Jul 20, 2008

"Statistics is for sissies!"

There are times I really hope knife crime is as out of control and as common as certain tabloids are making it out to be. Because then pricks like Keith Waterhouse might get stabbed in the hand and be unable to write such incredible bollocks.*

We are among sceptics

he says as he goes on to give another pronoucement amount knife crime. Now, remind me again, is scepticism that thing where you've already made up your mind about something and fix the evidence around the policy? No? My mistake then, carry on.

There are lies, damned lies, statistics, figures in a plain brown envelope slipped to Pollyanna Toynbee in an alley round the back of the Home Office, graphs and tables from the British Crime Survey and estimates from the National Guesswork Authority.

What a delightful looking strawman you have there, Mr Waterhouse. May I enquire as to its precise purpose?

Stripped of all its flim-flam, the British Crime Survey is about as informative - for one thing, no one under the age of 16 allowed to give evidence, so that loses a whole stratum of new-type knife victims - as a questionnaire on your favourite shampoo.

Oh? And do you have any figures to support this?

The plain fact is that nobody believes a blind word the Government, or its minders and manipulators, say any more. Truth is not a commodity in which they wish to have further dealings.

I'll take that as a "no" then. What then do you base your assessment on?

A few nights ago, as an 18-year-old stab victim lay in a pool of blood awaiting his statistical turn to become the 21st teenager to die violently in the streets of London this year, we learned that crime statistics are dropping dramatically.

Oh, emotional appeal. Interesting, and obviously tragic for the poor lad. However, it doesn't actually disprove the government's figures. Would you like to try something else?

Addresses of interviewees, we're told, are drawn at random from post office lists.

Which means that if you live in a crime-free area, you're going to give crime-free replies.

Well, yes, that's kind of how it works. You know, if there is no crime around, then of course it won't get reported. As opposed to the Daily Mail's wonderful method of statistical analysis, where they used police reports of knife crime to claim that there had been a 35% rise in some parts of the country - which is true - but also failed to report that these were in fact in the minority as crime had dropped in most of the country. Could you please comment on this discrepancy?

And meanwhile the altogether different police figures, based on whatever they have managed to dig out of their hectares of paperwork, are - how shall I put this? - altogether different.

Yes, you could say that. They show that while 4 police forces have seen a huge rise in knife related crime, 18 have seen a drop. All in all, in 20 forces in the country, knife crime has either stayed level or gotten better, whereas in 14 forces in total, it has gotten worse. Now, which of these numbers is larger 14, or 20, 14, or 20? Take your time.

If I was told by some government agency that tomorrow was Tuesday, I should check it against the calendar. Then I would shake the calendar to make sure it was still working.

Well yes. That's because you're a moron, as the above post so clearly illustrates. For those interested in a more thorough debunking of the knife crime panic, Five Chinese Crackers has the relevant material. Surprisingly, David Aaronovitch also has a good article.


*I don't really wish he'd get stabbed in the hand. Mainly because I know he'd dictate his article to someone else, and would probably never shut up about how he was the victim of a "horrifying knife crime!!212!" What about me, having to read this and be the victim of a horrifying logic crime? Never thinking of the little man, these bloody journalists....

No comments: