Activity Number:
|
284
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Survey Research Methods
|
Abstract - #305813 |
Title:
|
Analysis of Data from Complex Survey Samples: Propensity Scores and Survey Weights
|
Author(s):
|
Weiwei Cui*+ and Nan Liu
|
Companies:
|
National Institute of Statistical Sciences and Child Trends, Inc.
|
Address:
|
1990 K St NW Suite 500, Washington , DC, 20006,
|
Keywords:
|
complex survey data ; survey weights ; propensity scores ; observational studies
|
Abstract:
|
In observational studies, investigators have no control over the treatment assignment. Thus, the estimates of treatment effects may be biased because of the differences in the distribution of covariates between treatment and control groups. Propensity score matching is often used to balance the distributions of covariates. The propensity score matching and the statistical test of the treatment effect become more complicated when complex survey data are used. Appropriate survey weights should be incorporated into the calculation of propensity scores and the single-level statistical analysis conducted after matching. This study compares the use of survey weights in calculating propensity scores and statistical analysis of treatment effect after propensity scores matching. Simulated and read survey data are used to demonstrate the essence of the accommodation of survey weights.
|