The impact of distributed homework on student outcomes

Distributed practice improves learning by requiring the brain to expend extra effort retrieving prior learning after a time delay. I examine whether repeating the most troublesome homework question on the next assignment improves exam performance within one large upper-level undergraduate economics course. I compare exam outcomes of students enrolled in Fall 2017 as my control group (N = 136) with those of the intervention group in Spring 2018 (N = 163). Adjusting for differences in student characteristics, the intervention was associated with a statistically significant (at the 90% level) increase of 2.44% in final exam scores, with raw average scores of 84.6% versus 81.7%. No difference was found post-intervention in overall course scores, while small increases for midterms. Subgroup analysis suggests the benefits may accrue more to the strongest and weakest performers. Findings suggest that repeating troublesome problems could improve learning in economics.