Environmental Preference

Abstract

This article examines four domains of variables to assess their relative merit in explaining environmental preference. Within each of the domains, between three and seven specific attributes were measured, for a total of 20 predictor variables. The study site includes small forested areas, agricultural land, and fields, with little topographic variation. Preference ratings of 59 scenes representing the area serve as the dependent variable. Taken together, the 20 attributes accounted for 83 percent of the preference variance. Taken separately, the Physical Attributes lacked predictive power. Of the Informational variables, Mystery was the only significant contributor. The Land Cover types proved effective, with Weedy Fields, Scrubland, and Agriculture all significant negative predictors. Finally, the Perception-based variables were most powerful, with Openness and Smoothness particularly useful predictors. The results point to the importance of using different predictor domains, rather than relying exclusively on any one, since their role in different environmental contexts is likely to vary.