This paper examines the effect of attending elite private high school on college placement using admissions data from the most selective high school in a large metropolitan area. To overcome omitted variable bias, we limit the sample to admitted applicants and control directly for the scores assigned by admissions based on in-depth analyses of the applicants and their families. In addition, we control for a wide set of covariates including student and family characteristics and entrance exam scores. Results indicate that attending selective private high school rather than other public and private high schools causes students to attend more selective universities. Effects are driven by gains for girls and students from lower-income neighborhoods. (Contains 5 tables.)