Digit accuracies in age data for developed regions with old pattern of digit preference error in developing world

Barun Kumar Mukhopadhyay, Ex Population Scientist Indian Statistical Institute

Turner (1958) considered a hypothetical ranking from 0 to 9 digits in age reporting, such as '0' the maximum ranking '5' the next, '2', '4' and, etc., and later compares with the observed ranking by Spearman's coefficient of rank correlation in order to compare the quality of age reporting among different countries of the world around fifties. In the present time his method may not be suitable for developed countries following almost a progressively declining series in their raw single year age data. As such present author tries to develop some theoretical ranking only for developed countries based on life table stationary population by ten digits within a broad age band being depleted by death only assuming no effects of fertility and migration. The findings are very encouraging as a "case study" to study the "digit accuracy index", among developed countries. However, Turner's ranking is most suitable for developing countries with heaping still exists.

  See paper

Presented in Poster Session 5: Contexts