Wednesday, October 11, 2006
Then I read that it was published in The Lancet and I lost all interest. This is the journal that gave us the infamous MMR-causes-autism study and that published a similarly discredited tally of Iraqi casualities before the last American election. In the ranks of medical journals, I place them on a par with The Guardian.
This time, however, the media is on to them:
Robert Blendon, director of the Harvard Program on Public Opinion and Health and Social Policy, said interviewing urban dwellers chosen at random was “the best of what you can expect in a war zone.”
But he said the number of deaths in the families interviewed — 547 in the post-invasion period versus 82 in a similar period before the invasion — was too few to extrapolate up to more than 600,000 deaths across the country.
Donald Berry, chairman of biostatistics at M. D. Anderson Cancer Center in Houston, was even more troubled by the study, which he said had “a tone of accuracy that’s just inappropriate.”
Kudos to the two New York Times reporters for taking the time to run the study by a couple of statisticians.
Color the Washington Post skeptical, too:
And neither does Michael E. O'Hanlon of the Brookings Institution, which also tracks Iraqi deaths.
"I do not believe the new numbers. I think they're way off," he said.
Other research methods on the ground, like body counts, forensic analysis and taking eyewitness reports, have produced numbers only about one-tenth as high, he said. "I have a hard time seeing how all the direct evidence could be that far off ... therefore I think the survey data is probably what's wrong."
The full survey is here. The researchers spent two months canvassing households in various regions of Iraq asking about deaths in the family. Sometimes they were able to confirm the reports with death certificates, sometimes they weren't. They didn't ask if the dead were combatants or non-combatants. They were afraid to ask that question. Afraid for themselves and for those they were asking. They interviewed 40 households in each of their selected regions, then extrapolated the 600,000 figure from the number of deaths they had recorded in their interviews. The margin of error of +/-200,000 speaks for itself. It's not reliable.
And sorry, but the defense that it's as soundly designed as can be expected for these kinds of public health surveys is a weak one. Retrospective, interview-based studies like this are poor designs. It may be the standard way of gathering data in the public health field, but that doesn't make it the best methodology, and it certainly doesn't make its statistics sound. For too long the field of public health has relied on these types of
But what do the Iraqi's think? Here's one who is particularly livid:
I wonder if that research team was willing to go to North Korea or Libya and I think they wouldn’t have the guts to dare ask Saddam to let them in and investigate deaths under his regime.
No, they would’ve shit their pants the moment they set foot in Iraq and they would find themselves surrounded by the Mukhabarat men counting their breaths. However, maybe they would have the chance to receive a gift from the tyrant in exchange for painting a rosy picture about his rule.
They shamelessly made an auction of our blood, and it didn’t make a difference if the blood was shed by a bomb or a bullet or a heart attack because the bigger the count the more useful it becomes to attack this or that policy in a political race and the more useful it becomes in cheerleading for murderous tyrannical regimes.
UPDATE: From Dani in the comments section, the editor of The Lancet, expressing his opinion, to which he is certainly entitled. However, his obvious passion (is it necessary to shout when using a microphone?) casts more than a shadow of doubt on his ability to be unbiased in selecting articles for publication that cover the same topic.
UPDATE II: The Lancet podcast defending the survey.
UPDATE III: Much more via Tim Blair, including Lancet editor Richard Horton's assessment of peer review, the annals of The Lancet's various controversies, a statistician's analysis of the study, criticism from anti-war epidemiologists.
UPDATE IV: More here.
posted by Sydney on 10/11/2006 10:51:00 PM 36 comments
aww you forgot to mention all the statisticians in those articles who support the methodology and the numbers. aww poor baby with a short memory.
By 10:34 AM, at
So Iraq has almost the same rate of population loss as Japan had during WWII? Considering the last study these folks did right before the 2004 election, this is beyond BS. It borders on election tampering.
Given that conflicts are not evenly distributed, and that the independent variable wasn't really independent (households on the same street are likely to have many casualties if they have any), the external validity of this study is near zero.
What a lucid summary of everything that's wrong with the Lancet's latest attempt to meddle in American politics. (And, yes, I am one of those who thinks that the timing makes this a political attack.)
By 12:34 PM, at
1) The Lancet claims that their active survey is more accurate than a “passive” system of counting media reports, morgue reports or other lists of the dead (and generally totalling about 50,000 dead in Iraq), which are often grossly incomplete in a war zone. This seems reasonable.
Silly, hey...anyone have any thoughts on this. My best friend sent it to me and I'm curious. Sorry to get off topic, didn't see anything from Sid on the subject.: http://www.technologyreview.com/read_article.aspx?id=17597&ch=biotech&ch=biotech
By 3:14 PM, at
Sorry, the link appears to be giving me trouble. It's about a new nano blood clotting technology. See below.
By 3:16 PM, at
By 3:20 PM, at
By 4:16 PM, at
So this is where we're at today: Johns Hopkins/MIT release a peer-reviewed study in the one of the oldest and most respected peer-reviewed medical journals in the world, using standard accepted best practices. They study comes to conclusions that are somewhat counter-intuitive, and very politically unpalatable.
By 4:37 PM, at
"Then I read the first sentence and saw that the number was gathered by public health researchers and it lost some credibility."
By 4:41 PM, at
Who mandated John Hopkins U to estimate the death toll of the war in Iraq? Can't Iraqi count?
By 4:57 PM, at
It is election tampering. The lead authors of the 2004 study are the same. One of them came right out and said the timing of that article was deliberate, according to the AP.
By 5:08 PM, at
Red right hand,
Yeah, I saw that later after a second read. I posted a comment to that affect, but for some reason, it didn't show. Here's the mea culpa:
By 5:32 PM, at
The researcher sounds like a man with a small n.
I have to say I quit reading the Lancet when it got into the habit of releasing press reports before the actual articles. I would say this was around ten years ago or so.
By 6:40 PM, at
Also, since I am in a surgical specialty, the Lancet became less relevant to me as the years rolled by. In fact, I just finished up my board recert, and not one article came from there. Out of scores of articles.
By 6:43 PM, at
Wow. This is a rather good selection of bogus arguments, considering the small number of comments. We've got the 'it's a biiig number!', 'where're all the bodies', 'previous article was discredited', ...
By 8:53 PM, at
Here is another example of why the Lancet is more than a little biased. Watch the editor, Robert Horton, at his finest. Seriously, we're supposed to think this article is legitimate?
By 9:15 PM, at
Dani, beat me on that YouTube video of Richard Horton's raving at an antiwar rally three weeks ago.
By 10:21 PM, at
Woo hoo! North Korea will be a slam dunk. If we've managed to kill 655,000 Iraqiis while trying to help them, imagine what we'll do when we invade a country and actually TRY to kill everyone.
Also, since I am in a surgical specialty, the Lancet became less relevant to me as the years rolled by.
By 2:25 AM, at
Well, ibc, if you couldn't assume I knew nothing becaue of my cutesy moniker, now you can assume I know nothing because I am a surgeon!
By 9:12 AM, at
The point of my posts has been to highlight the poltical orientation of the Lancet, a position which is both unethical and unparallelled among modern peer-reviewed journals, and as such makes all of their articles suspect. Instead of responding to the evidence I have provided, you attack me.
By 1:21 PM, at
By 3:32 PM, at
The Lancet has problems, and it is perfectly reasonable for me to attack the messenger.
By 4:03 PM, at
"According to the CIA world factbook, roughly 2,465,149 americans died last year. One question: Where are the bodies? Ha! Put that in your pipe and smoke it Mr. Statistical Researcher Epidemiologist."
By 5:01 PM, at
Well, this pediatrician trusts The Lancet a great deal more than I trust you, and The Lancet is a great deal less biased than you are.
By 5:47 PM, at
In case anyone hasn't re-read the original post, Sydney has put up some updates.
By 10:53 AM, at
"Given that conflicts are not evenly distributed, and that the independent variable wasn't really independent (households on the same street are likely to have many casualties if they have any), the external validity of this study is near zero."
By 3:59 PM, at
I'd like to say that I truly feel sorry for anyone defending the methodology of this survey.
Xeno: I'm curious.
By 12:58 PM, at