Survey response behaviour and the dynamics of self-reported health and disability: an experimental analysis

Abstract

Disability research often uses survey questions asking whether respondents have a long-standing health problem, but longitudinal repetition of such questions produces implausibly high empirical transition rates into and out of ill-health. We exploit a repeated experiment in the Understanding Society Innovation Panel to identify reasons for these high transition rates, and to assess the common practice of using such questions to control who is asked further questions on difficulties with daily activities. Our results reveal ambiguity in the concept of a long-standing health problem and indicate significant biases in commonly used measures and multivariate analyses of health dynamics and disability.