Tourists and others who are unfamiliar with an environment may be sensitive to environmental cues when choosing their walking routes. In this study, we combined inductive and deductive approaches to evaluate walkability cues. We defined a set of walkability variables by analyzing the narratives of par- ticipants, who walked along one of 19 diverse routes. These cues were then supplemented with environmental and walkability variables from Mehrabian and Russel scales (Russel et al., 1981) and SPACES (Pikora, 2003), resulting in a total of 48 descriptors. Using the 48 descriptors, 60 diverse photos of streetscape views were rated, and we identified 14 relevant walkability corre- lates. Using principal component analysis, we identified six components that best predicted walking decisions: safety from traffic, comfort of walking area, environmental appearance, activity potential, shade and exploration. These results suggest that real walkers make more finely grained walking judgments than those measured by current, conceptualized walkability scales.
