A central tenet of Bowlby's attachment theory is that early child-caregiver attachment is reflected in the quality of the child's interpersonal relationships throughout life. Schneider, Atkinson, and Tardif (2001) conducted a meta-analysis of studies conducted up to 1998 to corroborate that contention. They found a significant but small to moderate effect size (r = .20). Their finding that studies of friendship bonds had higher effect sizes than studies of other interpersonal relationships has important theoretical ramifications. The present brief report is a meta-analysis that covers research conducted for the same purpose since 1998. The sample consists of 44 studies with a total of 8505 participants. The overall effect size r of .19 (adjusted r = .12; 95% confidence interval, .08-.17) in the current study was similar in magnitude to the effect size reported in the 2001 meta-analysis, documenting consistency in the predictive power of attachment theory. However, we failed to replicate the moderating effect of friendship. One possible explanation for these findings is that the friendships of school-age children and adolescents no longer invoke very high levels of intimacy. Effect sizes are higher in studies conducted outside North America than in U.S.--and Canada-based studies.