|
Activity Number:
|
150
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #306863 |
|
Title:
|
Bayes Methodology Accounting for Uncertainty of Commonality in 'Random Effects' in a Linear Mixed Model
|
|
Author(s):
|
Guofen Yan*+ and Joseph Sedransk
|
|
Companies:
|
University of Virginia and Case Western Reserve University
|
|
Address:
|
1942 Powell Creek Court, Charlottesville, VA, 22911,
|
|
Keywords:
|
hierarchical model ; partition
|
|
Abstract:
|
We describe and evaluate Bayesian methodology to improve inference for linear mixed effects models by embedding in a more general model the usual specification that the random effects come from a single normal distribution. If the usual specification is incorrect, standard inferences for the variances and random effects will be unsatisfactory. We demonstrate this analytically through a set of examples. The new methodology may be of special value when the random effects correspond to "small areas."
|
