|
Activity Number:
|
465
- Probabilistic Record Linkage: Better Assumptions, Scalable Inference, and Accounting for Uncertainty
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Social Statistics Section
|
|
Abstract #329148
|
|
|
Title:
|
Multiple Imputation of Probabilistic Linkage of Employers in Survey and Administrative Data: Creating CenHRS
|
|
Author(s):
|
Dhiren Patki*
|
|
Companies:
|
University of Michigan
|
|
Keywords:
|
Probabilistic record linkage;
Multiple imputation;
Wages ;
Employer size
|
|
Abstract:
|
CenHRS is a new data source that enhances the Health and Retirement Study (HRS) with administrative data from the Census Business Register (BR). The dataset is constructed by probabilistically linking records on employed HRS respondents with employer characteristics from the BR. An important aspect of the CenHRS is that employer information is derived by multiple imputation (MI) to account for uncertainty in record linkage. In addition to documenting the record linkage procedure used to construct CenHRS, this paper re-examines the relationship between employer size and worker wages to illustrate the trade off between bias and variance implied by the MI linkage method in comparison to alternative methods.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|
