Does Allowing Multiple Attempts On Online Homework Benefit All Students?

Abstract

Determining if assignments with multiple attempts have an impact on student learning is essential in designing an effective online learning environment. Researchers conducted the study on various online learning systems in a diverse range of courses: economics, Archer; physics and electronics, Zhang et al.; macroeconomics, Rhodes & Sarbaum; geology, Brevik; chemistry, Faulconer et al. While some found evidence of multiple attempts increasing student learning, others saw it as a pathway for grade inflation and reduced learning. Methods used in the studies were ANOVA, t-test, and regression analysis. This poses the question if the number of attempts influences student learning and exam scores or if there are other hidden factors at play. In this experiment, data from four classes in a university course was taken, including homework and test scores. Multiple attempts were provided as an option on the homework assignments (with a student trying 17 attempts), and tests were limited to a single attempt. A linear regression test was used to measure the relationship between the number of attempts and exam scores. Students tended to take up to 3 attempts, with fewer taking more retakes. The mean exam score for students who took 1 attempt max was 72.91%, whereas the mean exam score for 3+ attempts taken was 84.93%. The coefficient of determination increased by the number of attempts up to 7 attempts, where the maximum correlation was found. Any attempts after 7 showed a much smaller correlation, suggesting a lack of student learning and an objective for higher grades, or grade inflation. The result supports previous research (Archer and Brevik) to limit the maximum number of attempts and provides a suggested value of a maximum of 7 attempts.