|
Activity Number:
|
132
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #304412 |
|
Title:
|
Dangers of Transforming 'Noninformative' Prior Distributions
|
|
Author(s):
|
John W. Seaman, III*+ and John W. Seaman, II and James Stamey
|
|
Companies:
|
Baylor University and Baylor University and Baylor University
|
|
Address:
|
One Bear Place #97140, Waco, TX, 76798-7140,
|
|
Keywords:
|
Prior Distributions ; Bayesian methods ; Non-Informative Prior ; Induced Prior
|
|
Abstract:
|
"Non-Informative" priors are widely used in Bayesian inference. Often, priors are placed on parameters that are components of some function of interest. That function may, of course, have a distribution that is highly informative, in contrast to the joint prior placed on its arguments, resulting in unintended influence on the posterior for the function. This problem is not always recognized by users of "non-informative" priors. We consider several examples from the statistical literature ranging from priors for covariance matrices to priors used in surrogate marker assessment. We also offer guidelines to avoid problems with induced priors.
|
