Children growing up in disadvantaged neighborhoods perform much worse in school compared to children growing up under more favorable circumstances. We examine a program ("Coaching for Teaching") targeted at ten lower secondary schools in Sweden's most disadvantaged city districts. The program's aim was to enhance student performance by strengthening the teachers' professional development. Using a difference-in-differences design we estimate effects on several outcomes. We find a positive and statistically significant impact on student performance on standardized tests in English. Estimates for several other outcomes are also positive, but not statistically significant. For test scores in Swedish there is no improvement. An analysis of a survey among pupils supports the idea that the intervention enhanced the teaching and the classroom climate.