Thursday, July 31, 2008

Crime and cognitive biases

In the course of my day at work I go through a lot of statistics for one reason or another. This one surprised me, from the police recorded crime statistics (XLS):

...

...

In 2007/08 the police recorded 1,899 incidents of rape of a child under the age of 13.

That's far higher than I would have expected, the high hundreds at most. At work, I asked around and everyone else was equally shocked. The number is particularly high when you bear in mind that it is recorded crime and, as we are talking about a sexual offence, almost certainly a severe underestimate.

Now, a common theory for why the public see crime as such a serious issue is that they overestimate its prevalence because of sensationalist news reporting. We see the reports and assume they represent a broader pattern, fall victim to the distortions of the availability heuristic.

If an entire office of people with plenty of exposure to the right-wing press could make the opposite error about a crime that is clearly, in a horrible way, sensational doesn't that raise some broader questions about that hypothesis?

7 comments:

James Schneider said...

That figure is way above what I'd expect. I figured 500, similar to the murder figures. I asked the four nearest people to me in the office and they gave (4373, 1000, 1000, 12) as their figures. Giving an average of 1377.

That stat is giving me much to think about.

RobW said...

You have to differentiate here though. How many of the crimes are committed by a family member? Or at least someone known to the child? I bet it's a lot.

While I bet the sensationalised 'Paedo on every corner' only makes up a small percentage of this.

Matthew Sinclair said...

James,

Yeah, around 500-1000 was my first instinct. 12 would have surprised me less than 1,899. I'm struggling to understand why most of us are lowballing it as well to be honest.

TBRRob,

I think you might well be right. These are a lot of low level, horrible stories that people would rather not hear about, rather than the big cases with a clear-cut evil stranger.

Gracchi said...

My first thought was that most of it was statutory rape- but then I saw you were talking about under 13. Just as a question how many 13 year olds having sex with 16 year olds does the figure include- it would be interesting to know.

Having said that it does seem incredibly high. I agree with tbrobb- there are probably more nasty things happening when the doors are closed and the family shut in on itself than we would guess.

Ruthie said...

Now, a common theory for why the public see crime as such a serious issue is that they overestimate its prevalence because of sensationalist news reporting. We see the reports and assume they represent a broader pattern, fall victim to the distortions of the availability heuristic.

This is called "mean-world syndrome" and it's one of the few mass communication theories that I've done primary research on. I don't think it's limited to sensationalist news reporting-- in our study we found that consumption of violent programming of any kind (video games, CSI, the news, you name it) and the propensity of that person to overestimate the actual percentage of victims of violent crime in their city, state, or country were strongly positively correlated.

That said: holy crap on the number of incidents of child rape. I never would have guessed.

This may be one of those crimes that's not reported as often because of the sensitive nature of the crime-- the victim's youth, the family's desire for anonymity.

Interesting post...

Dave Cole said...

How many people work in your office? Are they a valid sample?

xD.

Matthew Sinclair said...

Probably not, that's why I opened things up to the blog :)