We present an analysis on the differential effects of incentivizing homework in an introductory mathematics course at the United States Military Academy. We found that including homework as part of a student's overall course average (incentive) led to a significantly higher performance (achievement) on homework assignments. However, doing homework only led to a significantly higher exam performance for medium-ability students. One possible explanation for our results is high-ability students may be able to grasp material with minimal out-of-class practice; medium-ability students benefit significantly from solving practice problems; and low-ability students may have difficulty in understanding even if they are actively practicing problems.