Look, cards on the table - I don't like feeling I might be stupid.
So I've been digging further into the crime stats, trying to discover whether we should perhaps take them a bit more seriously, just like the Prog Con says. So far I haven't found anything that convinces me, but I have stumbled across a couple of interesting snippets.
1. The X-files theory
Prog Con spokesman Mark Easton reckons sceptics like me are in the grip of an X-files style conspiracy theory. Apparently we believe:
"...the figures are lies, manufactured by corrupt politicians. This conspiracy theory requires us to believe that literally thousands of professional statisticians, police officers, academics and civilian staff are fiddling the data for no personal gain and in some cases risking professional suicide. I have met no-one who has produced a shred of evidence that the numbers have been got at."
Hmm. Assuming Easton actually believes this, and isn't just wheeling out a Sixth Form Debating Soc straw man, he's clearly never spoken off the record to a serving police officer. Maybe he should check out the world weary Inspector Gadget discussing what happens in the real world of police crime stats:
"Almost exactly one year ago, I wrote about Section 5 of the Public Order Act (popularly known as Threatening Behaviour).
Section Five is classified by the Home Office as a ‘violent crime’.
When the Government want to reduce crime, it is important that we stop arresting offenders for this, and use Drunk & Disorderly arrest powers instead. D & D does not show up on Government statistics.
Thanks to this (and other Spanish Practices) the Government can now show a reduction the amount of crime.
It now becomes important to show an increase in the detection of crime. We have now been told to start arresting for Section Five again. We can detect this with one of the laughable new Penalty Notice tickets.
So now we suddenly find that Section Five is back in fashion. It is the new ‘black’. ‘Senior Officers are just insisting on it darling!’
Is that a "conspiracy"? Tens of thousands of police officers gaming the recorded crime stats in order to advance their professional careers?
I prefer to think of it as what inevitably happens when politicos try to manage everything from Whitehall via stats and targets, rather than allowing local communities to control their own policing.
But it does mean you should never trust what the recorded crime stats say.
2. Can you spare a few minutes... er, hours?
We hear a lot about the British Crime Survey - the government's opinion poll - but how many of us have actually seen it? I decided to take a peek.
Blimey! It turns out to be huge - 285 pages of questions. 285 pages!
So you're sitting at home watching the Midsomer Murders.
You shuffle to the door.
"Whatever it is I'm not interested. Go away before I set the wife on you."
"No, hah-hah, I'm not selling anything. No. I'm from the government and I wondered if you could spare the next three hours to answer 285 pages of questions, many of which are extraordinarily repetitive and/or grossly intrusive?"
"Why, that's different. Yes, of course, I'd be delighted to help. Do come in."
So who do they get to answer? They say there's an overall response rate of 75% (2006-07), which sounds reasonable. But who are the ones who don't respond? I've certainly never come across anyone who's done it. Why should you? It's not compulsory and it takes nearly an hour.
They say that reponse rates are highest in rural areas, and lowest in inner cities, especially London, where the overall response was only 62%. Which surely means they under-represent the crime blackspots, doesn't it?
Well, yes, of course they can adjust their sample weights to compensate, but you can never be absolutely confident you've compensated properly for information you don't have. The Home Office puts a +/-3.5% statistical confidence interval around its estimate of total crime (an uncertainty margin that rarely gets mentioned by ministers). But that is not the sum total of the uncertainty.
As always, you don't know what you don't know. What if there's a hidden relationship, which seems quite likely, between willingness to respond and experience of crime? Surely then crime will be understated in the survey. Or maybe overstated. Or maybe... in truth, we have absolutely no idea.
To be continued...