Peers play a significant role in affecting students' academic, social, psychological and labor market outcomes. Although there is an extensive literature on peer influence on adolescents' academic outcomes, existing literature failed to examine in-depth peer influence processes, and there is still very little cross-national comparative and long-term perspective research on peer influence on academic outcomes. In this dissertation, I conducted three empirical studies to address each of these three above-mentioned limitations in the peer influence literature, using nationally representative, longitudinal datasets. I first explored peer influence at different levels and the relationship between the effects of different peer group attributes. I found that the average GPA of students' friends have significantly positive effects on the academic outcomes of high school students of all races/ethnicities and that the effect of the racial heterogeneity of friends on student's academic outcomes is mediated in part through the effect of the average GPA of students' friends. I then conducted a cross-national comparative study of peer influence on eighth-graders in the U.S. and three East Asian countries/regions. Results from my HLM analyses show that school-level peer achievement is positively and significantly associated with eighth-graders' test scores in all four countries and that there is very slight US Ì¢ âÂ' East Asian difference in peer effect at the school level. At the individual level, results from analyses of the pooled data using interactions by region (East Asia versus the U.S.) show that friends' attitude on doing well in Math/Science is less negatively and significantly related to East Asian students' achievement than with that of the U.S. students. Yet further analyses using pooled data with interactions by country (with the U.S. as reference group) and with country-specific data reveal that the differences are more country-specific than region-specific. Finally I examined the long-term influence of high school friends on 10th-graders' higher educational attainment process. Results from my logistic regressions show that friends' pro-education attitude at 10th-grade increases the probability that respondents expected to need a college degree for their jobs at the age of thirty, and that they planned to go to college right after high school and applied to at least one college at 12th-grade. I also found that the number of friends who were to attend a 4-year college at 12th-grade increases respondents' likelihood of going to a 2-year or 4-year college (versus no PSE) and to eventually attain a Bachelor's degree as of 2000, eight years after their high school graduation.