Activity Number:
|
37
- Combining Data and Use of Administrative Lists
|
Type:
|
Contributed
|
Date/Time:
|
Sunday, July 30, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Survey Research Methods Section
|
Abstract #324549
|
View Presentation
|
Title:
|
Supplementing Record Linkage with Statistical Matching for Non-Consent Bias Reduction
|
Author(s):
|
Jonathan Gessendorfer* and Jonas Beste and Jörg Drechsler and Joseph Sakshaug
|
Companies:
|
Institute for Employment Research and Institut für Arbeitsmarkt- und Berufsforschung and Institute for Employment Research and The University of Manchester
|
Keywords:
|
Statistical Matching ;
Record Linkage ;
Non-Consent Bias
|
Abstract:
|
Record linkage has become an important tool for increasing research opportunities in the social sciences and is likely to become even more important in the "big data" era. Surveys that perform record linkage are often required to obtain informed consent from respondents prior to linkage. A major concern is that record non-consent can introduce biases. One strategy to solve this missing data problem is statistical matching. The missing administrative data of a non-consenter is estimated using the data of a statistically similar individual from the administrative dataset. To evaluate statistical matching in that regard, we use data from two major German panel surveys - the National Educational Panel Study and the Panel 'Labour Market and Social Security'. Both were linked to process data of the German Federal Employment Agency on almost the complete German population of employable age. Our results show that non-consent biases in marginal distributions can be reduced. Biases in correlations and coefficients of regression models, using variables from both survey and administrative data can sometimes even be worse after supplementing biased record linked data with statistical matches.
|
Authors who are presenting talks have a * after their name.