Statistics


"The third world war has already started. It is not George Bush's rhetorical "war on terror", but terrorism itself. In other words, terrorism is the new war. A recent analysis of the casualty statistics of global terrorism shows that they follow the pattern previously observed for conventional conflicts ranging from small local skirmishes to the second world war...

"But how can a single simple statistic such as the number of people killed in attacks tell us anything meaningful about events and conflicts conducted in completely different places for what seem to be totally different reasons? Isn't this like expecting to understand a country's culture by counting its population?

"That depends on what you are looking for. When he first studied the statistics of "deadly quarrels" 80 years ago, the British physicist Lewis Fry Richardson wanted to understand why wars happen. Richardson, a Quaker who served as an ambulance driver in the first world war, hoped that such insight could promote world peace. He decided first to find out how wars were distributed according to their size.

"In the 1920s he plotted the fatality statistics for 82 wars fought since 1820 on a graph showing the size of the conflicts on one axis and the number of conflicts of that size on the other. He found that the data fitted onto a smooth curve which, when the numbers were plotted as logarithms, became a straight line. This sort of mathematical relationship is known as a power law. The line slopes "downwards" because there are progressively fewer conflicts of ever greater size: little wars are common, big ones are rare.

"The power law continued to hold as the data embraced conflicts such as the second world war and Vietnam. Richardson's discovery of power-law statistics of conflicts has been followed by the recognition that power laws govern all sorts of "social" statistics, from the sizes of towns to the fluctuations of economic markets and the structure of the world wide web.

"Power-law statistics of event sizes are also found for natural phenomena that occur close to points of instability, such as earthquakes...

"There are two different power laws--one that fits the figures for terrorist attacks in industrialised (G8) nations, and another for attacks in the rest of the world. The slope of the straight-line plot was steeper in the latter case, indicating that attacks in industrialized nations are more rare but more severe when they do occur. The attacks of September 11, 2001, indicate that, as do the London bombings...

"Johnson argues that while the conventional approach of political analysts is to look for micro-explanations of the course of a conflict in terms of the motivations of the groups concerned, that statistical analysis suggests that the outcomes are much more to do with "the mechanics of how people now do war"...

"The team's conclusion supports the assertion of Mary Kaldor, a political scientist at the London School of Economics, that "the ongoing war in Iraq is a new type of war." She says US military action in Iraq has been predicated on the assumption that they are fighting an "old war."...

""This is immensely dangerous," Kaldor says. That, it seems, must also be the message for any global "war on terror"--it is not one that can be won by military might, but by new strategies. In "new wars", she says, military forces should be deployed for law enforcement, and "forces are needed that combine soldiers, police and civilians with the capacity to undertake humanitarian and legal activities"

"But if, as Johnson's work suggests, these conflicts have indeed turned into a form of terrorism, they will not be over soon. According to Clauset, the power-law statistics of terrorism show that it "is an endemic feature of the modern world and is likely not something that can be completely eradicated. Instead it should be considered in a similar way to other endemic problems, such as crime and natural disasters"" (Philip Ball. "Statistically, this means war." Guardian Weekly, Aug. 12, 2005: 19).


Colby Glass, MLIS