If cycling on the road were inherently dangerous then the more miles a cyclist travels annually the more likely he or she should be to have an accident. The exact opposite is true. Those who travel fewer miles per year, such as elementary school children and adults attending college, are far more likely to be involved in an accident than the more experienced cyclist who covers many more miles per year, frequently under more adverse and testing road conditions.

This is illustrated in John Forester's book Bicycle Transportation, 2nd Edition published by MIT Press in 1994. Forester reviewed four cyclist accident surveys and used them to compare different cyclist populations and their accident rates.

Table 5-1 of General Accident Rates shows that a cyclist belonging to the League of American Bicyclists (LAB) cycled on average 2,400 miles per year with an accident rate of 113 accidents per million miles. By contrast, a college-associated adult who cycled on average only one-quarter this distance annually had an accident rate almost 4.5 times as high. Elementary school children cycling on average only 580 miles per year had an accident rate over 6 times greater than the average LAB cyclist.

This is an important issue, in view of the push for both off-road mixed use trails and on-road bike lanes, and the widespread perception that the existing road system is unsafe for bicyclists.

Prof. William Moritz, a long-time member of LAB and emeritus professor at the University of Washington in Seattle, did a detailed survey of LAB members and published his results in 1996. This confirmed earlier surveys of LAB cyclists showing that roads were generally considerably safer than off-road facilities and that sidewalks were generally the most dangerous facility of all.

In fact, in relation to a major road with no special facilities for bicyclists, the sidewalk is typically almost 25 times more hazardous for cyclists, unpaved off-road trails nearly 7 times more hazardous, and mixed-use recreational trails twice as hazardous.

One conclusion of interest is that just signing a road as a bike route lowers the risk by 23% compared to the standard risk on a major road. This raises an intriguing question: Why would such a minor difference make such a significant improvement? One suggestion is that signage may alert motorists to the presence of bicyclists, leading to more cooperative behavior. Another is that such signage is only applied on roads already fairly safe.

Another conclusion fueling the debate over bike lanes is that providing a bike lane was observed to lower the risk compared to a major road by 38%. Proponents of bike lanes take this as support for providing such facilities. Opponents counter that bike lane striping is rarely applied in isolation of other road improvements, so that the effect of the bike lane is obscured. In addition, they are unlikely to be applied on an inherently dangerous road. And missing from the survey was any reference to wide curb lanes, which are preferred to bike lanes by some cycling advocates.

There is an understandable preoccupation with car-bike collisions, which are believed to account for the majority of bicyclist fatalities. However, study after study shows that cyclists suffer serious injuries in far greater numbers from causes which never involve a motor vehicle at all.

Here we focus on the 16% of urban bicycle crash causes which typically get the most attention in the media and elsewhere: car-bike collisions. (It means we are ignoring, at least for the moment, the reasons for the other 84% of bike crashes and how to mitigate them.)

The major crash types for car-bike collisions are crossing movements, turning movements, parallel path collisions, and swerving. In urban areas these together make up 85% of all car-bike crashes, of which crossing and turning movements combined account for no less than 75%. Parallel path collisions account for only 7% of the total, yet these are the ones causing the most fear among cyclists and is arguably the main rationale for bike lanes and off-road bike facilities.

The data is based on a comprehensive study by Cross and Fisher undertaken for the National Highway Traffic Safety Administration and published in 1977. A detailed analysis of the study appears in John Forester's Bicycle Transportation, 2nd Edition, Table 5-8.

We'll just pursue the parallel path car-bike collision bogey a little further in an effort to finally allay cyclists' fears. However, doing so won't be easy because such fears are deep-rooted. I know, because for many years my greatest concern when cycling on the road was being hit by a car from behind.

That fear was the reason I became a vocal advocate for off-road bicycle facilities while living in England and did not abate until I took a job at McDonnell Douglas, St. Louis, in May, 1980, and met Bob Soetebier, a passionate bicycling advocate. Bob introduced me to the reality of cycling on road, and in particular to the work of John Forester, author of the seminal work on the subject, Effective Cycling, published by MIT Press and now in its sixth edition.

When we pick apart parallel path collisions in urban situations the data shows that, out of the 7% of all urban car-bike collisions falling into this category, less than half (3.3% of all car-bike collisions) are due to a motorist hitting a bicyclist from behind. As a percentage of all car-bike collisions, 1.4% is due to a motorist passing too closely, and 1.9% is attributed to a motorist failing to see the cyclist.