Activity Number:
|
225
- Using Administrative Data to Improve Surveys
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, July 31, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Business and Economic Statistics Section
|
Abstract #322799
|
|
Title:
|
Utilizing Indirectly Linked Federal Administrative Records for Survey Nonresponse Bias Evaluation and Adjustment
|
Author(s):
|
Joseph Sakshaug* and Manfred Antoni
|
Companies:
|
The University of Manchester and Institute for Employment Research
|
Keywords:
|
administrative data ;
nonresponse bias ;
record linkage
|
Abstract:
|
Auxiliary data sources are being sought for general population surveys to correct for possible nonresponse bias in key survey estimates. While several auxiliary data options exist in large-scale survey settings, questions about their usefulness for detecting and correcting for nonresponse bias remain. Federal administrative databases offer a potentially rich source of auxiliary information useful for nonresponse purposes, but linking them to general population survey samples (including the nonrespondents) is usually limited to surveys which draw their samples from population registers, and/or possess a unique identifier that can be directly linked to the relevant administrative database. We report on a case study in which a large federal administrative database in Germany was indirectly linked to respondents and nonrespondents in a general population survey using non-unique and error-prone identifiers. We examine the usefulness of the linked data for detecting nonresponse bias in the survey. We further examine the usefulness of these data for informing nonresponse bias adjustment procedures.
|
Authors who are presenting talks have a * after their name.