Featuring contributions from well-known international experts, this book addresses the methodological issues involved in administrative data research as well as the various concerns users face with administrative records such as issues of privacy, confidentiality, and legality.  The book illustrates numerous real-world examples of administrative data research from various countries, culminating in a comprehensive guide to administrative data in statistical surveys.  This historical and international perspective provides readers with a better understanding of the practical applications and approaches needed for improving the quality of surveys, controlling the cost of survey data collection, and integrating administrative data with other data obtained from surveys and censuses. This book also features detailed coverage on the advanced statistical techniques for the control of data quality and reduction in total survey error.  The first part of the book focuses on the theory of total administrative records error and provides relevant practices case studies tied to the survey life-cycle, while the second part of the book features the technical issues of processing and linking administrative data with multiple sources of data in multimode data collection.  Bayesian approaches are linked to real-world applications for the use of administrative data in surveys and censuses over the survey life cycle, and the relevance of how these cutting-edge techniques can affect administrative records research is illustrated in key sectors of health, economy, and education.  In addition, these technological and statistical innovations are used to advance the systematic integration of administrative data, improve the survey frame, reduce nonresponse follow-up, and assess coverage error.  Topical coverage includes: pandata systems to enhance survey and census systems; integration of survey and administrative data for statistical purposes; evaluation of the quality of administrative data; measurement of data quality in register-based statistics; cleaning and using administrative lists; assessing uncertainty; record linkage and assessment of date in health sciences; methods to improve small area estimation; administrative records for imputing nonresponse; Bayesian use of administrative records around census life cycle; use of administrative data in official statistics; application of administrative data in health science; using linking survey and administrative data to improve economic surveys; and administrative sources for censuses with demographic and social statistics.

Asaph Young Chun, PhD, is Research Chief for Decennial Directorate at the U.S. Census Bureau. He is the author of over 110 journal articles and received his PhD in Sociology from the University of Maryland.

Michael D. Larsen, PhD, is Professor in the Department of Statistics and Director of the Graduate Certificate Program in Survey Design and Data Analysis at The George Washington University. He received his PhD in Statistics from Harvard University.

Section 1: Fundamentals of Administrative Records Research and Applications

1.            On the use of proxy variables in combining register and survey data, Li-Chun Zhang, Statistics Norway and University of Southampton

1.1.         Introduction

1.1.1.     A multisource data perspective

1.1.2.     Concept of a proxy variable

1.2.         Instances of proxy variable

1.2.1.     Representation

1.2.2.     Measurement

1.3.         Estimation using multiple proxy variables

1.3.1.     Asymmetric setting

1.3.2.     Uncertainty evaluation: a case of two-way data

1.3.3.     Symmetric setting

1.4.         Summary           

1.5.         References

2.            Disclosure Limitation and Confidentiality Protection in Linked Data, John Maron Abowd, U.S. Census Bureau and Cornell University, Ian M. Schmutte, University of Georgia; and Lars Vilhuber, Cornell University.

2.1.         Introduction

2.2.         Paradigms of protection

2.2.1.     Input noise infusion

2.2.2.     Formal privacy models

2.3.         Confidentiality protection in linked data: Examples

2.3.1.     HRS-SSA

2.3.2.     SIPP-SSA-IRS (SSB)

2.3.3.     LEHD: Linked establishment and employee records

2.4.         Physical and legal protections

2.4.1.     Statistical data enclaves

2.4.2.     Remote processing

2.4.3.     Licensing

2.4.4.     Disclosure avoidance methods

2.4.5.     Data silos

2.5.         Conclusions

2.6.         References

2.7.         Appendix: Technical Terms and Acronyms

2.7.1.     Data

2.7.2.     Other Abbreviations

2.7.3.     Concepts

Section 2: Data Quality of Administrative Records and Linking Methodology

3.            Evaluation of the Quality of Administrative Data Used in the Dutch Virtual Census, Piet Daas, Eric Schulte Nordholt, Martijn Tennekes, and Saskia Ossen, Statistics Netherlands

3.1.         Introduction

3.2.         Data sources and variables

3.3.         Quality framework

3.3.1.     Source and Metadata hyper dimensions

3.3.2.     Data hyper dimension

3.4.         Quality evaluation results for the Dutch 2011 Census

3.4.1.     Source and Metadata: application of checklist

3.4.2.     Data hyper dimension: completeness and accuracy results

3.4.3.     Discussion of the quality findings

3.5.         Summary

3.6.         Practical implications for implementation with surveys and censuses

3.7.         Exercises

3.8.         References

4.            Improving input data quality in register-based statistics: The Norwegian experience, Coen Hendriks, Statistics Norway

4.1.         Introduction

4.2.         The use of administrative sources in Statistics Norway

4.3.         Managing statistical populations

4.4.         Experiences from the first Norwegian purely register based Population and Housing Census of 2011

4.5.         The contact with the owners of administrative registers was put into system

4.5.1.     Agreements on data processing

4.5.2.     Agreements on cooperation on data quality in administrative data systems

The forums for cooperation

4.6.         Measuring and documenting input data quality

4.6.1.     Quality indicators

4.6.2.     Operationalizing the quality checks

4.6.3.     Quality reports

4.6.4.     The approach is being adopted by the owners of administrative data

4.7.         Summary

4.8.         Exercises

4.9.         References

4.10.      Appendix: Example of a quality report for registered persons in the Central Population Register

5.            Cleaning and Using Administrative Lists: Enhanced Practices and Computational Algorithms for Record Linkage and Modeling/Editing/Imputation, William Erwin Winkler, U.S. Census Bureau

5.1.         Introductory comments

5.1.1.     Example 1

5.1.2.     Example 2

5.1.3.     Example 3

5.2.         Edit/Imputation

5.2.1.     Background

5.2.2.     Fellegi-Holt Model

5.2.3.     Imputation Generalizing Little-Rubin

5.2.4.     Connecting Edit with Imputation

5.2.5.     Achieving Extreme Computational Speed

5.3.         Record Linkage

5.3.1.     Fellegi-Sunter Model

5.3.2.     Estimating Parameters

5.3.3.     Estimating False Match Rates

5.3.4.     Achieving Extreme Computational Speed

5.4.         Models for Adjusting Statistical Analyses for Linkage Error

5.4.1.     Scheuren and Winkler

5.4.2.     Lahiri and Larsen

5.4.3.     Chambers and Kim

5.4.4.     Chippenfield, Bishop, and Campbel

5.4.5.     Goldstein, Harron, and Wade

5.4.6.     Hof and Zwinderman

5.4.7.     Trancredi and Liseo

5.5.         Concluding Remarks

5.6.         Issues and some related questions

5.7.         References

6.            Assessing Uncertainty when Using Linked Administrative Records, Jerome P.  Reiter, Duke University

6.1.         Introduction

6.2.         General sources of uncertainty

6.2.1.     Imperfect matching

6.2.2.     Incomplete matching

6.3.         Approaches to accounting for uncertainty

6.3.1.     Modeling matching matrix as parameter

6.3.2.     Direct modeling

6.3.3.     Imputation of entire concatenated file

6.4.         Concluding Remarks

6.4.1.     Problems to be solved

6.4.2.     Practical implications

6.5.         Exercises

6.6.         References

7.            Measuring and Controlling for Non-Consent Bias in Linked Survey and Administrative Data, Joseph W. Sakshaug, University of Manchester, United Kingdom, and Institute for Employment Research, Nuremberg, Germany

7.1.         Introduction

7.1.1.     What is Linkage Consent? Why is Linkage Consent Needed?

7.1.2.     Linkage Consent Rates in Large-Scale Surveys

7.1.3.     The impact of Linkage Non-Consent Bias on Survey Inference

7.1.4.     The Challenge of Measuring and Controlling for Linkage Non-Consent Bias

7.2.         Strategies for Measuring Linkage Non-Consent Bias

7.2.1.     Formulation of Linkage Non-Consent Bias

7.2.2.     Modeling Non-Consent Using Survey Information

7.2.3.     Analyzing Non-Consent Bias for Administrative Variables

7.3.         Methods for Minimizing Non-Consent Bias at the Survey Design Stage

7.3.1.     Optimizing Linkage Consent Rates

7.3.2.     Placement of the Consent Request

7.3.3.     Wording of the Consent Request

7.3.4.     Active and Passive Consent Procedures

7.3.5.     Linkage Consent in Panel Studies

7.4.         Methods for Minimizing Non-Consent Bias at the Survey Analysis Stage

7.4.1.     Controlling for Linkage Non-Consent Bias via Statistical Adjustment

7.4.2.     Weighting Adjustments

7.4.3.     Imputation

7.5.         Summary

7.5.1.     Key Points for Measuring Linkage Non-Consent Bias

7.5.2.     Key Points for Controlling Linkage Non-Consent Bias

7.6.         Practical implications for implementation with surveys and censuses

7.7.         Exercises

7.8.         References

Section 3: Use of Administrative Records in Surveys

8.            A Register-Based Census: The Swedish Experiences, Martin Axelson, Anders Holmberg,  Ingegerd Jansson, and Sara Westling, Statistics Sweden

8.1.         Introduction

8.2.         Background

8.3.         Census 2011

8.4.         A register based census

8.4.1.     Registers at Statistics Sweden

8.4.2.     Facilitating a system of registers

8.4.3.     Introducing a dwelling identification key

8.4.4.     The census household and dwelling populations

8.5.         Evaluation of the census

8.5.1.     Introduction

8.5.2.     Evaluating household size and type

8.5.3.     Evaluating ownership

8.5.4.     Lessons learned

8.6.         Impact on population and housing statistics

8.7.         Summary and final remarks

8.8.         References

9.            Administrative Records Applications for the 2020 Census, Vincent Tom Mule, Jr., Andrew Keller, U.S. Census Bureau

9.1.         Introduction

9.2.         Administrative Record Usage in the United States Census

9.3.         Administrative Record Integration in 2020 Census Research

9.3.1.     Administrative Record Usage Determinations

9.3.2.     NRFU Design Incorporating Administrative Records

9.3.3.     Administrative Records Sources and Data Preparation

9.3.4.     Approach to Determine Administrative Record Vacant Addresses

9.3.5.     Extension of Vacant Methodology to Non-Existent Cases

9.3.6.     Approach to Determine Occupied Addresses

9.3.7.     Other Aspects and Alternatives of Administrative Record Enumeration

9.4.         Quality Assessment

9.4.1.     Micro-Level Evaluations of Quality

9.4.2.     Macro-Level Evaluations of Quality

9.5.         Other Applications of Administrative Record Usage

9.5.1.     Register-Based Census

9.5.2.     Supplement Traditional Enumeration with Adjustments for Estimated Error for Official Census Counts

9.5.3.     Coverage Evaluation

9.6.         Summary

9.7.         Exercises

9.8.         References

10.          Use of Administrative Records in Small Area Estimation, Andrea L. Erciulescu, National Institute of Statistical Sciences, Carolina Franco, U.S. Census Bureau, Partha Lahiri, University of Maryland

10.1.      Introduction

10.2.      Data Preparation

10.3.      Small area estimation models for combining information

10.3.1.   Area-level models

10.3.2.   Unit-level models

10.4.      An Application

10.5.      Concluding Remarks

10.6.      Exercises

10.7.      Acknowledgments

10.8.      References

11.          Using Administrative Records to Control for Nonresponse Bias, Asaph Young Chun, Statistics Korea

Section 4: Use of Administrative Data in Evidence-Based Policymaking

12.          Enhancement of Health Surveys with Data Linkage, Cordell Golden, Lisa B. Mirel, NCHS

12.1.      Introduction

12.1.1.   The National Center for Health Statistics (NCHS)

12.1.2.   The NCHS Data Linkage Program

12.1.3.   Initial Linkages with NCHS Surveys

12.2.      Examples of NCHS health surveys that were enhanced through linkage

12.2.1.   National Health Interview Survey (NHIS)

12.2.2.   National Health and Nutrition Examination Survey (NHANES)

12.2.3.   National Health Care Surveys

12.3.      NCHS health surveys linked with vital records and administrative data

12.3.1.   National Death Index (NDI)

12.3.2.   Centers for Medicare & Medicaid Services (CMS)

12.3.3.   Social Security Administration (SSA)

12.3.4.   Department of Housing and Urban Development (HUD)

12.3.5.   United States Renal Data System and the Florida Cancer Data System

12.4.      NCHS Data Linkage Program: Linkage Methodology and Processing Issues

12.4.1.   Informed consent in health surveys

12.4.2.   Informed consent for child survey participants

12.4.3.   Adaptive approaches to linking health surveys with administrative data

12.4.4.   Use of alternate records

12.4.5.   Protecting the privacy of health survey participants and maintaining data confidentiality

12.4.6.   Updates over time

12.5.      Enhancements to health survey data through linkage

12.6.      Analytic considerations and limitations of administrative data

12.6.1.   Adjusting sample weights for linkage-eligibility

12.6.2.   Residential mobility and linkages to state programs and registries

12.7.      Future of the NCHS Data Linkage Program

12.8.      Exercises

12.9.      Acknowledgments and Disclaimer

12.10.    References

13.          Combining Administrative and Survey Data to Improve Income Measurement, Bruce D. Meyer, University of Chicago, and Nikolas Mittag, Charles University

13.1.      Introduction

13.2.      Measuring and Decomposing Total Survey Error

13.3.      Representation Error

13.4.      Item Non-response and Imputation Error

13.5.      Measurement Error

13.6.      Illustration: Using Data Linkage to Better Measure Income and Poverty

13.7.      Accuracy of Links and the Administrative Data

13.8.      Conclusions

13.9.      Study Problems

13.10.    References

14.          Combining Data from Multiple Sources to Define a Respondent: The Case of Education Data, Peter Siegel, Darryl Creel, James Chromy, RTI International

14.1.      Introduction

14.1.1.   Options for defining a unit respondent when data exist from sources instead of or in addition to an interview

14.1.2.   Concerns with defining a unit respondent without having an interview

14.2.      Literature Review

14.3.      Methodology

14.3.1.   Computing weights for interview respondents and for unit respondents who may not have interview data (useable case respondents)

14.3.2.   Imputing data when all or some interview data are missing

14.3.3.   Conducting nonresponse bias analyses to appropriately consider interview and study nonresponse

14.4.      Example of Defining a Unit Respondent for the National Postsecondary Student Aid Study (NPSAS)

14.4.1.   Overview of NPSAS

14.4.2.   Useable case respondent approach

14.4.3.   Interview respondent approach

14.4.4.   Comparison of estimates, variances, and nonresponse bias using two approaches to define a unit respondent

14.5.      Discussion: Advantages and disadvantages of two approaches to defining a unit respondent

14.5.1.   Interview respondents

14.5.2.   Useable case respondents

14.6.      Practical Implications for Implementation with Surveys and Censuses

14.7.      References

14.8.      Appendix: NPSAS:08 Study Respondent Definition