We investigate the effect of attending selective secondary schools belonging to the "Bicentenario" schools program in Chile, a free education option that is intended to give vulnerable students everywhere in Chile a unique educational opportunity, in the period from 2011 to 2014. By using propensity score methods, we find that attending a "Bicentenario" school improves students' performance by a range of 0.35 to 0.23 SD in language and 0.5 to 0.35 SD in math. Also, we show that a proportion of this effect is due to the outstanding performance of the new "Bicentenario" schools as opposed to the restructured version of the program. We prove the robustness of the previous results through falsification, changes-in-changes, and a more demanding control group. This paper adds new evidence to analyze the effect of selective schools in developing countries like Chile, where the quality of public education is low.