Reliability and validity of selected responses to the visual properties of an architectural environmental display

In a study of the ability of a simulation technique (color photographic slides) to elicit descriptive and affective responses comparable to those elicited by a more direct experience of an architectural setting (guided walk) as measured by the same adjective rating scales, several disturbing findings emerged to challenge not only the use of simulation techniques, but the general usefulness of subjective ratings of environments. Comparisons of responses from one group who viewed the slides to responses from another group who took the guided walk demonstrated an unprecedented degree of agreement (validity) for both descriptive and affective responses. However, further comparisons of those data to responses from each of two control groups, one of whom was not exposed to either the slides or the guided walk, also revealed virtually no difference. That the agreements among the four groups were so extremely high leads to the conclusion that such data are determined mainly by instrument bias, or instrument variance. The implications of such findings are most clearly put into perspective by considering the large percentage of current and past environment-behavior research that has employed such a simple, direct, subjective rating method.