Many educational policymakers consider attendance as a tool to induce learning. Researchers also agree that attendance has a positive effect on learning; however, there are few empirical studies that measure the nature and significance of that effect. The authors analyzed the effect of class attendance on academic performance and evaluated the existence and importance of a minimum attendance requirement. Using student data from a sample of public primary schools in Chile, and considering for endogeneity and sample selection bias, they found two important results. First, attendance had a relevant and statistically significant effect on educational performance. Second, the existence of a threshold was identified, but educational performance did not continue to decrease after a certain number of absences, which seems to contradict policies that have a minimum attendance requirement. (Contains 4 notes, 5 tables, and 1 figure.)