Association for Public Policy Analysis & Management

Additional instructional time has been a popular, yet controversial, policy to improve achievement of low-performing students. This policy is potentially appealing because instructional time is a fundamental resource in education, and not surprisingly, researchers have long described the theoretical importance of instructional time for student outcomes. Additional instructional time for low-achieving students certainly has intuitive appeal - struggling students may simply need more time to learn. On the other hand, additional instructional time policies are expensive, and the costs may outweigh the potential gains in achievement if the benefits of additional instructional time are mitigated by low teacher quality, disadvantaged school settings, or failure to efficiently use the additional time.

Empirically, we know very little about the effectiveness of additional instructional time. Many studies have explored the correlation between additional instructional time and achievement, but schools that implement these policies are likely very different from others in terms of observable and unobservable characteristics. For example, high-performing schools may have more resources, and these resources allow high-performing schools to provide more instructional time, and simple comparisons would overstate the effect of additional instructional time on achievement. Similarly, additional instruction time policies are typically tied to other school-level interventions, making it difficult to isolate the causal effect of additional instruction time on student outcomes.

In this study, we make two contributions to the literature on the effects of additional instruction time on student outcomes. First, we address these potential sources of bias in a regression discontinuity framework by exploiting school-level administrative cutoffs that determine whether students receive additional instructional time in a Florida education program. Second, we conduct an unprecedented analysis of heterogeneous treatment effects using unique administrative data that links school records with student birth records. In particular, we examine the differential effects of additional instruction time along a multitude of student characteristics including student poverty, gender, race/ethnicity, nativity, and language spoken at home as well as maternal characteristics such as maternal education, mother’s marital status, mother’s age during pregnancy, and mother’s place of birth.

We find significant benefits of additional instruction time on reading test scores. In particular, regression discontinuity estimates suggest that students enrolled in schools whose reading accountability scores fell right below the AIT cutoff score roughly 0.03σ to 0.09σ better in reading compared to students enrolled in schools on the other side of the cutoff. These differences in reading achievement correspond to approximately 1 to 3 months of learning. In contrast, there are no significant discontinuities in math performance at the AIT cutoff. We also find that the effects of additional instruction time vary considerably across student subgroups, with the benefits on reading achievement mainly concentrated among students from disadvantaged backgrounds. For instance, we find that additional instruction increases reading achievement by 0.07σ for persistently poor students, by 0.05σ for students whose mothers has a high school degree or less, and by 0.10σ for students whose mothers were teenagers when the student was born, while we find no significant effects for students from more advantaged backgrounds.