|
Activity Number:
|
369
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #305613 |
|
Title:
|
Order-Restricted, Randomized Designs for Linear Models Using L1 Norm
|
|
Author(s):
|
Shannon Markiewicz*+ and Omer Ozturk
|
|
Companies:
|
The Ohio State University and The Ohio State University
|
|
Address:
|
955 Mount Pleasant Ave., Columbus, OH, 43201,
|
|
Keywords:
|
ranked set sampling ; order restricted randomized designs
|
|
Abstract:
|
Ranked set sampling is not well-suited for the design of experiments because it requires more experimental units and the role of randomization is not well-defined. To resolve these concerns, order-restricted, randomized (ORR) designs have been developed recently and their properties discussed in the literature. In this talk, we will develop nonparametric statistical inference for linear models in the context of order-restricted, randomized designs based on L-1 norm. We will derive the asymptotic distribution of the parameter estimates and develop a drop and Wald test for the contrast parameter. It will be shown that ORR design performs better than classical designs in linear models. A simulation study will show that test and estimators perform reasonably well, even for moderately large sample sizes.
|
