Beyond Binary: Using Propensity Scores to Account for Varying Levels of Program Participation in Randomized Controlled Trials.

Elizabeth A. Stuart, Siri Warkentien, Booil JoSociety 2011

Abstract

• Originally proposed by Rosenbaum and Rubin (1983). • Type of matching method used with observational data to mimic randomized experiment with respect to the observed covariates. • Treatment and control groups matched without using information on the outcome variables (Stuart, 2010). • Researcher can conclude that any difference post-treatment was caused by treatment. • Main benefit is reduction of selection bias due to observed covariates (Stuart & Rubin, 2008). • Recently extended beyond the traditional binary treatment condition and into multiple doses and continuous dose levels (Imai and van Dyk, 2004; Joffe and Rosenbaum,1999; and Imbens, 2000).
BibTeX
@inproceedings{stuart2011beyond,
  author = {Elizabeth A. Stuart and Siri Warkentien and Booil Jo},
  title = {{Beyond Binary: Using Propensity Scores to Account for Varying Levels of Program Participation in Randomized Controlled Trials.}},
  booktitle = {Society},
  year = {2011},
}

Loading related content...

Figure