Analysis of Pedestrians’ Behavior: A Segmentation Approach Based on Latent Variables
Peer reviewed, Journal article
MetadataShow full item record
Crash statistics indicate that the number of pedestrian fatalities has been increasing at an alarming rate in Iran. Crossing over non-designed places is a main cause of traffic crashes among pedestrians. This study aimed to investigate how perceptions about walking facilities and risk-taking affect pedestrians’ crossing behaviour. A stated preference questionnaire was designed and a random sample of 390 pedestrians were interviewed face-to-face in two regions of Tehran with three options for pedestrians to cross (overpass, zebra crossing, and non-designed places (NDP)). Exploratory factor analysis (EFA) resulted in three latent dimensions: risk-taking/conformity, pedestrians’ perception of overpass, and NDP. Then, data were classified based on latent variables using K-means cluster analysis. Clustering resulted in four groups: group 1 (Cautious; negative perception of overpasses; positive perception of NDP), group 2 (Cautious; negative perception of overpasses; negative perception of NDP), group 3 (Cautious; positive perception of overpasses; negative perception of NDP), and group 4 (Risk-taker; negative perception of overpasses; negative perception of NDP). Finally, a Multinomial Logit Model (MNL) was developed for four groups of pedestrians. The results show that pedestrians’ behaviour differentiate based on latent variables. It was found that being accompanied by a child increases the probability of using an overpass even for pedestrians in group 4 with high risk-taking propensity, but it was more important for pedestrians in group 3 who held positive perceptions of overpasses and negative perceptions of NDP. Also, during congestion, group 4 was more inclined to cross at NDP. It was concluded that in the first group, unsafe choices among student respondents could be associated with their facility perceptions rather than their risk-taking/conformity. Results of this study can be helpful in selecting more appropriate locations for overpasses and crosswalks installation based on pedestrians’ behaviour.