Several thoughts on this interesting monograph:

Importantly, this study does not argue or imply that violence perpetrated by Sunni or Shi’a extremist groups is carried out for reasons that are inherent to one or the other tradition of Islam. There is no evidence sought or identified in this work contending that historical patterns of terrorist attacks by organizations identifying themselves as Shi’a, for example, are pre‐determined by Shi’a theology or philosophy. What this study does suggest, however, is that those Shi’a organizations that have conducted terrorism in non‐combat zones display several preferred operational patterns that contrast markedly from their Sunni counterparts.
Indeed, most of the Shi'ite cases relate to Iran, Hizbullah, or both. Given this, and the conclusion above, I wonder whether its really useful to speak of a mode of "Shi'ite terrorism" at all, given that the events data is largely driven by two (rather rational) actors.

Related to this is the issue of how terrorist "incidents" are coded:

  • Only terrorism in "non-combat zones" is counted. From an Iranian point of view, however, the GCC states certainly would have been considered a combat zone in the 1980s, given active and extensive Gulf financing of the Iraqi war effort.
  • Similarly, by any definition Lebanon was an active combat zone in the 1980s, given the Israeli occupation—although it is not listed as one. It is not clear to me whether Amal violence against civilians during the War of the Camps (1985-87) is counted.
  • I'm equally unclear why some combat zones are excluded in analysis of waves of Sunni terrorism—if I read Table 2 correctly, Chechnya doesn't count, but Algeria does. On the other hand, Map 3 shows only 168 terror attacks in North Africa in the 1980s and 1990s. Given that this was the monthly (or even weekly) toll at the height of the Algerian civil war, I can only assume that Algeria is excluded (although it is not shown as such).


I'm always wary about data drawn from large-N data quantitative sets without a very full description of what is coded, how, and what the collection biases might be.