BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:BSU Seminar: “Building Representative Matched Samples with Multi
 -valued Treatments in Large Observational Studies” - Professor Jose Zubi
 zarreta\, Harvard University
DTSTART:20190411T130000Z
DTEND:20190411T140000Z
UID:TALK120076@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:In observational studies of causal effects\, matching methods 
 are widely used to approximate the ideal study that would be conducted und
 er controlled experimentation. In this talk\, I will discuss new matching 
 methods that use tools from modern optimization to overcome five limitatio
 ns of standard matching approaches. In particular\, these new matching met
 hods (i) directly obtain flexible forms of covariate balance\, as specifie
 d before matching by the investigator\; (ii) produce self-weighting matche
 d samples that are representative of target populations by design\; and (i
 ii) handle multi-valued treatments without resorting to a generalization o
 f the propensity score. (iv) These methods can handle large data sets quic
 kly. (v) Unlike standard matching approaches\, with these new matching met
 hods\, usual estimators are root-n consistent under usual conditions. I wi
 ll discuss connections between matching and weighting. I will illustrate t
 he performance of these methods in a case study about the impact of a natu
 ral disaster on educational opportunity.
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR