|
Activity Number:
|
33
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #305245 |
|
Title:
|
Efficacy of Post-Stratification in Complex Sampling Designs
|
|
Author(s):
|
Ismael Flores Cervantes*+ and Michael J. Brick and Mike Jones
|
|
Companies:
|
Westat, Inc. and Westat, Inc. and Westat, Inc.
|
|
Address:
|
1650 Research Blvd., Rockville, MD, 20850,
|
|
Keywords:
|
Calibration ; stratification ; poststratification
|
|
Abstract:
|
Post-stratification is a calibration estimator that is often used to reduce the variance of the estimates (efficiency) and to reduce bias due to noncoverage or nonresponse. In this paper we examine the efficiency of post-stratification in the full response and coverage situation. Virtually all results on the efficiency of post-stratification in the literature assume a simple random sampling. We expand this to look at the efficiency in complex designs including stratified and multi-stage cluster samples. In addition, we show that the efficiency of post-stratification differs substantially depending on the type of estimate. In particular, the reduction in variance for totals and is very different than the reduction for means and ratios. We attempt to describe the conditions where the post-stratification produces the largest gains in efficiency.
|
