One additional hour of literacy instruction per day = significant improvement in reading test scores

Instructional time is a fundamental educational input, yet we have little causal evidence about the effect of longer school days on student achievement. 

This paper uses a sharp regression discontinuity design to estimate the effects of lengthening the school day for low-performing schools in Florida by exploiting an administrative cutoff for eligibility. 

The results indicate significant positive effects of additional literacy instruction on student reading achievement. In particular, a 0.05 standard deviations of improvement in reading test scores for program assignment in the first year, though long-run effects are difficult to assess.