We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: Contrast Specific Propensity Scores

Abstract: Basic propensity score methodology is designed to balance multivariate pre-treatment covariates when comparing one active treatment with one control treatment. Practical settings often involve comparing more than two treatments, where more complicated contrasts than the basic treatment-control one,(1,-1), are relevant. Here, we propose the use of contrast-specific propensity scores (CSPS). CSPS allow the creation of treatment groups of units that are balanced with respect to bifurcations of the specified contrasts and the multivariate space spanned by them.
Subjects: Methodology (stat.ME)
Journal reference: Biostatistics & Epidemiology 2021
DOI: 10.1080/24709360.2021.1936421
Cite as: arXiv:2007.01253 [stat.ME]
  (or arXiv:2007.01253v1 [stat.ME] for this version)

Submission history

From: Shasha Han [view email]
[v1] Thu, 2 Jul 2020 17:02:52 GMT (162kb)

Link back to: arXiv, form interface, contact.