No Robust Evidence for Cross-Lagged Effects Between Self-Esteem and Externalizing Problems: A Longitudinal Meta-Analysis
摘要
The longitudinal association between self-esteem and externalizing problems has been well-documented, yet the direction of this relationship remains inconsistent across studies. This meta-analysis synthesized longitudinal studies on their bidirectional links, aiming to explore whether self-esteem predicted subsequent externalizing problems or vice versa, and to identify sources of heterogeneity by examining moderating effects. A comprehensive search of six databases, followed by title/abstract screening and application of inclusion/exclusion criteria, yielded a total of 27 articles (25 independent samples) involving 38,489 participants (49.54% female; mean age 12.04 years). Bivariate correlation analyses revealed weak negative associations between self-esteem and externalizing problems over time (r = − 0.121, 95% CI = [ − 0.167, − 0.075], p < 0.001; r = − 0.092, 95% CI = [− 0.143, − 0.041], p < 0.001). However, once baseline levels were controlled for in the cross-lagged panel model, neither prior self-esteem predicted later externalizing problems (B = − 0.016, p = 0.438), nor did prior externalizing problems predict later self-esteem (B = 0.014, p = 0.479). These non-significant reciprocal effects remained robustly consistent across all demographic and methodological moderators. These findings help clarify the nature of the longitudinal relationship between self-esteem and externalizing problems. Educators should exercise caution in targeting self-esteem enhancement as an intervention for externalizing problems, nor should they over-rely on reducing externalizing behaviors as the primary means of promoting adolescent self-esteem.