This article describes an independent assessment of the Fast ForWord Language computer-based training program developed by Scientific Learning Corporation. Previous laboratory research involving children with language-based learning impairments showed strong effects on their abilities to recognize brief and fast sequences of nonspeech and speech stimuli, but generalization of these effects beyond clinical settings and student populations and to broader literacy measures remains unclear. Implementing a randomized field trial in eight urban schools, we generated impact estimates from separate intent-to-treat and treatment-on-the-treated analyses of the literacy outcomes of second- and seventh-grade students who were more generally at risk for poor reading and language outcomes. There were some problems of implementation in the field setting, and the Fast ForWord Language program did not, in general, help students in these eight schools improve their language and reading comprehension test scores. (Contains 6 notes and 10 tables.)