Deadliest jobs
Q: What proportion of fatal car crashes involve an alcohol-impaired driver
A: I can’t find the NZ figures, but according to the Centers for Disease Control and Prevention, in the US it’s about 1 in 3
Q: Since everyone involved is sober in 2/3 of crashes, does that mean it’s safer to drive drunk?
A: Why would you ask such a stupid question?
Q: Because I’m just a rhetorical foil, helping you to externalize a really complicated situation on the blog?
A: Perhaps. But we nearly digress.
Q: Can you elaborate on why it’s such a stupid question?
A: Of course. Drunk driving is dangerous because each individual is more likely to be involved in a crash than if they weren’t drunk. The number of crashes divided by the number of drunk drivers is larger than the number of crashes for sober drivers divided by the number of sober drivers.
Q: So if we said farming was the deadliest job in NZ, we would mean that the proportion of deaths among farmers was higher than the proportion of deaths among, say, forestry workers?
A: Yes. Unless we were Fairfax NZ, when we would use the number of deaths, without accounting for the population size.
Q: Is this pointless statistical pedantry?
A: Not entirely. The article compares ‘dangerous’ farming to relatively ‘safe’ forestry, so it really is important that there are a about 15 times as many farmers (industrial classification A01) as forestry/logging workers (A03). The 39 deaths among farmers compares to 3 in forestry, which is a bit less than a 15:1 ratio.
Q: Could you also go on endlessly about Poisson variation and one year not being long enough to see trends?
A: Yes, but it’s a holiday, so I won’t.
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 »
Hi Thomas,
Another way to look at the relative risks is to check the ACC (Accident Compensation Corporation) levies for different occupations (PDF). There the levy for people working in forestry ($5.30) is higher than for any other occupation in agriculture (except for shearing). Of course this includes all accidents (not only fatal ones), but it is a good start.
XKCD had a good comic on reference population http://xkcd.com/795/ (read the mouseover).
Couldn’t the stat of the week (or the summer) competition an example of missing the reference to the population size? Thus, we are highlighting the outrageous cases without taking into account that there could be many cases of correctly reported stats in the media.
Nice meeting you in Kiama.
Cheers,
Luis
12 years ago