Advocacy & Safety - Scientific American article on helmets, passing clearance, etc.

The last page of Scientific American, October 2007, has a brief article, “Do Helmets Attract Cars to Cyclists?”. It cites the recent study by Ian Walker in in Accident Analysis and Prevention (abstract LINK (http://www.ncbi.nlm.nih.gov/sites/entrez?Db=pubmed&Cmd=ShowDetailView&TermToSearch=17064655&ordinalpos=2&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum)) that received popular press coverage.

R. Swart, founder of the Bicycle Helmet Safety Institute, thought that Walker's study may mislead cyclists since “The cars were already giving him ... a very wide passing clearance already.” He noted that most vehicles typically stayed well over three feet, rendering the 3.35 inch discrepancy insignificant.

In response, Walker re-analyzed his data. He found that that 23 percent more vehicles came within one meter (presumably a risk) when he wore a helmet.

The Scientific American article goes on to mention a few other recent studies on helmets.

Unfortunately, it is difficult to have a thoughtful discussion about such complex safety issues, particularly on an Internet forum. Helmet manufacturers may favor helmet use or legislation but I suspect would resist efforts to make the helmet standards more stringent or limit color choices to ANSI orange or lime-yellow (it's challenge for me to find choices of solid yellow color helmets). Individuals or institutions that have formed opinions or made a decisions on either side of the question are often unwilling to reconsider new findings or approaches. For example, if helmets increase the percent of close passes, would measures such as a horizontal flag (e.g. a FlashFlag) counter the effect?

Teasing out what are the causes and effects is also challenging because it may require a careful multivariate statistical analysis (e.g. ) and the relevant denominator (number of people exposed) is not easy to measure.

Finally, Walker's paper has several interesting findings besides the helmet data that received popular press attention. For example, see the lane positioning versus passing distance data in Figure 1 attached below.

Walker discusses this. “... the most likely origin of the idea that riding further out leads to greater passing leeways can perhaps be inferred from Fig. 1 where, although the space between passing vehicles and the rider dropped as the rider’s distance from the kerb increased, it did not drop as much as 25 cm for each 25 cm shift in rider position. This means that on average, as the rider moved further towards the centre of the road, the absolute position of overtaking motorists indeed shifted outwards – a 1mshift in riding position led to a 73 cm shift in driving position – even though relative to the rider they came closer. This seems the most likely cause of the misperception common amongst bicyclists.”
________________________________________________________________________________
"For every complex problem, there is a solution that is simple, neat, and wrong,"
H.L. Mencken

"I have yet to see any problem, however complicated, which when you looked at in the right way, did not become still more complicated."
Poul Anderson (in New Scientist, 9/25/69)

This paper has been much discussed on these pages and also had mainstream coverage by ABC TV, and USA Today.

Scientific American had an article on the web (still active)

http://www.sciam.com/article.cfm?articleId=778EF0AB-E7F2-99DF-3594A60E4D9A76B2&chanId=sa013&modsrc=most_popular

In the article, Walker notes that when it comes to riding in traffic, motorists are the real problem.

Randy Swart, founder of the Bicycle Helmet Safety Institute (BHSI), says that studies such as Walker's run the risk of misleading cyclists as to the effectiveness of helmets but that point is addressed by noting Dorothy Robinson, a patron of the Bicycle Helmet Research Foundation and a senior statistician at the University of New England in Armidale, Australia, publishing a 2006 review article in the BMJ (British Medical Journal) about regions in Australia, New Zealand and Canada that introduced legislation that spurred an over 40 percent increase in bicycle helmet use among their populaces. ...The newly instituted (MH) laws, she found, did not have a significant effect on bicycle accidents resulting in head injuries... Her conclusion was "helmets are not designed for forces often encountered in collisions with motor vehicles"

I'd say that Randy Swart misleads people about the effectiveness of helmets if he implies that the helmets cyclists wear provide effective protection in collision with motorists.

If drivers treated cyclists with due care and respect and cyclists treated the rules of the road in the same fashion, we will all end up keeping everyone much safer.

Too many people get caught up in the "to or not to" wear a helmet angle, but that's just a hook. The point of the article is to determine what gets drivers to give safe distance to cyclists, because its is motorists that are the real problem for cyclists.

That anomaly at 1.00 m is interesting. I didn't read the full Accident Analysis & Prevention article, so I might've missed something.

Did anyone read Walker's journal article? (I'd rather read the journal than SciAm). A few things I'd like to mention in regards to it:

- While Walker mentions that accident data is fairly consistant between the United Kingdom, the United States, and New Zealand, that doesn't necessarily correlate to consistent driving behaviour and/or passing distance.

- Concerning the above point, SUV drivers (including trucks) passed with the same (safest) distance as other cars, whereas commercial vehicles passed with the lowest safety margin. In the UK, the SUV/Truck:Car ratio is far different than from the US. Even here in Canada, SUVs/trucks make up almost half the traffic, whereas that might be different in the UK (though that's just my perception). And, SUV drivers are stereotypically (and justly so) thought of as being road rage maniacs. Here, it's moreso with pickup trucks.

- Walker said that relative position (concerning lane width) couldn't be determined and wasn't very relevant. However, commercial vehicles (the large ones) were consistently noted as having a smaller passing margin. Depending on the types of roads (British roads are smaller than American roads, and may be in certain cases contraflow), larger vehicles might not be able to pass with as much of a margin as they would like. When I bike, buses change lanes to pass me (when it occurs...I'm usually too fast for that!), whereas the study noted buses as being the most notorious offenders. However, the taxi data seems reasonable...crazy nutters. In his discussion, Walker does not that it typically takes a much longer time for a larger vehicle to overtake the cyclist (4 s as opposed to 0.5-1 s with a car).

- Walker actually dressed in a long-haired wig to look like a female. Haha.

- Walker used a Trek hybrid bike and a white Giro helmet.

- There is no mention at how faithfully Walker kept his laser-marked distance from the kerb. However, such data would only be qualitative, and quantitative anaylsis would have been complicated based on the "dynamic" nature of the kerb (such as with the presence of parked cars, road debris, etc.).

- Walker was actually hit twice (both by commercial vehicles, both with a helmet).

It's an interesting read, and I'd like to read up on some of the sources Walker cites. But, I've got exams to study for :-(.

Whats also interesting is that the further away from the curb you get, whether wearing a helmet or not, the closer the pass.

What does this say about taking the lane? Maybe nothing because the data have to do with passes and the whole idea with taking the lane is to avoid a pass.

My anecdotal evidence is that when I take the lane, most (but not all) drivers wait until oncoming traffic is clear and then go completely in the next lane to pass, treating me as if I were a car. The closer passes are when I stay to the right and the cars stay in the right lane and pass.

Whats also interesting is that the further away from the curb you get, whether wearing a helmet or not, the closer the pass.

What does this say about taking the lane? Maybe nothing because the data have to do with passes and the whole idea with taking the lane is to avoid a pass.

My anecdotal evidence is that when I take the lane, most (but not all) drivers wait until oncoming traffic is clear and then go completely in the next lane to pass, treating me as if I were a car. The closer passes are when I stay to the right and the cars stay in the right lane and pass.Well, he was on a contraflow/2-way road.


"The data showed that drivers of buses and heavy goods vehicles
passed the rider much closer than other drivers, and indeed
the author was struck by both classes during the experiment.
This accords with the high number of bus overtaking incidents
reported in the OxCam Survey (Walker and Jones, 2005) and
the effect cannot simply be a product of driving for professional
reasons, as light goods vehicles left significantly more room.
The most likely explanation arises from vehicle characteristics.
Owing to their length and poor acceleration, buses and heavy
goods vehicles take much longer to pass a bicyclist than shorter
vehicles (our data showthat a standard bus typically takes around
4 s to overtake, in contrast to 0.5–1.0 s for a car). This means that
to pass safely, a driver must encroach onto the oncoming traffic
lane for a long period (even with a bicyclist riding towards the
road edge, thanks to the width of these vehicles). We suggest it
is an inculcated reluctance to remain out-of-lane, coupled with a
lack of lengthy gaps in oncoming traffic and vehicle design issues
which put bicyclists out of sight long before overtaking is complete,
that often cause drivers of long vehicles to pull back across
before it is safe to do so, hence creating the close proximities and
frequent conflicts." (Walker, 423).

Interesting. I would like color tested this way. Now that it's winter I'm riding in see-me green jacket and super see-me green helmet cover. My wife says she can see me from a mile away. Almost always I'm passed in the complete oncoming lane with a very wide distance, while I'm riding along the right tire track area. I ride at that position because the road is clean there and I avoid the bad paving towards the edge, all too common on our country lanes. Unfortunately folks will give me the whole lane into blind corners and over blind hills. Occurred to me several times that I'm getting passed better since I put the helmet cover on.

I used to ride with a traffic spacer. It never got hit, although I got bumped at times riding without it. City commuting and transportational work. So testing with one of those would be interesting, too.

I've become convinced that the lurid jerseys of many colors are really "camo" for cyclists and will phase mine out in favor of yellow and see-me solid colors.

I am gonna design and market helmets which look like heads, so the problems
will go away.

I am gonna design and market helmets which look like heads, so the problems
will go away.

Make that heads with long hair, to get the most benefit...

I've found that my PB superflash makes drivers really not want to get too close to me. Don't know if it helps the efficacy of my helmet, but it can't hurt. I don't get too many close passes.

I haven't analyzed the numbers, but for one thing the effect shouldn't persist at night when conditions are most dangerous, because drivers probably don't get a good view of your headwear in the dark. I also believe the bulk of the danger while riding is from people who haven't seen you. So while you may get some more close passes with the helmet, they can't be that oblivious if they see at least see you. Also, this applies to passes from the rear which are (relatively) a lower risk than other interactions like a good t-bone at an intersection.

So in the end, I'll keep the helmet. Though as closetbiker says, they ain't perfect. But I'll take my chances with one rather than without.

Varying distance from curb from 1-4ft is only going to give data about clearance of encouraged passes. 4ft is still right far biased for many lane widths.

It also seems the person who did the study has a tendency to ride agains traffic flow - at least for photo ops. ;)
http://www.bikeforums.net/showpost.php?p=3046201&postcount=30

Al

Emiprical results obtained during contraflow cycling surveys are arguably irrelevant to vehicular cyclists who know how to ride in the proper direction. Contraflow is a splendidly dumb idea.

Emiprical results obtained during contraflow cycling surveys are arguably irrelevant to vehicular cyclists who know how to ride in the proper direction. Contraflow is a splendidly dumb idea.
Whenever I walked on a street facing traffic, I always got closer encounters. When walking with my back to traffic, motorists were usually much more careful to pass with a wide margin.

I haven't tested this on a bicycle yet. ;)

(I'm curious how one would "take the lane" riding against traffic.)

This paper has been much discussed on these pages and also had mainstream coverage by ABC TV, and USA Today...

I finally found the previous threads by using the recently improved Yahoo Advanced search, domain limited to www.bikeforums.net (http://www.bikeforums.net), after failing to find them with domain limited Google Advanced or this forum's search (in part because I restricted it to this subfourm). The threads are Thread1 (http://www.bikeforums.net/showthread.php?t=227980) Thread2 (http://www.bikeforums.net/showthread.php?t=298995) Thread3 (http://www.bikeforums.net/showthread.php?t=297486) Thread4 (http://www.bikeforums.net/showthread.php?t=315419)

Unfortunately, the popular press articles seem based on the U. of Bath's press release (http://www.bath.ac.uk/news/2006/9/11/overtaking110906.html). Press releases and subsequent popular accounts all too often get important details wrong or omitted. It is best to read the actual research publication, which is usually difficult or expensive unless you have an academic affiliation.

As far as Walker having a predetermined outcome; you are supposed to have a hypothesis to test. In fact, the results in Fig 1 (see attachment in this thread's first post) were the opposite of what he expected". This finding is contrary to what many experienced bicyclists believe (including the author prior to this study). That the data so clearly proved this belief wrong is, as well as anything else, an excellent illustration of the need for objective research over subjective experience when seeking to understand traffic behaviour." (page 422)

To me, one of the most surprising things about the study is absence of similar previous studies. The technology is not that expensive (the $695 Massa M-5000/95 distance sensor is more accurate and precise than needed; you could make an adequate one for less). Statistical software is not that expensive (R (http://www.r-project.org/), for example, is free but does have a bit of a learning curve). Alternatively, a data set such as this would be attractive to a second year statistics class or one could find and send the data to an interested cycling statistician.

Some questions I have about the paper is there is no mention of randomizing when he cycled at a particular distance or helmet status. Simply rolling a die or a few coin flips could suffice.

A second comment is that the implied statistical model is a collision occurs in the tail end of the square root transformed overtaking distance distribution (e.g. overtaking distance less than the width a large motor vehicle side mirror). Data transformation is a standard approach, but caution is needed about how close it really is to a bell-shaped Gaussian distribution far out in the tail. For example, I use the internationally recommend approach for my profession which involves parametric transformation to approximate a Gaussian distribution, comparison to non-parametric approach, and bootstrapping to locate where a small portion (typically 2.5% or less) of the sampled population's tail should be. The percent of overtaking distances less than the collision distance is much less than 1% and thus assuming a Gaussian distribution is much more problematic.

This should not affect the analysis in the paper which deals with the means (averages) of data; the central limit theorem (http://en.wikipedia.org/wiki/Central_limit_theorem) keeps those distributions a Gaussian bell-shape.

Walker does not miss there is a more central "comfort zone" passing distance: "Another factor which could profitably be explored in future is bicyclists’ perceptions of motorists’ overtaking behaviours. If we are to promote safer overtaking to motorists, it would be useful to know what leeways bicyclists feel comfortable with, and whether there are individual differences in these. It would also be useful better to understand the phenomenology of bicyclists being overtaken too closely, as this may well impact upon efforts to promote bicycle use, particularly in cities (Lingwood, personal communication)." This brings up the interesting possibility that decreased numbers of cyclists sometimes associated with mandatory helmet laws might not be people not cycling just because the helmet campaign makes people think biking is too risky; if helmeted cyclists are more frequently closely passed, fewer people may choose to cycle. Since there is evidence of a "safety in numbers" effect, fewer cyclists then increases the risks of cycling (see, for example, LINK) (http://www.transport.sa.gov.au/pdfs/personal_transport/bike_direct/cycling_strategy.pdf).

Finally, the the data could be graphed better. Figure 2 in particular is not more informative than a small table. Using quantile plots and other techniques that give informative displays of a large amount of data would be better (see publications by Tukey, Tufte, Cleveland, etc.).

All in all, I think Walker's paper is thoughtful and well worth reading. I hope it encourages even better studies in other countries with more cyclists, perhaps using a factorial design to include variables such as color and horizontal flags. Other posters suggestions of doing this type of study at night (Dinotte or Planet Bike lights, reflective vests anyone?) is a very good one.

If anyone needs more detail on the paper, please private messenger me. Attached is Fig 3 which confirms some anecdotal impressions of other Bike Forums posters regarding motor vehicle type vs. passing clearance.

Finally, Walker's paper has several interesting findings besides the helmet data that received popular press attention. For example, see the lane positioning versus passing distance data in Figure 1 attached below.

Walker discusses this. “... the most likely origin of the idea that riding further out leads to greater passing leeways can perhaps be inferred from Fig. 1 where, although the space between passing vehicles and the rider dropped as the rider’s distance from the kerb increased, it did not drop as much as 25 cm for each 25 cm shift in rider position. This means that on average, as the rider moved further towards the centre of the road, the absolute position of overtaking motorists indeed shifted outwards – a 1mshift in riding position led to a 73 cm shift in driving position – even though relative to the rider they came closer. This seems the most likely cause of the misperception common amongst bicyclists."

Interesting. There is an alternative (but very simple) way to analyse this data, which may shed light on why cyclists feel safer riding further out, even if the passes are closer.

You can see that if I ride 25cm from the kerb, I am likely to be passed with a margin of around 146cm (read from the graph, non-helmeted data). That gives me 171cm of road to ride in whilst being passed.
If I ride 125cm from the kerb, cars pass me closer (118cm, again from the graph). But that gives me 243cm of road to ride in. Whilst the driver might be closer, I've got a heck of a lot more road to take evasive action in. And to me that feels safer - if the car starts drifting towards me, I've got more chance of getting out of the way. I wonder if this extra road is what makes it feel safer, which is attributed by cyclists (who are not kitted out with proximity measurement devices, on the whole!) to being passed less closely?

It still leaves open the question of whether it actually is safer, of course...

(I'm curious how one would "take the lane" riding against traffic.)

Only after having completed ones final will and testament

Varying distance from curb from 1-4ft is only going to give data about clearance of encouraged passes. 4ft is still right far biased for many lane widths.

It also seems the person who did the study has a tendency to ride agains traffic flow - at least for photo ops. ;)
http://www.bikeforums.net/showpost.php?p=3046201&postcount=30

Al

I think, from that photo, he is turning into a narrow sidestreet (say, 7ft kerb-kerb). These are common in the uk, and usually so narrow that only one car at a time can pass down them. Parked cars have to pull up onto the curb, as you can see in the photo.

It is very important to stay well out on these narrow streets, or cars simply won't notice you.