Nonparametric inference for interventional effects with multiple mediators
This work provides a means for researchers in statistics and causal inference to leverage modern statistical learning techniques in mediation analysis, though it is incremental as it builds on existing parametric estimators by extending them to nonparametric settings.
The authors tackled the problem of estimating interventional direct and indirect effects in mediation analysis by developing nonparametric inference methods that allow for flexible, machine learning-based estimation, establishing weak convergence results for closed-form confidence intervals and demonstrating multiple robustness properties with adequate small-sample performance in simulations.
Understanding the pathways whereby an intervention has an effect on an outcome is a common scientific goal. A rich body of literature provides various decompositions of the total intervention effect into pathway specific effects. Interventional direct and indirect effects provide one such decomposition. Existing estimators of these effects are based on parametric models with confidence interval estimation facilitated via the nonparametric bootstrap. We provide theory that allows for more flexible, possibly machine learning-based, estimation techniques to be considered. In particular, we establish weak convergence results that facilitate the construction of closed-form confidence intervals and hypothesis tests. Finally, we demonstrate multiple robustness properties of the proposed estimators. Simulations show that inference based on large-sample theory has adequate small-sample performance. Our work thus provides a means of leveraging modern statistical learning techniques in estimation of interventional mediation effects.