Using Dempster Shafer theory to aggregate usability study data

Abstract

This paper describes an attempt to develop a mixed data aggregation technique applicable for a groupware usability study. We show the formal parallelism between the evidence combination and decision making problems and that of ranking alternatives according to their usability. After the parallelism is established we explain how the ranking of alternatives may be approached by a combination of techniques from Dempster-Shafer and multi-criteria decision making (MCDM) theories and illustrate the presented technique on actual data gathered during a usability study conducted by the Ambient Technology Group in Middlesex University.