Socio-economic and demographic variation in health and in its measures: the issue of reporting heterogeneity.

True health state is an unobservable concept. Researchers and practitioners now have access to a large variety of tools to measure the health state and health related quality of life by self-reports. Socio-demographic variation in these measures is usually interpreted as variation in health. However, building on several measures simultaneously (multiple indicators), true health might be better represented, so that socio-demographic variation in any indicator can be decomposed into variation in the estimated true health, and measure-specific variation, holding true health constant. The latter variation is referred to as "reporting heterogeneity". Using structural equations models, the paper provides an empirical assessment of reporting heterogeneity in three popular measures of health and health related quality of life: the number of chronic conditions (CHRON), the SF-36 instrument and the visual analogue rating scale. Considering a large array of socio-economic and demographic characteristics from an Israeli health survey, the results indicate the existence of age-related reporting heterogeneity in the CHRON; income-related heterogeneity in the rating scale measure; and age, sex, income, ethnic origin and religiosity-related reporting heterogeneity in the SF-36 tool, in particular in its mental component scale. The main implication of reporting heterogeneity on the common uses of self-reported health measures is the need to adjust the measures not only for the determinants of health but also for the determinants of reporting heterogeneity.