News Archives

Scale Invariance in Global Terrorism

October 11, 2005

  • Date: Tuesday, October 11, 2005 
  • Time: 11:00-12:15pm. 
  • Place: Woodward 149

Aaron Clauset Department of Computer Science University of New Mexico

The traditional analysis of global terrorism holds that rare but catastrophic events are qualitatively different from common but less severe events. In this talk, using the set of terrorist attacks between 1968 and 2004, as compiled by the National Memorial Institute for the Prevention of Terrorism (MIPT), we will show that such events are actually part of a global statistical pattern. That is, we will show that the statistics of terrorism, like other kinds of both manmade and natural disasters, e.g., wars, forest fires, floods and earthquakes, have a mathematical form like a power law, P(x) ~ x^(-alpha) where alpha is the scaling exponent that describes the relationship between the frequency and severity of events. Notably, in such heavy-tailed distributions, events that are orders of magnitude larger than the average are actually relatively common.

The focus of this talk will be on describing the several global trends in the statistics of terrorism that we have discovered, as well as the computational and statistical tools used to support them. We will briefly contrast our approach with the traditional one for conflict analysis, and discuss its relative strengths/weakness. Finally, we will close with a brief discussion of the policy implications and outstanding questions raised by this work.

This is joint work with Maxwell Young, and has been covered in the popular science press by The Economist, The Guardian (UK), Die Welt (Germany), Nature, and the Institute of Physics.

The preprint is available at http://arxiv.org/abs/physics/0502014 .