BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Stable Weights that Balance Covariates for Causal Inference and Es
 timation with Incomplete Outcome Data - José Zubizarreta\, Columbia Unive
 rsity
DTSTART:20150409T150000Z
DTEND:20150409T160000Z
UID:TALK58822@talks.cam.ac.uk
CONTACT:20082
DESCRIPTION:Weighting methods that adjust for observed covariates\, such a
 s inverse probability weighting\, are widely used for causal inference and
  estimation with incomplete outcome data. Part of the appeal of such metho
 ds is that one set of weights can be used to estimate a range of treatment
  effects based on different outcomes\, or a variety of population means fo
 r several variables. However\, this appeal can be diminished in practice b
 y the instability of the estimated weights and by the difficulty of adequa
 tely adjusting for observed covariates in some settings. To address these 
 limitations\, this paper presents a new weighting method that finds the we
 ights of minimum variance that adjust or balance the empirical distributio
 n of the observed covariates up to levels prespecified by the researcher. 
 This method allows the researcher to balance very precisely the means of t
 he observed covariates and other features of their marginal and joint dist
 ributions\, such as variances and correlations and also\, for example\, th
 e quantiles of interactions of pairs and triples of observed covariates\, 
 thus balancing entire two- and three-way marginals. Since the weighting me
 thod is based on a well-defined convex optimization problem\, duality theo
 ry provides insight into the behavior of the variance of the optimal weigh
 ts in relation to the level of covariate balance adjustment\, answering th
 e question\, how much does tightening a balance constraint increases the v
 ariance of the weights? Also\, the weighting method runs in polynomial tim
 e so relatively large data sets can be handled quickly. An implementation 
 of the method is provided in the new package sbw for R. This paper shows s
 ome theoretical properties of the resulting weights and illustrates their 
 use by analyzing both a real data set and a simulated example.
LOCATION:MR11\,  Centre for Mathematical Sciences\, Wilberforce Road\, Cam
 bridge
END:VEVENT
END:VCALENDAR