Evaluating Interventions and High-Stakes Decisions: Causal and Robust Modeling Approaches

This session centers on statistical methods that support evidence-based decisions in programs and high-stakes settings. Talks cover causal evaluation under continuous or multi-phase treatments (dose–response and cumulative effects with noncompliance), robust modeling for fractional outcomes with boundary inflation, and measurement models that improve inference when naive metrics can mislead. Together, the papers advance practical tools for assessing what works, for whom, and under what conditions—aligned with "community in action."