The impact of part-time employment on adolescent functioning remains unclear because most studies fail to adequately control for differential selection into the workplace. The present study reanalyzes data from L. Steinberg, S. Fegley, and S. M. Dornbusch (1993) using multiple imputation, which minimizes bias in effect size estimation, and 2 types of propensity score matching, to account for selection effects. In this sample (N = 1,792; Grades 10-11, M = 16.26), youth who begin working more than 20 hr per week evince declines in school engagement and increases in substance use and delinquency compared with youth who remain unemployed. Conversely, working 20 hr or less a week has negligible effects, positive or negative, on academic, psychological, or behavioral outcomes. (Contains 4 tables and 4 figures.)