Terrorist Attack Data Analysis – Business Is Business

Here’s an interesting paper that analyses over 3000 terrorist attacks since 1968, correlating frequency of attacks against a terrorists group’s levels of experience over time. There’s quite a bit of academic statistical equations and language to wade through, but it’s worth it for the nuggets within. The authors posit that there is an analogy between terrorist groups and manufacturing companies, demonstrating that organisational growth theories applied to a commercial company making widgets can also apply to a terrorist group “making” political violence. This is a new idea, attractive, I think, but I have a few reservations about the data.

Firstly, 3000 attacks isn’t actually a very big sample for the period. My friends at HMS have collected about 150,000 in the last 8 years alone. The main data source for this study appears to be from MIPT but this has some flaws, with regard to its definitions specifically the recording of international terrorist events where a politically correct definition in the past has hampered various studies.

Secondly, the 70’s in Europe were very different terrorist playgrounds from Iraq of 2003- 2007. There may be a chance of comparing apples with oranges there. Thirdly some of the data is a little soft, such as the estimates of group size which is only available by surveying “experts”.

Those quibbles aside, several interesting factors that do come out are things like:

• The mean delay between the first and second fatal attacks of a terrorist group is almost six months
• After 13 attacks, the mean delay between attacks is less than one month.
• The severity of the attack appears to be independent of group experience. (I think this is quite significant)
• Older groups however remain significantly more lethal overall because they attack much more frequently than small groups, not because their individual attacks are more deadly.
• Timing of attacks appears driven by internal factors.
• Most groups never mount more than 100 attacks (there are some exceptions of course)
• Two key reasons for the acceleration of frequency of attacks (following the business analogy)
  • Organisational learning (learning by doing, getting better at planning, conduct of operations and coordination by experience
  • Organisational growth (by recruiting greater capacity) 

Comments

[robi]

the link does not seem to work.

[Roger]

The artcice appears to have been removed from source. If you click through http://de.scientificcommons.org/41491120, you get the same negative result. It was ceratinly up there yesterday. Apologies , i will try and find another reference.

[Roger]

Link appears to be working again.