Stability of externalizing problem behaviors with onset in early childhood: a meta-analytic review

Abstract

A meta-analysis was conducted to examine the magnitude of stability of externalizing problem behaviors with onset before age 6 and the variables affecting the stability effect sizes. Gender difference in the stability was also investigated. Seventy empirical research reports, representing 12,111 non-referred children assessed before age 6 drawn from 72 independent aggregated samples and 27 pairs of matched gender samples, met inclusion criteria. Stability was coded as correlational effect sizes for the relationship between externalizing behaviors at Time 1 and Time 2. Results showed great variability in the weighted mean stability effect sizes ranging from.12 to .52 with most of the effect sizes around .30. Boys’ externalizing behaviors were more enduring than girls’ externalizing behaviors. The effects of informants and subtypes of externalizing behaviors were the most robust findings: the stability effect sizes were larger if Time 1 and Time 2 measured the same subtypes of externalizing behaviors and used the same type of informants. The stability of children’s externalizing behaviors decreased as time intervals between measurement points increased. Children assessed before age 3 and from low socioeconomic status (SES) and Caucasian backgrounds had less stable externalizing behaviors. Low SES had differential effects on boys and girls: externalizing behaviors were less stable for boys from low SES families than girls. The findings suggest that externalizing behaviors in young children are not as stable as those in school age children and the information on externalizing behaviors in early childhood alone is insufficient to predict later antisocial behaviors accurately.