BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Being Realistic About Unmeasured Biases in Observational Studies -
  Paul Rosenbaum (University of Pennsylvania)
DTSTART:20260120T160000Z
DTEND:20260120T170000Z
UID:TALK241789@talks.cam.ac.uk
DESCRIPTION:Observational studies of the effects caused by treatments are 
 always subject to the concern that an&nbsp\;ostensible treatment effect ma
 y reflect a bias in treatment assignment\, rather than an effect actually&
 nbsp\;caused by the treatment. The degree of legitimate concern is strongl
 y affected by simple decisions that&nbsp\;an investigator makes during the
  design and analysis of an observational study. Poor choices lead to&nbsp\
 ;heightened concern\; that is\, poor choices make a study sensitive to sma
 ll unmeasured biases where&nbsp\;better choices would correctly report ins
 ensitivity to larger biases. Indeed\, perhaps surprisingly\,&nbsp\;unambig
 uous evidence of the presence of unmeasured bias may increase insensitivit
 y to unmeasured&nbsp\;bias. These issues are discussed with the aid of som
 e theory and a simple example of an observational&nbsp\;study.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR