In a November 3 Slate column,
http://slate.msn.com/id/2090636/
Daniel Drezner claims that the Center for Public Integrity "has no evidence to support its allegations" that the contracting process in Iraq and Afghanistan has favored Bush cronies and campaign contributors. Drezner backs up his claim by analyzing CPI data on 70 firms that have received Afghanistan and/or Iraq reconstruction money.
According to Drezner, "If the corruption argument is true, then the size of campaign contributions should be strongly and positively correlated with the size of government contracts...The bad news is, the correlation coefficient turns out to be 0.192 and not statistically significant."
Drezner is dead wrong. A better analysis of the CPI data that Drezner used shows a STRONG POSITIVE correlation between contributions to Bush and the size of reconstruction contracts. The flaw in Drezner's analysis, as footnoted in his story, is "the wide variation in both contract size (from $2.3 billion to $10,000) and campaign expenditures (from zero to $8.8 million)."
When dollar sizes vary this widely, a correlation analysis that compares dollars will be skewed by the handful of firms that with the largest contributions and/or contracts. This select group includes Halliburton and Bechtel, but also General Electric and Northrup Grumman. The latter companies top the contributors list but are well down the list of contracts, possibly because our government is not building a lot of weapons systems in Iraq or Afghanistan.
Instead of using raw dollars, it's better to rank the contributors and contractors according to their dollar amounts. Hmm...the top 20 contractors include 12 of the 21 firms that gave Bush at least $100,000. Furthermore, an analysis of the two sets of rankings produces a correlation coefficient of .559. This would be considered a moderately high correlation in an experimental setting; it's a very strong correlation for non-experimental, real-world data.
Now, I'm not going to claim that this PROVES that CPI's argument is correct. As Drezner himself notes, it would be better to have data for all of the firms that could have been considered for these reconstruction contracts. But I DO know that Drezner's argument, to put it mildly, is incorrect.