Co-Principal Investigator: Morgan, Paul

Purpose: The purpose of this project is to analyze secondary data from two extant longitudinal studies to examine the mechanisms of the development of comorbidity between externalizing and internalizing disorders across the school-age period (5-6 to 12-13 years of age). Although externalizing and internalizing disorders are distinct forms of behavioral maladjustment, rates of comorbidity are substantial, especially among high-risk populations. It remains unclear, however, for whom, how, and why externalizing and internalizing disorders co-develop, how these two domains of behavioral disorders and other domains of development (e.g., academic, cognitive, social) transact over time, and whether these transactional processes operate differently across developmental stages and across population subgroups (e.g., gender, race/ethnicity). By evaluating competing theoretical hypotheses regarding the mechanisms of externalizing/internalizing comorbidity, the project aims to advance the field's limited knowledge base and help inform intervention efforts to serve children with or at risk for multiple functional difficulties or disorders.

Project Activities: The research team will conduct secondary data analyses using data from the Early Childhood Longitudinal Study- Kindergarten Cohort of 2010-11 (ECLS-K: 2011) and the Family Life Project (FLP).

Products: The research team will produce evidence on the mechanisms of development of externalizing-internalizing comorbidity across the school-age period. The project will also result in peer-reviewed publications and presentations, and additional dissemination products that reach education stakeholders such as practitioners and policymakers.

Structured Abstract

Setting: This project will use secondary data from ECLS-K: 2011 and FLP. The ECLSK: 2011 is a nationally representative study of children who attended kindergarten during the 2010-11 school year. The FLP is a 12-year multi-site longitudinal study of children recruited at birth in 2003-04 from six low-income, rural counties in North Carolina and Pennsylvania.

Population/Sample: The ECLS-K: 2011 includes a nationally representative sample of approximately 18,000 kindergarteners enrolled in about 970 schools in 2010-11 and followed through the fifth grade. The FLP includes a sample of 1,292 children representative of the population in six counties in North Carolina and Pennsylvania who were followed from birth to the seventh grade. The FLP oversampled low-income families in both states and African American families in North Carolina. Of the 1,292 children who were in the first wave of data collection, 57% were African American and 42.5% were non-Hispanic White.

Malleable Factors: This project will examine externalizing and internalizing behaviors as factors that Influence – and are influenced by – each other and by other domains of development such as academic (e.g., academic achievement, retention) and social experiences (e.g., relationships with teachers and peers) in school. Early childhood individual and environmental factors will also be examined as potential antecedents of the development of externalizing and internalizing disorders (e.g., parenting, preschool quality, temperament).

Research Design and Methods: The ECLS-K: 2011 collected data in the fall of 2010, fall and spring of 2011 and 2012, and spring of 2013, 2014, and 2015. The FLP conducted parent interviews and home visits at 2, 6, 15, 24, 36, 48, 69, and 90 months of child age, and completed teacher interviews and school visits annually starting preschool.

Key Measures: Externalizing and internalizing problems were measured by teacher reports using the Social Skills Rating System in the ECLS-K and the Strengths and Difficulties Questionnaire in the FLP. Other key measures include academic achievement, executive functions and self-regulatory skills, teacher-child relationships, peer relationships, early childhood temperamental traits, and socio-demographic characteristics.

Data Analytic Strategy: To adequately model within- and between-person levels of associations between externalizing and internalizing behaviors, the research team will use longitudinal modeling approaches that combine latent curve modeling and cross-lagged panel modeling. These include random-intercept cross-lagged panel modeling and its extended variant, autoregressive latent trajectory modeling with structural residuals.