Adversity in adolescence: Modeling risk for different purposes

Abstract

Policymakers and practitioners are increasingly leveraging research on the links between adversity and wellbeing in childhood and adolescence. However, conceptualizations and analytical approaches focused on these connections vary across disciplines, with implications for empirical results, interpretation of findings, and how those findings guide policy and practice. This article demonstrates the importance of researchers matching study aims to analytic approach when modeling relations between adversity and problems signifying poor outcomes. We modeled the Philadelphia 2019 Youth Risk Behavior Survey using three approaches to estimate relations among indicators of adversity and problems along with demographic factors in adolescence (N = 1,217). The analytic approaches involved two variable-centered approaches (cumulative risk score and individual main-effects modeling) and one person-centered (latent class analysis) approach. Findings across the three approaches varied in specificity, strength of causal inference, and interpretation. Results demonstrate the importance of congruity between analytical approaches, research questions and study conceptualization to ensure that empirical findings are accurately interpreted and translated into policy and practice.