The concentration index of a binary outcome revisited

Health Econ. 2011 Oct;20(10):1155-60. doi: 10.1002/hec.1752. Epub 2011 Jun 15.

Abstract

The binary variable is one of the most common types of variables in the analysis of income-related health inequalities. I argue that while the binary variable has some unusual properties, it shares many of the properties of the ratio-scale variable and hence lends itself to both relative and absolute inequality analyses, albeit with some qualifications. I argue that criticisms of the normalization I proposed in an earlier paper, and of the use of the binary variable for inequality analysis, stem from a misrepresentation of the properties of the binary variable, as well as a switch of focus away from relative inequality to absolute inequality. I concede that my normalization is not uncontentious, but, in a way, that has not previously been noted.

MeSH terms

  • Data Interpretation, Statistical*
  • Health Status Disparities*
  • Health Surveys
  • Humans
  • Models, Statistical
  • Social Class*