Different ways of seeing stats
Numbers used to be cold, hard and clinical - a clincher in any argument. Now look at them: fluffy, soft and left meaningless by woolly-headed thinking, says Michael Blastland in his regular column.
Numbers lack warmth. Cold as last year's love, they sit counting their fingers. Think of numbers and what do you see? Dust and ledgers and the yellow fingers of a parched accountant.
No longer. Numbers have had the mother of makeovers. No ordinary scrubbing up, shiny PR or new logo, this transformation is complete: they have turned into their opposite.
Once they stood aloof. Now they gush. Now you can't shut them up for heart-felt passion. They cry and cheer and sneer and shout. Formerly a counterweight to emotion, they are often now nothing but. The one thing they don't do is count.
A number in the news is no longer a cold fact, it is a killer fact, with all the murderous zeal that word implies. Journalists everywhere know the meaning of the phrase, the dagger of detail that runs the opposition through: the 23% up! The £16m wasted! The 140,000 children!
'5.0, 50... how many fingers am I holding up?"
For an example, try 271. I have it on reliable authority that this is the percentage increase in the money supply in the US in five months. There, he's gone cold on me again, you're thinking. But stay a moment. This is the killer fact in a piece by a smart, famous former senior advisor to President Clinton.
The point is obvious. The economy is in weirdness overdrive, things are out of control, cash is flooding through nook and cranny, but barely a cent is spent.
Feel the peril. 271%!
This was no typo. That the writer really believed the magnitude is apparent when he says that 271% is nearly three times higher.
But mull that number. Let it count its fingers a while. This is actually an increase of nearly four times (£100 increased by 271% is £371. A 300% increase is four times bigger, not three). Imagine nearly four times as many dollars in five months. Four times, seriously? What would your salary look like x4? Or GDP?
The actual number in the US statistics was 27.1%. He had moved the decimal point. This is wrong with bells on, but with no evidence of any reflection on the absurdity. I have a number therefore I need not think.
A moment for the number to do what numbers are supposed to do, to count and take seriously the counting, must suggest - mustn't it? - that an increase in the money supply of 271% in five months might be credible in a basket-case inflationary hell but not in America.
My favourite example of this tendency is the heavyweight newspaper report that an increase in retirement age for men from 65 to 67 would mean one in five who formerly lived long enough to claim a pension would now die before they were eligible.
This column aspires to sobriety. The facts might be shocking, the tone will not. Today is an exception.
One in five? Twenty per cent of men die in the space of two years? They'd be turning up their toes all over the golf course, 100,000 of them more than usual. Being 65 would be more dangerous than the front line in Afghanistan.
This column aspires to sobriety. The facts might be shocking, the tone will not. Today is an exception. Today, I'll be frank, the column is like its subject - a great Neanderthal grunt.
It will, from time to time, no doubt be wrong too. But these don't feel like mistakes, they feel like a reality-check failure, a breathless dash into print without passing the corner shop of normal life, all because the number appears to fit the mood.
So 271 doesn't mean 271, it means "wow!" The word "million"' likewise lost its currency years ago. "Billions" might go the same way. As with cliches, so with figures, inflationary use means they buy you less impact. Debts used to be bad, now they're toxic, regardless of whether they might make a good return. A plain old bad debt will never sound as problematic again. A billion will seldom sound so large.
Numbers become, in other words, the very thing numbers are supposed not to be: fur balls of undigested intellectual gunk. Writers who use them this way might as well say: "The government admits the figure is, like, you know, cor! oooooh!"
How big is huge?
Or here's an idea. Why not replace figures with emoticons - a smiley or a frown? These numbers do not measure. They are arrows pointing limply, like vague adjectives, fuzzy conveyors of warmth or fear that beam at you with pride, or scowl with cynicism, expressive of how the writer feels but not of the world he or she claims to feel about.
Is the end nigh, if your colleagues present you with one of these?
Is this a scandal? Not really. Our treatment of numbers is not like our treatment of the old. It's hard, except in rare cases, to be outraged by abused statistics. And in truth there's nothing new here. But we'd do well to remember that numbers often represent people. To be careless of how we describe them in figures is to be careless of the truth about them, which in turn is not to care.
In a new book about energy, Professor David MacKay sees a problem in another subject many care about: the environment. He notes a news report about CO2-reducing LED traffic lights. The energy savings, the report said, could be "huge". Here are adjectives as if they were numbers. "Huge" being how big, exactly?
Professor MacKay found that traffic lights account for about 0.03% of UK energy consumption. Let's say the huge savings cut that by a third. That's 1/10,000th of the total.
One response is that every little helps. Mr MacKay replies that every big helps rather more - but it's no good being dim to the distinction.
How do we get away with it? In the first case because we have a number - and that beats thinking any time; in the second because we don't - and so elide the evidence entirely.
Numbers are often used without a sense of proportion. Proportion is often invoked without a sense of the numbers. But isn't proportion what numbers are for? Perhaps it's time to take out some of the heat.
Below is a selection of your comments.
I've always felt that "billion" should be replaced (in the US) with "thousand million." More money, more words. Not that anyone has taken up my suggestion...
Chris, Kansas City, Missouri USA
As for replacing figures with emoticons - I recently encouraged the amendment of a presentation that had "meaningless" figures (i.e. meaningless except to technical experts) to have smiley faces next to good figures, average faces for average, and sad or angry faces for bad and very bad figures. The original figures were left in for reference and transparency (well, for anyone who would know what they meant), but the interpretation was very useful for those who would not know whether a given figure was good or bad. My aim was not propaganda (besides, there was the risk of being patronising) but rather to just provide a genuinely useful interpretation of technical figures.
As a physicist we get trained from the start that every number must have scale and comparison - something that you rarely see in the media. Pet favourite of mine is the standard of quoting percentages on values where the population is less than 100 (ie action X up 50% year on year - when last year it occurred twice and this year three times).
The classic misuse of numbers for sensational effect is in the phrase "children as young as...". It tells you nothing meaningful except they've managed to find one child of that age. All the others could be any higher age. The BBC is as guilty of this as anyone.
Colin McKenzie, London UK
I think that the increasing abuse of numbers by the press over the past 10 years to perpetrate their profitable new form of "shock-and-awe" journalism has been atrocious. One in five people does this! 80% of people don't know that! No context, no further investigation, and certainly no counter-point to the main story: today's reporter often seems to just grab a number, jump to a conclusion, and then embark on a page of sensationalisation, in the knowledge that the unthinking masses will buy their paper/magazine/whatever just to find out how bad it really is.
Adam the accountant, Edinburgh
Proportion and context, this has been the missing element all throughout this 'downturn' in the economy. The press have helped whip up the public into an ill-informed frenzy due to the lack of proportion and context for which they report a lot of these numbers which has scared us into recession.
Mark Johnson, Leeds, England
As an example of numbers that are often used without a sense of proportion, take the substantial budget deficit the UK government will run in the coming year. The actual numbers in billions of pounds will sound alarming - and they are. But as a percentage of our national income, the new borrowing will be around 10% - and in more "normal" recessions such as the early 1990s recession, the proportion reached 6% of GDP. And we survived that and even prospered in the years thereafter.
So OK, debt as a proportion of GDP will be higher than for decades; but dangerously higher? That's a question we can discuss, but it's better to discuss the relative proportions than merely to talk of the absolute figure. The other example where proportion is used without a sense of the numbers is in the latest medical research. "If you eat this food or drink this drink, your chances of falling prey to some terminal illness is increased x%."
But what were the chances of falling prey to that particular illness? Often very low. To get a sense of the actual risk, ask the researcher what the average life expectancy for the group in question would be if they followed the researcher's advice, and what it would be if they didn't. We might find out that the difference would be a matter of months, rather than years - in other words, the increased risk is hardly worth worrying about.
Richard, Beckenham, Kent