A growing body of empirical research has provided provocative evidence that competition from private schools improves student achievement in neighboring public schools. However, this uniform conclusion has been based on fundamentally different empirical specifications. This study examines the importance of these different specifications by presenting new evidence on the relationship between public school quality and competition from private schools. This evidence is based on a unique data set that contains consistently defined high school graduation rates for the unified school districts in 18 states. The results indicate that empirical strategies which rely exclusively on ordinary least-squares (OLS) can lead to misleading inferences because of omitted variables bias and the simultaneous determination of the demand for private schools and public school quality. Nonetheless, two-stage least-squares (2SLS) estimates indicate that competition from private schools does have a positive and statistically significant impact on the high school graduation rates of neighboring public schools.