Holiday road toll
Here are the data, standardised for population but not for the variation in the length of the period, the weather, or anything else
As you can see, the numbers are going down, and there’s quite a bit of variability — as the police say
“It’s the small things that often contribute to having a significant impact. Small decisions, small errors..”
Fortunately, the random-variation viewpoint is getting a reasonable hearing this year:
- Michael Wright, in the ChCh Press “But the idea that a high holiday road toll exposed its flaws may be dumber still. A holiday week or weekend is too short a period to mean anything more.”
- Eric Crampton, in the Herald: “People have a bad habit of wanting to tell stories about random low-probability events.”
Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »
On average the risk of a car crash is low. But not everyone has that low risk, people make decisions that change the risk – drink driving, speeding, fatigue, dangerous maneuvers etc.
People want to work out what are the characteristics that change a driving event from low risk to high risk.
So I’d change EC’s comment to
“People have worthy intentions in wanting to explain the non-random elements of what are low-probability events on-average.”
10 years ago