JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Activity Details


CE_14C Mon, 7/30/2012, 8:30 AM - 5:00 PM HQ-Indigo D
Propensity Score Matching in R — Continuing Education Course
ASA , Section on Statistics in Epidemiology
Instructor(s): Ben Hansen, University of Michigan
In an observational study, the researcher attempts to understand effects of an intervention upon people, without controlling which of them receive the intervention. Absent random assignment, findings of a difference between groups receiving and not receiving the intervention will inevitably be equivocal, explicable as consequences either of the intervention or of preexisting differences between the groups. Propensity score matching aims to strip observed covariates of their ability to confound a comparison, in this way disambiguating findings the study might produce. It is particularly beneficial for nonrandomized comparisons in which there are many measured, potentially important baseline differences, but nonetheless relatively simple and widely accessible analytic procedures are desired. Rich diagnostics are available both to guide the selection of the match and to enrich the final presentation of results. Although it leaves open the possibility of confounding in terms of _unobserved_ variables, its handling of observed covariates is sufficiently compelling to make it a quite popular tool in observational research -- the 1983 paper introducing it now has, according to Google, nearly 7,000 citations. This workshop introduces propensity score matching and related ideas conceptually and methodologically, with complementary programming exercises using R and its "optmatch" and "RItools" add-on packages.



2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.