A test of the correspondence between perceptual spaces based on pairwise similarity judgments collected with and without the inclusion of explicit ideal objects

Abstract

The use of multidimensional scaling to create a joint space based on pairwise similarity judgments collected with the inclusion of an ideal object raises concern that the resulting perceptual map might be distorted by the affective overtones of the ideal-point ratings. To investigate this possibility, 44 student Ss were randomly assigned to 2 experimental groups of 22 Ss each. Group 1 provided pairwise similarity ratings for 12 real female singers. Group 2 rated this same set of singers plus a 13th ideal singer. Comparison of the intersinger distances in the resulting perceptual spaces, using the statistical test proposed by L. J. Hubert and J. R. Levin (see record 1977-04798-001 ) suggests that the maps for Groups 1 and 2 are highly congruent so that the potential contamination resulting from the inclusion of ideal objects in pairwise similarity judgments may not be a serious problem in practice. (16 ref) (PsycInfo Database Record (c) 2025 APA, all rights reserved)