The subject report, Development of the Bicycle Compatibility Index: A Level of Service Concept, is another in the series of attempts by the United States Department of Transportation to find scientific justification for its bicycle transportation program of bikeway cycling. The purpose of this investigation is unknown, because its authors do not choose to clearly state it, probably for reasons that they think to be good.
Here are some of their words about purpose. "The goals of the United States Department of Transportation (USDOT) as stated in the National Bicycling and Walking Study are: 1) to double the number of trips made by bicycling and walking, and 2) to simultaneously reduce by 10 percent the number of pedestrians and bicyclists killed or injured in traffic crashes. Meeting the first of these goals will require a substantial increase in the number of trips made by bicyclists using on-road or shared facilities. This increased exposure could, in turn, jeopardize the second goal of improved safety unless careful consideration is given to the needs of both bicyclists and motor vehicle operators in the enhancement of existing roadways or development of new roadways. To develop or improve roadways for shared use by these two modes of transportation, one must begin by evaluating existing roadways and determining what is considered "user friendly" from the perspective of the bicyclist." (Background paragraph)
Here are more similar words: "The primary objective of the current study was to develop a methodology for deriving a bicycle compatibility index (BCI) that could be used by bicycle coordinators, transportation planners, traffic engineers, and others to evaluate the capability of specific roadways to accommodate both motorists and bicyclists. The BCI methodology was developed for urban and suburban roadway segments (i.e., midblock locations that are exclusive of intersections) and incorporated those variables that bicyclists typically use to assess the "bicycle friendliness" of a roadway (e.g., curb lane width, traffic volume, and vehicle speeds)." (Objectives and scope paragraph)
Note the weakening of the stated purposes of this study as each paragraph progresses. The first paragraph wekens from "safety", to "needs", to accommodation for "shared use", to "user friendly from the perspective of the bicyclist." The second paragraph weakens from "capability ... to accommodate" to "variables that bicyclists typically use," to "bicycle friendliness".
Another paragraph states: "In the current study, the bicycle compatibility index (BCI) reflects the comfort levels of bicyclists ... " (Bicycle Level of Service paragraph)
Still another paragraph discusses this change in purpose. "During the initial stages of this study, considerable thought and discussion went into the development of the rating scale to reflect the bicyclists' level of comfort. The project team ultimately decided that the study participants should be evaluating their level of comfort as it related to their perceived level of risk. Thus, a six-level scale was used in the pilot study, where a rating of one implied that the condition of interest (e.g., lane width, speed of traffic, or traffic volume) represented virtually no risk and a rating of six implied that the condition represented unacceptably high risk. Discussions with the participants in the pilot study after the survey was completed indicated that the majority of the subjects thought that `perceived risk' accurately reflected their comfort level. However, a few of the more experienced bicyclists preferred a scale with less of a safety consideration. The one term that the more experienced riders seemed to like was `tolerance'. A rating of one in this case would indicate that the condition of interest would be tolerated for an unlimited amount of time while a six would indicate that the condition of interest would not be tolerated for any length of time.
"The preferred alternative, however, was to simply use `comfort level' as the rating term with some qualifying statements to indicate that comfort does not refer to the smoothness of the ride. Thus, for the full-scale data collection effort, a six-point scale incorporating `comfort level' was developed and used (see chapter 3)." (Conclusions paragraph, chap 2)
The actual instructions for rating roadways given to the subjects who made the ratings are: "I want you to indicate how compatible or good you think the road is for bicycling. In other words, I want you to rate each roadway with respect to how comfortable you would be riding there, where comfort is defined by how likely you are to ride on a given roadway under the speed, volume, and lane widths shown. ... roadway and traffic conditions being such that you would not hesitate to ride there ... rate the roadway as a `1'... conditions being such that you would never ride there ... rate the roadway as a `6'."
This changing of objective or purpose calls into question the value of what is supposedly an engineering study.
Many years ago, the USDOT's bikeway program was created (originally in California) as a method of making motoring easier (by removing bicycles from the motor-traffic lanes), while excusing this as making cycling much safer, and regardless of the added danger, difficulty, and inconvenience to cyclists. Somewhat later, the bikeway program was promoted as an anti-motoring environmentalist program, because anything that made cycling as much safer as was claimed would cause many motorists to change from motoring to cycling for many trips. The only bit of evidence for the original promotion, and the only bit of evidence that now exists for it, is that a bikeway aids motorists by getting bicycles out of the motor traffic lanes. There never was any evidence for the claims of making cycling much safer, or of reducing the level of skill for safe cycling, or of making traffic operations much smoother, or of reducing motoring, and no such evidence has been discovered in the thirty years since. Rather to the contrary, the present evidence is that bikeways make cycling more dangerous, requiring more skill, and make traffic operations more difficult, while most people still drive for most trips. Faced with these facts, USDOT has given up formal claims for these supposed safety and engineering benefits, although it still relies on the popular superstition that they exist.
However, USDOT refuses to give up its bikeway program. Motorists see that the bikeway program serves them (a motive that it is not politically acceptable to admit), while environmentalists have the contrary vision, politically very acceptable, that, someday, the bikeway program will reduce motoring. Therefore, the USDOT has been reduced to promoting its bikeway program on the meanest of scientific excuses, the public fear of same-direction motor traffic. If same-direction motor traffic were the greatest danger to cyclists, then the bikeway program would be justified. Since the public believes that superstition, the public supports bikeways. Since the USDOT wants bikeways (for whatever reasons its members might have), the USDOT supports this superstition. It does so, even though its own accident facts proved in 1978 that same-direction motor traffic is only a minor danger to cyclists. To justify its bikeway policy that has no other support, the USDOT pays for academic studies that appear to provide scientific support for it. That's why this study was made. Its measure of "comfort level" has no connection with safety, accident reduction, accident rate, roadway capacity, traffic movements, ability to accommodate several types of traffic, or any other engineering characteristic of a transportation system. This measurement of "comfort level" is merely an attempt at using the public's superstitious fear of same-direction motor traffic to justify the creation of bicycle lanes.
The bait and switch maneuver regarding the condition being measured, from its initial concept of risk to the final concept of comfort, has resulted in a scoring absurdity. The Bicycle Compatibility Index ends up being the Bicycle Incompatibility Index, or, more accurately, the Bicyclists' Fear Index, for high scores indicate supposedly dangerous incompatibility while low scores indicate supposedly high safety and compatibility. This can confuse the reader of any analysis of this study. When possible, I intend to discuss all measures in terms of what was apparently measured and scored, cyclist fear.
Given the purpose, valid or not, of trying to measure the relative strengths of fears generated in cyclists by different conditions of same-direction motor traffic, it is appropriate to evaluate how well this was done. Well, there was no measurement of any physiological condition indicative of fear while a cyclist was riding in any traffic at all. The investigators ruled out any cycling in traffic because they feared that some cyclist might get injured or killed. Therefore, the indirect approach of viewing videos of traffic was used. Cameras were set up on the right-hand curb or line of parking stalls, looking downstream of traffic flow. Measurements were made of lane width, volume of traffic, and speed of traffic, and segments of the films taken were selected as being typical of that traffic. Then subjects were shown those filmed traffic segments and asked first, to imagine that they were cycling in that traffic and, second, to imagine what their level of fear (or of discomfort, or of comfort, or however one chooses to define this amorphous feeling) would be when doing so, and to record the intensity of that feeling on a scale of 1 to 6. The investigators tried to validate this system. In the pilot study to determine whether or not this system was valid, they took the subjects who had seen the films to the points where the cameras had been positioned (or would be positioned for the half of subjects who had location first and films second), and asked them to rate the traffic that they saw according to the same system. The investigators claim that the ratings of film and traffic views were sufficiently similar to validate the method.
Really? What the investigators had demonstrated is the very old knowledge that, given even elementary precautions, motion picture camera and projector systems, using either film or tape, present lifelike images. The investigators never attempted to demonstrate that the feelings imagined by the subjects while viewing the images had any connection with the feelings of cyclists riding at the time and locations shown in the images. Later on, the investigators tried a similar test by comparing the ratings made by the subjects of the full study for similarity. The investigators found that there was considerable agreement about which traffic was rated high and which low. That indicated that most viewers felt that more traffic, faster traffic, and narrower lanes were worse than less traffic, slower traffic, and wider lanes: another significant discovery, that.
"[T]he survey was conducted in three cities that range in population and are geographically distributed, namely Chapel Hill, NC; Olympia, WA; and Austin, TX. One important criterion for a city to be included in the survey was the availability of both experienced and casual cyclists to participate in the survey. All of the cities selected have a number of commuting bicyclists as well as casual and recreational bicyclists. The survey participants were recruited in each city through newspaper advertisements, radio announcements, posted flyers, and announcements at bicycle club meetings. Each participant in the study received a $20 payment upon completion of the survey. The total number of participants from all three cities was 202." (Video survey paragraph, chap 3) This is not a random sample of the American population, but it might well be a reasonable simulacrum of the population of interest, constituting those most likely to be attracted to bicycle transportation by a bikeway program.
All three cities are university cities with considerable bikeway programs. Persons already participating in cycling were recruited by appeals directed at such persons. Persons with no current cycling experience were excluded, presumably because they would have no knowledge of the subject being investigated. Presumably, those chosen came from a population that has been exposed to more bikeway propaganda, and has paid more attention to it, than the typical American. However, it is not known whether this would make them more afraid of same-direction motor traffic, or less afraid of it, than the typical American. The fear appears to be widespread throughout American society, except among the best-informed cyclists. On the one hand, cycling experience tends to reduce the fear of same-direction motor traffic, while on the other hand greater interest in bicycling affairs, as commonly presented, would tend to increase the fear, or at least the importance, of same-direction motor traffic. On balance, I would not conclude that a distinct bias existed in those recruited, either for or against bikeways, or for greater or smaller fear of same-direction motor traffic.
It is reasonable to suppose that the cameras had no significant effect on the motorists passing them. Certainly, that was what the investigators assumed, because if the cameras had an effect on the drivers passing by, they would have had to invalidate their study. However, ironically, the very fact that the cameras were assumed to have no effect on passing traffic also invalidated the study. Car-bike collisions of the type that the investigators speculated about, and which are the source of the fears being measured, are extremely rare under the conditions being studied, urban streets in daylight. Therefore, over the period of the filming, it would have to be assumed that all the motorists observed would have ensured that they overtook the cyclist, who was presumed to have been present, without hitting him. However, with the camera out of the motorists' track, in a position in which the cyclist could not possibly be riding (and also probably unnoticed by the motorist), there was no reason for any motorists to change their path of travel. Therefore, the viewers saw a stream of motorists passing by them as if they were not there (as, of course, they weren't there). Furthermore, the relative velocity between motor vehicle and camera was not the relative velocity that would have existed between motor vehicle and cyclist, had the cyclist been there. The apparent velocity with which the filmed images overtook the static viewer was too high by the velocity with which the viewer cyclist would have been riding, a velocity which probably would be different for each viewer, and possibly also in each scene. These errors are just the right prescription for stimulating whatever fears of same-direction traffic the subjects already possessed. It is difficult to justify that the scores produced as the measurement of "comfort" have much agreement with the real world.
The investigators segregated roads with bike lanes from those without. "As a starting point, bicycle lane/paved shoulder facilities were separated from standard/wide curb lane facilities." (Site selection paragraph, chap 2, emphasis in original)
The results showed that fear scores increased with increases in motor traffic volume and speed, and with decreases in width. There was no significant difference in average scores between cyclists of the three cities in the survey, Olympia WA, Austin TX, and Chapel Hill NC. There was little difference between the scores of the self-assumed groups of "experienced commuter cyclists", "experienced recreational cyclists", and "casual recreational cyclists", except that the casual recreational cyclists were somewhat more fearful than the other two groups.
The investigators had available to them information on the following characteristics for each site: Roadway lanes, total number; Traffic direction, single or two-way; Width of curb lane; Stripe, presence of; Width of bike lane or shoulder; Motor vehicle volume, curb lane; Motor vehicle volume, other lanes in same direction; Motor vehicle speed, 85%ile; Speed limit; Parked cars, more than 80% of spaces occupied; Gutter pan, presence of; Width of gutter pan; Median, presence of; Two-way-left-turn lane, presence; Driveway density in 100 meters; Sidewalk, presence of; Sidewalk separated from roadway; Adjoining structures, residential or other. The investigators ran regression analyses using these as the dependent variables and the overall fear levels as the response variables. This was done for all subjects, and for each of the three classes of subject. Some dependent variables were discarded as having insignificant effect. The resulting "model" for predicting cyclist fear on different types of urban roadway, for all the subjects combined, is:
3.67 (a pure number)
-0.966 if stripe present
-0.410 times bike lane or shoulder width in meters
-0.498 times width of outer motor traffic lane, in meters
+0.002 times motor vehicle hourly volume, outer lane
+0.0004 time motor vehicle hourly volume other lanes
+0.022 motor vehicle 85%ile speed, kph
+0.506 if parking spaces occupied more than 30%
+0.264 if adjacent buildings are not residential
There are minor differences in the values for each of the three categories of cyclist considered alone. For the experienced commuter cyclists, the variables of bike-lane width and type of adjoining structures were insignificant. "The range of conditions included in the development of the model should be representative of most urban and suburban roadway conditions. However, since the sites included in the development model contained a limited range of widths, volumes, and speeds, the model should not be extrapolated beyond the values shown in table 9." (Model development Chap 4) Those values are:
Outer motor vehicle lane width: 3.0-5.6 m
Bike lane or shoulder width: 0.9-2.4 m
Motor vehicle outer lane volume: 90-900 vph
Motor vehicle speed, 85%ile: 40-89 kph
The conclusion of the paper is the predictive formula described under Calculations. Note that all factors are linear; presumably, it was assumed that the fear levels (as measured by the shaky method of this investigation) changed linearly with the condition. The paper provided no investigation to determine linearity or nonlinearity of response to input. It is likely that the elimination of the factor of cyclist speed produced nonlinear response of fear to motorist speed. The fact that no investigation was made into the linearity, or nonlinearity, of response to input is just one more reason for strongly doubting the formula produced on the assumption of linearity.
More significant is the confusion between the width of the outer motor vehicle lane and the total width available for the outer motor vehicle and a bicycle. One of the longstanding questions about bicycle lanes is whether it is better to divide a wide outside curb lane with a bike-lane stripe or to retain it as one lane. The given factors favor retaining a single wide curb lane instead of dividing it into a motor-vehicle lane and a bike lane. This is because the fear score drops by 0.498 for each additional meter of width of outside traffic lane, but only by 0.410 for each additional meter of width of bike lane. For that matter, if there is a bike-lane stripe or striped shoulder, why should the cyclist bother at all about the width that is available to motorists?
When first looking at the fear formula, it looks as though the -0.966 for the presence of a stripe is the strongest factor. However, this is not so. For example, the permitted range in width (see paragraph Calculations) of the motor vehicle lane is 2.6 meters. 2.6 multiplied by its factor of 0.498 equals 1.29. Therefore, variation in width of the motor-vehicle traffic lane has 34% more effect than the presence of the stripe. The permitted variation in motor-vehicle volume times its factor produces a change of 1.62 in the fear score. The permitted variation in motor-vehicle speed produces a change of 1.08 in the fear score. In short, the stripe factor is not as significant as it appears at first glance. The paper concludes with the assignment of Level of Service letters, with the conventional A to F names, to specific ranges of fear scores. The authors provide a fake mathematical argument for the specific fear scores for each LOS letter, but there is no logic to that argument, no physical reality, except that it looks nice and neat.
The paper, Development of the Bicycle Compatibility Index, specifically asserts that which obviously is the official position on this subject, the subject for which the USDOT paid its research money that supported this investigation. "The primary objective of the current study was to develop a methodology for deriving a bicycle compatibility index (BCI) that could be used by bicycle coordinators, transportation planners, traffic engineers, and others to evaluate the capability of specific roadways to accommodate both motorists and bicyclists." (Objectives and scope) "This research effort was undertaken to develop a methodology for deriving a bicycle compatibility index (BCI) that could be used by practitioners to evaluate the capability of specific roadways to accommodate both motorists and bicyclists." (Abstract from the Technical Report Documentation Page)
However, reading of the paper, which would not normally be done by those using or evaluating its results, discloses that all that is being measured, in an inaccurate way, is the relationship between fear when cycling and same-direction motor traffic. It must be presumed, because this is the way in which similar research contracts are managed, that USDOT knew, at the time the contract was let, the type of research to be done.
Therefore, this paper is evidence that it was written to justify USDOT's policy regarding the nation's bicycle transportation program. This paper is equally the evidence that USDOT does not want this policy to become known, but wants, instead, to be able to pretend that its bikeway policy is based on such engineering characteristics as the "capability of specific roadways to accommodate both motorists and bicyclists." The way that the results were written up appears to show that the presence of a bike-lane stripe is the strongest factor in reducing that fear. "[T]he variable having the greatest effect on the BCI is the presence or absence of a bicycle lane (or paved shoulder." (Model development) "The most dramatic effect occurs with the addition of a 1.2m bicycle lane to the existing facility; this change reduces the BCI value by almost 40 percent to 2.22." (Model sensitivity) The authors admit that only the presence of the stripe enabled the viewers to tell whether the width available indicated a bike lane or paved shoulder. "First, the two facility types are indistinguishable from each other on video unless the pavement markings or signs designating a bicycle lane are visible in the video frame." Bike lane signs could not be used because they are spaced "at very large intervals or are at locations where the filming could not be done." Therefore, it was the presence of the stripe that made the difference, as has been argued by others for decades.
Considering that the investigation did not measure any of the characteristics which are claimed for it, the only remaining issue concerns the purpose for which the USDOT chose to pay for such a false investigation. The answer is rather obvious. The USDOT has found that the only politically acceptable basis for its bikeway program is the public's fear, when riding a bicycle, of same-direction motor traffic. All the proposed justifications that have some degree of scientific validity, such as traffic accident statistics, such as standard traffic-engineering knowledge, such as roadway efficiency, have failed to justify the bikeway program. Indeed, most of them have demonstrated that the bikeway program will not work. Since the only remaining excuse for the bikeway program is the public's fear, when riding a bicycle, of same-direction motor traffic, the USDOT decided to provide what would look like a scientific support for that excuse. This study demonstrates that that excuse is a lie.
It is considered inappropriate to criticize scientific papers for their method or purpose. As Bernard K. Forscher stated in Rules for Referees (published in Science and adopted by, among many others, the Transportation Research Board), there are two forbidden topics. He states these as:
"1) Experiment design. `I would have gone at the problem a different way' is one type of comment that has no place in refereeing. If the author has demonstrated in a valid way the point he set out to demonstrate, he has done his work. ... " "2) Scope or Goal. The investigator selects his objectives; if the finished manuscript reports on a project of a magnitude consistent with the policy of the journal, it should not be criticized on the basis of what could be done in addition to or instead of what was done. ... " However, when one is evaluating the effect of a paper in producing political results, criticizing its method, scope, and purpose is essential. The BCI paper is obviously defective according to scientific standards. However, one reason for its deficiencies is that the investigation was commissioned, performed, and written up to support a political motive that is contrary to scientific and engineering standards, or, phrased differently, to support an hypothesis that has practically zero probability of being correct. Under the circumstances, it is appropriate, even necessary, to criticize both the report's scientific errors and the misuse that is being made of it on the assumption that it is essentially correct.
The nation's strategy for bicycle transportation has, for thirty years, been based on the assumption that some proper design of the highways will enable bicyclists who are without significant training to travel around town in safety. The further assumption is that bikeways provide that ability. The result is best summarized, although the policy makers don't, as being a strategy of incompetent cyclists using bikeways. This bikeway strategy was conceived and promoted by motorists, and the bikeways were designed by motorists at UCLA's Transportation Institute under contract from the state of California, 1971. While these motorists probably believed that same-direction motor traffic was the greatest hazard to cyclists, they also saw that bicycle traffic using the travel lanes was the greatest bicycling problem for motorists. For them, the bikeway strategy neatly combined public virtue, bike safety, with private convenience, clearing bicycles from the travel lanes. This strategy promised three key advantages. First, it pleased motorists. Second, its key feature was construction of physical facilities, which required only training of engineers to follow the design standards. Third, it therefore avoided trying to train the general bicycling population to ride competently and safely, which would require correcting the general cycling population's superstition about the great danger of same-direction motor traffic, the superstition that had long been key to motorists' view of bicycle traffic.
This strategy was opposed by lawful, competent cyclists as soon as they discovered that it was to be implemented. They knew that cyclists fare best when they act and are treated as drivers of vehicles, and that bikeway cycling was less convenient and more dangerous than roadway cycling in accordance with the traffic laws for drivers of vehicles. These cyclists managed to prevent implementation of the most dangerous parts of this strategy. Although the bikeway strategy is accurately described as a strategy for motorist convenience, its most vociferous present support comes from anti-motoring environmentalists. This support developed only a year or so after the original establishment of the bikeway strategy. The environmentalists believe that same-direction motor traffic is the greatest hazard to cyclists, and therefore that bikeways make cycling much safer. The environmentalists also believe that bikeways make cycling safe for beginning cyclists and others who do not want to learn to operate a bicycle as a driver of a vehicle. The environmentalists also believe that there is a great unmet desire to use bicycle transportation in place of motor transportation. Therefore, the environmentalists vociferously advocate the bikeway strategy as being necessary to save the nation from the evils of motoring.
The nation's bicycle transportation strategy is not a big program, but it is big enough to largely direct the little that the nation spends on bicycle transportation, so that every state and location produces largely the same facilities. The basic highway appropriations acts have required that bicycle transportation facilities be included in all highway projects unless there is demonstrated clear indication that they would not be used. These bicycle transportation facilities are defined in the acts as "lanes, paths, or shoulders for use by bicyclists." There is a host of documents specifying the design and selection of these "lanes, paths, or shoulders", including AASHTO's Guide for Bicycle Facilities and the FHWA's Selecting Roadway Designs to Accommodate Bicycle Traffic. Long-range bicycle transportation planning is required because it is a required part of the transportation plans for all major urban areas. There is an effort to persuade at least one university in each state to teach these design and planning skills. There is the Office of Bicycle and Pedestrian Transportation in the FHWA, which has been under John Fegan for years. That office, with others of similar bent, have, over the years, contracted for research attempting to justify the government's program for bicycle transportation. This BCI Development paper is just one of the latest of these. All of these, together, ensure that whatever monies the nation does spend on its program of bicycle transportation largely produce the same results everywhere, and that almost all of those funds have gone, and will continue to go, for bikeways. Even when the USDOT NHTSA convened a national bicycle safety conference, the conference proceedings recommended a strategy in which the bulk of any funds spent on bicycle safety be spent on bike lanes.
Thirty years ago, when the bikeway strategy was first being implemented, two claims were made for this strategy. First, bikeways would greatly reduce the cyclist accident rate. Second, bikeways would reduce the level of skill that is required for safe cycling. These two claims were not stated in any accurate sense; the proponents did not think sufficiently clearly to do so. The accident reduction claim was typically expressed by phrases such as bike-safety, and the intent was to get cyclists out of the path of motor traffic. Clearly, the proponents believed that same-direction motor traffic was the greatest danger to cyclists, so that getting cyclists out of its path would greatly reduce the cyclist accident rate. The skill reduction claim was never stated explicitly, but bikeways were always touted as being specially good for beginning cyclists. The specific phrasing of reducing the accident rate and reducing the skill level was developed by those who questioned the claims, as a means of understanding them. To these two traffic-engineering claims, the environmentalists quickly added a third claim. This was that making cycling safe for beginners and others who chose not to ride in the vehicular manner, by building bikeways, would greatly stimulate the transfer of many trips from motoring to bicycling.
To this day, bikeways are frequently referred to as "bicycle safety facilities", as particularly good for "beginning" or "casual" cyclists, and as important parts of environmental strategy to reduce motoring. These purposes are stated in the Background paragraph of the subject paper. "The goals of the United States Department of Transportation (USDOT) as stated in the National Bicycling and Walking Study are: 1) to double the number of trips made by bicycling and walking, and 2) to simultaneously reduce by 10 percent the number of pedestrians and bicyclists killed or injured in traffic crashes."
It is therefore reasonable to evaluate the effect of the bicycle transportation strategy, including the subject BCI paper, in terms of its intended results.
The premise that same-direction motor traffic was the major hazard to cyclists was disproved during the time that the bikeway strategy was being designed. Cross's first study of car-bike collisions in California, commissioned by California's Office of Trafffic Safety, was suppressed because it disproved that premise. Cross's second study, commissioned by the NHTSA, also disproved the premise on a national scale, in 1978. Since same-direction motor traffic is not a major hazard to cyclists, and since bikeways protect only against same-direction motor traffic, bikeways cannot significantly reduct the car-bike collision rate. Many other studies have shown that car-bike collisions constitute only about 12% of accidents to cyclists. If bikeways have any effect on other accidents to cyclists, they increase accidents. Therefore, bikeways cannot significantly reduce the cyclist accident rate.
In the thirty years that America has had a bikeway program, and despite efforts to demonstrate that those bikeways have reduced the car-bike collision rate, no study has succeeded in making that demonstration. Those studies that appear to have demonstrated an area-wide reduction in car-bike collision rate have never demonstrated that this was caused by bikeways.
Those who claim that bikeways make cycling safe for beginning cyclists have never tried to support this claim in a scientific manner. They have never analyzed the skills that are required for safe cycling, so they could never demonstrate which elements of skill are not required. The required traffic-cycling skills have long been known. Comparison of those skills with bikeway systems has shown three things. First, any practical bikeway system cannot be ubiquitous. Therefore, it still requires that the cyclist possess the normal traffic-cycling skills for cycling on normal streets. Second, any practical urban bike-path system must consist largely of sidepaths. Urban sidepath systems create very dangerous conflicts between bicycle traffic and motor traffic. Therefore, they either require that cyclists have very high skills, and travel slowly, to avoid those dangerous conflicts, or they require duplication of traffic signals and their phases to prohibit those conflicts, thus delaying both motor and bicycle traffic. While the latter has been done elsewhere, for example in The Netherlands, it has never been done in the U.S.A.
The preferred system in the U.S.A. is the bicycle lane. The absolute best in bike lanes, installed in the most appropriate locations, still requires that the cyclist use all the normal traffic-cycling skills. However, every city has locations that are inappropriate for bike lanes. Therefore, the choice is either to do without the bike lane, or to design the bike lane in an inappropriate manner. The first choice requires that the cyclist have all the normal traffic-cycling skills, while the second choice requires that the cyclist add to those the skills of outwitting the errors of the bike-lane designer.
Therefore, it is impossible that any practical bikeway system will allow a cyclist to travel safely without possessing the normal traffic-cycling skills. Therefore, it is impossible that the bikeway program will reduce the level of skills that is required for safe cycling.
The subject of the determinants of bicycle transportation volume, of the choice to use a bicycle for a transportation trip, is very complex. The possible determinants range from history to today's weather, and include such things as the shape of the urban transportation system, the ages of the population, the level of education, the level of income, the proportion of women in the workforce, the availability of motor-vehicle parking, the proportion and types of linked trips, the level of motor-vehicle congestion, the utility of the mass transit service, the topography, and more besides. The environmentalists' claim that bikeways will produce a transportationally significant shift from motoring to bicycling is based on two assumptions. The first is that there is a great unmet demand for bicycle transportation. The second is that this demand is unmet because of the fear of cycling in traffic. Reducing the fear by making bicycling safe by building bikeways will, therefore, according to the assumptions, result in a great switch from motoring to bicycling. There is probably a connection between a person's fear of same-direction motor traffic and that person's fear of motor traffic as an environmental danger. Their other actions and arguments make it quite clear, in any case, that the environmentalists have a great emotional investment in opposing motoring. So it may be that they overstate both the magnitude of the fear of same-direction motor traffic and its effect upon transportational decisions.
Does this postulated great unmet demand for bicycle transportation exist in the industrialized world? There are many more signs that it does not than that it does. Motoring is increasing everywhere that motoring is practical, from either the economic or the transportational aspects. Bikeways have never been shown, certainly not in the U.S.A., and generally not in the rest of the world, to produce that significant switch from motoring to bicycling. It appears that most of any switch to bicycling has come from mass transit use, which is logical because bicycling is most competitive with mass transit use. Environmentalists present several arguments against motoring, in addition to the basic environmental arguments. They argue that motoring has inherently little utility to the individual, that the individual would be able to live a better life if he did not use motoring, or used only very little. They argue that motoring has been forced onto the population against public desire, by the combination of motor manufacturers, oil companies, highway construction companies, and residential developers, so that the individual has to depend on motoring. Two empiric facts remain. First, whatever may be the causes, the production of bikeways, where that has been done, has not produced the transportationally significant switch from motoring to bicycling. Second, the popularity of motoring has been growing wherever it is practical. Therefore, it is unrealistic to assert that there is a great unmet demand for bicycle transportation in the industrialized world. If there is no such demand, then reducing the fear of traffic by building bikeways will not produce any transportationally significant switch from motoring to bicycling.
The bikeway strategy, over the thirty years that it has been implemented, has therefore failed to produce any of the three claims made for it. It has failed to reduce either the cyclist-accident rate or the car-bike collision rate. It has failed to reduce the level of skill that is required for safe urban travel by bicycle. It has failed to produce a transportationally significant switch from motoring to bicycling.
Despite this evaluation of the national strategy for bicycle transportation, the nation has committed itself to two goals in bicycle transportation. These are referenced as being in the USDOT Strategic Plan 2000-2005, but they have been quoted extensively elsewhere. The subject study quotes them as "The goals of the United States Department of Transportation (USDOT) as stated in the National Bicycling and Walking Study are: 1) to double the number of trips made by bicycling and walking, and 2) to simultaneously reduce by 10 percent the number of pedestrians and bicyclists killed or injured in traffic crashes."
It is quite clear that the national bikeway strategy cannot produce these results. The goals require that the car-bike collision rate be reduced to 45% of its previous value, while no part of the plan has any hope of reducing the car-bike collision rate at all. One wonders whether the authors of this plan considered the mathematics of their goals?
Bikeways do not reduce the car-bike collision rate. Bikeways do not reduce the level of skill that is required for safe bicycle travel. Without those motivations that would have been objective were they to have occurred, that leaves the justification of the bikeway strategy dependent solely upon the fear factor. That is, upon the fear of same-direction motor traffic that is possessed by the general population. That fear is completely unjustified, contrary to the facts of both accident statistics and vehicular operation. However, since that fear is the only factor that would appear to justify the government's bikeway program, the government has contracted to have that fear measured under varying conditions of same-direction motor traffic, to show which conditions of same-direction motor traffic increase that fear and which conditions decrease it. That is the BCI study which is the subject of this review. Notice the deliberate limitations that are built into this study. It did not investigate ways of reducing the cyclist accident rate. It did not investigate ways of lowering the skill required of cyclists. It did not investigate ways of increasing the skill level of cyclists. It did not even investigate ways of reducing the level of fear experienced by cyclists. It did not even measure the fears of traffic that in general would be felt by cyclists operating normally, and compare them with the actual hazards. All of those investigations might have been done and most could have led to useful results. All that the BCI study investigated was the level of fear caused by same-direction motor traffic and the extent to which that would be reduced by changes in the relationship between the cyclist and that same-direction motor traffic. For any person who had considered the bicycle transportation controversy, there was no doubt about the direction of the results; the only doubt concerned the magnitude of each of the factors. Therefore, several conclusions are most reasonable.
The better strategy is to do the best for those people who choose to use bicycle transportation. This is well summarized in the vehicular-cycling principle: "Cyclists fare best when they act and are treated as drivers of vehicles." That means accepting that persons who ride bicycles according to the rules of the road for drivers of vehicles, as the law requires, are both legitimate and welcome roadway users. Such persons have a much lower accident rate than today's general cycling public, and they feel much more confident when using the roads. Therefore, with more of such people there is likely to be both a reduction of accidents and an increase in bicycle transportation. It is likely that such cyclists, with more confidence and better knowledge, would be more likely than average persons to use bicycle transportation when it would be useful for them, to the benefit of themselves and of society.
The policy should advocate three levels of action. One is to train cyclists in proper cycling behavior. The Effective Cycling Program is an example of such a program with scientifically demonstrated results. Another is to improve the roads with particular emphasis upon accommodating cyclists who operate as drivers of vehicles. Standard traffic-engineering knowledge, with a few additions that are already known, provides the appropriate design changes. Still a third is to produce public acceptance of lawful and competent cyclists. Practically all the effort to date has been to the contrary. There are no examples of successful political or social programs along this line, but this level of activity can wait till last, and is likely to be made easier by the success of the first two levels.
The USDOT has funded this study because all the other, more rational and more scientific, arguments for bikeways have failed to demonstrate that bikeways make cycling much safer, or lower the level of skill that is required for safe cycling, or produce a significant transfer from motoring to cycling. The USDOT persists in following its bikeway strategy despite the evidence that this is unethical and useless, and despite the evidence that a vehicular-cycling policy would produce much better results for cyclists and for society.
The superstitious fear of same-direction motor traffic is the only remaining excuse that the USDOT can find to justify the bikeway program that it persists in conducting, against the evidence. Therefore, the study's authors have clothed this work in scientific obfuscation, so that it might be taken as a scientific justification for the bike-lane strategy. That effort has been exposed.