Empirical research supporting the contention that insecure attachment is related to internalizing behaviors has been inconsistent. Across 60 studies including 5,236 families, we found a significant, small to medium effect size linking insecure attachment and internalizing behavior (observed d = .37, 95% CI [0.27, 0.46]; adjusted d = .19, 95% CI [0.09, 0.29]). Several moderator variables were associated with differences in effect size, including concurrent externalizing behavior, gender, how the disorganized category was treated, observation versus questionnaire measures of internalizing behavior, age of attachment assessment, time elapsed between attachment and internalizing measure, and year of publication. The association between avoidant attachment and internalizing behavior was also significant and small to moderate (d = .29, 95% CI [0.12, 0.45]). The effect sizes comparing resistant to secure attachment and resistant to avoidant attachment were not significant. In 20 studies with 2,679 families, we found a small effect size linking disorganized attachment and internalizing behavior (observed d = .20, 95% CI [0.09, 0.31]); however, the effect size was not significant when adjusted for probable publication bias (d = .12, 95% CI [-0.02, 0.23]). The existing literature supports the general notion that insecure attachment relationships in early life, particularly avoidant attachment, are associated with subsequent internalizing behaviors, although effect sizes are not strong.

The central premise of attachment theory is that the security of the early child-parent bond is reflected in the child's interpersonal relationships across the life span. This meta-analysis was based on 63 studies that reported correlations between child-parent attachment and children's peer relations. The overall effect size (ES) for child-mother attachment was in the small-to-moderate range and was quite homogeneous. ESs were similar in studies that featured the Strange Situation and Q-sort methods. The effects were larger for peer relations in middle childhood and adolescence than for peer relations in early childhood. ESs were also higher for studies that focused on children's close friendships rather than on relations with other peers. Gender and cultural differences in ESs were minimal. The results for the few studies on father-child attachment were inconclusive.

No abstract