CMI Afternoon Seminar: Pseudo OLS for linkage-data regression by Prof Li-Chun Zhang
| Dates: | 2 October 2018 |
|---|
| Times: | 15:00 - 16:15 |
|---|
| What is it: | Seminar |
|---|
| Organiser: | School of Social Sciences |
|---|
| Who is it for: | University staff, Adults, Alumni, Current University students, General public |
|---|
|
Speaker: Prof Li-Chun Zhang
Abstract:
Unless a unique identifier exists for this purpose, linkage of separate datasets will generate errors that can cause bias of the subsequent analysis, if the linked data are treated as if they were truly observed. In this work we take on the perspective of secondary analysts, who we assume to neither have full access to the linkage key variables nor the details or tools of the actual linkage procedure, but at most are only provided with some non-disclosive linkage comparison data about the record linkage precision or how the records compare to each other. We discuss several existing approaches to statistical analysis, and their respective theoretical and practical difficulties. This include the maximum likelihood estimation, the frequentist approach to regression adjustment, and some Bayesian approaches that have been proposed.
Focusing on linear regression as the case-in-point, we develop a simple method of Pseudo OLS for linkage-data regression, where the analyst is only given the linked dataset, but not any of the unlinked records. We do not assume that the true matches are confined to within the units associated with the linked dataset, nor that the linkage error probability is a constant for different units. Moreover, we develop a diagnostic test for the assumption of non-informative linkage errors (NILE), which is needed for all the existing methods of linkage-error adjustment. Our approach will be illustrated by simulation and application to real data.
Tea/coffee and cakes from 2.45.
Join us for this event, which is part of the CMI Afternoon Seminar Series. All welcome. No registration necessary.
The Cathie Marsh Institute for Social Research (CMI) provides a focal point at The University of Manchester for the application of quantitative methods in interdisciplinary social science research in order to generate a world class research environment.
Travel and Contact Information
Find event
CMI Seminar Room, 2.07
Humanities Bridgeford Street
Manchester
