|
Activity Number:
|
54
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 3, 2003 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #301468 |
|
Title:
|
Rank-based Estimation in Linear Models with Clustered Data
|
|
Author(s):
|
Suzanne R. Dubnicka*+
|
|
Companies:
|
Kansas State University
|
|
Address:
|
101 Dickens Hall, Manhattan, KS, 66506,
|
|
Keywords:
|
rank regression ; random effect ; pseudo-sample
|
|
Abstract:
|
Consider a study in which clusters of individuals are observed. Examples of such clusters include siblings, littermates, and repeated environmental measurements taken at the same location. Whatever the origin of the cluster, responses from individuals within a cluster are considered to be correlated while responses from individuals in different clusters are not. A rank method is presented for estimating regression parameters in the linear model when observations are correlated. We account for this correlation by including a random effect term in usual linear model. The estimation procedure is an iterative process which involves (1) the estimation of the regression parameters using rank regression and (2) the estimation of the random effects using pseudo-samples. The advantage of this method is that it does not require specific distributions of the random effects or the errors. Simulation results show that this method performs quite well relative to existing methods under a variety of distributional assumptions. An example will be presented to illustrate the method.
|
- 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 2003 program |
