Saturday, November 10, 2007

Why is sponsored research into cycling safety so flawed?

Bicycle advocates,

Many times in my career as a traffic engineer and bicycling advocate, I've had other professional engineers tell me that they do not support doing something for cyclists that I have proposed either because they are convinced that it is not safe for the majority of cyclists or they do not know how to do it in a way that is safe for cyclists. For example, some years ago I was speaking with the chief traffic engineer for Caltrans District 4 about bicycle access to the Richmond-San Rafael bridge. She told me that she was responsible for the safety of all road users and that she absolutely refused to allow bicyclists to use the ramps necessary to access the bridge. In another case, engineers for the Fort Ord Reuse Agency told me that they did not provide for cyclists on the reconstructed 12th Street interchange to State Route 1 because they did not know how to do it safely. More recently, the chief of the Electrical Systems Branch at Caltrans told me that she did not believe that bicycles could be reliably detected using inductive loops even though I had just given her a detailed presentation showing how it could be done.

I have been asking myself why these professional engineers are so ignorant of bicycle traffic engineering and I now have a theory. In each case, the engineer has placed the burden on proving my assertion on me, but I (and others like me) do not have access to the resources necessary to perform the research to prove our assertions. And those who do have access to those resources are either not asking the right questions or not interested in the answers.

In reviewing research studies on bicycle traffic engineering, the recurring theme that I have found is that the research has not been done or what research has been done is either biased or flawed. Here are three such examples:

A 2005 study for Florida DOT, Sidepath Facility Selection and Design, looked into the characteristics of sidepaths that make them safer than the parallel street. It used a regression model that, when I showed it to my brother (who is a professor of statistical psychology at Kansas State University), turned out to be inadequately documented. My brother recommended that I obtain the raw data for the study. When I asked the contractors who performed the study for the raw data, they said that they had discarded it and to contact FDOT. Inquiries to FDOT revealed that they have a policy against releasing crash data to members of the public. So far, my attempts to obtain the data have been unsuccessful. My brother told me that the standard among researchers in psychology is to keep raw data for 5 years in case of a request from a peer for a review of the data. The policy at the Transportation Research Board (TRB) is for its contractors to keep raw data for 3 years. Without the raw data, there is no way to know whether the research was performed correctly.

A 1999 study performed for the Federal Highway Administration (FHWA), A Comparative Analysis of Bicycle Lanes Versus Wide Curb Lanes, compared conflicts on streets with wide outside lanes vs. similar width streets with bike lanes. But the comparison turned out to be faulty because the streets with wide outside lanes vs. those with bike lanes had unequal levels of traffic as well as different traffic control characteristics and cycling populations. Also, what they called conflicts included ordinary negotiation between cyclists and motorists approaching intersections. Furthermore, several of the researchers' conclusions were not supported by the data. At the end of the report, the authors baldly state that even though streets with wide outside lanes and streets with bike lanes had similar operating characteristics, bike lanes are preferable because they attract more bicyclists, despite the fact that bicyclist preferences were not in the study's scope.

A couple of years ago, the chief of the Caltrans Electrical Systems Branch helped write a problem statement, Bicycle Detection and Operational Concept at Signalized Intersections, for research now being done by California Partners for Advanced Transit and Highways (PATH) at UC Berkeley. Inductive loops, however, are not in the scope of the study. Furthermore, the researchers are being asked to develop a procedure to discriminate between bicycles and motor vehicles for the purpose of providing additional minimum green, ignoring the concerns expressed by bicycle advocates for the length of the all-red clearance interval. When asked about these inconsistencies, the principal investigator replied that he was performing the research that Caltrans had requested. It was this comment that led to my presentation to the Caltrans Electrical Systems Branch where the chief told me that she did not believe bicycles could be reliably detected using inductive loops. Of course, she has no research to support her belief, and she has not asked for any because she already knows the answer.

It appears that the social bias against cycling that Bob Mionske describes in his new book, Bicycling and the Law, extends into bicycle traffic engineering research. That is personally disappointing to me, because I was trained as a transportation researcher and I can see when the standards for such research are being violated.

So here is what I think is happening: (1) Professional engineers are trained to base their decisions on data; (2) Bicycle advocates make assertions about bicycle traffic engineering that they believe are true but cannot prove because they do not have the resources to perform the appropriate research; (3) Sponsored research on those assertions either is not done or is done improperly; which (4) Leads professional engineers to make uninformed decisions about bicycle traffic engineering.

I would be interested if anyone has any evidence that will support or refute my theory. In particular, I am interested in any direct evidence that sponsored research into bicycle traffic engineering is either biased or flawed, or is simply not done because the sponsors believe they already know the answers.

Robert M Shanteau, PhD, PE
Consulting Traffic Engineer
13 Primrose Cir
Seaside, CA 93955-4133
Voice: (831) 394-9420
FAX: (831) 394-6045


Ed W said...

Thanks for posting this, Fritz. I've forwarded it to the state and local advocacy group, as well as a knowledgeable planner and an engineer.

Eccentric guy on a bike said...

Critical thinking is dead. Arguing that correlation does not imply causation is seen as negative cynicism these days, even in higher education. A "researcher" who doesn't arrive at the correct conclusions isn't likely to get more funding for further research.

The Mike Judge movie "Idiocracy" might be realized even in our lifetimes. Today, soccer moms are driving diesel powered monster trucks that barely fit in existing lanes or parking spots, so they're widening. But a handful of eccentric guys riding bikes who don't spend a lot are a whole lot easier to ignore.

cafiend said...

Sad but true: sponsors of research tend to get the results they're paying for. This is my cynical and unverified assertion, but I bet I'm right.

Has anyone in the bike industry tried to get studies done to support the establishment of usable cycling infrastructure? Does any advocacy group have the funds to bankroll such studies? If not, you'll get what you're getting now: highway and traffic research paid for by motor vehicle interests to support their desires.

Ed W said...

I've been working through the Thunderhead Benchmarking Report. Now, realize first that a good deal of Thunderhead's funding comes from the industry, and the manufacturers exist to sell more bikes. In their view, anything that helps sell more bikes is a Good Thing. In their conclusions, TA simply says,
"Examples abound and local advocates and officials
know that a new trail or path built where few biked and walked
before more often than not will dramatically increase bicycling
and walking in that particular corridor. Results of this study
suggest what these advocates and officials know to be true:
when you build it, they will come."

The underlying assumption is that the 'build it and they will come' approach will help to sell more bikes. There is no evidence to support this, of course, and in reality things have worked out quite to the contrary. Bicycle sales have been essentially flat while public spending on bicycle and pedestrian facilities went from $22.9 million in 1992 to $394.9 million in 2006. Don't misunderstand me - I'm not saying that money was wasted, but it may be disingenuous to suggest that on a national scale spending more money leads to more cyclists.

cafiend said...

My key word was "usable" cycling infrastructure. In other words, not the usual sidepath to nowhere, that amounts to a concentration camp for cyclists and forces them to use their bike as a toy, separate from the practical portion of their lives. The industry just wants to sell bikes, true. But they are the only deep pocket on the cycling side of the equation. Where else would funds come from? Advocacy groups? I can't afford to belong to one. When I did, they ignored my letters to them and barraged me with requests for more funds.

The best cycling advocacy an individual cyclist can perform is to go out there and ride, relentlessly. Recruit other people to ride.

Our local rail trail gets lots of traffic, even though it is lame. So the industry is right that if people believe they have a safe place to ride, they will do so. This may not appear as a jump in sales, because a lot of people own bikes already. And they aren't as in love with technology and new new new stuff as the industry has been trying to believe.

Anonymous said...

Too many traffic engineers fail to incorporate variables that are difficult to measure into their analysis for obvious reasons.
The auto and truck industries are well represented while independently minded cyclists are easily overlooked. Recreational bike paths are being funded while critically needed bike lanes, wider streets, proper signs, drivers' education,... are being ignored and unfunded.
Cities that have successfully integrated more cycling in their tansportation system are great examples of what works. Why some cyclists continue to deny the obvious is hurting all cyclists.

Fritz said...

I appreciate the comments, all. Note that this was posted by Dr Robert Shanteau, a traffic engineer in Monterey who has also been a tireless cyclist advocate since the 70s in the SF Bay Area. Please welcome Bob Shanteau to Cyclelicious.

SiouxGeonz said...

I'd agree that the bias affects the researchers who might vaguely remember what scientific research is, while most of the folks are in the "what are you paying me to find?" category :( Hmmm... how to do research to support that?