assessments

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.