Prior research suggesting that longer bilingual experience benefits inhibitory control and monitoring has been criticized for a lack of control over confounding variables. We addressed this issue by using a propensity-score matching procedure that enabled us to match early and late bilinguals on 18 confounding variables--for example, demographic characteristics, immigration status, fitness, extracurricular training, motivation, and emotionality--that have been shown to influence cognitive control. Before early and late bilinguals were matched (N = 196), we found early active bilingual advantages in flanker effects (in accuracy), global accuracy, and sensitivity (d') on the Attention Network Test for Interaction and Vigilance and global accuracy on the saccade task. After matching (n = 113), many of the early active bilingual advantages that had been identified before matching were either attenuated or disappeared. However, we observed that early active bilingual advantages in flanker effects (in response time) were strengthened after matching. These results stress robust early active bilingual advantages in inhibitory control and highlight the importance of matching language groups on nonlinguistic covariates.