Title
|
Room
|
* ! Statistical Matching and Record Linkage: Implementation and Evaluation
|
H-Jefferson
|
Date / Time
|
Sponsor
|
Type
|
08/05/2001
4:00 PM
-
5:50 PM
|
Social Statistics Section*, Section on Government Statistics*, Section on Survey Research Methods*, ENAR
|
Topic Contributed
|
Organizer:
|
Carole Popoff, U.S. Census Bureau
|
Chair:
|
Bonnie Ghosh-Dastidar, RAND
|
Discussant:
|
5:25 PM - Yves Thibaudeau, Census Bureau
|
Floor Discussion
|
5:40 PM
|
Description
Statistical matching and record linkage are essential methodologies for assembling information from diverse sources into databases used in policy analysis. We consider issues arising in construction, analysis, and validation of such matched databases. Two papers critique existing matching methods and propose a new one. The third considers model-based analyses of linked files. The last two papers present case studies of creating and validating matched files, based on population surveys of health and of income and poverty dynamics.
|