Shocking

Today's Guardian has an absolutely shocking story about swine flu. Shocking because someone - the subs presumably - doesn't know the difference between could and will. It is nevertheless good news.

The headline says that swine flu won't be as dangerous as was thought. But the subheadline gets all mixed up:

The estimate of the number of Britons who will die of swine flu this winter has fallen dramatically after health experts admitted the virus is less lethal then they feared

When it was estimated in July that that up to 65,000 people could be killed across the UK, a few tabloids made a meal of this worst case scenario but it was generally clear that that was what it was. Now the official estimate of the number of Britons who could die this winter from swine flu is to be reduced substantially to roughly 20,000.

In both of these cases, I've quoted from the Guardian article itself, which in both cases talks about the number of people who could die. The article also quotes Scottish health minister Nicola Sturgeon as saying "that that (sic) official worst case scenario had been revised downwards".

"Worst case scenario", "could" and "could". Which bit of that did the sub on this story not get? Thankfully it is a story that is about the numbers being revised downwards, which lessens the impact. If the story was that more people "will" die such sloppiness would be outrageous.

Could does not mean will

The Times reports today that waiting five minutes before drinking your tea or coffee "could save your life".

It's another one of those media stories about medical research that gets confused over relative risks. Drinking hot tea may increase fivefold or eightfold your risk of cancer of the oesophagus, depending on how hot you have it. Apparently this kills "more than half a million people around the world every year".

It doesn't take a lot of maths to work out that this is a pretty minimal risk, multiplied by eight or otherwise. With such a small risk, the idea that letting your drink cool down could "save your life" you life is a bit far-fetched.

Spin doctor at work

A Downing Street political adviser overrode the "forcible" concerns of the NHS's chief statistician, who refused to sanction the publication of controversial figures on knife crime and hospital admissions, MPs were told today.

So a spin doctor overruled the exports to make the government's case more compelling. Where have we heard that before?

It all adds up

Philip Johnston in the Times says the knife crime figures scandal shows that "you can never believe a Labour statistic". People in the government don't seem to realise how counter-productive it is to acquire this kind of reputation for spin. Apparently, having briefed the media:

suspicions were aroused when the Home Office refused to publish the detailed statistics, bizarrely citing "police confidentiality".

The truth soon emerged: the official statisticians did not want any figures released because they were incomplete and had not been subject to proper methodological scrutiny.

Obviously, this raises further questions about arresting people for leaking - at the Home Office. Labour leaks when it suits it and then gets on its high horse about national security, confidentiality and the workings of government when that suits it instead.

Whose fault is that?

The RAC is saying that

The cost of motoring has fallen by 18% in real terms over the past 20 years

although you'd never believe it, listening to the media. The BBC story on this seems to be getting a bit confused about statistics.

more than half of drivers think Britain's roads will be gridlocked in the next 20 years - despite the average annual mileage of British drivers dropping from 10,200 in 1988 to 9,070 in 2008.

Obviously, if drivers are spending less time on the road, roads won't be so busy (assuming they are the same roads), unless there are more drivers...

The number of households with a car has gone from 14m in 1988 to 19.5m - an increase of 39%

The number of women drivers on UK roads has gone from 10.2m to 15.3m

Locking up the truth

Meanwhile the Evening Standard reports that a report into Metropolitan Police's handling of the aftermath of the shooting of Jean Charles de Menezes is being held back until after the London Mayoral election.

In the same paper Andrew Gilligan says that Met officers at local level are stopping people reporting crime, in order to keep numbers down.

It doesn't add up

If you believe all the press and t.v. reports over the last few weeks, prices have gone through the roof. But the annual rate of inflation remains at 2.5% - pretty low by historic standards.

Meanwhile, the Guardian says that:

Downing Street today promised further action to help borrowers as new figures showed that house prices are falling at their fastest rate since records began 30 years ago.

In fact, as the BBC says:

The Royal Institution of Chartered Surveyors' (Rics) said that 78.5% more surveyors reported a fall than a rise in house prices in March.

This was the gloomiest reading since Rics began the survey in 1978.

Which is not the same thing. It's tempting to say that journalists twist statistics but I think the truth is that most don't understand them.

Spin, lies and statistics

A good piece from Philip Johnston in the Telegraph about government abuse of official statistics and how a new dawn of transparency might not be as shiny as it seems. Johnson argues that "ministers have already flunked their first big test of this supposed new era by refusing to relinquish their privileged and extravagantly premature access to the statistics."

I think he probably has Freedom of Information in mind when he says:

This is typical of this Government. It has a good idea, which is to make the release of official statistics subject to independent monitoring and parliamentary accountability. But once it has garnered the plaudits for having made the pledge, it then begins to fudge it.