The role of differential patterns of responses in the relationship between preference and complexity

Abstract

Two major questions relevant to a cognitive conceptualization of personality were studied: (1) the identification of differential patterns of responses associated with individuals' preferences for complexity, and (2) the identification of factors associated with preference responses to stimulus complexity. Existing conceptualizations of the construct of preference for complexity were found to be simplistic and to ignore problems illustrated in the present study. Preference responses on the Revised Art Scale where complexity was defined in terms of informal abstract drawings were not consistent with preference responses to complexity as defined in terms of Attneave and Arnoult-type random polygons. This inconsistency was attributed both to domain and sampling differences between the two sets of stimuli. Previous research described a general statement of preference as an inverted u-shaped function of increasing stimulus complexity. By contrast, the present study, using inverse factor analysis, found preference-for-complexity functions involving four different preference responses for a-sided, random polygons. These were represented by four discrete groups of Ss. The first two groups only, of 21 and 4 Ss respectively, were described by a linear relation between unidimensional scales of preference and complexity; the larger group preferring complexity and the smaller group preferring simplicity. The relationships were more complex in the other two groups, each of 7 Ss. In the first two groups the same semantic differential scale factor was related to the unidimensional scales of preference and complexity, whereas in the latter groups, preference and complexity were related to different semantic factors. Preference judgments, for each of the groups, were best described by a unidimensional scale. Interpretation of preference dimensions derived from a multidimensional scaling analysis of the forms in terms of semantic differential scales and physical measures was not possible. Whereas the preference responses could be clustered into four different groups, all of the !S exhibited a similar strategy in evaluating complexity (as defined by Attneave and Arnoult-type forms). Three judgment dimensions were determined from a multidimensional scaling analysis which, interpreted psychophysically, were designated "Jaggedness, Perceived Sidedness" and "Asymmetry."