The last Thursday in September always marks the publication of the full Recorded Road Casualties Great Britain results by the Department for Transport
and today they released the 2015 version. The report has moved on from being a series of charts and tables regurgitating data and now offers much more insight into why trends are occurring and inject a good amount of statistical techniques for good measure. I have previously commented on their insight into
weather and the
bottoming out of casualty reductions and now I want to consider if there is a regional trend that can be reliably analysed.
It is clear to see for anyone casually observing the trend line that since 2010 there hasn’t been a significant fall in the number of people killed on the roads. In fact the DfT say so themselves:
“There has been no clear trend in the number of fatalities since around 2011 (see front page chart). Prior to that, and particularly during 2006 to 2010, the general trend was for fatalities to fall. Since that point, though, most of the year on year changes are either explained by one-off effects (for instance, the snow in 2010) or natural variation. The evidence, points towards Britain being in a period when the fatality numbers are fairly stable and most of the changes relate random variation.”
The use of the term ‘random variation’ when referring to someone’s life is a little unfortunate but policy makers need to think carefully, and objectively when making decision about delivering interventions, whether they are based in engineering, enforcement or education. The problem is that when you look at small sample sizes, like annual number of people killed on local roads, then numbers to go up and down a bit.