On Thursday 18 October, the eve of getstats‘ second anniversary, BBC Four will be showing a documentary on statistics and probability. ‘Tails You Win:the science of chance” will be presented by David Spiegelhalter, Royal Statistical Society fellow, getstats board member and long-time champion for greater statistical literacy in society.
We spoke to David about what it was like to make the BBC documentary, and how to engage teenagers in the numbers behind calculating probability.
How long did it take to make the documentary?
We first put in the proposal in September 2011 so it’s a year from the pitch to the BBC Four showing. We started filming in January – there’s footage of me buying a lottery ticket in thick snow, and you will notice during the programme that the seasons change!
We’ve heard you did a skydive for the programme to illustrate a point about risk – how did that go?
I looked at the calculations, and it’s only about a weeks’ worth of average acute risk – and if you think of death from all causes it’s only about an hours’ worth for someone my age. So you think, come on, you can’t get too concerned about this. But when you’re in a plane, being pushed through a small hole into nothing at all, you’re suddenly confronted with your vulnerability.
So you won’t be decreasing your life expectancy by doing any more skydives?
I’m not going to take it up, no! The point is that it’s very safe as a novice doing a tandem jump, because you’ve got a big chute and you’re strapped to someone who’s done thousands of jumps. It’s quite natural that as people get more into it, it actually can get riskier because they do more daring things such as coming down faster by using smaller chutes.
What other aspects of risk do you look at in the programme?
We look at lots of different aspects of risk and the way it affects our lives. As well as extreme sports we look at weather forecasting, natural disasters such as earthquakes, gambling, health and economic risk. We had to do slightly nerve-wracking things like go into the Bank of England and talk to the Chief Economist, in the room where the monetary policy committee meet.
We were really talking about putting numbers on risks and how much we can, in a way, tame it by trying to quantify it. So it’s very much from a statistical perspective rather than a psychological perspective. We don’t cover the other types of risks such as social risks or reputational risks, how daring we are in the way we live our lives. That’s a whole other ball game.
How challenging was it to make the programme accessible to non-statisticians?
Always there’s a tension in science programming, particularly when there’s some complicated maths involved, about whether you’re allowed to have equations. You will often see equations presented in a blurry, mystical way, floating around in the background – in fact, you might see one or two examples of that in the programme! But I can see what they’re doing, they have to clearly give the impression there is something mathematical behind all this, but not for that to get in the way of the story.
My concern in the programme was not really my colleagues or even statisticians. I was aiming it at people who say, ‘I never could do maths’ and enthusiastic school students – teenagers – who might find this a little bit inspiring.
How important is it for people to understand the numbers behind probability?
There is absolutely no reason why everybody should understand probability – I have real difficulty with it myself – but I think it’s very valuable to be aware of the way that numbers are presented and the tricks people play with them. This goes along with the work of the getstats campaign which is aimed at improving people’s basic numeracy.
So for example, when people say, ‘there’s a risk, therefore we must bring in this law’, that’s not enough. You have to ask, ‘how big is that risk?’. You can’t always measure these things precisely, but you should be able to approach it in a systematic way.
Enormous decisions are made on the basis of numerical calculations. And there are very skilled people out there to do the modelling and calculations. But those sorts of skills should not be left just to professionals buried in offices. I’m a huge believer in public participation in science, of having a public critique of what’s going on. And it’s happening – people can’t get away with using numbers as badly as they have done in the past. There is more criticism and I think there is increased demand for accountability when using numbers.
Your other big interest is teaching probability and risk in schools – how do you go about that?
It’s absolutely vital that probability is taught in schools. I’m working with the Millennium Mathematics Project in Cambridge on trying to get more inspired material into schools. For example, we might look at getting insurance for mobile phones. They know that mobile phones get lost and stolen, so they have to work out whether, if you have to pay this premium charge, how likely do I think this will happen, will it really be worth it and can I cover my costs? There are enormously subtle ideas concerned with risk, expectation, and utility you can teach through that basic idea. And of course there is ‘Deal or No Deal’ as a continual inspiration.
After ‘Tails You Win’ is broadcast, are there any plans to follow it up with any more programmes or talks?
No, this was a complete one-off. And I already do fifty talks a year, I’m trying to cut back on that! I would much rather other people blog and tweet and get out there. I think what’s really important is getting a new generation of people talking about statistics. That is why I’m a huge believer in the Young Statisticians [a section of the Royal Statistical Society]. Statistics isn’t just a subset of maths, it’s a technology that lets people tackle really important and tricky problems.