This is an account of how superstitious fear interferes with scientific judgement and corrupts the scientific process. The particular scientific discipline is cycling transportation engineering, which studies the facts of bicycle operation to make recommendations about how individual cyclists should act and what policy society should follow regarding cyclists. The data and reasoning support the vehicular-cycling policy: cyclists fare best when they act and are treated as drivers of vehicles. (Rather than repeating the words "data and reasoning" throughout, I use the term "evidence" to include both data and reasoning.) However, government supports the contrary policy based on the cyclist-inferiority superstition: cyclists are merely rolling pedestrians who should operate on bikeways and pedestrian walkways. To place the conflict in perspective, you need to know some history. Two normal criteria for a transportation system are the safety and convenience of the traveling public. Optimizing the safety and convenience of cyclists should be the goals of any national cycling policy. If we optimized these (as far as conflicting requirements permit), individual cyclists would be treated reasonably well and society would benefit from an optimum amount of cycling transportation.
Almost from the beginning, bicycles were considered vehicles and cyclists were considered drivers. Cyclists were required to operate their bicycles according to the rules for all drivers of vehicles. This is the vehicular-cycling principle. When high-speed motoring on freeways was foreseen, the nation started to discriminate against cyclists. In 1944 the Uniform Vehicle Code redefined bicycles as devices, as part of a change that included the prohibition of bicycles from most of the road surface, from any road near which was a path that cyclists could use, and from any road that did not have driveways. Cyclists had no part in these decisions; they were made by motorists. These rules meant that cyclists had to stay out of the way of motorists just as much as was physically possible, just in case a motorist might happen to come along and be delayed. However, the rules still maintained that cyclists had the rights and duties of drivers of vehicles in all other respects. In other words, where cyclists were permitted the legal standard for the operation of bicycles remained the vehicular-cycling method.
This discrimination was created by motorists. Logically, they did this for their own convenience, but such a bald-faced grab for dominance would not have been acceptable. Therefore, the motoring establishment covered up by saying that this was to make cyclists safe by preventing them from being run down by motor traffic. At the time, there were no data about cycling accidents to support this claim. The predominant danger to cyclists could have been falling off, or motorists running stop signs, or any one of several other types of accidents. The motoring establishment, for its own convenience, asserted that the predominant danger to cyclists was being hit from behind by fast motorists, merely for being in the way by using the roads. This is the cyclist-inferiority superstition. The first data bearing on this issue were published 30 years later by Ken Cross (see below).
With the restrictive laws, government established bike-safety programs. While their ostensible purpose was to prevent car-bike collisions, actually these taught child cyclists to ride dangerously and unlawfully by following the cyclist-inferiority superstition. The most charitable explanation is that the authors of these programs assumed that cyclists were children who had to be limited, for their own safety, to the simplest maneuvers and concepts. Right-of-way and yielding were the difficult concepts that were eliminated. Looking over one's shoulder and moving away from the curb were the difficult maneuvers that were eliminated. These eliminations produced several results. Cyclists were not told what to do at stop signs. (Stop, yes, but not the important yielding action.) Because of that lapse, it was impossible to teach them that stop signs in their favor protected them. Cyclists were told that extending the left arm had the magical (that is, supernatural, not explained by natural causes) effect of keeping the cars away. With that action, cyclists were taught to turn left from the curb lane without looking behind. In summary, children were taught that bicycles operated by a different set of laws and safety principles than cars did, and that their position was inferior.
These programs were never tested to see whether they produced safe and practical results. Had the authors tested their own words by riding in traffic while operating according to their own instructions, it is likely that they would have discovered their errors. After all, if you make many left turns from the curb lane without looking behind you are likely to get smashed. The only reasonable conclusion is that the authors, lacking any real cycling experience, relied entirely on the cyclist-inferiority superstition, believing that cyclists who acted frightened of cars, and inferior to them, were least likely to incur car-bike collisions. These teachings were so foolish, so dangerous, and so unlawful, that they could not be explained in any rational manner. Therefore, the authors had to rely on fear to make it credible. The message was, "Ride according to these rules or the cars will get you." In the minds of the child recipients, such messages produced a stronger version of the cyclist-inferiority superstition: the cyclist who rides in traffic will either delay the cars or, if the cars choose not to slow down, will be crushed. The first is Sin and the second is Death, and the Wages of Sin is Death.
This situation continued for 25 years. Child cyclists and adult motorists believed the superstition, while the few adult cyclists were ignored. Then, about 1970, a sudden increase in the number of adult cyclists struck fear into the motoring establishment. These cyclists, so motorists feared, would plug up the roads that they considered their own. Therefore, the motoring establishment set out to ensure that cyclists would be prohibited from many roads, by building bikeways to give concrete reality to the cyclist-inferiority superstition as the national bicycle policy.
California took the lead in this. The California Highway Patrol and the Automobile Club of Southern California persuaded California to adopt a three phase program of restricting cyclists. The first item was a set of bikeway standards from UCLA. The second item was Ken Cross's first study of car-bike collisions. The third phase was the California Statewide Bicycle Committee to implement the results of these first studies by advising the legislature to tighten the restrictions on cyclists, including enacting the mandatory bike path law (which California had not yet done). The bikeway standards were largely copied from Dutch practice and were horribly dangerous. Ken Cross's study confounded the motoring establishment by demonstrating that most car-bike collisions were caused by turning and crossing maneuvers, which bikeways would aggravate, while only 0.5% were motorist-overtaking-cyclist collisions of the type used by the motoring establishment to justify its policy. The California Statewide Bicycle Committee (which had accepted me, by a logical misunderstanding, as the only representative of cyclists that it allowed to be a member) tried to restrict cyclists by a series of eight to one votes.
Information about these events reached California's cyclists, aroused many of them, and the proposals were largely stopped. The mandatory-bike-path law was never enacted, although the side-of-the-road law was strengthened and its words copied into a new mandatory-bike-lane law. As soon as the Office of Traffic Safety understood that Ken Cross's study contradicted all the arguments for bikeways and for restricting cyclists, that study was concealed and no further copies were made available. The bikeway standard was killed by my engineering analysis that demonstrated that it conflicted with both traffic-engineering knowledge and Ken Cross's statistics, but a new committee was formed to produce a standard that would be politically acceptable.
This California Bicycle Facilities Committee was run by representatives of the organizations who would build and operate bikeways, with one cyclist added (John Finley Scott, professor of sociology at UC Davis, an accomplished cyclist with transportation interests and libertarian leanings). The majority suggested many designs of facility intended to keep the way clear for motorists, but Professor Scott, backed up by the California Association of Bicycling Organizations, managed to kill all those that were known at the time to be so obviously dangerous that the operating organization would be found liable for accidents caused by the built-in hazards. The final standard was formed more by the threat of liability suits than anything else. That standard kept cyclists out of the way of motorists in only those ways that were not so obviously dangerous that government would run serious risks. However, despite its advocacy of separating bicycles from motor traffic, the standard specifically recommended against the use of bicycle paths alongside urban streets, because of the dangers that these create. The present national bikeway standard and the laws for cyclists in the Uniform Vehicle Code are largely copied from this work in California.
Over the same period the Federal Highway Administration ran a research project to create a national standard for bikeways. The product was Safety and Location Criteria for Bicycle Facilities. The first volume contained the research reports on which the other two volumes relied for their planning and engineering standards. The research was faulty. Only one study tried to compare accident rates on streets with bike lanes against streets without bike lanes. It used data from one unusual city (Davis, CA) and applied a method of analysis that has never been used elsewhere, without control for confounding factors, and, when corrected for statistical variation, its range of probable results showed that bike lanes could either increase or decrease car-bike collisions. Another study tried to show that speed differences made bicycles incompatible with motor vehicles. However, the data for motor vehicles, when the same analysis was done to them, showed that motor vehicles were incompatible with motor vehicles, clearly a false conclusion. Another study tried to show that intersections with bike lanes produced fewer car-bike conflicts than normal intersections. The analysis considered only same-direction conflicts, whereas Cross's accident statistics showed that more than half of car-bike collisions at intersections are produced by traffic from conflicting directions. In any case, 27% of the assignments of movement pairs as conflicting or nonconflicting were defective. When corrected, it showed that intersections without bike lanes produced less risk of car-bike collision than intersections with bike lanes. I published these and other criticisms in Cycling Transportation Engineering (later condensed in Bicycle Transportation, p335). The federal government withdrew its proposed standard and accepted that which had been produced in California's second effort.
So far this appears to be a tale of pure political power without much scientific or engineering content; cyclists got run over by the highway establishment's steamroller. Certainly, Ken Cross's first study was scientific. However, it was not merely ignored but squelched. The results of two different bikeway research programs were criticized on scientific and engineering grounds, and they were both withdrawn by their governmental sponsors. However, if this were all the tale, I think that the results would have been far more favorable for cyclists. Government and the motoring establishment would not have succeeded in overpowering and endangering cyclists on the basis of political power alone, because other political forces would have sprung to the aid of cyclists. Environmentalists, Parent-Teacher Associations, liberals of several stripes, and others, would have opposed the motoring establishment. To counter this potential opposition, government and the motoring establishment ensured their success by continuing their strategy since before 1944. They swore that their motive was to prevent accidents to cyclists, to make cycling safe and popular, especially for children. This ploy attracted many groups who were otherwise quite unsympathetic with the motoring establishment: residentialists, traffic reformers, environmentalists, parents, child-safety advocates, bicycle activists, and the like.
Thus, both the motoring establishment and the cyclists who opposed its bicycle policy agreed that accident rate was an important criterion. The cyclists also considered convenience to be important, particularly in the sense of short door-to-door trip times. Both accident rate and travel times are susceptible to scientific study. The scientific question concerned which system best served the safety and convenience of cyclists (without adversely affecting motorists to any significant extent), the vehicular-cycling system or the cyclist-inferiority system. The convenience question was easily decided. In general, bikeway systems are less convenient than the existing road system because they often require out-of-direction travel at slower speed. The car-bike collision data showed that bikeways could not significantly reduce car-bike collisions. Traffic-engineering knowledge showed that cycling on bikeways creates more car-bike conflicts that are more difficult to overcome, and are more dangerous, than does cycling on roadways in the vehicular-cycling manner. Despite all the effort that government had exerted to find evidence that bikeways were safer than roadways, no such evidence had been discovered. Therefore, the evidence overwhelmingly supported the vehicular-cycling principle and denied the cyclist-inferiority superstition.
Rather than accept the vehicular-cycling principle, the original bikeway advocates were joined in a new argument by many people who called themselves bicycle activists. The new argument had two aspects. Vehicular-style cycling requires elite abilities possessed by only a very few people, the so-called professional cyclists. Therefore, saving the world from environmental destruction requires bikeways. The environmental argument brought in many activists, but it should be peripheral at this point. That is because the issue should be the safety and convenience of those who cycle, regardless of their motives. Therefore I discuss only the safety question of whether most people could, if they choose, cycle in the vehicular manner.
The argument of bikeway activists that vehicular-style cycling requires elite abilities has no reasonable basis. The only elite characteristics ever advanced have been aggressiveness, high speed, and risk-taking. Aggressiveness in a cyclist will neither hold off cars nor frighten motorists, while no amount of cycling speed will enable cyclists to outpace cars. Therefore, the activists' argument required, although they never said so, that vehicular-style cyclists possessed the ability to bounce cars off their backs. This argument cannot be true, but it has been made for years in all sincerity. You may have considered that my reference to Sin and Death, in my description of the cyclist-inferiority superstition, was somewhat overwrought. Here we see that the bicycle safety activists have elevated the vehicular-style cyclist to the level of a priest who, through arduous and dangerous training, has acquired the magical power (that is, supernatural, not explained by natural causes) to handle the taboo objects, in this case cars and trucks, without being harmed. While the tenets of the cyclist-inferiority superstition have no objective reality, in the sense of being related to traffic operations, the description of the superstition shows its similarity to many other entities studied in sociology and cultural anthropology. In that sense it is a valid description.
More scientific data were discovered in the next few years. Ken Cross's second
study of a national sample of car-bike collisions, made under the auspices of
the National Highway Traffic Safety Administration, largely confirmed his first
study of Santa Barbara cyclists and provided more detailed analyses of car-bike
collision types and proportions, as shown in Table 1, Rank Order of Urban
Car-Bike Collision Subtypes.
|1||Cyclist on proper side of road runs stop sign||9.3||C|
|2||Motorist turning left hits cyclist head-on||7.6||A|
|3||Motorist restarting from stop sign hits wrong-way cyclist||6.8||T|
|4||Cyclist turns left in front of overtaking car||6.1||T|
|5||Cyclist hit on light change||5.9||A|
|6||Motorist turns right||4.8||A|
|7||Cyclist exits residential driveway||4.3||C|
|8||Motorist restarts from stop sign||4.2||A|
|9||Cyclist exits commercial driveway||3.9||T|
|10||Cyclist on sidewalk turns to exit driveway||3.0||C|
|11||Wrong-way cyclist runs stop sign||2.6||T|
|12||Wrong-way cyclist hit head-on||2.6||T|
|13||Motorist exits commercial driveway, hits cyclist on sidewalk||2.4||C|
|14||Uncontrolled intersection collision||2.2||T|
|15||Motorist exits commercial driveway||2.1||T|
|16||Cyclist runs red light||2.1||T|
|17||Cyclist turns left from curb lane, hits opposing car||2.1||T|
|18||Motorist turns right on red, hits wrong-way cyclist||1.9||T|
|19||Motorist overtaking does not see cyclist||1.9||T|
|20||Motorist exiting commercial driveway hits wrong-way cyclist||1.5||T|
|21||Motorist turning right hits wrong-way cyclist||1.5||T|
|22||Motorist overtaking too closely||1.4||A|
|23||Cyclist swerves left||1.3||C|
|24||Wrong-way cyclist swerves right||1.3||C|
|26||Motorist turning left, hits cyclist on sidewalk from opposite direction||1.2||C|
|27||Cyclist hits slower car||1.1||A|
It also added an analysis of the ages of cyclists involved in each type of car-bike collision. That analysis enabled me to demonstrate that car-bike collisions are not a random matter of motorists driving into cyclists (as many people argued). Instead, cyclists learn with experience to avoid car-bike collisions, starting with those types that are simplest to understand and progressing to those that are more difficult to understand.
Jerrold Kaplan's study of club and commuting cyclists gave accident rates per bike-mile, broken down by type of accident (Table 2, Accident Types and Frequencies, Cycling Club Members), type of facility, trip purpose, environmental conditions, experience of cyclist, and sex of cyclist. These data showed that car-bike collisions were only 18% of accidents. They also showed the enormous effect of experience and skill; those cyclists with the most experience and those cyclists who habitually rode in the most adverse and supposedly dangerous conditions had the lowest accident rates. They also showed that the accident rate per bike-mile on bike paths was 2.6 times that on average roads.
|Type||Percent of all accidents||Percent of serious accidents|
|Collision with moving motor vehicle||18||26|
|Collision with moving bicycle||17||13|
|Collision with moving dog||8||10|
|Collision with parked car||4||2|
|Collision with pedestrian||1||1|
Two studies from the National Safety Council, one of schoolchildren and the other of college-associated adults showed that car-bike collisions were only 10% of the cycling accidents of children in elementary school and only 16% of the cycling accidents of college-associated adults. (These were not all students; some staff and some faculty were among the subjects.) Comparison of these studies against Kaplan's again showed the enormous effect of cycling skill; their accident rates were 4 and 5 times those of the club cyclists in Kaplan's study, despite the generally more adverse conditions under which Kaplan's cyclists operated,. as shown in Table 3, General Accident Rates
|Type of Cyclist||Miles per Year||Accidents per Million Miles|
|League of American Wheelmen||2,400||113|
|Cyclists' Touring Club||2,000||66|
Using the same observational method that I had used for testing students I compared the cycling behaviors of cyclists in three university cities, Davis, Berkeley, and Palo Alto, and of a group of club cyclists based in Palo Alto. This demonstrated (with 95% and 99% confidence intervals) that cyclists in cities with bike-lane systems committed many more dangerous traffic errors than cyclists in cities without bike-lane systems, and that the type of errors in each city fitted the design of the system in that city. It also showed that club cyclists made substantially no errors while traversing cities with different types of bike-lane systems. See Table 4, Percent Defective Movements, by Cities With Differing Bike-lane Systems.
|Avoid Right-Turn-Only Lane||7.1||96.7||0.01|
|Right Side of Moving Car||0.0||5.3||0.05|
There was also a project, one that apparently never was completed, to analyze the skills necessary for bicycle riding, using the skills analyses that had been developed for motorists and motor-cyclists as models. These analyses showed some differences between the skills of car driving and of motor-cycle driving, but very little difference between the skills for motor-cycle driving and bicycle driving.
In 1977 I produced Cycling Transportation Engineering, which was a handbook establishing the discipline of cycling transportation engineering by summarizing and discussing all the previous work. Over the years that book progressed to Bicycle Transportation, published by the M.I.T. Press in 1983, and to its forthcoming second edition from the same publisher.
In 1984 the British Cyclists' Touring Club produced a survey of the usage and accidents among its members. This study was intended to show which parts of the British road system were most dangerous for cyclists, but its design did not permit drawing conclusions about that subject. However, it again showed the enormous effect of experience. Reductions of 75% and 80% in accident rate occurred with increased experience, either in years or in miles.
The second Cross study confirmed the theory that most car-bike collisions are caused by turning and crossing movements that bikeways make more difficult and more dangerous, while very few car-bike collisions are of the type that would be ameliorated by either cyclist-inferiority style cycling or by bikeways. That study also contributed to the information about skills, as did all the other studies listed above. Cross's data showed that cyclists learned how to avoid car-bike collisions. Kaplan's and the NSC studies showed that car-bike collisions are only a small minority of accidents to cyclists, and their comparison, with the CTC study also, showed the enormous effect of experience in developing the skill of avoiding accidents. My teaching studies showed that it was possible to teach, in reasonable time, even children to behave in traffic as experienced cyclists did, far better than average adult cyclists. My teaching studies also showed that the skill of operating in traffic (by bicycle or by car) largely depends on following only five basic traffic principles. Quite clearly, the scientific evidence disproves the hypothesis that traffic cycling requires elite capabilities possessed by only a few.
The question of convenience has rarely been studied for cyclists. However, it is known from general commuting studies that travel time is of great importance. Cycling in those European cities where there is high cycling volume is done for very short distances at a speed about twice walking speed. Such a method is useful in those cities. In sprawling American cities, cyclists travel much further, much faster. I measured commuting cyclists on the road at an average speed of 16 mph with an 85 percentile speed of 18.5 mph. Others have measured speeds on bike paths as much lower. It is well recognized that urban bikeway systems cannot (with exceptions for specific places) reduce travel distances, but typically increase them by requiring more out-of-direction travel to reach them. To summarize, the ubiquitous road system, by allowing direct travel at high speeds, is far more convenient for cyclists than any practical system of bikeways.
There have been some studies that present peculiar data. In 1993 the Consumer Product Safety Commission produced Bicycle Use and Hazard Patterns in the United States, and Options for Injury Reduction. That study includes comparisons of accident rates on different facilities that contradict all other data. For instance, it claims that riding on residential roads has an accident rate eight times that for riding on bicycle paths, yet it agrees with other studies in saying that only 10% of accidents are car-bike collisions. Making those data agree requires assuming that roads have more surface defects, worse bicycle traffic, more dangerous pedestrians, more things to get caught in the spokes, and the other conditions known to cause bicycle accidents, than do paths. That is obviously an incorrect conclusion. The study's classification of data obviously is defective, but why that defective classification produced such an obvious error is unknown at this time.
Over the years 1992 to 1994 the Federal Highway Administration has been issuing 24 case studies of bicycling and walking. These studies contain few data with scientific value. Many of them have obviously been done by people ignorant of their subjects. There is very little, and unfocused, discussion of the safety and convenience of the cycling public.
There aren't any other major studies in the field. Those European nations with extensive bikeway systems have even less scientific information than we have; their technology is about where ours was in the early 1970s. The scientifically acceptable knowledge in the field can be summarized in one sentence. Cyclists fare best when they act and are treated as drivers of vehicles. Here is greater detail. The traffic behaviors required of drivers (right turn, stop sign, left turn, etc.) also work better for cyclists than any other method. In general, all practical urban bikeway systems increase the number of conflicts between motor and bicycle traffic and increase the difficulty of handling those conflicts safely. Because of those dangers, practical urban bikeway systems require cycling with greater care and skill, at reduced speed, and with more delays. Nearly all car-bike collisions are caused by one or the other party disobeying the rules. Very few car-bike collisions occur when a motorist comes up behind a lawful cyclist and hits him from behind. (The statistic is 2% of urban, daylight, car-bike collisions, or 0.3% of accidents to cyclists.) Cyclists ten years of age can be taught in reasonable time to behave properly on multi-lane roads, cyclists of eight years on two-lane roads. Bicycle traffic causes very little of the total delay that motorists experience, and most of that occurs on two-lane roads that are overloaded because of (generally recent) rapid increases in traffic volume. In short, proper cycling on the roadway with the rights and duties of drivers of vehicles provides safer and quicker bicycle transportation than any practical system of urban bikeways.
There it is. No study published anywhere that I know of demonstrates that practical urban bikeway systems reduce accidents to cyclists who are traveling at normal speed, or shorten the distances to be traveled sufficiently to compensate for the lower speed required for safety, or make cycling safe for the incompetent. This is a scientific controversy in which all the scientifically acceptable evidence supports one side. Yet that is the side that has lost the argument, so far at least. Government policy assumes that cyclists are, and will remain, incompetent to ride in traffic, for which they should ride on bikeways instead of roadways. So far as I know, the only people in America who support the vehicular-cycling principle are experienced cyclists who have experienced the effectiveness and safety of proper cycling technique either through self-discovery or by being taught.
It is important to repeat this point. The only people who support the vehicular-cycling principle, that principle that is supported by all the scientific evidence available, are those who have experienced its value by having cycled successfully, according to its precepts, for some time in heavy traffic. Teaching experience, with cyclists, with traffic engineers, with legislators, with teachers, demonstrates that Americans sufficiently mature to be interested in such concepts don't come to the concept with clean minds, no tabulae rasae where cycling is concerned. They are afraid of motor traffic, afraid that the moment that they get out into traffic some motorist will hit them from behind, and that therefore their behavior must be to do whatever they feel provides the best protection from behind. The typical traffic engineer can listen to two days of lecturing, can even say that he agrees that the data are sound, yet the moment that he gets on a bicycle for a test ride or he has to do original design work he reverts to the cyclist-inferiority superstition that he started with. He does so because he understands nothing else. He can intellectually accept that the data are correct, but he cannot accept the most reasonable conclusion indicated by the data. Some kind of mental block prevents that understanding from entering.
The only means that we have discovered of changing people's minds is teaching them vehicular-cycling technique. We have also discovered that trying to hurry the course doesn't work. Technically, mature students who have driving licenses have very little to learn; proper traffic-cycling technique is very like driving a car. A trained traffic engineer has practically nothing to learn because he has already been applying the principles. Yet hurrying the process simply drives them away. The technique that works is to expose the students to traffic of gradually increasing intensity, practicing their skills at each level, until they are ready to make ten left turns in a row on a road that carries 40,000 cars a day at 45 mph. Once they have accomplished that, they have a complete change of mind. That which seemed impossible has become possible, and they have done that which, before, they thought required superhuman abilities.
The former psychological condition fits the simple definition of phobia: phobia "differs from ordinary fear by being related to events and experiences that normally would not excite such a reaction, and it cannot be relieved by logic or reason." (Encyclopedia Britannica, 1973) Phobia is often connected with feelings of guilt; the cyclist may be delaying the cars, which is sinful. The modern treatment of phobia is repeated successful exposure, without harm, at gradually increasing levels of intensity, to the feared object or condition. I had all of these characteristics of the cyclist-inferiority emotion worked out and published, including the description that included the characterization of cars as territorially-jealous demons and the sense of sin and death, long before I had any technical knowledge of phobias and their treatment. I propose, therefore, on the basis of this independent analysis, that many Americans, perhaps even typical Americans, suffer from a phobia about being chased down and attacked by the cars who own the roads, whenever they might ride a bicycle in traffic.
I propose, also, that phobias of this type are the results of a normal, and formerly beneficial, psychological process. We live in a dangerous world. Those children who pick up on their parents' fears are more likely to avoid the dangers and survive than those who do not. In pre-verbal times (and speech may have come late in the evolution of the hominids) it had to be done by nonverbal clues. If a phobia represents a true danger, the person who is in a position to violate that phobia is risking his or her life and may well not survive. The selection for phobias that resisted all forms of verbal rationalization was probably severe. By several means, but particularly by bike-safety programs and the way that they are taught, we have inculcated in our children a phobia about bicycles among moving cars. That phobia ruins our public and scientific judgements about questions of bicycles and traffic.
The judgements made by the public and by politicians are simple and unexamined. Because they believe that the roads are too dangerous for cycling, they advocate any means of getting cyclists off the roads. Since this accepted and popular opinion fits their fears, they don't question it. Those persons who do question it are called aggressive, risk-taking elitists who should not direct public action. That settles the matter for the public and politicians, unless the scientific advisors provide advice that they find compelling.
Those who administer public policy, however, at least give lip service to scientific standards and procedures. Here is a small example. The Federal Highway Administration's position is that since the proponents of the vehicular-cycling hypothesis had provided insufficient data to prove their case, the FHWA adopted the cyclist-inferiority, bikeway policy. This statement violates standard scientific procedures in three ways. First, one should not abandon a working paradigm, the vehicular-cycling principle, unless there is effective evidence against it and for the new paradigm. There was, and is, no evidence supporting the cyclist-inferiority, bikeway superstition. Second, when deciding between two competing hypotheses one needs to weigh the evidence supporting each. The FHWA did not weigh the evidence. Third, it is impossible to prove any scientific theory; disprove, yes; prove, no; only support is possible.
One can reasonably and legitimately consider the question of whether data that may be discovered in the future are likely to overturn the vehicular-cycling hypothesis in favor of the cyclist-inferiority hypothesis. Scientific research rarely occurs in a context-neutral environment; different investigators have many different motivations. While we must criticize bias that affects the results of an investigation, we have to accept that bias, in the form of one hypothesis or another, forms the decision to undertake a particular investigation. The probabilities of each of several competing hypotheses being correct may be evaluated by considering the resources devoted to investigations that would support each of them versus the results achieved. If comparatively simple research supports hypothesis A while complicated research into the same phenomena fails to support competing hypothesis B, it is not likely that further research will discover sufficient evidence to make B the preferred hypothesis.
The government has devoted considerable effort to searching for support for the
cyclist-inferiority hypothesis. The contract cost of the FHWA bikeway research
project of the 1970s was $289,000, added to which are internal costs. In one
legislative act, the federal government spent $9 million on demonstration
bikeways that produced no useful data at all. The FHWA has recently spent $1
million on 24 studies largely aimed at supporting the cyclist-inferiority
hypothesis. Forty thousand dollars were spent in a study to see whether cyclists
could look over their shoulders, a matter that cyclists already knew. The second
Cross study cost over $200,000. Added to these are the sums spent by state and
local authorities, of some unknown amount. The two leading California committees
that worked on traffic law and bikeway designs probably cost $200,000 (estimate:
10 persons, 200 days, $100 per day). The first California bikeway standards list
two principal professors, two consultants, and eleven graduate students as
authors, with nine other professors and eleven more graduate students as
advisors and contributors. These are small efforts in terms of high-energy
physics or cancer studies, but they are enormous in our cycling world.
In contrast, all the work on the vehicular-cycling hypothesis has been done by a very few amateurs. Jerrold Kaplan, John Allen, Diana Lewiston, myself, and some others. The actual cost can be only a few thousand dollars. The time spent can't be more than a few person-years. We used knowledge of traffic engineering that had been developed by many people over many years, but that knowledge was available to all parties. We have developed some data ourselves, but many of the data that we used were collected by government, largely with the intent of supporting the cyclist-inferiority hypothesis.
Government used great resources, produced no evidence for the cyclist-inferiority hypothesis, but considerable data that supported the vehicular-cycling hypothesis. Cyclists used very few resources, but produced consistent data and reasoning for the vehicular-cycling hypothesis. Given the ratios of resources applied to scientific evidence produced in each case, it is highly unlikely that at any time, with any amount of resources, sufficient evidence will be discovered to overturn the vehicular-cycling principle in favor of the cyclist-inferiority hypothesis. The evidence shows that the cycling-inferiority hypothesis is merely a superstition that is supported by public emotion and the self-interest of the motoring establishment.
Since cyclists consider this a scientific question and since government refers to scientific justification for its actions, one would consider that government should avail itself of scientific advice. The scientific committee charged with advising government in this matter, and the only scientific body that I know with a significant and comprehensive interest in cycling affairs, is the Bicycling Committee of the Transportation Research Board, which is an arm of the National Academy of Sciences. This committee operates according to the standard process of peer review. I have both reviewed papers and presented papers for its sessions. This committee has rarely considered the two hypotheses.
The Bicycling Committee held a debate on bikeways in 1978. Much of the evidence considered above was available at that time. Several speakers described the social setting of the bikeway controversy, a subject which is not relevant here. Robert Theissen, of Seattle, said that one Seattle bikeway had reduced the accident rate, but the evidence for that was poor and has not been confirmed. I presented a short, less sophisticated version of the above discussion. Michael Everett ridiculed my claim that the vehicular-cycling hypothesis had overwhelming support from the data that were available by ridiculing the idea that any scientific hypothesis could achieve such support. He also argued that he had seen no evidence that cycling on roads with high-volume, high-speed traffic was not extremely dangerous. Of course, that begged the real question. That is, which is more dangerous, and which more efficient, cycling on such a street with the rights and duties of drivers of vehicles or cycling on such a street under the cyclist-inferiority laws? The debate occurred, but no decision was reached.
The committee presented two studies rating the effectiveness of two cyclist training programs. In one, the evaluator of a California State training program (a recognized academic expert in the driving of motor vehicles) judged the effectiveness of the training by how close the cyclists rode to the curb. The discussion from the floor closed with my comment that I could evaluate cyclists because I knew how to ride, while the investigator could not properly evaluate cyclists because she did not know how to ride properly. In another, the evaluator judged the effectiveness of the training by how closely the students obeyed the instructions of the program. In the first case, the evaluator explicitly used the cyclist-inferiority superstition as the criterion, even though that was not the criterion that the program used. In the second case, the evaluator failed to investigate the suitability of the instructions, many of which were erroneous. In other words, getting the students to follow bad instructions earned a high rating. These presentations did not explicitly consider the question of how cyclists should act and how government should treat them, but they show that the majority of those members who reviewed the proposed presentations failed to recognize that these studies were based on incorrect methods of operation for cyclists. I was one of those who reviewed the second study. While I mentioned its failings I remember that it was accepted.
The paper that brought the controversy to the attention of the committee was a paper that I submitted in 1978 and again, revised, in 1982. It showed that cyclists in cities with bike-lane systems made more traffic errors than cyclists in cities without bike-lane systems, and far more than club cyclists riding through several cities with different bike-lane systems. The average score of cyclists in the city without bike lanes was 84%, in the cities with bike lanes were 66% and 58%, while the club cyclists, who rode across cities with different bike-lane systems, scored 98%. The minimum acceptable level is 70%. There were 22 differences that were statistically significant; seven were significant at the 5% level, 15 at the 1% level. The significant errors had to do with obeying traffic signals, yielding at stop signs, yielding before moving to the center lane before a left turn, and avoiding getting on the right-hand side of a car that can turn right. These errors were related to the particular designs in each city, and the significant errors were of the types that, without bike lanes, cause 30% of car-bike collisions. (Statistics from Ken Cross's second study.) I concluded that cities with bike-lane systems at least delayed the development of proper cycling skill, perhaps discouraged it, and probably had a higher car-bike collision rate.
The bike-lane paper was rejected by eight reviewers. That sounds as though it were a very bad paper, until one examines the content of the objections. Several reviewers objected that I earned money through some of my traffic-engineering activities ("Forester's presentation could be useful to advertising sales of his materials."), although every one of them was professionally involved in the field. They objected that I exhibited an anti-bikeway bias, although all of them exhibited a pro-bikeway bias, and although none of them identified which statements against bike lanes were not based on data and reasoning. While most accepted my data, they disapproved of reasoning based on the data. A few questioned the data, although they could give no reason for thinking it was inaccurate beside their disbelief ("the meager and biased data"). When a study of differences in peoples' behavior shows statistical confidence at the 5% and 1% levels, the data can hardly be meager. The reviewers many times violated the Rules for Referees, by Bernard K. Forscher of the Mayo Clinic, originally published in Science, which the Transportation Research Board distributed as its guide. Two aspects of the paper received the most criticism.
The first aspect was the system for scoring the behavior of cyclists and the demonstration by club cyclists that that standard could be met almost perfectly. The scoring system was based on the vehicular-cycling principle, which the reviewers thought was an elite standard that few could meet. Therefore the club cyclists, who did meet it, must have been some of the elite few. One reviewer compared them to motorists who had taken two-week courses in high-performance driving. Another described them as aggressive and high-speed, although there were no data about either aggressiveness or speed. None of the correct methods for the movements with significant differences in proportion correct (see above) depended on aggressiveness or speed or special training. Yielding to traffic, which is required in most of the correct behaviors and whose failure characterized most of the errors, is a very unaggressive response that we require of ordinary drivers. Staying behind a car that may turn right, instead of pushing alongside it to be endangered if it does turn right, is the exact opposite of aggressiveness. The reviewers typically asserted that the club cyclists were an improper control group, although I had stated that they were not a control group and described, at length, why it was impossible to produce a control group in this type of study. One gets the impression that the reviewers disliked club cyclists and would try any argument to get the data about club cyclists removed.
The reviewers had one point. The club cyclists and the general population cyclists were treated differently; the club cyclists knew that they were being observed, while the other cyclists did not. Therefore, claimed the reviewers, the club cyclists rode better than they normally did, and had the general population cyclists known they were being observed they would have ridden better also. The question of whether the general population cyclists could have ridden better is irrelevant; transportation studies must be based on normal behavior. Whether the club cyclists rode better is a question of fact that I discussed. I had ridden with this group many times and I observed no difference on the ride on which I measured their behavior. Other facts also apply: the ride took several hours through considerable traffic, and most of the time the cyclists were conversing among themselves in their normal way. It would have been substantially impossible for them to maintain an artificially high standard of driving during this ride with the distractions that were present. Indeed, several of them said after the ride that they had forgotten that I was observing them. Reason enters also. The reviewers' complaints maintain that both groups of cyclists knew how to ride properly and safely, but the club cyclists did so only because they knew they were being observed. That says that people who know the danger of turning left from the curb lane without looking behind, or of running through stop signs without looking either way, or of getting on the right-hand side of cars that may turn right, ignore their own safety and ride properly only to satisfy public opinion. One has to have some very strong psychological bias to believe such foolishness. The much simpler explanation is that the club cyclists rode properly and safely because they understood, while the general population cyclists rode improperly and dangerously because they did not understand.
The second aspect that was widely criticized was the discussion of, first, the pre-existing state of knowledge about bike lanes and, second, to evaluate the effect of the new data on several contemporary hypotheses about bike lanes. Not all scientific papers contain such material; perhaps only a minority. When scientists are doing what Kuhn calls "normal science," everybody in the field accepts the ruling hypotheses, what Kuhn calls paradigms; they don't have to be mentioned. However, when the field is undergoing a scientific revolution and paradigms are doubtful, then any paper that is intended to test paradigms, as mine was, must discuss them.
I will be quite open about my bias in this matter; I believe that I have been leading a scientific revolution in cycling transportation. This is not a revolution in which an older scientific paradigm is replaced by a newer. This is the first revolution in a new field, in which what has been mere superstition is replaced by the first scientific paradigm. My characterization of the preceding belief as superstition is justified because, as I discussed above, the considerable effort to conduct research based on it produced no data supporting it. Belief without foundation is superstition. Because of the difficulties I have encountered, much of my attention has been devoted to the hold that superstition has upon the mind, even the scientific mind.
The reviewers demanded that the discussions of state of knowledge and of hypotheses be deleted as mere anti-bikeway rhetoric that should not exist in a scientific paper. One reviewer, discussing another paper, showed that he or she believed that scientific papers contained only the methods and data sections. Because that reviewer also wrote that people never possessed skills but only exhibited behavior, I deduce that he or she was trained in the behavioralist school of psychology. Such a limited view of the scientific process might be appropriate in that field, but it is not generally either useful or correct.
Rather than dismissing the material as rhetoric, the reviewers should have considered the accuracy and validity of the arguments, and some did in some cases. For example, two contemporary hypotheses were (and still are) that bike lanes teach beginning cyclists to ride properly, and that bike lanes enable incompetent cyclists to use the roads safely. The data clearly showed that neither hypothesis was correct, and I so wrote. Nevertheless, one reviewer admonished me in adverse criticism: "Mr. Forester must remember that facilities do not teach, and are not expected to teach." That reviewer was so intent on removing anti-bikeway rhetoric that he forgot that bikeway advocates frequently advanced that argument and that my conclusions supported his belief (stated elsewhere) that bike lanes are purely intended to increase the capacity of the roadway; that is, for the convenience of motorists.
Other papers offered to the committee have touched on the controversy between vehicular cycling and cyclist-inferiority cycling. So far as I know, the committee has avoided discussing the issue whenever it might have arisen. Few, if any, members of the Bicycling Committee are trained scientists, although many have masters degrees. Several work in research settings for universities or for traffic institutes. There is a professor in economics among them, who prefers ridicule to reasoned argument. However, the committee members are performing a scientific function, advising the government about the science of cycling transportation. In doing so, they have consistently supported superstition, when, at the very least, they should have maintained that a significant controversy existed that placed in doubt all current governmental actions about cycling.
The actions of the committee members described above are only one example of a general condition that I have observed many times in cycling affairs, among both the technically trained and the general public. This phenomenon appears to be more generally applicable. Beliefs that have been inculcated by making the subject frightened of some entity create a phobia that requires specific actions that are intended to preserve life. If this belief is generally prevalent throughout the population, it becomes undetectable as a phobia because everybody suffers from it; it becomes believed as a law of nature. When data arise that contradict the premises of that phobia, those who possess the phobia tend very strongly to reject those data. While the victims may accept the fact that the data exist, they do not take the meaning of the data into their thinking. Their reasoning processes remain fixed in the pattern set by the phobia, until the phobia can be corrected. Correcting the phobia cannot be done through reasoning; it must be done by psychological conditioning. Only then do its victims become ready to accept reasonable discussion of the subject and to come to a new understanding.