« Those Pesky Details | Main | Correction... »
March 27, 2005
The Lancet Study
Joe Katzman at The Winds of Change has more on the highly questionable study that claims that 100,000 Iraqis have died since the beginning of the war.
This study, published in the British medical journal The Lancet, was conducted by a research team at the Bloomberg School of Public Health at Johns Hopkins University and published less than a week before the presidential election. Following the debate on this study gets complicated, but the overall weakness of the study is clear. I think the most important point to observe is that the much-referred-to "CI", the "confidence interval", must lead one to conclude the study is worthless. On the day the study was released online, Fred Kaplan at Slate wrote (emphases mine):
That number [of deaths since the war began] is 98,000. But read the passage that cites the calculation more fully:
We estimate there were 98,000 extra deaths (95% CI 8000-194 000) during the post-war period.
Readers who are accustomed to perusing statistical documents know what the set of numbers in the parentheses means. For the other 99.9 percent of you, I'll spell it out in plain English—which, disturbingly, the study never does. It means that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000. (The number cited in plain language—98,000—is roughly at the halfway point in this absurdly vast range.)
The significance of the CI number was never explained in the published study. That's bad enough; add to that the very small sample size of reported deaths discovered by the interviewers. From a Winds of Change commenter who analyzed the study (emphases mine):
Roberts et al. surveyed households in 33 randomly-chosen clusters in Iraq. 988 households were chosen, most were successfully interviewed. Interviews in the 32 clusters outside of Fallujah turned up reports of 9 deaths due to Coalition action. Interviews in the Fallujah cluster yielded reports of 52 deaths due to Coalition action.
And I think the trumpeting of this study by the left misses the larger point: Even if civilian deaths in Iraq did increase, that by no means proves that we shouldn't have invaded the country and removed Saddam. It's certainly analogous to looking at the terrible conditions that existed in Europe in the winter of 1946.
UPDATE: Commenter "Brian" at the Chicago Boyz blog gives a devastating (and non-technical) argument against the validity of the Lancet study: To get to the stated "100,000 deaths" number, over 5,000 Iraqis would have to have been killed every day month for the fifteen or so months between the start of the war and the summer of 2004. Given that any large car bombing that resulted in casualies was covered relentlessly in the MSM, it beggars belief that 5,000 deaths could have passed unremarked upon.
Trackback Pings
TrackBack URL for this entry:
http://www.thebernoullieffect.com/cgi-bin/mt/mt-tb.cgi/76
Listed below are links to weblogs that reference The Lancet Study:
» Miscellanea: Fun With Numbers Edition from Decision '08
Jeff at the Bernoulli Effect takes apart a study some anti-war folks are using as ammo. I suggest you take a look... [Read More]
Tracked on March 28, 2005 10:24 PM
Comments
The significance of the CI number was never explained in the published study.
The readers of the Lancet are British medical doctors; oddly enough, you don't have to explain to them what a confidence interval is every month.
Posted by: dsquared at March 29, 2005 02:53 AM
That is a valid point, and it leads to other big-picture questions concerning the ways in which difficult science gets reported in the media, or explained to juries, etc.
I haven't drilled down to the lower depths of the comments discussion at Winds of Change, and I, like a lot of others, had one stats course years ago. But for me, the 100,000 number just doesn't pass a "sanity check"--I believe that Iraqi society would be shattered by that number of deaths, the MSM would be all over the story, and again: where are the bodies?
Posted by: Jeff at March 29, 2005 10:39 AM
