Comparability of Self-Assessed Health Across Countries

This study aims to establish whether people from different countries respond to self-assessed health (SAH) questions differently, regardless of their health state. Reporting heterogeneity might cause results from SAH studies to be biased. Using anchoring vignettes and data from the first two waves of the Survey of Health, Ageing and Retirement in Europe (SHARE), this thesis investigates reporting heterogeneity in SAH across different EU countries. Also, there is a comparison made of the results between two waves of data, to look into possible differences in reporting heterogeneity over time. In this thesis SAH is divided among six health domains: pain, sleeping problems, mobility, memory, breathing problems and depression. Estimating an ordered probit model with the vignettes as the dependent variable shows that there indeed exists a difference in response scales when comparing countries across the six health domains. The answers to the vignettes are then used to correct the six SAH measures for reporting heterogeneity, using a HOPIT model. When comparing the results for corrected SAH across two waves of SHARE data, it can be seen that conclusions for SAH differences between countries are not constant over time. These results imply that reporting heterogeneity in SAH across countries exists and that results for these country differences are prone to changes over time when comparing two waves of panel data.