|
Activity Number:
|
14
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #303869 |
|
Title:
|
Summary of Methods and Preliminary Assessment of the SIPP Synthetic Beta, Version 5.0
|
|
Author(s):
|
Gary Benedetto*+ and Martha Stinson and Melissa Bjelland
|
|
Companies:
|
U.S. Census Bureau and U.S. Census Bureau and Cornell University
|
|
Address:
|
4600 Silver Hill Road, Department of Commerce, Suitland, MD, 20746,
|
|
Keywords:
|
confidentiality ; disclosure protection ; imputation ; partial synthesis ; SIPP ; public use data
|
|
Abstract:
|
This paper summarizes the methodology and quality assessment of the most recent version of the SIPP Synthetic Beta (SSB v5.0), a public use data set that combines long earnings histories and benefits information from administrative data sources with the detailed demographic data collected in the SIPP. We use multiple imputation to create partially synthetic microdata that contains no confidential data but should preserve many covariate properties of the underlying confidential data of interest to analysts. This method of disclosure protection has the benefit of not requiring any unusual software or sophisticated techniques to run traditional analyses and perform proper statistical hypothesis testing. In addition to describing our methods, we attempt to assess the analytic validity of the synthetic data and determine the disclosure risk of making such data available to the public.
|
