Attachment in institutionalized children: a review and meta-analysis

Abstract

In this article we review the literature on attachment patterns in institutionalized children and then perform a meta-analysis on data from 10 attachment studies involving 399 children in institutional settings. We computed the overall attachment distribution of secure, insecure, and disorganized rates and explored the effect of a set of moderating variables (i.e., country of institutionalization, attachment assessment procedure, age at entry, and age at assessment). To overcome bias related to the small number of studies, we conducted both classical and Bayesian meta-analysis and obtained comparable results. Distribution of children's attachment patterns was: 18% secure, 28% insecure, and 54% disorganized/cannot classify. Compared to their family-reared peers, children living in an institution were found to be at greater risk for insecure and disorganized attachment, with a similar medium effect size for both distributions (d=0.77 and d=0.76, respectively). The following moderating variables were associated with insecure attachment: representational assessment procedures (d=0.63) and Eastern European countries of origin (d=1.13). Moderators for disorganized attachment were: Eastern European countries of origin (d=1.12), age at institution entry before the first birthday (d=0.93), and age at assessment under three years of age (d=0.91). Implications for child development and policies are discussed.

Keywords: Attachment; Bayesian approach; Institution; Meta-analysis.