Cross-European Consent to Data Linkage in the Ageing Population of 28 Countries

Survey researchers increasingly seek to enhance their data by linking respondents' answers to external records such as health records or pension databases. This practice requires informed consent, which can vary widely across contexts. Using data from the Survey of Health, Ageing and Retirement in Europe (SHARE), we examined cross-national variation in consent to data linkage among individuals aged 50 and over. We analysed real-life consent rates from 10 countries (N = 44,446) and hypothetical consent preferences from a pilot study in 28 countries (N = 1,963). Real-life consent rates at the first ask were highest in the Netherlands, Luxembourg, and Denmark (91% to 97.5%), and lowest in Girona and Italy. Repeated requests for consent had minimal impact. Hypothetical consent questions used standardised wording for all countries and covered four data domains: employment, pensions and social security contributions, income and taxes, and health. Hypothetical consent varied considerably, with the highest rates in Belgium, Denmark and Estonia (75%, 72%, and 71%) and the lowest in Italy, Hungary and Poland (11%, 14% and 18%). Some multilingual countries showed marked differences between language groups. Preferences for linking different types of administrative data varied across countries , although linkage to income and tax information was almost universally least popular. Multilevel regression found real-life consent was significantly associated with medium or high education levels, age, and being a minority language speaker, while gender had no effect. In contrast, male gender was the only significant predictor of hypothetical consent. Intriguingly, real-life consent rates were generally higher than hypothetical ones, suggesting that stated preferences may underestimate actual willingness. Our findings highlight the value of hypothetical consent questions as a predictive tool, and the need for a larger studies to better understand language group effects and country differences.