A0423
Title: Local instrumental variable curves without positivity assumptions
Authors: Luke Keele - University of Pennsylvania (United States) [presenting]
Abstract: Instrumental variables have become a popular study design for the estimation of treatment effects in the presence of unobserved confounders. In the canonical instrumental variables design, the instrument is a binary variable, and most extant methods are tailored to this context. In many settings, however, the instrument is a continuous measure. While standard estimation methods can be applied to continuous instruments, these methods require strong functional form assumptions. Recent work has developed more flexible methods for the estimation when the instrument is continuous. However, these methods require an assumption known as positivity that is unlikely to hold for many applications. Doubly robust estimators are developed for continuous instruments that do not require a positivity assumption. The methods use a stochastic dynamic intervention framework that considers a range of intervention distributions absolutely continuous with respect to the observed distribution of the instrument. Empirical process theory and sample splitting are used to derive asymptotic properties of an estimator under weak conditions. Estimation methods also allow for the use of nonparametric estimators for the nuisance functions. The methods are evaluated via simulation and demonstrate their feasibility using an application on the effectiveness of surgery for specific emergency conditions.
