The Tiger that Isn’t (2007) by Michael Blastland and Andrew Dilnot is a fantastic book about the way numbers should be used, how important they are, how little we know about the world in terms of numbers and how they are chronically abused by the press. Blastland and Dilnot started the superb podcast More Or Less that anyone interested in current affairs should listen to.
Blastland and Dilnot point out that many numbers can be made personal. The endless silly stories about how much cancer or some other risk is raised or lowered by X or Y could be very succinctly and well put if journalists adopted a rule of saying that if X or Y were done then the rate of Z would change from, say 4-5 out of every thousand or whatever. It’s simple, easy to work out and would greatly improve things.
They talk about how counting is straightforward, but what you count changes things so much that proponents and opponents of measures can almost always happily find examples to support their case even if the examples when examined show very little. Blastland and Dilnot look at how averages can show all sorts of things. One point they make that is interesting is that most people earn less than an average because the average is pulled out by a small number of very, very high earners.
There is a chapter on targets that looks at the NHS targets in particular and how they have wound up making all sorts of health professionals do silly things in order to improve averages rather than actually improve care. It’s really fascinating. There is a chapter on sampling that points out how easy it is for sampling to confuse, conflate and distort things. On the chapter on Data they show a quiz given to senior public servants and just how badly they know the big numbers about their own country. It’s well worth doing the examples and finding out just how you do.
In the chapter on Shock figures the environmental movement and AGW alarmists come in for criticism with their pushing of extreme figures in order to scare people. Blastland and Dilnot are AGW believers themselves, but point out that much of the environmental movement acts like a lobby group. In the chapter on comparisons Blastland and Dilnot look at UK league tables and talk about the problems with them and that they have caused. It’s food for thought. The book finishes with a chapter on causation and how easily people infer causation and find too many false positives.
The Tiger that Isn’t is a Great book, one that I’d recommend for everyone. It’s well written, about an important subject and is also pleasantly concise.