Do rank order and scale properties remain invariant under changes in the set of scaled stimuli?

Abstract
The question, whether rank order and scale properties remain invariant under changes in the set of scaled stimuli, was addressed in the empirical context of paired‐comparisons judgments. Three overlapping sets of stimuli were prepared and presented, one set to each of three independent groups of judges. Rank order and scale properties were compared for the stimuli common to the different stimulus sets. In one of three cases, rank order reversal and breakdown of scale invariance was found in relation to change in stimulus context. An hypothesis about judgmental processes is presented to account for these findings.