We used data from a representative sample of public and private schools in Lagos state, Nigeria, to explore the effectiveness of low- and medium- cost private schools in producing English language outcomes. We described the scale of low- and medium-cost private schooling in low- and- middle income countries and provide descriptive statistics regarding the characteristics of students in private and public schools in Lagos. We found students in low and medium-cost private schools perform better on English language assessments in Lagos and this is particularly true for English as a Second Language students. Using a series of linear regression models and the Kitagawa-Oaxaca-Blinder decomposition, we found that private schools both attract students more likely to do well and are better at transforming these student endowments into better language outcomes. However, we found the composition of classrooms – specifically the share of students who speak English as a First Language – gender, wealth and age yielded significant results while attending a private school did not. While no model can fully capture the complexities of the various predictors of learning outcomes, our results suggested the private school learning gap is at least in part reliant on the provision of access to a different type of peer group.