Abstract:
|
Causal inference addresses the fundamental question of how changing the level of a cause or treatment can affect a subsequent outcome. Many analyses applied in behavioral, social, biomedical, and other fields of science are aimed at understanding causal relationships. Over the last thirty years, studies on causal inference using potential outcomes primarily focused on methods with a binary treatment. In contrast, studies with a continuous treatment have been a relatively under-explored problem. In an observational study where the treatment is continuous, the potential outcomes are an uncountably infinite set indexed by treatment dose. To deal with this problem, we parameterize this unobservable set as a linear combination of a finite number of basis functions whose coefficients vary across units. This leads to new techniques for estimating the population average dose-response function (ADRF). In this talk, we describe these new methods and introduce the R package causaldrf.
|