There is empirical evidence that neighborhood environment characteristics influence individuals' self-reported affective walking experiences. However, much of the research investigates residents' affective walking experiences at the neighborhood level using revealed-preference methodologies, making it difficult to identify the separate impacts of characteristics. In addition, empirical studies have not shown enough evidence that individuals from different sociodemographic backgrounds have distinct affective walking experiences. Therefore, the objective of this paper is to explain how different groups of people perceive the characteristics of a neighborhood differently. To do this, this study conducts a conjoint experiment employing videos of virtual environments involving a sample of 295 respondents. Using a latent-class regression model and a multinomial logit model, we are able to determine how individuals and groups perceive neighborhood characteristics differently based on their different emotions. The results somewhat confirmed the findings of the empirical research, indicating that land use mix, connectivity, road size, open space, and greenery are related to a positive walking experience. The level of affective walking experience that individuals associate with neighborhood environmental characteristics is, however, considerably variable. Therefore, our results show that open space and road width are crucial for a walkable neighborhood since they are most helpful to individuals' subjective well-being.
Keywords: conjoint experiment; latent-class analysis; neighborhood environment; virtual environment; walking experience.