This one has been languishing in my drafts for quite some time, with both Vole O'Speed and As Easy As Riding A Bike sharing their own views on the issue in the meantime. Statistically speaking, cycling here in the UK appears to be surprisingly safe. Indeed, in the past I have focussed on this when talking to new and potential cyclists about their experiences of poor subjective safety. These statistics are also readily utilised by a vocal minority who are ideologically opposed to the use of separation by mode for the prioritisation and protection of cyclists. As is often the case with statistics they only tell part of the story; whilst it appears at first that they show cycling to be a 'low-risk activity,' what they literally show is that the current sub-section of the population who choose to cycle are doing so relatively safely.
Like most other people who choose to cycle in the UK, when I cycle I do so in a hyper-aware state; I always expect the worst from other road users, I pre-emptively hover over the brakes when I see a car approaching a give-way line where I have clear priority and I plan my escape route for when that BMW makes a sudden turn without indicating. I am relatively fit, fast, I cycle in the optimal gear and I know precisely how much force I can put into the brakes before the wheel locks up. Put simply, the bar it set much higher for cyclists than it is for other road users because the road environment is inherently hostile for cycling. Most people who drive motorised vehicles, which are significantly wider, faster and heavier than bicycles, do not do so in a similar state of hyper-awareness. This is because there is simply no need; the vehicles and road environment have been designed in such a way that their operators are largely protected from the limitations of their own ability. The bar has been set rather too low for such inherently dangerous machines.
I have often thought that if some of the greatest minds of the 1950s were put together in a room and given and nearly unlimited budget and the specific task of designing a road network to minimise the number of people choosing to cycle, without being permitted to explicitly make cycling illegal, the result would not be far off the current UK road network. The exceptional hostility for cycling which is designed into the UK road network is enough to prevent the vast majority of people from every wanting to cycle on it. The result is that those few who are willing to cycle on it are not at all representative of the general population; it is because this minority can cope with the road network as it currently exists that cycling appears to be a statistically safe activity.
In The Netherlands, cycling is statistically slightly safer than the UK. The difference is not as much as might be expected, which is often used as an excuse for opposing the construction of Netherlands-style dedicated cycle infrastructure in the UK. However, with a little context the safety statistics from The Netherlands start to appear much more impressive. By implementing road designs which are not inherently hostile to cycling, the section of the population choosing to cycle is much more representative of society as a whole. The majority of ordinary people, cycling without being in a hyper-aware state typical of UK cyclists manage to get around by bike and are still statistically more safe than the tiny minority of physically and mentally exceptional UK citizens who choose the bicycle. Next time there is a discussion about how safe cycling is, remember that in places such as the UK where cyclists are a tiny minority, the statistics don't tell you a great deal about how safe cycling is, only how safe cyclists are.