B0499
Title: A coarsened data perspective of counterfactual survival analysis
Authors: Benjamin Baer - University of St Andrews (United Kingdom) [presenting]
Abstract: The purpose is to introduce coarsening and the identification of full data functionals as observed (or coarsened) data functionals. Examples of coarsening mechanisms include measurement error, right censoring, causal selection, or combinations thereof. An assumption known as coarsening at random is introduced alongside several examples. After reviewing semi- and non-parametric estimators, influence functions and efficiency bounds are reviewed. The general theory of coarsening is then applied to the causal survival problem. Sequential and non-sequential coarsening at random is characterized, the class of influence functions is provided, and then semi- and non-parametric estimation is discussed.
