Successive intervals analysis of preference measures in a health status index

Abstract

The method of successive intervals, a procedure for obtaining equal intervals from category data, is applied to social preference data for a health status index. Several innovations are employed, including an approximate analysis of variance test for determining whether the intervals are of equal width, a regression model for estimating the width of the end intervals in finite scales, and a transformation to equalize interval widths and estimate item locations on the new scale. A computer program has been developed to process large data sets with a larger number of categories than previous programs.