Crime: up, down, shake it all around

This piece is quite interesting, and broadly correct, I think, but the really good stuff is in the comments below. Amid the general distrust of government statistics (often, any statistics) are some pretty strong statements about (a) the British Crime Survey (it can't be trusted) and (b) police recorded crime statistics, which are either the only reliable source - presumably because they are 'real' - or completely made up.

This chimes with something I was going to write in the wake of the latest Tory fast-and-loose interpretation of violent crime stats, until I was more or less beaten to it. But I think the point is worth repeating in some sense, because I really can't fathom what people's problem with the BCS is. As long as you bear the caveats in mind (over 16s only; personal/household crime only; probably misses the truly excluded) it has to be the most reliable record of crime victimisation we have. One simple reason: the sample size is truly gigantuan, itself a reflection of the relative frequency of some of the crimes it's trying to pick up, whereas so many of the objections seem to made on the basis of anecdote and/or common sense ('what we all know'). There are other reasons, but I think this on its own is a massive point.

How people can say they know crime is going up in the face of the weight of evidence the BCS provides is beyond me. Unless of course they really believe it's all a conspiracy. Which some probably do, but I don't get the sense that it's a particularly common viewpoint. It seems to be more a simple denial that the world could possibly be different to way people believe it to me. Which always, to me, begs the question in such debates - my conclusion that crime is (still) falling is based on a representative sample survey with 40,000 respondents carried out by a reputable survey company. And your assertion that is going up is based on ....... what, exactly?

The BCS is most certainly not perfect, but we can be more sure that we 'know' things about trends in and patterns of crime - as opposed to crime events - based on the data it provides than we can 'know' things based on what a bloke in the pub said, what we read in the press, and yes, what actually happens to us as individuals.

I know I'm being hopelessly naive about the possibility for rational debate on this subject on the interweb.

On the other hand, I think most people recognise using police recorded crime statistics as anything other than a record of police activity is problematic, to say the least. Which is not to say they're useless, by any means. But the idea that anyone should place more store on police data than BCS to look at trends over time, particularly, is frankly ridiculous. Not that this will stop people doing it, especially when it's much harder to fudge BCS data to make it suit their particular narratives.