Do cars pass with less space for bicyclists who wear helmets?

Question

This question is an off-shoot of from the discussion about helmet safety. There is an assertion that when a car passes a bicycle, it allows less space to those who wear a helmet than those without. While this is metric is only one part of the larger picture regarding bicycle safety, it is often cited by helmet skeptics as scientific evidence to show wearing helmets has a negative effect on safety.

The assertion seems to base on a single study by Ian Walker. He says

"Drivers also tended to pass notably closer to the rider when he wore a helmet than when he did not."

And he reasons that

"The helmet effect is likely the result of drivers judging cyclists’ skill levels from their appearance and adjusting their overtaking accordingly"

I'm rather skeptical of his study. The chief problem is, in the eyes of a driver, bicycle helmets are not a salient feature. The driver's cognitive capacity is already taxed in judging the relative speed and space required for passing and often have to make a move in a short time window. There is little capacity to check out other details like helmets, let alone to use this to make a judgment and consciously vary his driving strategy.

Has anyone taken a critical look at his study? Does anyone have another related result?

Answer 1

[An aside, before I start: I think your question about the mechanism is a reasonable one, but I trust your skepticism, upon gathering more evidence, will lead you to provisionally accepting the evidence of a peer-reviewed controlled scientific experiment over a gut feel, if it turns out that the study holds it own! In particular, it doesn't take many drivers to be held up (or at least to perceive it that way) behind a cyclist, who have time to register the helmet and to cut too close to the rider, to sway the statistics.]

I looked for any replication or criticism of the study. I started by finding the original study on Google Scholar. Then I skimmed through the 28 citations. (Google Scholar may not be the best Citation Index, but it is easy and free.)

I couldn't see any replications (although this instrumented bike project sounded like a promising source.) It was (positively) cited in a number of other papers (e.g.), including reference to the helmets, suggesting other peers had found the experiment valid.

I also discovered that Ian Walker had a personal blog. In one blog post, he explained he had achieved some infamy from the original article and highlighted a draft rebuttal by Dan Gutierrez (which includes an experimental result which, while not replicating the same helmet versus non-helmet experiment, suggested the helmet differences was minor compared to other factors.). Ian Walker then provided a defence to that rebuttal.

In conclusion, yes, Ian Walker's paper has provoked a reaction from the cycling community, but the scientific community have, by and large, accepted the result. I found no replications, but hope that the instrumented bike project might offer such replications in the future. This was a rapidly performed literature review; a more detailed one may provide more support to either side.

Answer 2

You may want to check out some psychology of risk literature. I remember chatting with a psychology professor who specialised in risk and he pointed out that filling cars with safety devices (like air-bags and seat-belts) seems to encourage risk taking behaviour on the part of drivers, because the environment is seen as being "safer".

In regards to your example I might imagine that car users who see a helmet wearing cyclist may feel that the situation is safer and thus they take more risk in passing.

I'm sure a psychinfo search could give you more. Be safe, or not.

Update: A quick search of psychinfo turns up this abstract:

There is evidence linking certain vehicle characteristics to crash involvement and one possible mechanism behind this relationship is that these vehicle characteristics influence drivers' risk-taking behaviour. In order to investigate this, we conducted a roadside observation survey and a questionnaire-based study. Both revealed a significant relationship between vehicle performance and drivers' risk-taking behaviour. The causal direction of this relationship has important consequences. If drivers' risk taking predicts their car choice, then it could be justifiably argued that individuals who take more risks when driving simply choose more powerful vehicles to facilitate their behaviour. However, if it is the case that vehicle characteristics adversely influence drivers' risk-taking propensity then this has implications for vehicle design. Results indicated that the causal pathway operates independently in both directions. Finally, we sought to determine which vehicle characteristics influenced risk-taking intentions independently of other confounded characteristics. We found that high vehicle performance and a greater number of safety features led independently to greater intended risk taking in general, while higher internal car noise led to closer car following and more risky gap acceptance, but not to greater speed. Vehicle smoothness and handling did not affect risk-taking intentions.

Source: The effect of vehicle characteristics on drivers' risk-taking behaviour 2002

Answer 3

The key source for analysis of risk-adjusting behaviour is the work of John Adams who has long argued that people have a certain tolerance for risk and tend to adjust their behaviour to (partially) compensate for safety devices. In the case of cycle helmets both car drivers and cyclists themselves might take more risks because the cyclist is wearing a helmet.

Adams is an academic at University College London who has published some seminal works on this topic. He has made strong criticism of the way governments have used evidence to exaggerate the benefit from safety rules such as compulsory seat belts in cars.

His basic idea is summarised well on his blog:

Because seat belts are undeniably effective at reducing death and injury in crashes there is, or was, a mystery. Why in country after country that mandated seat belts was it impossible to see the promised reduction in road accident fatalities? The most plausible explanation is “risk compensation”. It appears that measures that protect drivers from the consequences of bad driving encourage bad driving. The principal effect of seat belt legislation has been a shift in the burden of risk from those already best protected in cars, to the most vulnerable, pedestrians and cyclists, outside cars.

I quote this because exactly the same argument is relevant to the debate about cycle helmets. While it is obvious that a cyclist who has an accident will be less likely to suffer injury if they wear a helmet we also have to take into account whether they are more likely to have an accident if they wear a helmet. So the many studies and anecdotes that look at the population of injured cyclists with and without helmets are statistically irrelevant as they look at the wrong sample to actually answer the important question.

The relevance here is that riskier behaviour for both cyclists and drivers is perfectly plausible.