The Effect of Social Groups and Gender on Pedestrian Behaviour Immediately in Front of Bottlenecks

Pedestrian crowds are unlikely to be composed of identical individuals. To understand how pedestrians interact, it is important to investigate what effect differences between individuals have on the observed movement dynamics. Two key aspects that can distinguish individuals are their gender and their membership in small social groups, such as families or groups of friends. Here, I report findings from a controlled experiment in which a pedestrian crowd has to pass a narrow bottleneck. In one experimental treatment, participants are asked to move independently from others and in the other treatment, they are asked to move in groups of four individuals. Trajectories of individuals, as well as group membership and gender are recorded. Investigating egress times produces conflicting results across two separate crowds in different experimental locations and I therefore investigate microscopic pedestrian behaviour. I present statistical models that capture how pedestrian behaviour immediately in front of the bottleneck affects the time gap between consecutive pedestrians passing through the bottleneck. This provides a rigorous approach to test whether individuals from the same social group or gender interact differently to individuals from different groups or genders. My analysis suggests that being from the same or from different social groups does not have an effect on time intervals between pedestrians. This suggests that in my data, group membership does on average not affect behaviour immediately in front of the bottleneck. Interestingly, my analysis suggests a gender effect: men follow women more closely than vice versa. Possible explanations for this range from the way mixed-gender couples move to physical differences between genders. These findings demonstrate the potential of my approach to uncover detailed aspects of interactions underlying pedestrian behaviour.