The purpose of this study was to apply a propensity score method that could control for selection bias at both the student-level and school-level in an investigation of the causal effect linking participation in school-based extracurricular activity (SBEA) to adolescents' learning outcomes. The data for this study were drawn from the Education Longitudinal Study of 2002 (ELS: 2002) data set. The final sample comprised 12,247 10th graders; 6,026 (49.20%) were males. A propensity score method incorporating marginal mean weighting through stratification was implemented to analyze the data. Results showed that 10th graders who had proper intensity of participation in SBEA (6-15 hours a week) slightly outperformed peers who did not participate in SBEA on the performance of mathematics achievement in 12th grade. Regarding the link between SBEA participation and adolescents' long-term learning outcomes, results indicated 10th graders in 2002 with low to moderate levels of intensity (i.e., 1-15 hours) were more likely to achieve higher education credentials by the year 2012 when compared to non-participating peers.