Sara Stace from Link Place wrote a piece last week arguing that cycling is much safer than we tend to think it is. She used data from the 2015 National Cycling Participation Survey (NCPS) to estimate the fatality rate for cyclists in Australia:
Our best estimate is that – on a per kilometre basis – cycling in Australia is slightly more risky than being in a car. That is, for every 10 million kilometres cycled there are 0.073 fatalities compared to 0.047 for car passengers and drivers.
That makes cycling in Australia about 50% more dangerous – measured in terms of fatalities – than driving. That’s certainly lower than other sources suggest (e.g. see Is cycling more dangerous than driving?).
These sorts of exercises are always fraught because of data limitations. For example, the NCPS relies on respondents recalling and estimating in the course of a phone call how much time they spent cycling in the last week. An adult in the household estimates the time spent cycling by children under 15 years of age.
The NCPS can’t provide an estimate of distance so Ms Stace assumed an average 15 km per hour based on what happens in Copenhagen. The NCPS also can’t provide information on travel by other modes used for comparison, so these come from an unrelated data base (not specified in the article).
So is there another way to look at this question?
The newly published third phase of the Victorian Integrated Survey of Travel and Activity (VISTA) provides data on travel in 2012-13 by all modes, including bicycles. It has some advantages.
It uses a sample of 5,780 households in Victoria compared to NCPS’s 8,375 for the nation (and 491 in Victoria). Most importantly, it’s a formal travel survey, so households are formally enrolled in the program and use a diary to record travel by all modes.
On the other hand, a key limitation is the data I have access to only shows weekday travel i.e. excluding weekends and public holidays. It also only collects limited data on regional travel.
So I’ve confined my examination to weekday travel in Melbourne. That means my numbers aren’t directly comparable with Ms Stace’s estimates.
Here’s how I arrived at my estimates:
- VISTA show 243,500 trips by bicycle on each of 250 weekdays in Melbourne; av trip distance is 4.4 km.
- The Victorian Transport Accident Commission (TAC) road trauma statistics database shows on average three cyclists were killed on a weekday on Melbourne’s roads per year between 2012 and 2015.
- I took the same approach to estimate weekday travel and fatalities by car and walking.
My calculations put the Monday to Friday fatality rate for cycling in Melbourne at 0.112 deaths per ten million km cycled. That’s more than five times higher than the 0.021 fatality rate for weekday car travel.
So cycling on roads in Melbourne from Monday to Friday is considerably more dangerous for life than driving when exposure is taken into account. It’s much safer however than walking, which has a weekday rate of 0.402 fatalities per ten million km walked.
As noted, I don’t have data for weekends. However if it’s assumed travel is the same on weekend days and public holidays as during the week, the fatality rate for both cycling and driving over a full week is much the same as mid week i.e. 0.109 for cycling and 0.020 for driving. Walking drops more; to 0.363 fatalities per ten million million km.
It’s important to note though that the likelihood of dying while cycling on Melbourne’s roads is still extremely low. While on average there are three cycling deaths per year on a weekday, that’s over 268 million kilometres of weekday cycling. Of course cycling also offers individual and social benefits that driving doesn’t.
Bear in mind also that the long-term trend is showing a decline in on-road cycling fatalities at the same time as cycling is becoming more popular (see Why is the number of cycling fatalities falling? and Is it just vehicles or are MAMILs killing themselves too?).
Cyclists should worry more about getting seriously injured. In the 12 months ending August 2014, the TAC’s road trauma database shows 336 Melbourne cyclists were admitted to hospital (69 of them on the weekend) – see Is cycling on roads getting safer or more dangerous?.