Evaluations of landscape preference, complexity, and coherence for designed digital landscape models

Abstract

Few studies have investigated whether intentionally manipulating objective measures related to the informational concepts of complexity and coherence, within the context of natural and built settings, affects respondents’ preference ratings and estimations of complexity and coherence. Our study addresses this research need. Color digital landscape model views depicted three variations in the number of unique plant species (complexity), translated into Shannon’s information entropy bit values, and three methods of plant organization (coherence)—formal, clustered, and scattershot—within field, residential, and urban settings. Participants viewed and evaluated the digital landscape model views for preference (n=77) or estimated the presence of complexity (n=34) or coherence (n=38). Strong, direct correlations resulted between respondents’ estimations of complexity and designed entropy bit values. Respondents’ estimations of coherence inversely correlated with the number of regions depicted in model views that represent the methods of plant organization. Preference did not correlate with estimations of coherence or designed entropy values, but did correlate with estimations of complexity. Repeated measures ANOVA test results suggest that respondents’ preference for scenes increased as entropy values increased between two and four bits, and that plants arranged in clusters were liked more than scattershot or formal compositions. Moreover, respondents liked residential settings significantly more than urban settings, and preference increased significantly more with increasing entropy values in urban settings than in residential. The interaction between setting and entropy values for depictions of clustered and formal plant arrangements affected preference differently in each setting.