``Identification Analysis for Randomized Experiments with Noncompliance and ``Truncation-by-Death''''

 

  Abstract

Zhang and Rubin (2003) derives the bounds on the average treatment effect in randomized experiments with ``truncation-by-death.'' Imai (Forthcoming) proves that these bounds are sharp and derives the sharp bounds on the quantile treatment effect. In this paper, I extend their results to randomized experiments with noncompliance and derive the sharp bounds on the treatment effects for compliers. (Last Revised August 2007)

© Kosuke Imai
  Last modified: Fri Aug 24 13:43:41 EDT 2007