|
Activity Number:
|
502
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306406 |
|
Title:
|
A New Method for Weighting Survey Respondents
|
|
Author(s):
|
Feiming Chen*+
|
|
Companies:
|
Spectra Marketing Systems
|
|
Address:
|
9371 Notre Dame Drive, Apt. C, Indianapolis, IN, 46240-4172,
|
|
Keywords:
|
sample survey ; poststratification ; raking ; demographic weighting ; singular value decomposition
|
|
Abstract:
|
In sample survey, we often need to adjust the weights of survey respondents via poststratification. Raking (Iterative Proportional Fitting) is typically used to align marginal distributions of the sample to those of the population. However it may converge slowly when there are many categories or empty cells. In this paper we present a non-iterative weighting method based on Singular Value Decomposition. It is computationally fast and may be useful in complex alignment problems. We provide a free R library for interested users.
|
