Seb Paquet asks for ideas for swarms tackling real problems

Here's one I thought of a few weeks ago. As the war was gearing up in Iraq I started thinking about the following problem. Was it possible to solve the controversial question : is US agression towards Iraq motivated by oil?

I wondered whether some kind of smart mob could solve this through swarm statistics. Perhaps by collecting data about US military focus (either aggression or "friendly" military presence like US bases) on countries, and the estimated oil reserves of the country at the time.

Then to do some stats to see whether one is a significant predictor of the other. The smart mobbing would come in because the swarm would pool data, ideas as to what statistics to collect, actual number crunching, and would (constructively) criticise each model. Ideally it should include both those who believe that oil is and those who believe oil isn't, a factor.

My brother-in-law works in the oil industry and promised me some figures for the oil reserves of different countries, but I'm stumped for measures of US aggressivity / military presence. Any ideas? If I get a good one, I'll be tempted to start this.