Improving performance in upper-division electricity and magnetism with explicit incentives to correct mistakes

Abstract

This study seeks to determine whether giving an explicit incentive to students in an upper-division first-semester electromagnetism course (EM1), in the form of partial credit for reworking unit exam problems, will improve their problem-solving skills as measured by performance on identical problems on the final exam. Three problems-a primarily algorithmic problem, a primarily conceptual problem, and a problem that blended conceptual with algorithmic-were selected and analyzed over the course of three consecutive sections of EM1, for which each student could freely choose whether or not to rework mistakes on each one of the given problems in exchange for the aforementioned partial credit. Regression models were chosen for quantitative analysis, with the covariate being unit exam performance and whether the student reworked the problem. Results indicate that overall, students who choose to rework problems perform better on the final exam attempt, at a level that often does not correlate strongly with unit exam performance, whereas students who decline to rework problems have a stronger correlation between unit exam and final exam performances. The results show a clear difference between the two stages of problem-solving, namely, invoking the correct principles and applying the principles, where the latter showed a more significant effect with a much larger effect size. Qualitative analysis of a sample of students interviewed about their exam solutions showed that reworking an exam problem for some students did result in more expertlike problem-solving trends; that being said, there were instances of persistent novicelike trends regardless of reworking, as well as instances of students independently reviewing unit exam problems in preparation for the final even though the partial credit incentive was declined at the time of the unit exam. Thus, while the intervention showed overall benefits insofar as exam performance is concerned, the usage of more well-defined scaffolding, e.g., tutorials, may prove a more thorough and definite benefit for improving problem-solving skills.