"This research examines similarities and differences in aesthetic response asa function of knowledge structure. We had 60 architecture students and 60non-architecture students rate 40 homes (20 "popular" styles and 20"high" styles) on four scales. The scales included two measures of knowledgestructure (goodness-of-example, and familiarity) and two measures ofeasthetic response (preference, and interest). The groups exhibited similarknowledge structures but differences in aesthetic response. However, thepattern of responses was consistent with the knowledge structure model.Finally, we Identified relationships between scene attributes and bothknowledge structure and affect.The present paper describes a model of the process underlying the similari-ties and differences in environmental evaluatin (cf. Zube, Sell and Taylor,1982; Purcell, 1984) and tests the model in relation to emotional/aesthetic responses. The model posits environmental experience as resulting from similarities between the particular instance and the representation inmemory of past experience related to the place or building type (Mandler,1984). Similarities and differences in affect can therefore result from dif-ferences in either knowledge structures or in how individuals evaluate mis-matches to their knowledge structure.We expected to find similar knowledge structures across the groups but dif-ferences in how they evaluated mismatches; and we expected the knowledgestructure and affect to be linked to attributes of the scenes.Method: We interviewed 120 students (60 architects and 60 non architects)at the University of Sydney and had each assess 40 American single-familyhouses on one of the four 100-point scales. The scales included two meas-ures of knowledge structure, goodness-of example, and familiarity, and twomeasures of affect, interest, and preference (cf. Berlyne, 1974; Kaplan andKaplan, 1982; Mandler, 1984). Displayed in color slides, the houses in-cluded 20 "popular style and 20 "high" style houses (see Devlin and Nasar,1989).Results and discussion.The results support the model linking knowledgestrusture to similarities and differences between groups and they suggest anenvironmental bases for the responses. We found a close similarity betweenthe architects and non-architecsts in terms of goodness-of -example and fa-miliarity. Both groups judged the "popular style houses as the better ex-amples. The American houses differ from the typical house In Sydney (thestudy site), yet both groups judged the "popular" styles as the better exam-ples. Perhaps some basic house attribute common to both cultures affectedthe knwledge structure.For both groups, familiarity related directly to goodness-of-example. Fur-thermore, for a given level of goodness-of-example (particularly in themiddle of the range), both groups preferred the unfamiliar. Both groupsrated Interest as Inversely related to goodness-of-example.The groups differed In their rating of the interestingness of "popular"styles. As goodness of example improved, non-architect interest increasedand architect interest decreased. Overall both groups judged increments intypicality as less interesting, but this effect was more pronounced amongthe architects. The main differences, however, emerged for preference. Non archiect pre-ferece across the full set of houses increased with goodness of example,while the achitects preferred moderate to high discrepancies from the goodexamples. Similar differences emerged within the "popular" styles and the"high" styles separately. The non-architect response can be seen as a famil-iarity effect (preference for a prototype) ad the architect response as fit-ting the discrepancy model (preference for a discrepancy from the schema).The paper characterized the physical features of the scenes and found thekowledge structure and affect related to those features. For exampe, judge-ments of goodness of example related to rectilinear forms, asymetry, hiproofs and popular styles."