This paper examines a quasi-experiment in which we encourage student effort by setting various weekly incentives to engage in online tests. Our identification strategy exploits i) weekly variation in incentives to determine their impact on student effort, and ii) controlled cross-group variation in assessment weighting. Assessment weighting strongly encourages quiz participation, without displacing effort over the year. We estimate the return to a quiz at around 0.15 of a standard deviation in exam grade. Effort in our study increases most for students at and below median ability, resulting in a reduction of the grade gap by 8%.