An Exploration of Proxy- and Self-Reported Adolescent Health in Low-Resource Settings

Abstract

Proxy reports in survey research are widely used when the index individual is unavailable or somehow incapacitated by age or disability. Proxy reports are plagued by concerns about accuracy, however, and self-reports are generally preferred when objective measurement is not possible. This paper uses the Young Lives Study of International Child Poverty to assess the validity and utility of adolescent self-reported health (SRH) and the conventional parent's proxy report. Using multivariate regression models and the framework of convergent validity, I find evidence for the validity of both proxy and self-reports, although proxy reports appear to be slightly more robustly associated with available physical health information. Exploratory multiple imputation simulations suggest that researchers should request both proxy and self-reports in household surveys; having both substantially improves the imputation of one if it is missing or implausible. Along with a moderate correlation between the two reports, these results suggest that proxy and self-reports of adolescent's general health status are not interchangeable and may complement one another.