October 28, 2008 8:21AM

Terrorism and Elections

With elections a week off, should we be especially worried about an Al Qaeda attack? Writing last week in Slate, Brookings' Daniel Benjamin says yes, following in the footsteps of other terrorism experts. They could be right, but they have almost no evidence. Like the others, Benjamin supports his claim with a handful of past examples of Al Qaeda attacks that occurred around election time. But if Al Qaeda attacks occur when they are most convenient for the attackers, they will be randomly distributed throughout the year, meaning that a certain percentage, which will head toward 1/12 as years go by, will fall in the month before elections. Citing a few attacks that occured around election time is evidence of nothing.

What about specific attacks? Do they reveal qualitative evidence for the hypothesis that Al Qaeda tries to sway elections, such as attackers saying that this was their goal? Short answer: no. The only Al Qaeda attack that remotely fits this billing, and the one that all the experts cite, is the Madrid train bombings in 2004.

The attack came 72 hours before Spain's general elections. When the Popular Party, led by José María Aznar, made a clumsy attempt to blame the ETA, a Basque separatist terrorist organization, the Spanish elected the Socialists, who then pulled troops from Iraq. Because the terrorists evidently influenced the election, pundits tend to assume that this was their aim.  But evidence for this assumption is almost nonexistent. The main data point is that some plotters may have read a tract on a website saying to attack around election time. That's about it.

What's more, the plot was only an Al Qaeda attack in spirit. So far as we know from public sources, there was not central planning from the remnants of Al Qaeda in Pakistan or elsewhere. The perpetrators organized locally. Even if they meant to swing the election, their tenuous connection to other groups makes it hard to form conclusions about the movement as a whole.

Some might say that what matters about the Madrid attack is the lesson that other terrorists drew from it. The apparent success in influencing the election might provoke imitation. Maybe so. But there is scant indication that imitation has occurred.

Terrorist plots, especially those that occur in countries foreign to the plotters, are tough to pull off. They require considerable organization. Police and intelligence agencies are hunting jihadist groups. They are likely to organize attacks to maximize the odds of success rather than to fit US election cycles.

In general, we should be wary of analysis that talks about Al Qaeda as a unified entity. Al Qaeda is a loosely connected network of small organizations and groups of guys. Assigning overarching preferences or goals to them can be analytically useful, just as talking about a country as a unified actor can be. But this kind of theorizing exaggerates terrorist unity. Psychology tells us that we tend to see patterns in random events -- causality, order and centralization where none is. We overtheorize, experts in particular. They deserve skepticism.