Pair Comparison, with and without Ties

Publisher Website
Google Scholar
Abstract
A probabilistic model for subjective pair comparison, with and without ties (i.e., an intermediate no-decision response class), and in various practical settings, is described. From both theory and experiment (in the field of taste) the conclusion is drawn that to allow ties rarely helps, and often hinders, the test objective. A notable exception is the sampling of population preferences; there the "no difference" response is essential as it provides real information.

This publication has 1 reference indexed in Scilit: