PLACE: 319 Snedecor

SPEAKER: Professor Kathryn Chaloner
                   School of Statistics, University of Minnesota

TITLE: Some Developments in Practical Bayesian Design
 

ABSTRACT:

A Bayesian approach to experiment design will be reviewed.  Prior information
is often used in experimental design and a Bayesian approach to  design can
formally account for the prior uncertainty in parameter values.

Some results on Bayesian design for more than one prior distribution will
then be described. One situation is where one prior distribution will be used
for inference and another for design.  Another situation is where a collection of
rior distributions is available and the goal of the experiment will be to design
an experiment so that consensus will be reached with high probability.
The example is that of a clinical trial where there is diversity of opinion at the
design stage.
 

COFFEE:  3:45 p.m., 104 Snedecor Hall