Talk on causal inference with David M. Drukker
Most of the literature on causal inference assumes that the treatment assignment process is ignorable, after conditioning on covariates. This assumption is also known as unconfounded treatment assignment, exogenous treatment assignment and selection on observables. This talk discusses estimators for the average treatment effect (ATE) and the average treatment effect on the treated (ATET) when treatment assignment is not ignorable.
Similarly, many researchers assume that the process of who is in the sample and who is not is ignorable, after conditioning on covariates. This talk also discusses estimators for the ATE and the ATET that accommodate non ignorable sample-selection, which is alsoknow as nonignorable loss to follow up, endogenous sample selection or Heckman sample selection. All of the discussed estimators are available in Stata and the talk discusses the estimators using replicable Stata examples.
Instructor: David M. Drukker, Ph.D., Executive Director of Econometrics, StataCorp
Please register at https://goo.gl/eBTpNkContact person: Nicola Orsini