BriDGE the gap: Improving behavioral research by integrating DAGs and GAMs into experiments
RCTs establish that an intervention works; they are rarely designed to reveal why.
Understanding the mechanisms through which behavioural interventions work remains a critical challenge in behavioural science. Randomised controlled trials provide reliable evidence for intervention efficacy, but they are seldom designed to reveal the underlying causal pathways that drive observed outcomes. BriDGE combines directed acyclic graphs (DAGs), causal discovery algorithms, and generalised additive models (GAMs) to strengthen mechanistic insight in behavioural applications. The workflow spans DAG-based hypothesis formulation, modelling of nonlinear relationships with GAMs, and detailed mediation analysis, with bootstrapping and sensitivity checks to ensure robust detection of both direct and indirect effects. An accompanying open-source R package implements the full workflow to support adoption and reproducibility.
Read Article