Measuring Poverty Efficiently Using Adaptive Deprivation Scales

Abstract

Deprivation scales are becoming increasingly familiar in research and official statistics on poverty. Taking advantage of the basis of these scales in Item Response Theory, this paper proposes a more efficient approach to implementation using adaptive testing. This maximises information collected for a given amount of survey time by screening respondents on the basis of initial answers, and halting questioning where it is very unlikely any further information will be gathered. The paper illustrates various implementations using data from eight years of the UK’s Family Resources Survey (FRS). Results show that an adaptive approach collects more than 99% of the information from the UK’s official deprivation measure in half the survey time. In addition, the paper suggests improvements in the design of the UK’s official deprivation scale as well as lessons for the development of future deprivation scales more generally.