Yet again the BBC does something awful to the English language. What? I hear you say – surely that’s not possible, they are the bastion of good English and pronunciation. Possibly true, unless numbers and numerical concepts are involved.
The word that’s being wrecked this week is random. I heard a trailer for a programme called ‘Random Edition’. Apparently each week they are going to take a random edition of a newspaper and go back over the stories to explore the historical significance. The first random edition is September 4, 1939. So on the 70th anniversary of war breaking out they are expecting me to believe that they have managed to pick out the day after war was declared at random from all the other days they might have picked in the last seventy years. The odds of getting any particular day in that time period are more than 20,000 to 1
What random means is that all the possible choices are equally likely. I’m prepared to bet that all of the editions that are picked in this series will turn out to be momentous days in some way or another. I haven’t seen any publicity material about the programme, other than hearing a flyer for the first episode, but it wouldn’t surprise me if we happened – at ‘random’ - to get the day Kennedy was shot, the day the Berlin wall fell and other days like that.
Of course I’m cheating a bit because this is not the first series. The last series featured Lindberg’s first flight across the Atlantic, the first performance of Handel’s royal firework music, Queen Victoria’s golden jubilee, and the restoration of the monarchy under Charles II. Hardly random events!
In my somewhat biased way – I’ve spent my life involved with epidemiology after all – I think this is part of a pattern of lack of numeracy at the BBC. Twice in the last five years I’ve heard items on the Today programme say something along the lines that in properly conducted randomised trials of prayer to alleviate sickness, that half of the attempts seem to work. The protagonist for prayer then cheerfully says that if it works half the time there must be something in it. This like saying that if I toss a coin and get it right half of the time then I must be a clairvoyant or fortune teller. The stupidity of the statement ought to be obvious to everyone yet on neither occasion did John Humphreys challenge the speaker, presumably because he has no understanding of what a randomised trial is, or how it works
When we try out new drugs the concept of randomisation is important, we usually divide the patients in the trial into two or more groups that are selected at random before treatment. That means that every patient has an equal likelihood of getting the new drug or of getting some other drug or treatment that we want to compare the new one with. That way, when we add up the results of the different groups, if there is a difference between the treatments, we have good grounds for thinking that the difference is caused by the treatments and not by some characteristic of the patients or their circumstance. If they are really chosen at random then the samples are much less likely to be biased and the results are more likely to be true.
If we misuse the word random so that the public think that a random date just means an interesting date, picked so that more people will listen to the programme, then how on earth will we convince them that treatments that have been shown to be worthless in randomised trials are not worth taking, or hazards shown to be dangerous are best avoided?
Our whole society is becoming hopeless at making sense of risk, and part of that comes from lack of understanding of the mathematical concepts. I think it’s time that the media faced up to their responsibility in this.
It’s also true that advertisers have a lot to answer for. How do we let the lottery, which is supposed to be random, get away with saying ‘It could be you’? OK I guess it is true that it could be you, but it would be more honest to say ‘It’s really very very unlikely to be you.’
If I invented a new drug that had a one in a million chance of making you better then I’m sure I wouldn’t be allowed to say ’This could cure you’, but such a statement would be just as honest as the advertising for the lottery, it gives a lot of money to good causes. What I object to is a steady process whereby the population is deskilled in assessing risk.
The consequence is that people smoke, drink and feed themselves to an early death. I don’t mind if they really do make an informed choice to do that, but if the BBC and other media keep wrecking any chance that the population will understand statistical concepts then informed choice is very unlikely.
No doubt someone will point out that the BBC does produce one programme (‘More or Less’ on Radio 4) which is about making sense of these concepts, but by putting it in a separate box and labelling it as the programme that is about numbers they pretty much signal that this is the one to switch off. What is needed is a change in culture at the BBC so that they are as scrupulous about statistical and mathematical truths as they are about other issues in the integrity of newsgathering and reporting. That is obviously harder to bring off than dragging in a few geeks to make one programme a week about numbers.
Why am I banging on about this? I think it’s a public health issue. If no one understands risk, then too many people will die and be disabled by diseases they could have avoided, and public health professionals like me will be forever labelled as killjoys for trying to legislate our way to a healthier society. If no one understands risks then risky things have to be banned in order to make people safe. In the long run that is a bad idea, it makes us all even less able to deal with the next risk to come along.
A long and happy life does not happen by accident – it’s not random.