Terrorism, and inference

There's a reason why I haven't posted anything about the latest terrorist threat - it's almost impossible to bring rationality to bear on the subject. Here are four reasons why clear thinking is so difficult:
1. The missing counterfactual. Sam Leith asks:

What if, after the attacks on the World Trade Centre or the London Underground, the West had taken a difficult and strange course of action, and done nothing at all?...What if we hadn't invaded anywhere, hadn't, since we couldn't find our real enemies, invented others to strike at?

It's a fair question. But we'll never know for sure, because we can't re-run post-2001 history with a different foreign policy.
Of course, the problem of establishing counterfactuals is widespread in the social sciences. But our solutions are to look for natural experiments; to find some statistically robust correlations or testable mechanisms; or just to be non-dogmatic about our inferences. Few of these remedies, though, seem to be used by those who so confidently answer Sam's question either way.
2. Inference from few data. You might reply that there's a simple answer here: since the US invaded Iraq in 2003, al-Qaeda has attacked Indonesia more often than the US.
But the sample size  - one and zero - is too small to draw robust inferences from.
This problem is a widespread one. Even if we assume there are 1000 Muslim terrorists in Britain, this represents only around 0.06% of the Muslim population. How can we infer much about Muslims generally from such a tiny sample? That's like trying to understand the stock market by looking at it's behaviour one day every six years.
But of course, no-one thinks about sample size. Instead, they commit the availability heuristic - they draw inferences from big, emotionally important, salient events. This is like trying to understand the stock market by looking only at crashes.
3. Outgroup homogeneity bias. We are a diverse group of individuals. Foreigners are all the same. So we think there's a single "Muslim community" - an image fostered by rent-seeking "community leaders."
This leads to absurdities like Blair's call to "mobilise the Islamic community" to defeat terrorism.  To see the cretinism of this, ask: why doesn't Polly Toynbee to use her influence to curb the rhetorical excesses of the Devil's Kitchen? After all, they're both members of the secular liberal community, of which she's a prominent leader.
4. Bayesian conservatism. No-one updates their beliefs about terrorism in line with the data, in proper Bayesian fashion. For example, those who believe "poverty breeds terrorism" are immune to evidence (pdf) to the contrary.
Instead, writing about the terrorist threat, or it's implications for foreign policy or "community" relations - from pretty much any perspective - consists merely of the articulation of tribal prejudices.
This claim, of course, applies to this post as well.

Comments

[Bob B]

Note this confident generalisation:

"There is really no such thing as the 'Muslim community' but a series of distinct Muslim communities, none of which has produced leaders of wisdom, moderation or distinction."
http://www.thebusinessonline.com/Stories.aspx?StoryId=7F974379-D91B-4780-A410-775F06D95B04&page=3

What of all those Royal honours Blair was dolling out? Did none go to Muslims?

[james higham]

Oliver Kamm and the P-G currently have something to say about this issue. As far as I can see, there is a distinct and pervasive 'malcontededness' running right through their communities. Oh look at us - how oppressed we are. There are other communities which just get on with the business of living with one another. And this is not an over-simplification - it is a fundament.

[Bob B]

Too true. We don't get these recurring terrorist threats from the Chinese community settled here, the Hindus, Sikhs and Buddhists.

[chi flat iron]

CHI flat iron by Farouk system. Direct from the manufacturer, this genuine Chi ceramic iron comes with valid, one year warranty!