The measurement of health inequalities usually involves either estimating the concentration of health outcomes using an income-based measure of status or applying conventional inequality-measurement tools to a health variable that is non-continuous or, in many cases, categorical. However, these approaches are problematic as they ignore less restrictive approaches to status. The approach in this paper is based on measuring inequality conditional on an individual's position in the distribution of health outcomes: this enables us to deal consistently with categorical data. We examine several status concepts to examine self-assessed health inequality using the sample of world countries contained in the World Health Survey. We also perform correlation and regression analysis on the determinants of inequality estimates assuming an arbitrary cardinalisation. Our findings indicate major heterogeneity in health inequality estimates depending on the status approach, distributional-sensitivity parameter and measure adopted. We find evidence that pure health inequalities vary with median health status alongside measures of government quality.
health inequality categorical data entropy measures health surveys upward status downward status