Junior high school students were provided tablets with digitized learning modules that connected to school servers via local area networks. The intervention improved learning in mathematics, but not English, with intent-to-treat effects of 0.34 and treatment-on-treated effects of 0.46 standard deviations. Effects were greater for students with higher baseline “grit” and achievement. The small subset of schools that continued using the materials for a second year achieved effects of 1.6 standard deviations.