Beyond the ATE: Estimating the causal effects of binary, categorical, continuous, and multivariate exposures in R
Nick Williams, Kara Rudolph, and Iván Díaz
Modified treatment policies (MTPs) are a class of interventions that generalize static and dynamic interventions for categorical, continuous, and multivariate exposures. MTPs are hypothetical interventions where the post-intervention is defined as a modification of the natural value of the exposure that can depend on the unit’s history. This short course will introduce the lmtp R package for estimating the causal effects of MTPs in both point-treatment and longitudinal studies. We will discuss identification of MTPs, estimation with a targeted minimum-loss based estimator and a sequentially doubly-robust estimator, and provide guidance on estimator choice.
We will walk participants through applying these methods in point treatment and longitudinal settings and for each of the following treatment types: 1) static interventions (e.g., in estimating the ATE), 2) dynamic interventions, 3) continuous exposures, and 3) multivariate exposures.