Mixing it up: Safety implications when different transport modes interact

Ian Walker

Department of Psychology, University of Bath

11 July 2006


When researchers talk about ‘vulnerable road users’, they usually use the word ‘vulnerable’ to describe the transport modes with which people lack physical protection in the event of a crash. However, the vulnerability of groups such as pedestrians, bicyclists and motorcyclists goes beyond this, such that as well as being physically vulnerable in the event of a mishap, these people are also vulnerable because they are at elevated risk of accidents happening in the first place, when compared to the typical motorist. These groups have tended to be considered separately in the past; however, I believe there is actually much to gain from looking at them together and aiming to identify common themes in the causes and circumstances of their road accidents. Not only does spotting themes common to various vulnerable groups allow insights from one mode to provide benefits to the others, but also identifying accident causes that apply to more than one group makes it easier to provide countermeasures, as their introduction becomes more practical as the number of people who potentially benefit grows. Here, to illustrate these points, I briefly discuss a couple of major common themes that apply to all the vulnerable groups mentioned above

Drivers’ attention

When people have studied the circumstances of accidents involving bicyclists, pedestrians and motorcyclists, they have tended to consider the three groups separately. But looking at the various accident tallies that exist, it is possible to spot common patterns across the groups. This is not to say that the three classes of road-user all have the same types of accidents for the same reasons. For example, the majority of motorcyclists’ injuries arise from incidents in which no other person is involved; this is very much the opposite of pedestrians, whose slower speeds means they only really suffer single-party injuries in extreme circumstances, such as when footways are covered in ice.

One circumstance that very clearly does cut across the various vulnerable groups, however, is that accidents where another vehicle is involved, and where the vulnerable road-user is not responsible, typically take place near junctions. For pedestrians, this is most clearly seen outside the United Kingdom. Here, we tend to use such rigid segregation of pedestrians and vehicles that when a pedestrian crosses the road, they can be confident they will not share the space with a vehicle; similarly, drivers can turn corners knowing there should be no pedestrians crossing ahead. This is not the case in, amongst others, the United States, New Zealand, and many parts of Continental Europe. In these countries, a system is often used whereby motorists are permitted by traffic lights to turn corners at the same time that pedestrians are signalled to cross the road around the corner. The onus then falls on the motorist to be vigilant for crossing pedestrians as they round the bend and to pause if somebody happens to be crossing (one reason this is primarily an overseas phenomenon is that it particularly suits cities laid out in a grid arrangement, which are rare in the UK).

Perhaps unsurprisingly, this system often leads to collisions as drivers rounding corners fail to spot crossing pedestrians (I happen to be writing this in Athens, where I have been watching with interest the startled responses of fellow Britons as they spy cars driving unexpectedly towards them on pelican crossings). These collisions are not happening simply because there is no time for drivers to spot the pedestrians as they turn, because the accidents still happen often in ‘long turns’, where the driver crosses the opposing carriageway (equivalent to a right-turn in the United Kingdom), thus providing several seconds during which the pedestrian crossing is visible. These driving errors must, then, often have their basis in a failure of attention. At present there is no direct evidence on this, but the idea that attention is involved in these pedestrian accidents is supported by evidence of drivers’ attention being involved in accidents with bicyclists and motorcyclists, specifically in ‘looked-but-failed-so-see’ collisions, in which the motorist looked in the correct direction but still failed to become aware of the vulnerable road-user.

Although often seen by bicyclists and motorcyclists as a glib excuse employed by careless drivers, there is reason to believe that the looked-but-failed-to-see phenomenon has a real basis. Although experienced drivers have far fewer collisions overall than novices, they are more likely than novices to fail to spot vulnerable road-users as they negotiate junctions. Studies using eye-tracking equipment suggest this is because more experienced motorists look at fewer parts of the road as they prepare to manoeuvre at a junction, often completely overlooking the parts of the road where groups like bicyclists tend to be found.

The reason experienced drivers look at these areas less is that in driving, as with most tasks, people try to carry out the necessary steps as efficiently as possible. In order to do this, they use their understanding of traffic and junctions – an understanding based on their experience to date – to guide their attention to the places where dangers are likely to be found. To take an extreme example, although drivers often look left and right at junctions, it is rare for a driver to look upwards, as their experience tells them that there is nothing in the sky likely to be relevant. More subtly, we know experienced drivers tend not to look at the edges of roads when gazing around at junctions, focusing instead on the centre of the lane as they search for traffic. They do this because again, their experience tells them that this is the place where obstacles are likely to be found – an assumption that is often valid, as the majority of a driver’s on-road encounters are with other motor vehicles, and so visual search is optimised to seek these. Further evidence that drivers attend to the places where motor vehicles are found and overlook places where vulnerable groups tend to be comes from the finding that as the number of pedestrians or bicyclists in a particular city rises, the number of collisions drops. This is not just because there are fewer drivers on the road because more of them are walking or cycling; rather, the greater number of encounters drivers have with these groups means that they – largely subconsciously – become more aware that pedestrians and bicyclists might be encountered at any given junction and so include their likely locations in visual searches.

Social processing

In 2005, I published a study whose results are perhaps somewhat surprising. People were shown photographs of roads and asked to describe the scenes in their own words. This revealed a very strong tendency for the language people used to reflect fundamental differences in how various forms of transport are perceived. Cars, for instance, were almost always described using impersonal terms: people would write ‘a car is turning left…’ rather than ‘a driver is turning left…’ or ‘a woman is turning left…’. Bicyclists, in contrast, were nearly always described in personal terms: ‘a man on a bike is turning left…’ rather than ‘a bicycle is turning left…’. This reflects people’s perceptions of vulnerable road users as fellow human beings but motorists as non-human: effectively as machines.

This difference in perception seems also to affect road safety. Human beings are intensely social animals, specialized for living in groups. One consequence of this is that when we see human objects, our minds and brains treat them as qualitatively different to any other object we might see (indeed, you have probably just been made slightly uncomfortable by my using the word ‘object’ to describe people, so familiar are we with the idea of them being somehow special amongst all the things to be found in the world). Further experiments carried out in Bath, in collaboration with Mark Brosnan and Nigel Holt, revealed that retaining a human appearance on the road seemed to make vulnerable road-users more likely to have accidents. We used gaze-tracking to show that drivers, upon seeing a bicyclist on the road and trying to judge their intentions, tend immediately to direct their gaze to the bicyclist’s face, even if the rider is giving an arm-signal which is a much more useful source of information. We also found that making eye-contact with a bicyclist at a junction slowed down drivers’ decisions about what that bicyclist was likely to do next, and that when there was no formal arm-signal, drivers were divided in situations where they judged the intentions of a rider who was seen gazing across a junction. Just under half the participants consistently took this gaze signal to mean the rider planned to turn the corner whereas the rest took exactly the same gesture to mean that the rider was not going to turn. Obviously, it is highly undesirable for drivers to interpret a given signal in different ways. Moreover, whether they took the gaze cue to indicate a turn or not, all drivers were particularly slow to make a decision in this situation. The process of interpreting another person’s facial signals – a process which people probably engage involuntarily – was slowing down drivers’ judgements fairly dramatically.


There is much to be gained by considering the various vulnerable road-user groups together rather than in isolation. Doing so, we see that collisions involving these groups and a motorist frequently occur in the same circumstances: at junctions as a result of a drivers’ attention failures caused by their not expecting to encounter vulnerable groups. Moreover, the fact vulnerable groups are not cocooned by vehicles and thus retain a fundamentally human appearance whilst travelling has several effects on the speed and accuracy of drivers’ decisions because upon seeing what is palpably another human being, the drivers are forced to engage in mental processing related to interpersonal perception, which is time consuming; the perceptual processes’ aim of identifying other people’s intentions renders the drivers liable to make misunderstandings regarding what the vulnerable groups intend to do – misunderstandings which would not arise if the vulnerable groups did not retain their human appearance. Investigation is now needed into the extent to which driver training can help overcome these effects, and into actions vulnerable groups can take to reduce these effects.

Ian Walker