|
Activity Number:
|
342
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #308979 |
|
Title:
|
Identification of Influential Observations for Regression Using Complex Survey Data
|
|
Author(s):
|
Jianzhu Li*+ and Richard Valliant
|
|
Companies:
|
University of Maryland and University of Maryland
|
|
Address:
|
1218 Lefrak Hall, College Park, MD, 20742,
|
|
Keywords:
|
influential observations ; complex survey ; influential observations ; cutoff ; variance estimations
|
|
Abstract:
|
Even though survey data sets are often large, a few extreme observations can still distort regression estimates. This paper intends to extend the OLS diagnostic statistics such as DFBETAS, DFFITS, and Cook's Distance to the survey setting. Distributional properties and cutoff values will be determined in terms of appropriate variance estimation methods. Extended statistics will be applied to real survey data in a case study.
|
