When Goldman Sachs' David Viniar wibbled recently about 25 standard deviation events, we all laughed at him. But he was at least asking an intelligent question: how likely are the events we've seen? And this is the question people aren't asking when the try to draw lessons from the murder of Rhys Jones.
The fact is, it's very rare for youngsters to carry guns or knives, let alone use them (table 2.6 of this pdf). To draw lessons from Rhys' murder is therefore to make inferences from the far end of the probability distribution, without asking what the more general distribution looks like. This is the mistake of the availability heuristic.
But this is not the only cognitive bias afflicting those who want to take a message from Rhys' murder.
Those who do so seem to have had an amazing ability to see what they want to see. David Cameron sees it as evidence of bad parenting and the malign influence of music and video game companies. Warren Bradley wants zero tolerance policing. Everyone seems to be blaming gangs.
I'm not saying these views are wrong or right. But they look like the confirmation bias - the lessons we should learn are those we already knew.
These biases combine to have a pernicious effect. In focusing upon extreme, rare events they stop people asking: what is normal everyday life like? Could it be that - especially in areas where such crime happens - everyday life is one where good people are powerless to prevent social decay? And could it be that politicians just don't want to ask how to empower them? Or is this just my confirmation bias?
It's not just that good people feel powerless. They feel that if they step in to try to improve things, the power of the state will be wielded against them.
Posted by: dearieme | August 25, 2007 at 04:21 PM
Hi S & M
You're right of course. There's also another important contributory factor. The killing of children is publicly perceived as particularly abhorrent and the deliberate killing of a child by another child virtually incomprehensible. Such events are very rare as you say but their perceived gravity makes them appear more influential. After such events, there is a universal demand that 'this must never happen again'. Like the killing of three year old James Bulger in the 90s by two 12 year old boys, this tragedy seems to have been a random or mistaken killing. We can do nothing to prevent random events.
I live in a part of the country where there are gangs and there are guns. You hear kids talk about guns on the bus. Probably they're bragging but most likely, they at least know someone with a gun. Unless you foolishly take yourself into the middle of a dark park in the middle of the night or to a crack den, you are unlikely, in the normal course of events to see a gun, even in Hackney.
But in places where access to guns is much easier (the USA springs to mind), the incidence of kids being killed with guns is higher so it makes sense to limit access to them by whatever means is available. Political manipulation is inevitable. Politicians are expected to speak out about events like this. If they didn't the public would be screaming that they didn't care.
xxx
Pants
Posted by: Pants | August 25, 2007 at 05:58 PM
The killing of children by children, given the evident availability of lethal weapons, seems entirely possible to a parent, amoral little creatures that they are.
Politicians, as they progress (to my cynical mind) become less about changing things for the better and more about retaining power. Obvious, really - once it's your livelihood, it's about keeping your job, after all. So they become more and more obsessed with maintaining the appearance of "getting things done" and less and less about actually doing anything useful. They therefore panfder to the press, who exercise their influence as the politicians' need becomes more apparent. And we get positive feedback of the worst kind.
The Americans are probably further down this particular tawdry road, but we're hard on their heels. More fool us.
Posted by: Mike Woodhouse | August 25, 2007 at 09:18 PM
"When Goldman Sachs' David Viniar wibbled recently about 25 standard deviation events, we all laughed at him. But he was at least asking an intelligent question: how likely are the events we've seen? "
Wrong.
We laughed at him not for asking a new and novel question, but for pretending that he had no idea of the answer to a question that has been asked and answered many times before.
When someone goes to the casino convinced that their new method for betting on roulette based on the song currently playing in the casion will make them rich, leaves poor, and asks "what are the odds that I would lose in this game" we don't applaud them as some sort of visionary in the field of probability; we rightly peg them as an idiot.
Posted by: Maynard Handley | August 25, 2007 at 09:44 PM
S&M: "The fact is, it's very rare for youngsters to carry guns or knives, let alone use them"
Times are changing. In the Sunday press:
"The Home Office figures - which exclude crimes involving air weapons - show the number of deaths and injuries caused by gun attacks in England and Wales soared from 864 in 1998-99 to 3,821 in 2005-06. That means that more than 10 people are injured or killed in a gun attack every day."
http://www.timesonline.co.uk/tol/news/uk/crime/article2328368.ece
Posted by: Bob B | August 26, 2007 at 01:45 AM
Bob, did you even look at Chris's linked evidence? It was data from 2005 on how many kids carry knives and guns, and the answer was 'very few'.
Posted by: john b | August 27, 2007 at 11:12 AM
John B - However many kids own or have access to guns, we are confronted by the horrific Home Office figures showing, "the number of deaths and injuries caused by gun attacks in England and Wales soared from 864 in 1998-99 to 3,821 in 2005-06. That means that more than 10 people are injured or killed in a gun attack every day."
If not many kids have guns, they have evidently each been increasingly active.
Btw how do we know how many kinds own or have ready access to guns since they are unlikely to volunteer that information?
I read that in some places, those with the right connections can easily hire guns at moderately low cost. Presumably, that would be due to the prevailing balance between supply and demand
By comparison, we are more likely to have broadly accurate figures on the numbers of gunshot victims as they tend to end up either in morgues or in hospitals undergoing trauma surgery.
Besides the issue of gunshot homicides, the number of homicides has been increasing according to this report in the press last year:
"Cases of murder and manslaughter have risen by almost a quarter since Labour came to power, Home Office figures have revealed. Since 1997, the number of homicide victims, including solved and unsolved cases, has averaged 737 per year. In the period from 1990 to 1996, the average was 601. The number of homicide victims has averaged 737 per year since Labour came to power."
http://www.telegraph.co.uk/news/main.jhtml?xml=/news/2006/07/09/ncrime09.xml
Posted by: Bob B | August 27, 2007 at 12:52 PM
Forget about our host's recommendation to sample Table 2.6, instead, try Box 1.1 on page 13 which will tell you that the whole rag-bag of spurious statistics based on interviews with sundry 'youfs' and 'youfettes' is on about the same level of reliability as the immigration figures. Your average big city vicar is more worldly than the dopes and dupes who conducted these 'interviews'.
'Bob B', above, quite rightly paraphrases the words of an old-time copper, "Never mind the legalisms, just count the bodies!"
Our host, normally a man of shrewd judgement, in this case must be suffering from the effects of the unexpected sunshine because, of course, it is not the particular numbers that are important, it is the trend. And I don't care how many Home Office statisticians dance on the head of a pin, I can tell you for an absolute fact that hardly any children went around blowing away other children during the 1950s.
Also, the murder rate in 1960 was around 250 but today it is around 750. What are the chances of me being murdered by an adult or a child? Disappointingly tiny, some of you may think. You would be right, but there-in lies the liberal selfishness and lack of empathy. If it hasn't happened to me, why should I worry over much if it happens to some one else? Particularly if they live in some grotty northern city. Pass the Pino Grigio, darling.
Posted by: David Duff | August 27, 2007 at 10:06 PM
Bob (and others): if you haven't already, check out Unity's piece debunking the horribly misleading stats cited by David Davies:
http://www.ministryoftruth.org.uk/2007/08/27/facts-lies-and-gun-crime-statistics/
Duff: do you seriously think the direction the yoof would lie is *downwards*?
Posted by: john b | August 28, 2007 at 12:42 PM
John, the question that needs to be asked is: how in hell do you know *who* is lying? If, and only if, that can be answered with any degree of confidence is it worth trying to sort the boasters from the hard cases.
Face it: the whole thing's not worth a bucket of spit!
Posted by: David Duff | August 28, 2007 at 02:56 PM