On the Law of Comparative Judgment: Scaling with Intransitive Observers and Multidimensional Stimuli

Abstract

Intransitivities in observers' preference judgments for multidimensional natural landscape scenes are examined. Thurstone's Law of Comparative Judgment (LCJ) psychophysical scaling routine is used to scale landscape preference. Both the method of Rank Ordering (RO) and the method of Paired Comparisons (PC) are used to gather the raw data necessary for scaling. The preference metric derived from the PC data which allows intransitivities to occur, is compared to the metric derived from the RO data, which does not allow observer intransitivity. Analysis indicates that an observer's expressed ordinal preferences derived from his PC data differs from that obtained from his RO data. Analysis also indicates that intransitivity exist to a moderate degree in each individual's expressed preferences and that this intransitivity is related to the differences between the two sets of ordinal preferences. However, very little difference exists between the two, LCJ derived, interval preference metrics obtained from group PC data and group RO data.