February 20, 2007

The null hypothesis in Iraq

Those for and against the "surge" in Iraq have made predictions of what will happen if their recommendations are not followed. "But there's only one thing we know for sure," according to CISAC's Michael May. "When the collection of politicians and pundits we call 'Washington' makes predictions about countries in which the U.S. has 'vital interests'(and especially about countries with which we have had bad relations), the predictions--even when they contradict one another--are almost always wrong." The solution, May writes in the Los Angeles Times, is for Washington to take a tip from scientists and start with the "null hypothesis" to guard against bias in testing predictions. Policy makers should assume that no theory starts out with more predictive value than chance, and then use data to prove which theory is more sound.

• » Los Angeles Times: "A clue to Washington's cluelessness"