Activity Number:
|
286
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
Sponsor:
|
ENAR
|
Abstract - #309773 |
Title:
|
Testing Random Effects in the Linear Mixed Model Using Bayes Factors
|
Author(s):
|
Benjamin Saville*+ and Amy H. Herring
|
Companies:
|
The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
|
Address:
|
3106 McGavran Greenberg, Chapel Hill, NC, 27599,
|
Keywords:
|
random effects ; linear mixed model ; Bayes factors
|
Abstract:
|
Deciding which predictor effects may vary across subjects is a difficult modeling issue. Testing on the boundary of the parameter space changes the asymptotic distribution of some classical test statistics and causes numerical problems for Bayesian methods of approximating Bayes factors. Random effects also induce high-dimensionality, which can limit the performance of popular approximations to Bayes factors. We propose a simple approach for testing random effects in the linear mixed model using Bayes factors. We introduce factor loadings on the random effects and scale the random effects to the residual variance. We suggest default priors on the factor loadings, and integrate out the random effects and variance components using closed form solutions. We use Laplace's method to approximate the marginal likelihoods needed to evaluate the Bayes factor.
|