Figures Don't Lie, but Liars Do Figure

An article in the May 26 NYT contained the following paragraph:

But the tactical success of the surge should not be misconstrued as making Iraq a safer place for American soldiers. Last year was the bloodiest in the five-year history of the conflict, with more than 900 dead, and last month, 52 perished, making it the bloodiest month of the year so far. So far in May, 18 have died.

Here is a bar graph of American deaths in Iraq. You can enlarge the graph by clicking on it. (Raw data are available here.)



As you can plainly see, US casualties went way down after the surge troops finished arriving in June, 2007. So the the first (thesis) sentence of the quoted paragraph is an outright lie.

As for the rest, the NYT is making true but misleading statements designed to cause the reader to believe the false thesis. That is, it is true that 2007 was the bloodiest year of the war for American soldiers. But this is because of casualties suffered during the first half of the year (which in turn were caused by the wave of violence instigated by Al Qaeda). And that wave of violence, which started in 2006, was a reason for the surge, not the result of it.

And yes, April has been the worst month of 2008 so far in terms of American casualties. But any year will have a worst month. So regardless of whether the worst month of 2008 so far were January, February, March, April, or May, this would not tell us anything about how successful the surge was or wasn't in terms of reducing American casualties. So it is misleading for the NYT to publish this paragraph in a way that makes it appear that this sentence supports its thesis.

Hat tip: Countercolumn