A number of studies have shown that childhood speech impairments such as stuttering are associated with lower test scores and educational attainment. However, it is unclear whether this result is causal in nature or whether it can be explained by difficult-to-measure heterogeneity at the community, family, or individual level. Using data from the National Longitudinal Study of Adolescent Health and ordinary least squares, we show that stuttering is negatively associated with high school grades, the probability of high school graduation, and the probability of college attendance. However, empirical specifications with family fixed effects or controls for learning disabilities such as Attention Deficit Hyperactivity Disorder suggest that these associations can, in large part, be explained by difficult-to-measure heterogeneity.