|
Activity Number:
|
212
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #303968 |
|
Title:
|
Score Test--Based on GEL in the Presence of Weakly Identified Nuisance Parameters
|
|
Author(s):
|
Saraswata Chaudhuri*+
|
|
Companies:
|
The University of North Carolina at Chapel Hill
|
|
Address:
|
Department of Economics, CB 3305, Chapel Hill, NC, 27599,
|
|
Keywords:
|
efficient score test ; nuisance parameters ; generalized empirical likelihood ; weak identification ; projection ; Lagrange Multilplier test
|
|
Abstract:
|
Conventional methods of hypothesis testing can severely over-reject the true parameter value beyond their nominal level if the strength of identification is low for some or all parameters in a model. While it has been recently shown that the score test for subsets of parameters based on generalized empirical likelihood (GEL) does not over-reject the true parameter value irrespective of the strength of identification of the parameters of interest; absence of over-rejection cannot be generally guaranteed if the nuisance parameters are not strongly identified. We address this problem by proposing a new method of projection-based score test for subsets of parameters that guards against the uncontrolled over-rejection of the true value of the parameters of interest even when the nuisance parameters are not identified, while achieving asymptotic equivalence with the GEL score test otherwise.
|
