« Culture and the politics of marriage | Main | End of the Enlightenment »

February 09, 2005

Global patterns in terrorism

Although the severity of terrorist attacks may seem to be either random or highly planned in nature, it turns out that in the long-run, it is neither. By studying the set of all terrorist attacks worldwide between 1968 and 2004, we show that a simple mathematical rule, a power law with exponent close to two, governs the frequency and severity of attacks. Thus, if history is any basis to predict the future, we can predict with some confidence how long it will be before the next catastrophic attack will occur somewhere in the world.

In joint work with Max Young, we've discovered the appearance of a surprising global pattern in terrorism over the past 37 years. The brief write up of our findings is on arXiv.org, and can be found here.

Update: PhysicsWeb has done a brief story covering this work as well. The story is fairly reasonable, although the writer omitted a statement I made about caution with respect to this kind of work. So, here it is:

Generally, one should be cautious when applying the tools of one field (e.g., physics) to make statements in another (e.g., political science) as in this case. The results here turned out to be quite nice, but in approaching similar questions in the future, we will continue to exercise that caution.

posted February 9, 2005 10:30 PM in Scientifically Speaking | permalink


Congratulations on the (semi-) popular press taking notice of your work. I would tend to be cautious with this sort of analysis. Something in my gut has more confidence in say, power law distributions in number of sexual partners (as per the Swedish study a couple of years back), than in this. Also I'm curious about the definition of terrorism and how that influences the numbers.

Posted by: Joshua at February 10, 2005 05:39 PM

Thanks! We were a little surprised when PhysicsWeb picked up the paper from arXiv, but it's exciting to see the work get out there. Ultimately, I can only hope that this result stands as a rational statement in the current din of irrational nitpicking that constitutes national discourse on terrorism.

Incidentally, I've read the Liljeros study (which I think is the original sex-power law study - cond-mat/0106507), and I think our power law (more than 7000 events, over a little less than three decades, using maximum likelihood estimators) is more believable than theirs (2810 respondants, over about two decades of data, using linear regression (known to be a bad estimator)), at least from a purely statistical point-of-view. (As a side point, the idea that preferential attachment generates the sex-power law is just silly.) Unfortunately, while it seems ethical (even desirable) to want to collect more data about the sexual web, I can't say I want to see additional data in order to put our results on a more firm statistical footing. In analyzing the data we had, we were very cautious in concluding that it was more like a power law than some other heavy tailed distribution (hense the log-normal test). But whether you accept the power law or not, it's clear that the largest events are not really outliers, which is a significant and new idea, nor has terrorism changed dramatically in the past few years since the same scaling was present pre-1998, which is a somewhat heretical idea in the current political climate.

One last comment, about the definition of terrorism. This is certainly a weak point of our study, since it's not clear what methodological biases are present in the database in this respect. After all, one man's terrorist may be another man's freedom fighter... I doubt a defensible definition will ever be well established...

Posted by: Aaron at February 11, 2005 01:37 AM