After writing my critique of Cyclecraft a few weeks ago, I noticed that a thread had cropped up on Cyclechat discussing the post. A common research article is cited by those who oppose introduction of Dutch-style cycle infrastructure in the UK is, Bicycle Track and Lanes: A Before and After Study, most commonly linked in the form of a summary report. This is often linked to as "proof" that segregated cycle facilities are dangerous, which in itself is rather unimportant, for reasons to be discussed later. Interestingly, very early on in the introduction, the author writes:"Many studies of bicycle tracks have been undertaken in Northern Europe. A meta analysis of 11 studies shows a reduction of 4 percent in crashes, and the crash reduction is almost the same for pedestrians, bicyclists and motorists respectively."
The meta-analysis being referenced there is from, "The handbook of Road Safety Measures," by Rune Elvik. Meta-analyses are useful because they take a broader consensus from numerous studies, minimising the effect of any flaws or limitations in individual studies by looking for overall trends in the body of work as a whole. Picking a single piece of research which agrees with your own opinion whilst ignoring the wider consensus offered by the body of literature is called "Cherry-picking," and is generally frowned upon.
The study compares the numbers of cycle*-car and cycle-pedestrian crashes on roads with cycle tracks and with cycle lanes, to predicted crash figures based on figures for unaltered roads which have been altered to factor in the alteration to traffic volume and composition. The crash figures for junctions and straight road sections are treated separately, and the study finds that on roads with cycle tracks, cyclist crashes are decreased by 13% on straight sections, whilst at intersections they are increased by 24%. Overall, crashes involving cyclists increased by 10%.
This is the oft-cited percentage increase when discussing Dutch-style segregated cycle facilities with those who are vehemently opposed to them, and it is interesting to see how it is calculated; previously cyclist injuries at junctions had been measured as 353. After the installation of cycle tracks, the number of cyclist injuries at intersections was measured as 285, a reduction of 19% in absolute figures. However, the 24% increase figure is calculated from a predicted number of crashes figure for the after period, based on the changes to the traffic volume and mode composition, which predicted that at unmodified intersections with the same increase in cyclists, decrease in motorists and subject to pre-existing crash trends seen at the intersections which had been modified with cycle tracks, there should be 230 cyclist crashes. This is the figure which is used to generate the eye-catching 24% increase in crashes figure. The author of the paper also states that:
"The construction of bicycle tracks resulted in a 20 percent increase in [bicycle] traffic mileage and a 10 percent reduction in motor vehicle traffic mileage on those roads, where bicycle tracks have been constructed."
Taking intersections and straight sections together gives a figure of a 10% increase in crashes involving cyclists overall versus the predicted figures on un-altered junctions for the same traffic mode/volume composition (broadly speaking, a 10% reduction in motor traffic and a 20% increase in cycle traffic), a composition which is arguably only achievable where segregation is applied. The actual before and after numbers show a decrease in the absolute numbers of cyclist crashes of 29%. It is important to consider the effects of any pre-existing downward trend in crashes which could be contributing to this number, but also important to consider that this effect is seen contemporaneously with an increase in cyclists' mileage of 20% on these facilities.
"The bicycle facilities effects on traffic volumes are rather large. We do not know for sure whether these effects are a result of changes of route choice or transport mode choice or both. The magnitude of the changes in traffic volumes on the reconstituted streets, and the traffic volumes on parallel streets, however, do indicate that thousands of travellers in total must have changed their choice of transport mode. We do not know who have shifted mode - children, middle-ages or elderly, women or men, beginners or experienced, etc."
It is also interesting to note the large effect which the presence or absence of car parking restrictions on the adjacent road has on the number of collisions and injuries for cyclists and pedestrians which occur at intersections. Where parking restrictions were in place, there were more collisions due to the effect of motor vehicles parking on minor roads instead, resulting in more turning thus collisions.
At the beginning of this post, I stated that the safety effect of Dutch-style segregated cycle facilities is actually rather unimportant. Cycling, even on our hostile road network is actually a very low risk activity. A lot of people have invested a lot of time in trying to convey this message, that cycling is actually very safe, low risk and that the benefits from cycling hugely outweigh the risks a person is exposed to by cycling. It also featured as a common theme on the Cyclechat thread too, demonstrations of the statistically low risk which comes from in motor traffic and links to works such as the study discussed above (ignoring the wider consensus offered by the overall body of literature, which are even discussed in the introduction of this particular study).
Generally, the body of research shows that Dutch-style segregated infrastructure moderately decreases the risk to which cyclists are exposed, despite expanding the demographic itself from what is largely a small minority of experienced and vigilant hardcore cyclists under a vehicular approach, to include such disparate groups as teenagers chatting as they ride together or riding alone with earphones in, older people, parents with their children (either on their own bikes or on the parent's bike), children cycling to school without the need for supervisions and boozy revellers returning home from a pub or club by cycle. Despite the incredible broadening of the demographic, safety is still increased.
However, all of this is missing the point. Surprisingly, the main benefit and purpose of implementing Dutch-style segregated cycle infrastructure isn't just to reduce risk, it is to reduce fear. Increasing people's sense of subjective safety is a huge part of making the bicycle seem like an attractive and viable mode of transport to them. Another important factor is convenience. Both the need to feel safe whilst cycling, and the need for it to be convenient are provided where there are Dutch-style segregated facilities (and the extra options it opens up for reducing the speed, volume and permeability available to motorised traffic). All the statistics demonstrating the low-risk of vehicular cycling isn't going to change the average person's mind as long as it doesn't feel safe to them. People don't work that way
"Making these bicycle facilities must have contributed to benefits due to more physical activity, less air pollution, less traffic noise, less oil consumption, etc. [...] The positive benefits may well be much higher than the negative consequences caused by new safety problems." (My emphasis)
*In the study, numbers for cycles and mopeds limited to 30 km/h (which are legally permitted to use cycle tracks in Denmark) are bundled together. Make of this what you will.