The goal of this study was to assess the effectiveness of an intervention--Quality English and Science Teaching (QuEST)--designed to develop the science knowledge and academic language of middle grades English language learners studying science in their second language and their English-proficient classmates. Ten sixth-grade science teachers in 5 middle schools in a large south Texas district participated in the study. For each teacher, 2 sections were randomly assigned to the intervention, Project QuEST, and 2 sections were randomly assigned to the district curriculum. The sample of students included English language learners, former English language learners, and fluent English-speaking students. Treatment effects were tested separately for science knowledge and vocabulary using a 3-level multilevel analysis of covariance (students nested within section, sections nested within teacher, and teacher) with the analogous pretest serving as the covariate. Analyses included fixed effects of treatment assignment and the covariate. Treatment effects were tested at the level of the section. Results indicated that posttest differences favoring the treatment group sections were statistically significant for both science knowledge and vocabulary. (Contains 5 tables, 1 figure and 7 footnotes.)