|
Activity Number:
|
299
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 14, 2002 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
SSC
|
| Abstract - #301367 |
|
Title:
|
Use of Generalized Variance Functions in Multivariate Analysis
|
|
Author(s):
|
John Eltinge*+
|
|
Affiliation(s):
|
U.S. Bureau of Labor Statistics
|
|
Address:
|
2 Massachusetts Ave. N.E., Washington, District of Columbia, 20212, U.S.A.
|
|
Keywords:
|
Eigenvalues and eigenvectors ; Goodness of fit test ; Model check ; Quadratic form test statistics ; Rao-Scott correction ; U.S. Consumer Expenditure Survey
|
|
Abstract:
|
In the analysis of complex survey data, goodness-of-fit tests and other model checks often are based on quadratic form test statistics. These test statistics in turn generally make explicit or implicit use of variance-covariance matrix estimators. Under some complex sample designs, these matrix estimators are relatively unstable. This can cause serious degradation in the performance of the associated test statistics. This paper examines the extent to which generalized variance function methods can produce more stable variance-covariance matrix estimators, and thus lead to improved test statistics. Relationships between the resulting test statistics and Rao-Scott type test statistics are discussed. Some of the proposed methods are applied to data from the U.S. Consumer Expenditure Survey.
|
- 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 2002 program |
