This review should be read after reading about the National Bicycle Safety Conference itself. This review concerns largely the Planning and Facilities portion, which I attended.
The organizers chose Andy Clarke, executive director of the Association of Pedestrian and Bicycle Professionals (that is, the bike planners), as leader of the panel on Planning and Facilities, who also had the task of writing the opening position paper for the panel. Clarke and Wilkinson are the two principal architects of the government's bikeway program, for they wrote the documents that state the program's policy and describe the methods of implementing that program.
With the above facts and principles about bicyclist safety and bicycling methods of operation clearly in mind, one can understand the meaning of Clarke's paper.
Clarke's words are in italics (as are titles of publications). My words are in plain text.
The title of this paper, Bicycle Safety Facilities and Planning, asserts a falsehood, that bicycle facilities are safety facilities. They never were intended to be such, and no such demonstration has been made.
This [1990s activity] has resulted in the improvements of the physical aspects of bicycling safety (e.g. trails, bike lanes, etc.). Clarke is here begging the question, asserting that bikepaths and bike-lane stripes are safety improvements. No such demonstration has been made.
Clarke's assertions about the origin of, and depth of research supporting, the bikeway standards are false. Not until the paralyzing oil crises of the 1970s...Between 1972 and 1974...the dubious quality of these planning techniques and facility designs prompted the federal government to pursue a vigorous research program into bicycle user characteristics, bicycle crashes, bike lane design, and facility development. The research culminated in 1981 with the publication of the first national bicycle facility design manual by the American Association of State Highway and Transportation Officials. This is almost all false.
The first American bikeway standard was California's Bikeway Planning Criteria and Guidelines (Calif DPW 1972), ordered in 1971 from the UCLA traffic studies organization and delivered in 1972, long before the first oil embargo of 1973. The federal government did pursue research, in two general ways. NHTSA funded Cross's second study, the first national study, of car-bike collisions (Cross & Fisher 1977). Federal Highway Administration funded a study of bikeways, culminating in two volumes of design standards and one volume of research reports (FHWA 1975-6). While the federal government did perform this research, none of it was ever used. Cross's research about car-bike collisions was not used because it contradicted the desire for bikeways (just as his first report, on car-bike collisions in California, was suppressed by the California Office of Traffic Safety, for the same reason, once I had showed the consequences of his statistics). The FHWA research and the design standards created from it were never used, because I, Forester, demonstrated that the research was full of errors.
Instead, the AASHTO Guide was copied from California's second bikeway standard. Where did this come from? Once the first bikeway standard was delivered, the California Legislature established a Statewide Bicycle Committee with the intent to pass laws making the use of bikeways mandatory. I got onto that committee by false pretenses, and I then discovered the first bikeway standard and detected the intent to make the use of bikeways mandatory. By applying standard traffic-engineering reasoning, I demonstrated that the bikeways required by this standard would be extremely dangerous for cyclists. California then never issued that standard.
California then established a California Bicycle Facilities Committee to produce a second standard. As on the first committee, I did more than anyone else, although, having acquired my reputation, I was not allowed to be named to the committee. I did probably 3/4 of the writing done, trying to get that committee to design facilities that would ameliorate the kinds of accidents that we knew about. Cross's first study of car-bike collisions (Cross 1974) became available during the operation of this committee, and was presented at a conference held at the Sacramento Airport. At that conference I pointed out that Cross's value of only 0.5% of car-bike collisions occurring when a motorist overtook a lawfully proceeding cyclist completely denied the assumption that restricting cyclists to bikeways would make cycling safer. Nothing further was ever heard of Cross's study, except for the copies distributed at that conference. The committee absolutely refused to consider any highway designs that would tend to ameliorate the more frequent kinds of car-bike collisions.
My information, some garnered sub rosa, the rest obtained from observing at many committee meetings and legislative hearings, says that the sparkplugs for the California action were the California Highway Patrol and the Auto Club of Southern California. The obvious intent of all of this activity was to clear cyclists from the normal traffic lanes for the convenience of motorists, almost regardless of the dangers caused to cyclists in the process. The only ways that we managed to change the proposals of those organizations was to demonstrate that they were so dangerous that government would probably lose liability suits when cyclists were injured or killed through those dangers.
Therefore, the first edition of the AASHTO Guide, which was largely copied from the California bikeway standard, was the result of the effort of the motoring organizations to clear cyclists off the roadways for the convenience of motorists. I was there as the most active participant in these operations. My account has been published in Effective Cycling (Forester 1976-93) and in Bicycle Transportation (Forester 1977-94), and in many other places. Nobody has ever demonstrated that my account is inaccurate. Yet here we have Clarke giving his self-serving fiction as if it were true.
Clarke ascribes the slow rate of bikeway production in the 1980s as follows: the prevailing wisdom among organized bicyclists was that special (especially segregated) facilities threatened cyclists' right to use the road, and prevented them from behaving as if they were motor vehicles, the basic tenet of the Effective Cycling education program. This description is incorrect in a self-serving manner. The slow rate of government spending on bicycling during the 1980s was more probably caused by the disinclination of those administrations to spend money on that subject. However, ascribing cyclists' attitudes to the Effective Cycling education program is false, as is the description of those attitudes. Cyclists never intended to ape motor vehicles; that's just Clarke's self-serving rhetoric. The view of experienced and well-informed cyclists is commonly stated as the vehicular-cycling principle: "Cyclists fare best when they act and are treated as drivers of vehicles." That is, to be treated as drivers, just like other drivers, on an equal footing, not as bicycles or as cars. The Effective Cycling Program is merely a reflection of that attitude that used to be practically universal among well-informed cyclists, how to teach the behavior appropriate for safety and efficiency.
Clarke praises the recent implementation of more governmental bicycle programs. However, he uses the phrase the need for special facilities for bicyclists as a precursor to higher use and greater safety....the state-of-the-art in bicycle safety facilities and planning is quite well-developed. This is false. There has never been a demonstration that either sidepaths or bike-lane stripes either make cycling significantly safer or significantly reduce the level of skill required.
Clarke advances several reasons for the advances in bicycle safety facilities and planning: The first that he lists is the National Bicycling and Walking Study.
The claim that this set of studies concerns bicycle safety is false. I list the titles of the 24 studies:
The only one of those studies that even considers ways to reduce accidents to cyclists is #19, Traffic Calming, by Clarke and Dornfeld. Clarke's recommendations are, largely, ways of making cycling so dangerous that the cyclist has to slow down to avoid accidents. The benefits they bring...far outweigh any concerns about slowing cyclists down too.
As I wrote in my review of all these studies , it is unconscionable for the federal government to issue such recommendations in a study that is supposed to improve cycling. The added danger to cyclists is obviously evil; slowing down cycling is also counterproductive because the prime way to make cycling more useful is to encourage faster cycling.
Scientific studies are expected to discover new facts or to develop new theories or understanding about existing facts. Engineering studies are expected to develop guides for action based on scientific facts. When this series was first proposed, I wrote that these studies would fulfill neither of the scientific purposes because they were too misdirected, underfunded, and hurried to improve on what we already knew.
I feared that they would serve as guides for further ill-conceived governmental effort that harms cyclists. So far I have been proved correct. Most of the studies have been done by people who obviously don't know cycling transportation engineering and are totally ignorant of controversies that have been debated for years. The amount of new knowledge is minute. In many cases, the recommendations, if carried out, would do great harm to cyclists.
Clarke praises this current program. The problem is that the bicycling programs are founded on the wrong assumptions, without any scientific support for their claims of greater safety and against the scientific knowledge we have about cyclist safety.
Insofar as cyclist safety is concerned, there hasn't been any worth a damn since Cross's study of car-bike collisions, which is ignored in all governmental bicycle programs. It is ignored because it demonstrates the futility of bikeway programs as bicycle safety measures.
Clarke makes the following statements about research results. Poor cyclist behavior may be associated with a lack of facilities....wide outside lanes are associated with a greater incidence of wrong-way and sidewalk-riding than are bicycle lanes...non-standard turns appear to occur more often when wide outside lanes are used instead of bicycle lanes. In the UNC study [Hunter & Stutts] with these findings, the roads with bike lanes were consistently smaller and had lower traffic volumes than the roads with wide curb lanes, which were mostly wide, high-volume arterials. No attempt was made in that study to control for these very large differences. Further, the study was unable to conclude that bike lanes (even given the uneven comparison) were safer than wide curb lanes, and noted the importance of training cyclists in how to operate according to the rules of the road. (Paul Schimek's comments for this discussion)
In a study of bicycle commuters, streets with bicycle lanes or routes were considered to be more than two-fold safer than streets without such facilities. This is a mendacious account of Moritz's data (and of Kaplan's much earlier data). There was no experimental evidence, only historical data. Clarke ignores the fact that the bike planning process (in which he is an expert) had first placed the bike-lane stripes on the streets that were already considered safest for bicyclists. So Moritz's data demonstrate merely that the safer streets were the safer ones.
Motorists are less likely to encroach on the adjacent lane when passing a bicyclist riding on a paved shoulder or in a bicycle lane [Than on a roadway with neither]. The Harkey & Stewart study demonstrated many other things also, which Clarke ignores. For example, that motorists overtaking cyclists gave them less clearance when there was a bike-lane stripe than when there wasn't. H & S's data clearly indicate that the roads used were not of comparable width. Had they been so, in many cases the data would demonstrate that the motorist was driving in the bike lane. This is one more attempt to extract anything supportive of the bikeway superstition from an unpromising collection of data. The point that H & S make is that, on two-lane roads when there was no traffic from the opposite direction, motorists gave cyclists more clearance when there was no bike-lane stripe or paved shoulder. Clarke states this as a benefit to motorists, of reducing the probability of bicycle traffic causing head-on collisions between motorists. Of course, this effect was not observed when there was opposing-direction motor traffic.
Clarke makes only one reference in his white paper to his major research study, Bicycle-Safety-Related Research Synthesis, FHWA-RD-94-062. His reference is: Bicycling is the only mode of transportation for which people consistently express a desire to appreciably increase their use.
I reviewed this work in 1998 (maybe earlier). I started my review as follows:
"The Synthesis of Bicycle-Safety-Related Research is the supposed scientific justification for the cycling policy and program of the nation. It purports to contain a summary of all the important research since 1981 into cycling transportation. That information, when added to the valid portions of the work done before 1981, ought to support our nation's cycling policy and program. According to Clarke and Tracy, it does. However, they have made it do so only by ignoring half or more of what was done and by misquoting and maligning much or the rest."
That was my judgement then. It may be wrong. However, in Clarke's recommendation to the National Highway Traffic Safety Administration concerning a national bicycle safety policy, the only quotation he makes from his own summary of research about bicycle safety from 1981 to 1994 is a quotation about the unrealized, possibly potential, popularity of bicycle transportation.
Clarke wrote a substantial paper; he didn't report that there was no substantial bicycle safety research in that period. It is evident that he has now realized what is revealed in his NHTSA recommendation. That is, there is no bicycle safety research supporting his position that is worth discussing and will stand up to scrutiny.
This analysis of Clarke's current claims for governmental research about bicycle traffic supports my long-term consistent criticism of that research. The three statistical studies of car-bike collisions (Cross's first, for California Office of Traffic Safety, Cross's second for the National Highway Traffic Safety Administration, and UNC's recent but not as good replication of Cross's second), have been reasonable investigations with little bias. It is well to note that these were done for the offices with prime responsibility for traffic safety, and therefore less likely to be biased by other concerns. It is well to note, also, that Cross's first study was suppressed by the California highway organizations, and that the conclusions of all of these studies have been systematically ignored by the highway organizations.
Almost all of the other governmentally funded research about bicycle traffic, starting with the California contract to UCLA of 1971 and consistently since, has been designed (intentionally or unintentionally) to seek data that would support the bikeway superstition that is the prime justification of the government's bicycle traffic program. As I remarked above, when one is seeking data to support a hypothesis that has a very small probability of being a correct description of reality, one is most unlikely to discover such data. For 24 years, government has been trying to find data to justify its bikeway superstition, and in those 24 years of effort it has been unable to find such data. The most that it has been able to produce are statements that twist inconclusive data into supposedly supportive claims.
On the other hand, consider the results produced by the vehicular cycling advocates. Without any funding at all, they have combined the principles of traffic engineering and of human engineering with the accident data collected by others, and added the results of their own research into cyclist traffic behavior, cyclist traffic skills, and cyclist training to produce the clear scientific support for the vehicular-cycling principle.
When government cannot in 25 years of funded effort demonstrate the validity of the bikeway superstition which is the justification for its bicycle program, and when amateurs with few monetary resources have demonstrated overwhelming evidence for the contrary vehicular-cycling hypothesis, which is more likely to be correct?
That's correct, more and more people who have no guidance in making cycling safer or better, but who are guided by documents that do the reverse. Note Clarke's admission: Nor has bicycle safety dramatically improved; every year, approximately 800 bicyclists still die in collisions with motor vehicles. Why should anyone have predicted great improvement, or be astonished that it doesn't occur, when the programs have nothing to do with reducing accidents to cyclists?
Despite tremendous advances in knowledge and the application of measures to increase bicycle safety...bicycle safety hasn't appreciably improved in the last decade or more.
Why should any reasonable person expect such results, when none of the program items is directed at reducing any significant amount of the accidents that occur to cyclists? I point out that there are two principal items in the government's bicycle safety program, the bike-lane stripe and the urban bicycle sidepath. I point out that there is no reason to believe that either of these would make cycling significantly safer. Nobody has ever produced an analysis of accidents to cyclists that suggests how either of these would make a significant reduction in accidents to cyclists. Nobody has ever demonstrated that either of these has, in fact, by some means unrecognized, attained that objective.
Consider Clarke's following statement: Strategies designed to allay fear of traffic, such as the provision of safe cycle paths. That's an oxymoron. According to all the evidence we have, cyclists incur more accidents per mile traveled on paths than on roads. Ah, yes, but Clarke protects himself by saying that the paths were designed to allay the fear of traffic. Now consider this next of Clarke's statements: Other studies suggest that physical improvements such as bike lanes and trails do increase the perception of safety. Thus, the perception of safety, as understood in the minds of people who know nothing about how to be safe when cycling, has now become the guiding feature of the government's bicycle safety program.
I ask you, should the National Highway Traffic Safety Administration sponsor a bicycle safety program that is based on nothing more than the perception of safety by the ill-informed general public, or should it repudiate such programs and call for a fresh start along the lines that are known to reduce accidents to cyclists?
Changes in the physical environment are needed to enhance the perception of safety and increase use, which in turn is likely to improve rider safety. This is Clarke's actual safety argument, for which see Clarke's Conclusion below.
The following is the full text of Clarke's Conclusion, containing his final argument.
The physical environment is a critically important element of the perceived and actual safety of cyclists. While any other factors also affect bicycle safety, the conditions and feel of the local street network must be inviting, convenient, and accessible if bicycling is to thrive in the United States. We propose that a greater number of riders is needed before motorists will note their presence (making cycling safer) and willingly share the road with them.
Clarke is in the uncomfortable position of having recognized that there is no scientific support for the proposition that bikeways either reduce accidents to cyclists or reduce the level of skill that is required for safe operation. Therefore he is reduced to the argument that he gives, which has the following parts:
These are the recommendations of the panel on Planning and Facilities, as published almost a year later without either explanation or justification.
The recommendations of the panel on Planning and Facilities are severely defective.
There can be no doubt about the accuracy and authority of this charge. While the facts have been known for almost 30 years, they are now substantiated by the official documents written by Mr. Clarke in the last six years and recently in his official position paper for this conference. When Mr. Clarke writes that there is no evidence that the bikeway program of the last thirty years has reduced cyclist casualties, but then goes on to recommend a great expansion of the program as a safety measure, one can reach no other conclusion but that this is deceit, pure and simple deceit.
I sent a protesting paper to the two chief organizers, Dr. Richard Schieber of the CDC and Ms. Maria Vegega of the NHTSA. A copy of that paper is also on my website.
That paper outlined the state of scientific knowledge about cyclist safety and also outlined the scientific process for deciding between hypotheses. Much of the background information of that paper was copied from my paper The Bikeway Controversy, that had just been accepted by Transportation Quarterly. The paper urged the organizers to recognize the necessary scientific conclusion and to take the courage to modify the results of the Conference in accordance with that conclusion.
For a general statement of the results of the Conference, read the general report.
Forester, John; Bicycle Transportation; Cambridge, Mass; The MIT Press, 1983, 2nd ed. 1994. First issued as Cycling Transportation Engineering Handbook, 1977. This is the standard description of the discipline of bicycle transportation engineering. It is a handbook for government personnel with cycling responsibilities and those who wish to influence such persons. It contains such items as studies of accidents, cyclist skill, highway and bikeway design, government bicycle programs, history of bikeway standards, etc.
Forester, John; Effective Cycling; Cambridge Mass; The MIT Press, 5th ed 1993, 1st ed 1976. This is a handbook for cyclists. In addition to many other cycling subjects, it describes and teaches the skill of cycling in traffic in accordance with the rules of the road for drivers of vehicles.
California, State of: Bikeway Planning Criteria and Guidelines; April 1972. This is the first bikeway standard in the USA. It largely copied Dutch practice, and, while printed, was never issued because its designs were shown to be far too dangerous for cyclists.
California, State of: Planning and Design Criteria for Bikeways in California; 1976 to present. This is the present bikeway standard, initially produced by taking into account the safety objections of cyclists to bikeway discrimination, and the probable litigational consequences thereof. While never designed to reduce accidents to cyclists, it was produced by removing the most obviously dangerous designs that would have made liability probable. It served as the model for the first AASHTO Guide.
Federal Highway Administration; Safety and Locational Criteria for Bicycle Facilities; 3 vols. FHWA-RD-75-112 is the research report. Washington; Dated 1975, not issued until 1976. This was intended to be the federal bikeway standard. While printed, it was never issued because the research upon which it was based was demonstrated to be faulty. The Federal DOT then adopted the AASHTO Guide instead.
Association of American State Highway and Transportation Officials; Guide for the Development of New Bicycle Facilities; Washington 1981 to present. After the failure of the first California bikeway standard and the first Federal bikeway standard because of the dangers of successful personal injury lawsuits that they presented, AASHTO decided to adopt the second California bikeway standard because it seemed to have removed the obvious liability problems. This was later adopted by FHWA.
Wilkinson, W. C., A. Clarke, B. Epperson, R. Knoblauch; The Effects of Bicycle Accommodations on Bicycle/Motor Vehicle Safety and Traffic Operations; Bicycle Federation of America for the FHWA; 1994. This document is the research justification for the next one in the list. This document argues that 95% of cyclists are incapable of obeying the rules of the road for drivers of vehicles, and therefore we need bike lanes on all major roads, while minor roads may be neglected. The known scientific evidence contradicts both the assumptions and the conclusions. The main argument is that uninformed persons want this conclusion.
Forester has a review of this document.
Federal Highway Administration: Selecting Roadway Design Treatments to Accommodate Bicycles; FHWA Manual; 1992
Cross, Kenneth D & Gary Fisher; A Study of Bicycle/Motor-Vehicle Accidents: Identification of Problem Types and Countermeasure Approaches; National Highway Traffic Safety Administration; Washington; 1976. The first such study in the world (after Cross's initial study in Santa Barbara County). Statistically robust in most respects, careful investigation of accidents. Fails to consider some details that only later were understood to be important.
Wachtel, Alan, Diana Lewiston & Gayle Likens; Risk Factors for Bicycle-Motor Vehicle Collisions at Intersections; ITE Journal, Sept 1994. This demonstrates that even with the slow speed of cycling on sidewalks, the urban bicycle sidepath produces a very high car-bike collision rate.
Clarke, Andy & Linda Tracy; Bicycle-Safety-Related Research Synthesis: FHWA-RD-94-062; FHWA 1994. This is an attempt to provide a scientific justification for the government's bikeway program. It appears to do so only by ignoring half or more of what was done and by misquoting and maligning much of the rest, and by ignoring the most important aspects of traffic operations.
Forester has a review of this document
Forester, John; The Effect of Bikeway System Design Upon Cyclists' Traffic Errors; 1978, 1982. Now available at www.johnforester.com. Observations of how the bikeway system design influences the dangerous traffic errors that cyclists make, within 95% and 99% confidence intervals.
Kaplan, Jerrold A.; Characteristics of the Regular Adult Bicycle User; MS thesis, U of Maryland; FHWA; 1976; National Technical Information Service, Springfield VA. Account of the cycling experience, including type of cycling, mileage, accidents, of two groups of regular cyclists: LAW members, Washington Area Bicycle Commuters. Produces accident rates per mile traveled for varous conditions. Showed the importance of skill in preventing accidents, and the high accident rate on bike paths.
Moritz, William E. Adult Bicyclists in the United States: Characteristics and Riding Experience in 1996; Bicycling Committee, Transportation Research Board; Washington; 1998. A shorter survey of the Kaplan type, with much the same results.
Schupack, S. A. & G. J. Driessen; Bicycle Accidents and Usage Among Young Adults: Preliminary Study; Chicago; National Safety Council; 1976. Study of cycling and accident experience of college-associated adults (students, staff, faculty). Produces accident rate per mile traveled.
Watkins, S. M.; Cycling Accidents; Cyclists' Touring Club; Godalming UK; 1984. Study of cycling and accident experience of CTC members. Demonstrates the importance of experience in preventing accidents.
In these studies, comparisons of Kaplan's, Watkins's, and Schupack's data show the importance of experience, a substitute for evaluatingskill, in reducing the accident rate for experienced cyclists to about 20% to 25% of that of the general bicycle-riding public. See Forester: Bicycle Transportation.
Forester, John & Diana Lewiston; Intermediate-Level Cyclist Training Program: Objectives, Techniques, and Results. Bicycling Committee, Transportation Research Board; Washington; 1981. Now available at www.johnforester.com
Forester, John: Elementary-Level Cyclist Training Program: Objectives, Techniques and Results; 1981. Now available at www.johnforester.com. This and the preceeding study tell how the instruction was performed and demonstrated that the students performed on the bicycle driving test with class average scores of about 95%, when the population average scores of the surrounding adult cycling population was only flunking scores about 55%.
Forester, John; Effective Cycling Instructor's Manual; Palo Alto; Custom Cycle Fitments; 1977, 1980. This is a comprehensive book on how to teach cycling, including in the vehicular manner as well as for sport and matters of mechanical maintainance.
Pucher, John & Lewis Dijkstra; Making Walking and Cycling Safer: Lessons from Europe; Transportation Quarterly, Vol 54, No. 3, Summer 2000
Forester has a review of this document
Pucher, John, Charles Komanoff & Paul Schimek; Bicycle Renaissance in North America? Recent Trends and Alternative Policies to Promote Bicycling; Transportation Research Part A, Vol 33, Nos 7/8, 1999, pp 625-654. This and the preceeding document are praises of the Dutch sidepath system without any understanding of the dangers that this creates and how badly the Dutch have had to compromise bicycle transportation because of the dangers that their system creates for cyclists.
Forester has a review of this document.
Forester, John; The National Bicycling and Walking Studies; Chainguard, March 1994
Forester, John; The Bikeway Controversy, Transportation Quarterly, Vol 55 No 2, Spring 2001, p 7-17
1: Forester, John; The National Bicycling and Walking Studies; Chainguard, March 1994