Over the past decade, several state and federal policies have directed schools to mainstream English Language Learner (ELL) students into English-only instruction classrooms. While there is mixed evidence on the effects of these immersion policies on the ELL students, research examining potential peer effects on their non-ELL classmates is non-existent. This paper begins to fill in this gap by using a nationally representative longitudinal sample of children in early elementary grades. Results indicate that having an ELL classmate during kindergarten and first grade is associated with lower test score gains in reading but not necessarily in math for non-ELL students whose primary language is English, controlling for unobserved fixed school characteristics as well as individual characteristics. The negative peer effects on reading test score gain are sensitive to the frequency of within classroom ability grouping usage, gender, and household income level of non-ELL children. (Contains 5 tables and 1 figure.)