Pathways to Resilience: Novel approaches to understand Environmental, Economic, and Social Determinants of Child Development and Community Recovery in the Wake of Disaster

Abstract

Social phenomena are difficult to define and quantify by their nature, including multidimensional nature of exposures, heterogeneous effects across individuals, and outcomes unfold over time rather than at a single moment. Conventional approach that simplifies this complexity often obscure the very patters that matter most for policy. This dissertation applies methods designed to embrace such complexity rather than reduce it and examined how neighborhood environments, economic recourses, and community social capital shape child development and community resilience across three related studies. In Chapter 1, we applied a multi-phase modified treatment policy (MTP) framework to examine the association between neighborhood opportunity, measured by the Child Opportunity Index (COI), and child externalizing behaviors in the Adolescent Brain Cognitive Development (ABCD) study (n=7,063). By analyzing individual neighborhood indicators and accounting for co-movement among them, we identified which domains and specific indicators drive neighborhood effects, rather than producing a single summary estimate. We found that higher neighborhood opportunity was associated with lower externalizing behaviors, with the effects driven by the health and economic domains. Early childhood education accessibility emerged as a particularly important indicator. In Chapter 2, we used Bayesian Causal Synthesis (BCS) to examine heterogenous effect of unconditional cash transfers on children’s developmental outcomes in the Baby’s First Years (BFY) randomized controlled trial (n=853). While average effects were largely null, consistent with prior research, we identified meaningful heterogeneity across subgroups. Families facing specific structural barriers, including lack of paid maternity leave showed the greatest developmental benefits, suggesting that cash transfers address specific recourses gaps rather than providing uniform benefits. Lastly, in Aim 3, we examined whether pre-disaster social capital predicts neighborhood physical recovery trajectories following the 2011 Great East Japan Earthquake and Tsunami, using satellite imagery from the Iwanuma study (n=4,957). By modeling recovery as a longitudinal trajectory across 9 years rather than a single endpoint, we characterized 4 trajectory cluster membership, and identified community-level social capital may predict differential recovery patterns. Across all three aims, methods that reflect real-world complexity to some extent, yielded richer and more actionable insights that simpler approach would have allowed. These findings suggest that complex social phenomena require multi-dimensional policy responses, and that upstream interventions at the neighborhood and community level hold promise for supporting child development and community resilience even in the face of adversity.