Comparison of Survey and Register Data: The Swedish Case

Rand Corporation, Santa Monica CA

Investigators

Abstract

This research proposes to explore the benefits of combining register data and survey data. The results will inform an initiative to collect HRS-like data in Europe, entitled SHARE (Survey of Health, Aging, and Retirement in Europe), which is currently being prepared in eight European countries. Due to its similar set-up to the HRS, SHARE will greatly enhance the potential for the use of international comparisons in aging research. Our proposed research will not only inform the design of SHARE, but it may also have wider applications in the design and analysis of surveys and in the joint use of survey and administrative data. We propose to use Sweden as a case study and investigate various aspects of the linkage and comparison of register data and survey data. We will draw a sample of approximately 1000 individuals over the age of fifty from the registers and next survey them about a number of life domains relevant to aging. We will link the survey information to the information from the registers, and use the register information to gauge the reliability of the survey data thus collected. The research will: -Assess selectivity in survey responses in a number of important domains -Gauge reliability of selected survey measures including income and pension entitlements, by comparison with register data -Assess the value of linkage of survey data and register data -Estimate biases in a number of important empirical relationships (e.g. health and Socio-Economic Status (SES)) when using survey data rather than register data -Generalize from these findings to the limitations of international comparisons if register data are available in some countries, but not in others -Use the findings to evaluate the potential benefits of attempts to use register data in other countries.