Tests of rating models

Abstract

The study reports empirical tests of Anderson's, Haubensak's, Helson's, and Parducci's rating models when two end anchors are used for rating. The results show that these models cannot predict the judgment effect called here the Dai Prà effect. It is shown that an extension of Anderson's model is consistent with this effect. The results confirm the main contention of each model that ratings are linear measures of mental magnitude.