On nearly assumption-free tests of nominal confidence interval coverage for causal parameters estimated by machine learning

Lin Liu, Rajarshi Mukherjee, James Robins

Abstract: For many causal effect parameters of interest, doubly robust machine learning (DRML) estimators are the state-of-the-art, incorporating the good prediction performance of machine learning; the decreased bias of doubly robust estimators; and the analytic tractability and bias reduction of sample splitting with cross-fitting. Nonetheless, even in the absence of confounding by unmeasured factors, the nominal Wald confidence interval may still undercover even in large samples, because the bias of may be of the same or even larger order than its standard error of order .