Data journalism needs space to give readers context and it needs a correct understanding of data to back up claims, but today is the second day in a row in which The Irish Times serves up an article which gives us just half the story on cycling deaths and injuries.
Yesterday we covered how the paper got it wrong, and they are at it again today.
For example, there’s this gem of a generalisation: and it does not matter that they are accounting for population, mixing all levels of collisions from minor to deaths and all modes of transport is just nonsense, there’s no other words for it:
“The majority of collisions take place in Dublin – but when the population and volume of people is factored in, it is one of the safer counties. Louth, by contrast, has the highest number of collisions per 1,000 people.”
Again today we get more on cycling without mentioning that the only established trend on deaths is a decline. It does not cut it to say “though numbers have varied significantly year on year” — the article starts by talking about trends and with cycling deaths the trend is decline.
Just like yesterday, when it comes to cycling, the newspaper provided little or no regional context. Here’s an example of such context, as we reported on January 1, 2015:
IrishCycle.com: “Most [fatal] collisions happened outside Co Dublin, but 2014 was the first year in five years in which the cycling death rate in Co Dublin rose above a single death. These deaths included a woman who was struck by a bus in Dublin City in November; in July a man died when he was hit on the motorway near Swords in the Fingal County Council area; and a man was in a collision near City West in the South Dublin County Council area in January.”
And there’s also no basic context that a large bulk of people who use cycling as transport in Ireland commute into or around the Dublin City Council area. An area where the death rate was between none and one in the last five year — a decline from past years which has happened in the same time when the numbers of people cycling in Dublin City has had notable increases year-on-year every year in the same time frame. If the injury rate has increased in this area, please report it, that’s also worth noting.
There’s also no other context about the rise in cycling deaths nationally. For example, that a number of the deaths in 2014 were what seem like rare types of deaths — such as two people cycling killed in collisions with no other road users or vehicles involved, one killed on a motorway, one involving an oil spill on a decent and another involving a bollard. Those combined account for a large percentage of the sharp rise in deaths — but we need to look in detail at the collisions.
…which online at least (“Driver fatalities down on Irish roads, but pedestrians and cyclists at more risk”) is the same as the Tweet above, is not supported by facts. Risk of death in one year does not equal the same this year, and as we have explained above, there’s many factors to consider, including area/regional differences and possible rare risk.
We’re also told by The Irish Times that: “[cycling] Injuries, too, have increased, up from 300 to 400 a year prior to 2012 to almost double that number in recent years, according to statistics and informed sources.” …which statistics? What kind of sources? How does this correspond to the increase in transport, leisure and sporting cycling? What kind of injuries? Were most minor? Has there been a change in reporting?
Deaths and serious injuries are highy important, minor injuries are far less so. It must be remembered that at a population level that the health and mental health benefits of cycling far outweigh any risk of cycling — studies have found this to be true even for roads statically less safer than our roads and streets. So, that should be account for and the focus should be addressing collisions with potential for serious outcomes.
Looking in more detail across a number of years is something the RSA has failed to do. The Irish Times could have focused on this, instead, it mostly just repeated old RSA stats, with less context than the RSA and with conclusions not supported in the article.
This isn’t nitpick or anything like that, we need detailed data for a number of recent years and well-thought-out analyses of that data to figure out what are the main causes of cycling deaths and serious injuries are and how do we start to address those. Cycling deaths are relatively low over a number of years, but like a growing number of places we need a “vision zero” on deaths and serious injuries.
These new RSA stats are compiled from Garda Road Traffic Collision Reports.
The injuries are those reported by Gardaí, and are overwhelmingly minor injuries (e.g. 95% of injuries in 2012, the year of the injury spike).
The number of serious injuries in 2012 is very similar to two other maxima found over the last ten years, when cycling levels were lower. The number of minor injuries is unprecedented, which is the only genuinely remarkable finding in this report. The rest is as yet indistinguishable from normal swings you get with very small numbers.
It’s very misleading to talk about a rise in “death and injuries”, when your evidence for both is single data points in otherwise downward trends, and when the rise in injuries is predominantly minor injuries.
The idea that Dublin is especially dangerous (in both the IT editorial and the RSA report) is largely founded on the minor injuries too. They didn’t base it on fatalities or serious injuries, but on total injuries, which is again misleading.
Actually, the idea that Dublin is especially dangerous in the IT editorial seems to be from a time (single data points again?) when Dublin featured more heavily than usual in the fatality stats. Dublin doesn’t usually “generate half of all fatalities”.
In the RSA report, it’s based on a mismatch between the proportion of Dubliners cycling to work/school (6%) and the the proportion of reported injuries in the RTC reports represented by cyclists (18%). Again, these are overwhelmingly minor injuries, so what the report really shows for Dublin, if it shows anything, is that you’re more likely to suffer a minor injury cycling than driving or taking public transport. Which seems plausible, but is a funny definition of “dangerous”.
Thanx once again Cian for beginning to challenge loose analyses! It would help if a good statistician could gather the last 20 years of data for comparative purposes, just to illustrate how much things have actually improved for cyclists!
It is also disappointing in the extreme to read today’s Irish Times editorial, which broadly lays the onus on cyclists to ‘protect themselves’, while at the same time making no reference to the issue of motorists regularly breaking the speed limits, or their particular responsibilities.
Once again it is worth repeating again an and again, in relation to Road Safety – Cars Kill, Cyclists Don’t!