If you have not already done so, you may wish to read the
Introduction to Gnomica.
Statistics show that of those who contract the habit of eating, very few survive.
George Bernard Shaw (26 Jul 1856 – 2 Nov 1950)
I once referred to statistics as the ‘witchcraft of the twentieth century’. But they are a necessary tool for making sense of much of the world we inhabit. And I do believe that statistics as a discipline is important – certainly marketers do, as do pollsters, physicists and medical researchers. And among those types there is a great deal of intelligence at work. By the same token I have never forgotten what a math teacher once said in a college calculus class: it’s not that the mathematics of the statistics is wrong so much as that, even if the assumptions are error-free, they end up being hijacked by every Tom, Dick, and Harry who can get his hands on them in support of this or that favorite argument.
Each is perhaps not unlike Shakespeare’s devil “who can cite scripture for his purpose” (Merchant of Venice I.iii.97), substituting ‘statistics’ for ‘scripture’.
The little bit of Shavian tongue-in-cheek at the top does make a similar point in a more facetious vein. One can hardly fault its self-evident truth but is, of course, pulled up short upon a moment’s reflection on its self-evident vacuity. Of course nobody survives, right, and of course everybody eats, right, so of course there must be a statistical correlation between eating and dying.
I guess there is, but is it a helpful correlation?
In fact, is it even a correlation? Or are we in fact dealing here with a kind of post hoc propter hoc confusion? A coincidental correlation is hardly a causal one, at least not necessarily so. Most people (myself included) know very little if anything about the foundational mathematics underpinning statistics, but (or, perhaps, therefore) the very word – statistics – has acquired an almost religious sense of infallibility, complete – as in the case of any religion — with accompanying cadres of agnostics and atheists. Think the current back-and-forth in the public – and even scientific — discourse about ‘global warming’ and its reasons and even reality, all based on statistics pro and con!
It all depends on how one ‘cites scripture’ to suit a given purpose … the actual scripture as it were is unassailable. Consider some other horrors ‘proved’ by statistics here (medical) here (drugs) and here (epidemiology) and then consider that statistics don’t really prove anything at all except what the statistician(s) want(s) to prove to be the case – in and of themselves, they are just a springboard for further leaps.
My point here?
Well, maybe I wasn’t entirely wrong in my earlier analogy to witchcraft, but not entirely correct either. I guess for me what it amounts to is that whenever I read ‘statistics this and statistics that’ my bullshit sniffer automatically goes on high red alert … and I’ll think twice or more about the latest ‘news’ on the subject at hand.
Do the same!
But … it’s not the math that I suspect, it’s what’s done with the math.