| Literature DB >> 24409111 |
Abstract
We show that assumptions that are sufficient for estimating an average treatment effect in randomized trials with non-compliance restrict the subgroup means for always takers, compliers, defiers and never takers to a two dimensional linear subspace of a four dimensional space. Implications and specials cases are examplified.Entities:
Keywords: Average treatment effect; Complier average treatment effect; Instrumental variables; Principal Stratification
Year: 2014 PMID： 24409111 PMCID： PMC3881309 DOI： 10.1016/j.stamet.2013.07.003
Source DB: PubMed Journal: Stat Methodol ISSN： 1572-3127