Evaluating (weighted) dynamic treatment effects by double machine learning

Summary We consider evaluating the causal effects of dynamic treatments, i.e., mul-tiple treatment sequences in various periods, based on double machine learning to control for observed, time-varying covariates a data-driven way under selection-on-observables assumption. To this end, we make use so-called Neyman-orthogonal score functions, which imply robustness effect estimation moderate (local) misspecifications outcome and models. This property permits approximating models by even high-dimensional covariates. In addition total population, weighted that assessing specific subgroups, e.g., among those treated first period. demonstrate estimators are asymptotically normal $\sqrt{n}$-consistent regularity conditions investigate their finite sample properties simulation study. Finally, apply methods Job Corps