Methods: This study uses data from the Early Childhood Longitudinal Survey–Kindergarten Class of 1998-1999 (ECLS-K). The ECLS-K is a nationally representative cohort of children drawn from approximately 1,000 kindergarten programs. The data were collected via one-on-one assessment of children and parent and teacher reports using standardized questionnaires. In total, the analytic sample comprised 17,212 children over three waves of data collection. Our analytic plan consisted of four steps. First, we employed finite mixture analysis to identify homogeneous subgroups in the data based on indicators of self-control. Second, we validated these classes using a range of external covariates using one-way analysis of variance (ANOVA) and chi-square tests. Third, validation analyses were extended by employing polytomous logistic regression to predict membership in latent classes. Finally, structural equation modeling was used to examine key relationships between variables identified in polytomous logistic regression.
Results: Finite mixture analysis identified five classes consisting of a low impairment, teacher report (n = 5,047, 29.3%), low impairment, parent and teacher report (n = 2,889, 16.8%), moderate impairment, teacher report (n = 5,267, 30.6%), moderate impairment, parent report (n = 2,415, 14.0%), and severely impaired subgroups (n = 1,594, 9.3%). Parental stress, externalizing behaviors, and interpersonal skills deficits were important differentiators across class memberships. The severely impaired subgroup possessed an array of neurocognitive deficits, behavioral problems, and learning difficulties relative to other classes. Structural equation modeling indicated that learning problems at time one did not have a significant path association with self-control deficits at time two; however, its effect on interpersonal deficits at time two was significant (â = .30, p < .001). Interpersonal deficits at time two was also significantly associated with self-control deficits at time two with a large effect size (â = .69, p < .001). This relationship suggests that the effect of learning problems is fully mediated by interpersonal deficits.
Implications for Policy and Practice: These results not only add to an impressive knowledge base indicating that problems with impulse control, self-regulation, and behavioral inhibition are strong predictors of childhood aggression and antisocial behavior but also highlight the significance of intervening with early learning difficulties as a means toward preventing externalizing behavior problems in the future. Findings suggest that social workers in school setting can potentially identify mechanisms that can be targeted for prevention services early in the life-course.