Reliability of Respondent and Proxy Reports of the Social, Economic, and Health Status of Older Adults in Egypt and Tunisia

Cesar Rebellon, Emory University
Kathryn M. Yount, Emory University
Emily M. Agree, Johns Hopkins University

Reliance on proxies is common in research on older adults, yet the reliability of proxy and respondent reports is under-assessed in developing countries. Using data from the World Health Organization, we explore the reliability of self- and proxy-responses to items measuring the social, economic, and health status of adults aged 60+ in Egypt (n=941) and Tunisia (n=1129). Multivariate analysis is used to assess whether item reliability differs by respondent's gender and type of proxy, and whether effects of proxy type differ by respondent's gender. In both sites, items measuring availability of social support are highly reliable. Agreement on items measuring source of support is higher for males than females, particularly when proxies are spouses. Agreement on items measuring economic resources is low, with all proxy types more often reporting that respondents' economic needs are met. Agreement on items measuring health is moderate, with disagreeing proxies generally reporting better health.