There has been a controversy associated with low achievement scores of innovation schools. While the early dominant view in innovation school policy is that innovation schools reduce the achievement scores of students, recent empirical studies demonstrate that students in innovation schools gain more cognitive development than their peers in general schools. However, these findings have been limited to innovation schools in Gyeonggi Province among 17 cities and provinces implementing innovation school policy. In addition, these studies have not strictly controlled selection bias between innovation schools and general schools in estimating the effects of innovation school policy. Using propensity score to control selection bias between two groups and 3-level linear growth model, this study analyzes the effects of innovation schools in Busan to examine whether the previous findings are applicable to the effects of innovation schools in other regions. This study employs data from Busan Education Longitudinal Study(BELS), analyzing 2,708 students(331 innovation schools, 2,377 general schools) of 67 elementary schools(10 innovation schools, 57 general schools) in Busan. Result suggests that Busan innovation schools and general schools have no difference in cognitive development in Korean, English, and Mathematics. This leads us to reconsider the previous findings that innovation school policy has effects on students cognitive development.