5 studies of aesthetic preference for abstract designs varying in number of elements (n) are reported. Stimulus materials included designs which were random or composed, symmetrical or asymmetrical, and which were presented as drawings, photographs, or slides. Preference rankings or ratings (P') were obtained. For 4 sets of stimuli, but not random polygons, significant inter‐S concordance was found even with heterogeneous samples. Both group and individual trends gave an inverted‐U curve for P' on n, in most cases. In the main study, dispersion of P' scores increased with n. These results are largely compatible with the “information‐processing” approach to pr