|
Activity Number:
|
268
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
ENAR
|
| Abstract - #305012 |
|
Title:
|
Test of Trend for Clustered Data with the Use of Stochastic Ordering
|
|
Author(s):
|
Kyeongmi Cheon*+ and Aniko Szabo and E. Olusegun George
|
|
Companies:
|
The University of Memphis and Medical College of Wisconsin and The University of Memphis
|
|
Address:
|
373 Dunn Hall , Memphis, TN, 38152-3240,
|
|
Keywords:
|
clustered data ; test of trend ; stochastic ordering
|
|
Abstract:
|
Characterizing dose-response function is important in understanding and controlling toxic substances. The definition of dose related trend has been ambiguous in studies with clustered data. In this paper, we propose an approach which defines trend in terms of stochastic ordering of toxicology endpoints. To construct a likelihood ratio test for trend we device an EM algorithm of Lindsay, by adapting the procedure of Hoff. We address the problem of sparseness and augment data using the marginal compatibility assumption of Pang and Kuk. Our method is based on per whole cluster response, and does not require specifying a certain parametric models. Our test is flexible due to the use of stochastic ordering and allows for various forms of monotone increases, capturing linear increase in marginal mean response as a special case.
|
