Over the last decade, the federal government has directed schools to provide educational instruction for students with special needs in general education settings to the extent possible. While there is mixed evidence on the effects of these inclusion policies on the students with special needs, research examining potential spillovers of inclusion on non-disabled classmates has been scarce. There is particularly little research on the effects of inclusion policies on classmates during early elementary grades. This paper begins to fill in this gap by using a nationally representative, longitudinal survey of kindergartners. Cross-sectional results suggest that having a classmate with an emotional problem decreases reading and math scores at the end of kindergarten and first grade by over 10 percent of a standard deviation, which is one-third to one-half of the minority test score gap. To control for nonrandom sorting of students to schools, as well as students to classrooms, this paper uses school-level and then student-level fixed effects. Results from the preferred empirical models suggest a decrease of approximately 5 percent of a standard deviation in math and reading scores, though the reading results are less robust. The results also indicate moderate racial and gender differences in the effects. © 2010 by the Association for Public Policy Analysis and Management.