This paper tests whether increasing students’ exposure to “desirable difficulties” improves learning in real classrooms. In a year-long field experiment in Nigerian primary schools, interleaved math practice raised short-term test performance by 0.28 standard deviations but had no effect on cumulative end-of-year assessments. Gains were concentrated among lower-achieving students, while higher-achieving students saw little or no benefit. The results suggest that strategies that make learning more effortful can boost short-term mastery but may not produce durable improvements, highlighting the limits of applying laboratory-based cognitive interventions at scale.