|
Activity Number:
|
51
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 11, 2002 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biometrics Section*
|
| Abstract - #301655 |
|
Title:
|
Performance of a Statistic for Testing Model Adequacy in the Linear Mixed Model
|
|
Author(s):
|
Jean Orelien*+ and Lloyd Edwards
|
|
Affiliation(s):
|
Analytical Sciences, Inc./UNC and University of North Carolina
|
|
Address:
|
, , , ,
|
|
Keywords:
|
linear ; mixed ; models ; goodness-of-fit ; diagnostic
|
|
Abstract:
|
In linear mixed models, the Akaike information criterion (AIC) and the Schwartz Information criteria (SIC) are commonly used to assess goodness of fit. However, these statistics are limited in that they require that several models be fitted to the same data. Vonesh et al. (1996) proposed a more intuitive statistic denoted the concordance correlation coefficient (CCC) that can be interpreted as a measure of agreement between the predicted and observed values. The performance of CCC has not been demonstrated either analytically or in simulation. To that end, we conducted a simulation with the goal of determining: a.) the extent to which CCC can detect the adequacy of the conditional mean (fixed effect terms) given that the correct covariance structure for the random effect is specified; and b.) the extent to which CCC can detect adequacy of the covariance structure of the random effects given the conditional mean specified is the correct one. Our simulation results show that for fitting a regression type model, CCC is more sensitive to misspecification of the covariance structure than misspecification of the correct mean model.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |
