If we watch people walking along a street it is noticeable that they very rarely bump into one another. Furthermore, they seem to avoid each other with very little difficulty. This raises the interesting question of how collision avoidance amongst pedestrians is achieved. Several studies have been carried out to investigate the behaviour of pedestrians, and descriptions of these can be found in Hirsch (1970), Henderson and Lyons (1972), Morris (1967) and Stilitz (1970).
Such studies reveal that people tend to keep to the right of the pavement: thus some will be walking on the inside, away from the road, and others, walking in the opposite direction, on the outside and close to the road. Goffman (1972) has termed this behaviour ‘lane formation’, whilst Collett and Marsh (1981) have called it ‘pedestrian streaming’. Where lane formation occurs, those pedestrians who wish to quicken their pace may either weave their way through their own lane or along the interface between opposing lanes. On a crowded pavement, pedestrians frequently take to the road when lane formation is absent.
Although it may not be immediately obvious, some means clearly exist whereby lane formation is maintained and collisions avoided. It has been suggested that two processes are involved, which may be conscious or unconscious. Wolff (1973) has described how pedestrians scan the faces of those coming towards them and terms this behaviour ‘monitoring’. Goffman (1972) has called the other process ‘externalization’ or ‘body gloss’. This refers to observable body movements conveying information concerning someone’s likely behaviour to both those approaching and following. For example, where an area is very crowded, a characteristic avoidance movement involving a slight turn of the shoulders and a hardly noticeable side step has been observed. Wolff (1973) refers to this as ‘step and slide’.
Collett and Marsh (1981) were particularly interested in the relationship between collision avoidance and the processes of monitoring and externalization. They found that the best place to study collision avoidance was a controlled pedestrian crossing, since when the lights change and pedestrians cross they are obliged to find a way past each other. The movements of pedestrians were recorded over several days by means of a portable video-recorder and zoom lens from a building overlooking the crossing. The video-taped data were then analysed.
Several aspects of behaviour were studied, including passes between individuals, the direction of passes (to the left or right), the orientation of people when passing, and the position of their arms. The gender and approximate age of the individuals were also noted. Of particular relevance to the first exercise presented here are the passes that were made between single individuals.
When two adults are moving together with what Collett and Marsh term an ‘approach overlap’, they will collide at some stage unless one or both alters direction. If this is not done early enough then one or both pedestrians has to take avoidance action. Collett and Marsh found that men and women differed with respect to the way in which they orientated themselves to the person they were passing. Men tended to turn towards the person they were passing (termed an ‘open pass’) while women tended to turn away (termed a ‘closed pass’). This difference in collision avoidance behaviour between the sexes was found to be statistically highly significant. The difference is illustrated in Figure 1.
Further examination of the data showed that the type of pass was not simply a natural orientation according to which leg was forward at the time of passing. Even when a closed pass needed more effort than an open one, women were still significantly more likely to angle themselves in this way. Collett and Marsh’s explanation for this finding is that women wish to protect their breasts. Their explanation is supported by the movement of the arms during a pass: more women than men drew one or both their arms across their body when passing another person. Again, this difference was statistically highly significant.
Kendon and Ferber (1973) found similar behaviour in their analysis of greetings and salutations. Significantly more women were observed to draw an arm across their bodies immediately prior to a close salutation. However, Kendon and Ferber pointed out that this might simply be due to the transference of an article such as a purse or a newspaper from one hand to the other in order to prepare for the greeting. Collett and Marsh therefore excluded from their analysis all those who were carrying something. However, the sex difference remained. Thus the findings regarding arm movements do seem to support Collett and Marsh’s view that female behaviour is self-protective.
Other research, which may have relevance to the idea of open and closed passes, is that reported by Jenni and Jenni (1976) who investigated book-carrying behaviour in students. The researchers identified two ways in which books tended to be carried. In one, the books were held supported against the body, whilst in the other, the books were carried at the side of the body with the long edge more or less horizontal. Jenni and Jenni’s results indicated that approximately 90% of female students used the first method and approximately the same number of male students the second. This strong gender difference is reminiscent of the differences in collision avoidance behaviour between adult male and female pedestrians, and it is possible that the difference in book carrying behaviour is due to anatomical factors and, in women, represents a form of self-protection.
One further finding reported by Collett and Marsh is of interest. As mentioned earlier, Goffman (1972) and Wolff (1973) reported that pedestrian streams tend to form on the right. Collett and Marsh found that, when adult pedestrian density reached a certain level, four streams developed as though two pavements were placed side by side, and that there was a clear pattern of walking on the right. The researchers wanted to know whether pedestrians avoiding collision moved to the right when no lane formation occurred.
Of course, when two pedestrians are not directly opposite one another, the direction of pass is determined by the position of the people concerned relative to one another. Thus, if right shoulders are overlapping people will move to the left. However, when two people are approaching each other head-on, a choice between moving left or right can be made. Collett and Marsh found that, in such a case, almost twice as many pairs of pedestrians moved to the right as the left. This finding would seem to be well established: pedestrians tend to move towards, and walk on, the right. (Taken from Mcilveen, R., Higgins, L., & Wadeley, A. (1992) BPS Manual of Psychology Practicals. Exeter: BPCC Wheatons Ltd.)
The aim of the first exercise presented here is to attempt to replicate Collett and Marsh’s (1981) finding that adult men and women orient themselves differently when avoiding a collision. Replication is an important part of the scientific process, and an attempt to confirm Collett and Marsh’s findings would further our knowledge about pedestrian behaviour. On the basis of their research, it may be hypothesized that there will be a significant difference in the collision avoidance behaviour of adult men and women such that women will tend to make more closed than open passes and men more open than closed passes when avoiding other pedestrians. The null hypothesis predicts no significant differences in the type of collision avoidance behaviour of men and women.