Safety in numbers (SN) is a phenomenon, popularised by Peter Jacobsen in an article published in 2003 (abstract here: http://injuryprevention.bmj.com/content/9/3/205.abstract), that appears to show that as numbers of walkers and cyclists increase, that the number of injuries does not increase in line as you might expect i.e. if we double the number of cyclists on the road, the number of cyclists being injured does not double – it either increases by a small amount, stays the same, or, counterintuitively, leads to a fall in cyclist injuries. This has then been backed up by various studies and organisations coming out in support of SN – e.g. the BMA (British Medical Association). The suggested mechanism for this effect is that more cyclists on the roads means more awareness of cyclists, with motorists better able to predict their movements and more likely to be looking out for them. Of course, more cyclists means that motorists are also highly likely to be cyclists themselves, meaning a higher level of concern for their wellbeing.
I have seen it suggested in a few places on the internet recently that the ‘Safety in Numbers’ effect for cyclists is an effect caused by an improvement in road safety that leads the increase in cyclists. I do not believe this to be true, for reasons explained below, and think from looking at the data that the single biggest thing we can do to make the roads safer for cyclists is to get more cyclists on the road – obviously better infrastructure which reduces conflict, particularly on higher speed roads, will also reduce deaths and injuries and I would not suggest for a second that we do not need this infrastructure to allow all levels of cyclists to make safe journeys. In fact, I think that a marked growth in infrastructure, e.g. in towns, will lead to cyclists growing in confidence and then moving onto cycling between towns on rural roads. However, I do think that more cyclists on the road will reduce the number of cycling casualties and that the SN effect is valid.
Looking at this graph of death rates per distance travelled and distance travelled in the Netherlands, it’s clear to see that as the distance travelled increases from the 70’s, that death rates fall as levels of cycling increase. We know that it was in the 70’s that the Dutch started building their cycle lanes – so maybe it’s true that safety leads cycle numbers? It’s certainly true that a perceived lack of safety is the single biggest barrier to more people cycling, so it seems rational, right?
Take another look at the graph. What about before the 70’s? We can see a steep drop off in numbers cycling as cars take off, and the death rate per distance traveled soaring. There were cycle lanes then, but not in the same way – they were not connected and of very mixed quality. However, it still seems clear: when less people cycled, the roads became more dangerous for cyclists. There were less cars, sure – just as we’d expect now if more people cycled. As an aside, you might be interested to watch this video ‘how the dutch got their cycle lanes’: http://www.youtube.com/watch?v=XuBdf9jYj7o
Clear correlation, plus a plausible explanatory mechanism means that I think the evidence, on balance, supports the idea that more cyclists will reduce the number of deaths per mile cycling. To those who argue that the statistics seem to show that rising numbers of cyclists being killed here in the UK disproves SN, I would say that SN cannot trump all other factors. In the UK, historically we have low numbers of cyclists which means that we have lost the knowledge of cycling – from how to make sure your brakes work properly, to riding away from car doors to avoid being doored – apart from in a few very experienced, determined cyclists. We have a large number of new cyclists taking to the road who may not have ridden on the roads before, and who may not know someone that does to teach them – this means these new cyclists can be expected to have a significantly higher level of risk than the experienced cyclists; but the benefits of cycling still outweigh the risks many many times over, even for the newest cyclists (ref: http://cyclehelmets.org/1015.html).
To reiterate, I am not saying that this is an argument for not providing infrastructure. I believe that large scale, safe cycling comes from starting large numbers of people riding their bikes young, getting them cycling before they set up habits of driving everywhere. The road system is cycleable in safety currently, but to reduce the risk to a level on a par with driving requires a large amount of training and experience – thousands of miles of riding. How are people to know that this barrier is worth overcoming, if we don’t let them see the benefits of cycling? We cannot start hoards of children riding – to school, to the shops, to the park – without a network of cycle infrastructure.