A comparison of two models for scaling health indicators

Abstract

The measurement properties of a health status indicator are closely related to the scaling model on which it is based. The Thurstone and Bradley-Terry models are applied to paired comparisons data which had been used to scale the sleep category of the Nottingham Health Profile. The data in their original form are shown to be inadequately represented by either model. Weaknesses in the data are identified and the two scaling models are applied to the amended data. The results of applying two sets of scale values are compared.