Hiles > Essays > Listening >

Hiles, Jeffrey A. Listening to Bike Lanes. September 1996.

Chapter 7
Facilities of Fashion: From Bike Paths to Bike Lanes

How charming to control a complicated and ornery society by bestowing upon it rather simple physical goodies. In real life, cause and effect are not so simple (Jane Jacobs, 1961, p. 113).

This paper focuses on bike lanes for several reasons. One is that lanes have become the preferred facility for many bicycle advocates and advocacy organizations of late. Also, because of that popularity, a lot has been written lately in attempts either to justify or disparage bike lanes, and much of this literature provides excellent evidence that many contemporary bicycle program professionals remain “one-eyed” and pseudoscientific despite claims on both sides of the controversy that we now have the information and understanding we need to make wise decisions about these and other facilities. Finally, bike lanes are like ink blots. As we will see, their actual functions and meanings are fairly innocuous and ambiguous. People with strong feelings about bicycling tend to project their feelings into bike lane issues. And if we look beyond traditional clashes and cold statistics, if we dig for the roots of the feelings that run so strong, then we can put a more human face on our approach to bicycle facility design, planning, and research.

A holistic trend

In the 1960s and early 1970s, bike paths were all the rage. A 1973 bikeway plan for the Dayton, Ohio, area describes bike paths as “perhaps the most ideal type of bike facility” (Miami Valley Regional Planning Commission, 1973, p. 94). At the time, literature on bicycle facility design commonly put on an air of technical sophistication by labeling different types of bicycle facilities—bike paths, bike lanes, bike routes—with stodgy jargon: Class I, Class II, and Class III bikeways. This esoteric system may have given bikeway techies a rush, but otherwise hindered communication because the labels it used were, by themselves, meaningless. That is, unlike the descriptive term “bike lane,” the term “Class II” tells you nothing about the facility unless you study the manual. In another sense, though, the labels were not meaningless at all. They implied that bike paths (Class I) were first class facilities and everything else was second or third rate.

Fortunately, most modern literature has left the class system behind. (One exception is the California state design guidelines.) We now recognize more types of bicycle facilities than will fit into that simplistic scheme. Also, descriptive labels enable us to communicate more clearly than the old class system. Most importantly, we now realize that no one type of facility is “most ideal.” Each has places where it is most fitting and places where it is most ill-suited.

Although this chapter will focus on bike lanes, it is hard to understand bike lane issues without understanding the major alternatives: bike paths and wide curb lanes. So we will start with descriptions of these alternatives before we launch into bike lane issues.

Paths

Bike Path: A bikeway physically separated from motorized vehicular traffic by an open space or barrier and either within the highway right of way or within an independent right of way (American Association of State Highway and Transportation Officials [AASHTO], 1991, p. 3).

The question of fit is most important with bike paths. In some places, these facilities provide wonderful transportation and recreation for bicyclists of widely diverse abilities. In the wrong places, though, paths are dangerous impediments to bicycle travel. To make this more clear, we’ll look at two distinctly differently kinds of bike paths: side paths and rail trails.

Side paths

Also known as sidewalk bike paths, these facilities run along the sides of roads like sidewalks. Often they are sidewalks transformed into bike paths with “Bike Route” signs. As bicycle transportation facilities, side paths seldom allow bicyclists as much freedom of movement as the roadway. Frequently, side paths create more safety problems than they solve. Studies both in the U.S. and abroad have revealed high bicycle-motor vehicle accident rates on sidewalk bike paths, some as much as three times higher than the accident rates for on-street facilities, such as bike lanes (Clarke & Tracy, 1995, p. 85; Williams & McLaughlin, 1992, p. 7). A study in Palo Alto, California, found that, although streets with side paths carried only 15 percent of the city’s bicycle travel, those same streets hosted “70 percent of the reported bicycle/motor vehicle accidents on the bikeway system” (Zehnpfenning, Cromer, & Maclennan, 1993, p. 32). The State of Oregon Bicycle Master Plan summarizes some of the problems:

Sidewalks are generally unsafe because they put the cyclist in conflict with motorists using driveways and with pedestrians, utility poles and sign posts. Also, the cyclist is generally not visible or noticed by the motorist so that the cyclist suddenly emerges at intersections, surprising the motorist and creating a hazardous condition. Every attempt should be made to allow bicyclists to function as vehicle drivers, rather than as pedestrians (Oregon Department of Transportation, 1988, p. 24).

The American Association of State Highway and Transportation Officials’ Guide for the Development of Bicycle Facilities (1991), lists eight problems with side paths. Among them are encouraging wrong-way riding, increased conflicts at intersections, motorists blocking the paths while exiting side streets and driveways, bicyclists having to stop at every side street, and motorists harassing bicyclists who use the adjacent street (pp. 22-23).

Intersection complications

Wachtel and Lewiston (1994) carried out a very selective analysis of side paths in Palo Alto, California. The researchers’ goal was to compare the car-bike collision risks for cyclists on sidewalk bike paths with that of cyclists riding on streets adjacent to those paths. To accomplish this, the researchers did traffic counts of the bicyclists at three arterial intersections to determine the relative percentages of bicyclists on the sidewalk and on the street. Then the researchers compared those percentages with actual car-bike crash data for those intersections. If there was a match, if the percentage of observed bicyclists who were on the sidewalk matched the percentage of collisions that involved sidewalk cyclists, for example, then that would indicate no difference in risk between street and sidewalk. They found that bicyclists crossing intersections from sidewalk bike paths were 1.8 times more likely to collide with cars than were bicyclists crossing intersections while riding in the street.

 

Figure 5
Side path intersection hazards and risk factors from Wachtel and Lewiston 1995 study.

Figure 5: intersection hazards.

On closer examination, though, Wachtel’s and Lewiston’s most dramatic finding was not the difference between road and sidewalk, it was the difference between riding with traffic and against it. In fact, they found very little difference in risk between road and sidewalk for cyclists riding with traffic. But when riding against the flow, road riders had twice the risk and sidewalk riders had four times the risk of those riding the same direction as the motor traffic on their side of the street. Figure 5 illustrates two kinds of motorist maneuvers that are particularly dangerous to wrong-way cyclists, primarily because motorists focus their attention on traffic coming from the other direction.

The overall higher risk for sidewalk riders, then, was due to the high percentage of wrong-way riding on the sidewalk and to the high risk for those wrong-way riders. It would seem that as long as sidewalk cyclists ride with the traffic, they run no greater risk of an intersection collision than do road cyclists. If that is true, then right-way sidewalk cyclists, who are removed from the threat of mid-block overtaking collisions, appear to be better off than road riders. Once again, though, it’s not that simple.
For one thing, the study only dealt with roadway intersections. Sidewalk riders, even right-way sidewalk riders, probably run a higher risk at driveways; and they have to cross every driveway, including busy entrances to commercial parking lots, whereas road cyclists travel a track that’s beyond the point where motorists have to stop before entering the roadway.

In addition to the safety problem, sidewalk riding tends to be slow and confining. Riders move slowly and cautiously because of driveways, pedestrians, and stops at every intersection. In the rare cases where side paths cross few intersections or driveways, bicyclists may still have to deal with being struck on a facility that is not well connected with the streets that will take them where they want to go. Even to reach a destination on the opposite side of the street from a bike path, a cyclist may have to dismount, walk through wet grass, wait by the curb for a break in the traffic, then dash across the road. For a street-savvy cyclist, that’s a demeaning maneuver compared with a simple vehicular turn. In addition, any cyclist who approaches the street from the non-path side is apt to have an awkward time getting onto the path. A cyclist who both approaches and wants to exit from the non-path side is likely to forgo the path in favor of a more simple option.

Anyway, one particularly valuable aspect of Wachtel and Lewiston’s study is that the researchers compared statistics for bicyclists crossing the same locations during the same period of time. Bike-path safety evaluations that compare all crashes at a location before and after a facility is built can be misleading. Garder, Leden, and Thedeen (1994), whose own study of intersections with stop lights led them to conclude that bike paths increased accidents by 40 percent, remarked that before-and-after studies have yielded conflicting results, perhaps because of factors that inspired the paths to begin with:

Furthermore, countermeasures are usually installed where a large number of accidents have been recorded. That the number is large might be caused by the fact that the intersection actually is dangerous or by the fact that an abnormally high number of accidents occurred by chance. A consequence of this is that a number of countermeasures have seemed to be effective, even though they really have not had a positive true mean effect; rather, they may even have been counter productive. The explana-tion for this is usually called the regression-to-mean effect; i.e. that a randomly caused high accident number is probably followed by a lower number, even though nothing has changed. This has led to misinterpretation of several—and most older—nonexperimental evaluation studies.

before-and-after studies, without control for the regression-to-mean effect, usually have a tendency to overestimate the accident-reducing effect (pp. 430, 435).

The emphasis is theirs. We will see examples of this problem when we look at crash comparisons done on bike lanes.

So, state-of-the-art wisdom says that “bike paths are not a convenient way of simply ‘getting cyclists out of the way of motor vehicles,’ and should never be developed with that intention in mind” (Wilkinson, et al., 1994b, p. 63). Yet, once a side path is in place and city officials proudly point to it as an expensive gift to bicyclists, any suggestion that cyclists’ needs have not yet been met will likely fall on deaf ears and the road itself, which could afford cyclists the greatest freedom and flexibility, and often the most safety, will remain ill suited to bicycling.

Trails

Of course, if there are no intersections or driveways to cross, a cyclist on a bike path has no risk of colliding with a car. So paths that run through the country along abandoned railroad beds or that run along river banks beneath city bridges are not at all the same as most side paths. (For convenience, I’ll us the term “trail” for any path that is not primarily a side path.) Simply put, car-bike conflicts are few where intersections are few, particularly if what intersections do exist are simple, perpendicular road crossings without the complicating influence of motorists entering and exiting a road adjacent and parallel to the path.

Trails are the most grand and inspiring of bicycle facilities. They can run for many miles through scenic greenways. They can be used to preserve abandoned railroad corridors, along with historical tunnels and bridges. Hikers, joggers, skaters, people in wheelchairs, and bicyclists can commune with nature away from the noise and fumes of motor traffic.

From an urban transportation point of view, though, even these facilities are far from being “most ideal.” The League of American Bicyclists’ advocacy director, Noel Weyrich (1995), writes that trails “can poison the minds of planners, politicians, and other citizens in strange and dangerous ways.” He adds that “…many counties have put together their “comprehensive bike plans” by merely compiling lists of planned trail rights-of-way, ignoring potential on-road accommodations entirely.”

Clarke and Tracy (1995) have listed a set of “ground rules for accepting the legitimacy and value of separate facilities such as these.” The rules include dropping the term “bike paths” in favor of “multi-use trails,” which better describes the nature of these facilities that attract dog walkers, baby joggers, skaters, and wheelchairs. They also stress the need to avoid building trails where there are frequent conflicts with cross traffic at intersections and driveways. Authorities must not mandate trail use and must realize, Clarke and Tracy say, that trails may serve novice riders and children, but have “limited utility” for faster cyclists. Also, trails are “an addition to the highway system, not a substitute for it.” That is, the presence of a trail should not preclude making streets better for bicycling. “Indeed, bicycle use on roads adjacent to trails will frequently increase,” the authors add (p. 87).

Figure 6
A safe, isolated trail and a hazardous side path.

Figure 6: remote trail versus side path.

In practice, there are three kinds of bike paths: the best kind, the worst kind, and the kind you’re likely to meet in real life. Almost any trail has some intersections, for example. The corridor it follows may occasionally run next to a road. When such a trail runs through a city, the advantages of maintaining the continuity of the trail may outweigh the disadvantages of some clumsy side-path stretches. There is no absolute formula for making that call (Wilkinson, et al., 1994b, p. 64).

Forester (1993), on the other hand, says “by far the most dangerous facility is the bike path, with an accident rate 2.6 times that of the average roadway” (p. 262). That sounds very exact and scientific, pegged with a decimal point. But the number comes from a twenty-year-old study of cyclists who had to deal with some of the worst kinds of paths. Design guidelines have come a long way in two decades. If Forester’s assessment of bike paths in general seems unreasonably harsh, his slant on side paths looks downright phobic. “By actual measurement, during commuting traffic hours,” he says, “side path bikeways with most of their intersections protected by stop signs produced 1,000 times more serious car-bike conflicts than normal cycling on the same roadways at the same time of day.” Forester’s “actual measurement” consisted of a single ride he took on a four-mile stretch of side path and the “serious car-bike conflicts” were what he judged to be near misses, not actual collisions (Forester, 1994, pp. 100-101).

Williams (1993b) describes three factors that affect bike path safety. One is the “built environment”: the path’s width, nearby obstructions, curve radii, number and nature of intersections, etc. Another factor is “the bicyclists.” Paths attract less experienced riders, who crash more. That can skew crash data. Then there are “other users.” “In general, the more ‘mixed use’ there is, the more stressful for all users and, potentially, the more dangerous the path is for any given user type.” Williams points out that bike path crashes seldom get reported so we don’t have the data we would need to devise rules for how safe or dangerous a bike path might be, given a design and location. So, “we must rely on more intuitive and, possibly, less reliable measures.” “Generalized statements,” Williams says, “(such as ‘bike paths are X times safer—or more dangerous—than riding on the road’) do little to further understanding.”

Wide curb lanes

Wide Curb Lane: On highway sections without bicycle lanes, a right lane wider than 12 feet…(AASHTO, 1991, p. 14).

From a vehicular cycling point of view, wide curb lanes would be the “most ideal” way to accommodate bicyclists, at least in the city. Rather than attempt to separate bikes from motor traffic, just make sure there’s plenty of room on the road for motorists to comfortably pass cyclists. With wide lanes, bicyclists riding down thoroughfares don’t have to cross driveways and they have the right of way over drivers approaching from side streets. The cyclists are free to make vehicular left turns from a wide curb lane, even in the middle of a block, and aren’t forced to make awkward pedestrian-style turns at intersections. In heavy traffic, bicyclists can “negotiate” with motorists before changing road position.

Technically, anything wider than the standard 12 feet is a wide curb lane, although most guidelines recommend at least 14 feet of “usable width” (AASHTO, 1991, pp. 14-15) to qualify as a wide curb lane. Some design guides recommend lanes as wide as 17 feet (Clarke & Tracy, 1995, p. 77). If a lane is too wide, motorists reportedly double up in heavy traffic as if it were two lanes. Accounts vary as to the width at which this begins to happen, ranging from 15 to 17.6 feet (Wilkinson, et al., 1994b, p. 14). In any case, this effect limits how wide a shared lane can be and still function as a single lane.

Unlike bike lanes (described below), wide curb lanes are invisible. That is, there are no markings to set aside space for bikes or to affirm that bikes have a legitimate place on the road. Most significantly, wide curb lanes do not give bicyclists the sense of separation and protection from traffic that bike lanes or side paths appear to offer, however illusory that appearance may be. The feeling among many bike advocates these days is that where traffic is heavy, wide curb lanes do not adequately serve any but the boldest of bicyclists, and that something more is needed to make the more timid cyclists comfortable on city streets.

Bike lanes

Bike lane: A portion of a roadway which has been designated by striping, signing and pavement markings for the preferential or exclusive use of bicyclists (AASHTO, 1991, p. 3).

If bike paths were the darlings of the 1970s and wide curb lanes a fad for the ’80s, then bike lanes are a trendy choice for the ’90s. Realizing that bike paths work well only in a very limited number of places, and that wide curb lanes don’t give average bicyclists the sense of protection they desire, many bicycle advocacy groups and government bicycle programs have embraced bike lanes lately with wide-eyed optimism. “One-eyed optimism” may be a better term, though, to describe the alleged benefits being claimed for these facilities.

The Florida Department of Transportation’s Bicycle Facilities Planning and Design Manual (1995), for example, has two nearly identical photos on its cover. One shows a car passing a bicyclist in a wide curb lane and is captioned “1982–Wide Curb Lanes.” Next to it is the same picture with a bike lane stripe drawn between the car and bicyclist. Its caption reads “1992–Bike Lanes.” “Wide curb lanes,” the manual says, “are to be used in Florida only as a last resort” (p. 10). It also says that bike lanes “are to be used on future urban roadway sections, whenever right of way and existing curb/drainage sections permit” (p. 13).

The manual says that bike lanes:

Most of these “benefits” are questionable. As we saw in the chapter on how bicyclists behave, bike lanes have their share of wrong-way riding; in some studies they had more than their share. In fact, the lanes create a more comfortable space in which to ride the wrong way, and some bicyclists take advantage of this affordance.

According to vehicular cycling principles, the “correct riding position for bicyclists” is highly variable, especially at intersections, so fixed marks on the pavement have only limited use as guides. Also, we might ask how safe it is to encourage bicyclists to pass motorists stopped at signals, and could question whether the message to motorists might just as well be that bicyclists must stick to one small part of the roadway.

Of all the claims, the one that looks most impressive is that bike lanes “reduce serious bicycle crashes by up to 80 percent within some corridors.” This 80 percent reduction seems possible, with an emphasis on “serious” crashes within “some corridors.” We would expect bike lanes to reduce or eliminate overtaking crashes. In some corridors, namely on rural roads, Type 13 alone makes half of all car-bike fatalities (Williams, 1993); and fatalities are about as serious as crashes get. Furthermore, this 50 percent is just an average, so some kinds of corridors no doubt have above average frequencies of Type 13 crashes. if we add the four other types of overtaking crashes and perhaps some bicyclist ride-out crashes, it may be that in some corridors bike lanes could create an 80 percent reduction in serious crashes—just maybe. Unfortunately, the study from which the authors of the Florida manual got their 80-percent figure does not very well support the claim.

According to Mary Ann Koos (personal communication, June 23, 1995), the 80 percent figure comes from a study of car-bike crash reports in Gainesville, Florida, that Koos did while she was bike coordinator for that city. During the first year of the study period, the city had five fatal crashes. That’s a high number for a city that size, even considering Gainesville’s large bicycling population. During the following four years, Koos said, there were “one or two” fatal car-bike crashes each year.

The city first began painting bike lanes on major streets in the late 1970s, according to Koos, and continued to put bike lanes on more streets throughout the five-year study period. So the study, which began in 1984, was not a clean before-and-after assessment of the effects of bike lanes. Certainly, the mere fact that the city recorded five fatalities during the study’s first year and only one fatality in a succeeding year falls far short of demonstrating that bike lanes brought about an 80 percent reduction in serious crashes.

Acknowledging that the number of car-bike fatalities in a city naturally varies considerably from year to year, Koos spoke more proudly of the fact that non-fatal motorist-bicyclist crashes in Gainesville declined 16 percent during the study period. It is not possible to say how much these crash reductions are due to bike lanes, how much the city’s bicycle safety public information campaign contributed, or how much is normal variation unrelated to anything the city did.

Figure 7
Ohio car-bike collisions, 1980 through 1993.

Figure 7 offers some perspective. In 1987, there were 4,018 motorist-bicyclist crashes reported in the state of Ohio. After that year, the statistic dropped steadily to 2,541 in 1993 (Ohio Department of Public Safety, 1994). What’s more, fatalities fell from 42 in 1987 to just 15 in 1993. Why? Perhaps bike riding reached an all-time high in 1987. Maybe Greg LeMond inspired a surge. Maybe the weather had something to do with it. Certainly, there was no massive proliferation of bicycle facilities to account for the statewide drop in crashes. An increase in helmet wearing would decrease fatalities, but not the non-fatal count. And too few Ohio cyclists wear helmets to decrease fatalities that much.

In any case, it’s a documented fact that from 1987 to 1993 Ohio cut its reported car-bike crashes by 36.8 percent and its fatal crashes by 64.3 percent. The state’s Bicycle and Pedestrian Administration missed an opportunity in 1987 to launch a bicycle safety campaign. The program would have appeared to be wildly successful, even if it did nothing at all.

More claims for bike lanes

If the Florida Department of Transportation’s manual was an isolated case of reaching into the fringe to justify bike lanes, there would be no need to waste space here questioning the claims. But these days the ranks of bicycle program professionals seem riddled with one-eyed prophets extolling the virtues of painting thick white lines on pavement. This would not be so bad if they got their claims right. Bike lanes do have virtues. But there is as much myth as meat in a lot of what bike lane boosters are saying.

For example, the Bicycle Federation of America (1993a) reports that bike lanes in a San Diego community reduced car-bike crashes (p. 8). Forester (1994) disputes this, saying that the accident reduction came not from the bike lanes but from “the prohibition of parking motor homes and boats on the street” (p. 33). When cities install bike lanes, it is rarely just a matter of laying down stripes. They remove parking, widen streets, remove sight obstructions, and launch public awareness and safety campaigns. In some cases, an anomalous spike in bike crashes may be what prompts a city to lay down stripes. So even if crash reports do drop significantly after a city installs bike lanes, it may just be “regression to the mean.” It’s hard to know how much credit to give to a white strip and how much to these other factors. When Ronkin (1993), for example, says that there were 40 bicycle accidents reported in Corvallis, Oregon, the year before that city installed 13 miles of bike lanes, and only 16 the year after, the reasons for that drop are not as obvious as they at first appear.

Smith and Walsh (1988) conducted one of the rare before-and-after bike lane safety studies that takes into account a number of threats to validity. Madison, Wisconsin, had tracked its bicycle crash reports and had done regular bike traffic counts starting several years before the city installed bike lanes on Johnson and Gorham streets. Using this data, the researchers were able to control for fluctuations in ridership and to compare any changes in crash reports on the bike-laned streets with trends throughout the city during the study period of four years before and four years after the lanes were installed.

Essentially, Smith and Walsh failed to find a significant change in the crash rates on those two streets. There was a slight increase in crashes on Johnson, but only during the first year. The bike lane was installed on the left side of that one-way street to avoid conflicts with frequent right-turning traffic. Apparently, motorists who had grown used to having cyclists on their right on Johnson Street needed time to adjust to the new arrangement.

There are two conclusions we can confidently draw from bike lane safety studies: Bike lanes don’t always decrease car-bike crashes, and bike lanes don’t always increase crashes. Seattle, Washington, bike coordinator Peter Lagerwey calls the bike lane safety issue “a wash,” and Bicycle Forum editor John Williams says, “All in all, I don’t think you can either sell bike lanes or oppose bike lanes on the basis of data showing their effects on bike crash problems” (Williams, 1993c, p. 12).

Encouragement

Another benefit that proponents claim for bike lanes is that these stripes on the pavement encourage people to ride bikes. No doubt there is some merit to this claim. Nevertheless, a couple of studies that are frequently cited in support of this claim deserve a closer look.

One is a Harris poll commissioned by Bicycling magazine that, among other things, asked survey subjects if they thought they “would sometimes commute to work if there were safe bike lanes on roads and highways” (Pena, 1991, p. 44). A full 20 percent of all U.S. adults would answer “yes” to that question, according to the poll takers. There are two problems with this. First, what people think sounds like a good idea and what people will actually do are two different things. A more important flaw, though, is in the wording of the question. It asks about safe bike lanes. This makes it hard to tell if the subjects responded to the general concept of safety or if they specifically liked bike lanes. I would predict similar results from a question that just said “safe routes.” Surely, the response would have been far less enthusiastic if subjects were asked about “bike lanes, which may not affect your safety one way or another.”

Case Study Number 1 of the National Bicycling and Walking Study is another source of bike lane enthusiasm. Goldsmith (1993) studied twenty U.S. cities to find out what factors correlated with high levels of bicycling. He found that cities with “higher levels of bicycle commuting” have on average “six times more bike lanes per arterial mile.” The “presence of on-road facilities looms large,” he wrote (p. 1), and has often been quoted for it.

In their rush to justify their ideology, though, bike lane advocates invariably overlook Goldsmith’s caveat concerning his plot of bike lanes versus commuting:

Still one must be cautious in making inferences because of the numerous peaks and troughs evident in this chart. Moreover, innumerable other factors such as street layout, land use, and traffic patterns—not to mention the dubious quality of bike commuting estimates—may be confounding the picture. Lastly, it should not be discounted that in some instances the presence of bike lanes may be a product of an organized, vocal, bicycle community. Under these circumstances, highly visible bicycling facilities may be a result, rather than a cause, of high levels of bicycle commuting (pp. 41-42).

Channelization

Researchers have also done a number of studies on how bike lanes influence the tendency of motorists to veer left while passing bicyclists (McHenry & Wallace, 1985; Wilkinson, et al., 1994b, pp. 23-26, 69-77). The results of these studies are often used in defense of bike lanes. As with so much in bicycling, though, not all the implications of these studies are as obvious as they at first appear. Wilkinson, et al., (1994b) list the following conclusions from a study of the “lateral placement” of motor vehicles passing bicyclists:

  1. Motorists tend to slow down and move over when passing bikes on bike lanes and shared use lanes.
  2. There is less slowing down and less moving over at locations with marked bike lanes than there is at locations with shared use lanes.
  3. These behaviors are not correlated with bike lane width, shared use lane width, or parking lane width (p. 77).

In short, the researchers conclude that “a marked bike lane tends to direct vehicular traffic in a manner that produces less perturbance when a car passes a bike.” This, they go on to say, is of particular benefit to cyclists who are not bold, Effective Cycling types. “Less confident riders need to feel that traffic is not going to be driving in the same lane with them and will not be moving about from side to side—with the potential for misjudgment—as they pass. Bike lanes make traffic behavior more predictable and reliable” for less experienced riders (p. 23).

Surely, no one would dispute the fact that many bicyclists feel more comfortable knowing that motorists have been cordoned off to the side. But the research does not show that bike lanes create more separation on average between bikes and cars. It’s just that the amount of separation varies less from car to car with bike lanes than with shared lanes. What’s more, this difference between stripe and no stripe is “most pronounced on low speed streets (25 mph), less so on medium speed roads (40 mph), and nonexistent on 55 mph roadways” (p. 24).

Let me put this another way. Suppose you are riding down a street with 17-foot curb lanes. As cars pass you, some pass more closely than others. The difference between the distance from you to the cars that pass closest to you and the distance from you to the cars that give you the widest berth is the “variation in lateral location.” If you divide that 17-foot lane into a 12-foot car lane and a 5-foot bike lane, the average distance from you to the passing cars will probably be about the same as if you had no bike lane, but the “variation” will be less; the cars will drive a more uniform track. But this is most true at slow speeds. That is, motorists passing you in a shared lane differ more in how close they come to you when speeds are low than when they are high. In fact, at 55 mph the variation in a shared lane is not significantly more than the variation with a bike lane.

Now, something is strange here. Bike lanes, it is said, are desirable because they make the motorists’ behavior seem more predictable, thereby reducing some bicyclists’ fears of being rear ended. So far, so good. Bike lanes accomplish this fete by reducing the variation in lateral location. This, too, sounds fine. But bike lanes reduce variation most at low speeds and have virtually no effect on it at high speeds. If bicyclists fear variation so much, and enjoy bike lanes because they reduce variation, then it would seem that the slower the speed, the more bicyclists would want bike lanes in order to feel that motorists’ behavior was predictable. To put it another way, if variation is what bicyclists fear, then bicyclists in shared lanes should fear slow-speed (high variation) traffic more than high-speed (low variation) traffic. This is strange.

It’s not strange to assert that some bicyclists, and motorists, feel more comfortable with the delineated space that bike lanes provide. It is strange to think that these measures of variation describe some benefit that bicyclists receive from bike lanes. Yet Wilkinson, Clarke, Epperson, and Knoblauch are so confident of their premise as to conclude from the results of their study that bike lanes work best on streets “with a posted speed of 40 mph or less” (p. 31). The next chapter will put a different spin on variation studies.


Previous | Contents | Next