Gordon’s Stats Redux

by tristero

A couple of days ago, I asked the hive mind for help understanding the charts in this Michael Gordon article about the “surge”. A marvelous discussion ensued which was very informative. In addition, Mike the Mad Biologist posted an illustrated analysis of what conclusions can be gleaned from these charts. If I understand him, and the consensus of the comments here, correctly, there is not much of any conclusion to be had as there are so few data points. But from what one can tell, according to the data Gordon presents as the basis for his interpretation, there is no significant trend for the better or the worse. Therefore, Gordon’s analysis, quoted below, is, to be generous, overly optimistic, or to be bluntly accurate about it, misleading and untrue:

The most comprehensive and up-to-date military statistics show that American forces have made some headway toward a crucial goal of protecting the Iraqi population. Data on car bombs, suicide attacks, civilian casualties and other measures of the bloodshed in Iraq indicate that violence has been on the decline, though the levels generally remain higher than in 2004 and 2005.

By the way, it is very likely that the Times editors are aware of the problems with Gordon’s optimistic conclusions on the “surge.” In addition to numerous highly criticial editorials, the Times published a long article from reporters other than Gordon that contradicted his assessment.

If it seems rather odd that the Times so distrusts the ability of one of its most prominent foreign correspondents to report the truth in an objective fashion that it sends other journalists to check up on him, that’s because it is odd, at least when it comes to recent history. Had they only checked up on Miller.

Thanks, Mike, and everyone else, for taking the time to walk me, and the rest of us through these charts. We all need to get more statistically literate, and demand that more accurate and relevant data get printed. And also so that reporters with an axe to grind feel less likely bamboozle us and report spin on the data as a legitimate conclusion.