|
Activity Number:
|
37
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306504 |
|
Title:
|
Global and Hierarchical Linear Regression in Two-Stage Sampling
|
|
Author(s):
|
Dhirendra Ghosh*+ and Andrew Vogt
|
|
Companies:
|
Synectics for Management Decisions, Inc. and Georgetown University
|
|
Address:
|
1901 N. Moore Street, Arlington, VA, 22209,
|
|
Keywords:
|
cluster sampling ; least squares ; regression estimates
|
|
Abstract:
|
In a two-stage sampling design where the second stage units have a known independent variable and the survey collects data on a dependent variable, we compare different forms of hierarchical linear regression, as well as global linear regression (where the first stage units, or clusters, are ignored). We consider cases where each cluster in the sample has a large sample size with or without an auxiliary cluster variable; where each cluster has a small sample size and an auxiliary cluster variable; and where the cluster variable is unknown for clusters not in the sample.
|
