Activity Number:
|
149
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
IMS
|
Abstract - #301250 |
Title:
|
A General Asymptotic Result Useful in Models with Nuisance Parameters
|
Author(s):
|
Ma. Zenia Agustin*+ and Edsel Pena
|
Affiliation(s):
|
Southern Illinois University-Edwardsville and University of South Carolina
|
Address:
|
Box 1653, Edwardsville, Illinois, 62026-1653, USA
|
Keywords:
|
goodness-of-fit testing ; generalized residuals ; linear models ; Cox proportional hazards model
|
Abstract:
|
In statistical models with nuisance parameters, a general asymptotic result of test statistics obtained by substitution of estimators of the nuisance parameters for their unknown values in quantities whose asymptotic distributions are known is provided. Applications of the main result in obtaining asymptotic distributions of statistics formed from residuals of the classical linear model, as well as in goodness-of-fit testing and model validation for Andersen and Gill's (1982) multiplicative intensity model, are illustrated. The asymptotic results can be viewed as further extensions of those by Pierce (1982) in models where the plug-in estimators may not be fully efficient.
|