"Even if the data exists at the provincial level, the physical danger makes it hard for interviewers to collect it, as well as making people very reluctant to talk to, much less help, strangers."-MM
Good article, especially the part about the problems with bad data being worse than no data at all, and the closing exhortation on the need to remember that this equation has units of human bodies.
However, I was hoping for a bit more time spent on explaining the statistical problems with such surveys, in the same readable-to-nonexperts fashion that you presented them in some of your past blog posts. Is there any chance this article will have a "part 2"?
I place this again in front of the staunch "Iraq was a bad decision crowd"(especially since the Iraq campaign has had overwelmingly bi-partisan support to date). It seems to have been overlooked earlier.
Even in total hindsight, invading Iraq was a good idea. Even if I knew we would be right here at this time. The thoughtful anti-Iraq war Dems (as opposed to the loony anti-war left who are just insane) were and are still wrong.
But you had a comment on an earlier thread that fits this one also:
"It is only a good idea to remove a murderous dictator if this results in, on net, fewer people being murdered."
This basically sets the value of freedom at zero.
That there is no value to no longer living under a tyranny with no hope of ever being free. That the lives of those who live in North Korea are the same as those in the South. Under your equation, we should not have fought in Korea, nor liberated Kuwait. Nor for that matter fought the US Civil War or revolted from England in 1776. All these actions resulted in many more lives lost than if we continued to let tyranny triumph.
I thought America and the western liberal democracies would make a stand after 9/11. I thought the liberal left who always complain about our support of dictators (and still do) would be on board. Liberating 60 million people from two of the worst dictatorships on the planet is reduced to a cost benefit relationship on how many people have died? Whether or not 80,000 died (good, less than Saddam's estimated tally) vs. 150,000 died (bad, more than Saddam's estimated tally)?
So did you actually get around to reading the Lancet article finally?
When you wrote, "its authors, mostly researchers at Johns Hopkins University, have been accused of everything from bias to outright fraud," you failed to add that the accusations have only been made by non-scientists and by people like yourself who have felt free to comment on the article without actually reading it. Yet your article tries to take some high ground above those people "proving they had been right" even though you were expressing skepticism of the Lancet results without either reading or understanding the article in the first place.
Commentary like yours represents a whole new low for public discourse. I am amazed it ends up in the pages of the Atlantic, which was once a fine magazine.
Commentary like yours represents a whole new low for public discourse. I am amazed it ends up in the pages of the Atlantic, which was once a fine magazine.
Good to see that Mr. Giacometti, as usual, is leading the quality of discourse back upward via outstanding examples such as these.
[The Lancet article] authors, mostly researchers at Johns Hopkins University, have been accused of everything from bias to outright fraud," {and} you failed to add that the accusations have only been made by non-scientists and by people [who haven't read it.]
"Some scientists have subsequently challenged the validity of the Lancet study. Questions have been asked about the survey techniques and the possibility of "mainstreet bias".
"Dr Michael Spagat of Royal Holloway London University says that most of those questioned lived on streets more likely than average to witness attacks: "It would appear they were only able to sample a small sliver of the country," he said.
"Dr Spagat has previously conducted research with Iraq Body Count, an NGO that counts deaths on the basis of media reports and which has produced estimates far lower than those published in the Lancet."
Malignant Bouffant, while I wouldn't take the tone of IOZ's piece, I'm afraid I agree substantively with pretty much everything it said. Notably:
those like myself who said not simply that the invasion of Iraq would fail to achieve either its stated or its actual goals, but that it must fail to achieve them, were not arguing from some universalist principal of incompetence, but rather from the rich recent history of commensurate projects and their innate tendency to fail. Indeed, one of the most common charges levelled against war opponents by jingoes was that our historical analogizing was irrelevant. "Iraq is not Vietnam." "Iraq is not Algeria." "Iraq is not Afghanistan and the Soviets." Etc. And of course, Iraq turned out to be all of those things, and more--it could no more be otherwise than I could fly by flapping my arms. The people supporting invasion and occupation were the ones proposing that Iraq would be a single, vastly distant historical outlier, totally devoid of precedent or context, and those who thought that the mere incompetence of the administration, or its untrustworthiness, were substantive arguments against invading were almost universally people who wished (and still wish) to hedge their bets just in case it all worked out. McArdle says that only this latter sort are worth talking to because they were right for the right reasons, but the opposite is true.
Even though I supported the war until Megan decided not to, I think IOZ is sort of right here:
"People who supported the invasion of Iraq were fatuous, bloodthirsty, ahistorical, immoral, politically naive, callous, unthinking, reprehensible morons--to the man. The proper attitude is contrition, silence, and contemplation. Making a gaudy spectacle of having "supported" something so awful, even if only to show how smart you were to change your mind when you noticed things going south, is disgusting."
Spagat's view of "main street bias" is easily debunked. Unlike Megan, I hate to denegrate economists, but Spagat is out of his discipline on this study. It certainly is true, that the majority of critics of the lancet study are non-scientists, with little understanding of sampling deaths in war time.
I always hate to "unload" on people like you who I don't know, but facts are facts and I can't have erroneous comments like yours circulating on the nets without a (truthful) retort. So howsah:
First, Spragat was a fine guy and a great speaker--I heard him lecture many times at the "Tute"--but you can't hitch your waggon to only one "star." Me, I look to others--you must, too.
Second, I don't care what Megan says, she's wrong about the Iraq death count. I know more about it because, as you might know, I am a professional econoblogger. The total No. of deaths, i.e., stiffs, is four million, give or take a million.
Does that answer your questions? It sure as heck should.;
As you you know I am one of your most loyal posters. Believe me, I agree with you that the "death counts" in Iraq are completely exaggerated. My guess is that around 2,500 actually Iraqis have died--which is less than our American troops--for shame!
What we need to do--and you'll agree-is to increase the death counts--not decrease them. Question: How do you win a war? Answer: By killing the enemy.
Megan, the question shouldn't be, How many Iraqis have we killed?, but how many Iraqis DO WE NEED TO KILL to win the war.
Truth Seeker, that is an unfortunate misconception from the way the case has been reported in the media. Spagat and Hicks are not the only critics of Burnham et. al.; there's concern in the conflict epidemiology community as well. I'll be getting into that soon, I hope.
"Conversely, few of the study’s supporters expressed much pleasure at the news that an extra 450,000 people might be walking around in Iraq."
I wasn't exactly thrilled to learn that there are 450,000 zombie-Iraqis on the loose either. What if the virus spreads? Has anyone considered a quarantine?
Actually Rex, when the original study came out she blogged about it clearly without reading it. Her point of view was skewered in commments in her old blog and on blogs like Crooked Timber.
Indeed, her original comment about the study was that the number "seemed too big" to her, or some such utter nonsense.
If she finally has read it (which I still doubt), she has no clue about the methodologies used, and whether the methodology in the NEJM study is any better than the Lancet study or vice versa. She has no intellectual basis for her opinions.
I don't agree or disagree with any of the studies. I just think to comment on something you actually have had to have read it. That really is not too much to ask.
Anony_mouse appears with an equally ridiculous and air-filled statement ("My guess is that around 2,500 actually Iraqis have died.") Her ***guess***. Well I could guess that water is 2 parts Nitrogen and one part Fluffernutter, and I would be just as wrong as mouse. There are ways to learn things, and it usually begins by reading, not by guessing, and not by deciding out of a thin air that a number "seems" too big.
Funny how Megan missed something that was obvious to anyone who read it--the New England Journal study which gives a violent death rate four times lower than Lancet2 also gives a violent death rate which is essentially identical to that of Lancet1 for the first 18 months if the Fallujah outlier is excluded. That is, roughly 120 deaths per day.
That figure was about 4 times higher than the IBC number given at the time (60,000 vs 15,000, though IBC later added a few thousand to its total) and as I recall, that was why Lancet critics refused to believe it.
A competent honest reporter would have noticed this and informed the reader. That does not seem to have happened here.
You refer to the WHO study of Iraq mortality, preferring it to others conducted by an American university. I'm sure you have a very good reason for not mentioning that the data for the WHO survey was "checked and edited" by a Health Ministry controlled by Shia cleric Moqtada al-Sadr. (You can check the newspapers if you're not sure who he is.)
Unfortunately you forgot to give us the reason for this oversight. So what was it?
That's a brilliant finale paragraph. Thank you.
"Even if the data exists at the provincial level, the physical danger makes it hard for interviewers to collect it, as well as making people very reluctant to talk to, much less help, strangers."-MM
We do have Citizens, here, that avoid the Census.
http://www.census.gov/popest/estimates.php
Why should we think that it would be much different in Iraq, especially, now?
Also, links are a +
Good article, especially the part about the problems with bad data being worse than no data at all, and the closing exhortation on the need to remember that this equation has units of human bodies.
However, I was hoping for a bit more time spent on explaining the statistical problems with such surveys, in the same readable-to-nonexperts fashion that you presented them in some of your past blog posts. Is there any chance this article will have a "part 2"?
Just on this blog. Get ready to be bored to death.
Bored to death? I think it may get rather exciting on here.
Interesting article. I blogged about it.
Megan,
I place this again in front of the staunch "Iraq was a bad decision crowd"(especially since the Iraq campaign has had overwelmingly bi-partisan support to date). It seems to have been overlooked earlier.
Even in total hindsight, invading Iraq was a good idea. Even if I knew we would be right here at this time. The thoughtful anti-Iraq war Dems (as opposed to the loony anti-war left who are just insane) were and are still wrong.
But you had a comment on an earlier thread that fits this one also:
"It is only a good idea to remove a murderous dictator if this results in, on net, fewer people being murdered."
This basically sets the value of freedom at zero.
That there is no value to no longer living under a tyranny with no hope of ever being free. That the lives of those who live in North Korea are the same as those in the South. Under your equation, we should not have fought in Korea, nor liberated Kuwait. Nor for that matter fought the US Civil War or revolted from England in 1776. All these actions resulted in many more lives lost than if we continued to let tyranny triumph.
I thought America and the western liberal democracies would make a stand after 9/11. I thought the liberal left who always complain about our support of dictators (and still do) would be on board. Liberating 60 million people from two of the worst dictatorships on the planet is reduced to a cost benefit relationship on how many people have died? Whether or not 80,000 died (good, less than Saddam's estimated tally) vs. 150,000 died (bad, more than Saddam's estimated tally)?
Is that what it’s come down to?
How long before the calls to abandon Afghanistan?
Islam will change
So did you actually get around to reading the Lancet article finally?
When you wrote, "its authors, mostly researchers at Johns Hopkins University, have been accused of everything from bias to outright fraud," you failed to add that the accusations have only been made by non-scientists and by people like yourself who have felt free to comment on the article without actually reading it. Yet your article tries to take some high ground above those people "proving they had been right" even though you were expressing skepticism of the Lancet results without either reading or understanding the article in the first place.
Commentary like yours represents a whole new low for public discourse. I am amazed it ends up in the pages of the Atlantic, which was once a fine magazine.
Charles,
She read the study and blogged about it when it came out.
Commentary like yours represents a whole new low for public discourse. I am amazed it ends up in the pages of the Atlantic, which was once a fine magazine.
Good to see that Mr. Giacometti, as usual, is leading the quality of discourse back upward via outstanding examples such as these.
[The Lancet article] authors, mostly researchers at Johns Hopkins University, have been accused of everything from bias to outright fraud," {and} you failed to add that the accusations have only been made by non-scientists and by people [who haven't read it.]
Well, that's easy to prove false. Here...
http://news.bbc.co.uk/2/hi/uk_news/politics/6495753.stm
"Some scientists have subsequently challenged the validity of the Lancet study. Questions have been asked about the survey techniques and the possibility of "mainstreet bias".
"Dr Michael Spagat of Royal Holloway London University says that most of those questioned lived on streets more likely than average to witness attacks: "It would appear they were only able to sample a small sliver of the country," he said.
"Dr Spagat has previously conducted research with Iraq Body Count, an NGO that counts deaths on the basis of media reports and which has produced estimates far lower than those published in the Lancet."
While this dead horse continues to be beaten, do take a look at this.
Malignant Bouffant, while I wouldn't take the tone of IOZ's piece, I'm afraid I agree substantively with pretty much everything it said. Notably:
those like myself who said not simply that the invasion of Iraq would fail to achieve either its stated or its actual goals, but that it must fail to achieve them, were not arguing from some universalist principal of incompetence, but rather from the rich recent history of commensurate projects and their innate tendency to fail. Indeed, one of the most common charges levelled against war opponents by jingoes was that our historical analogizing was irrelevant. "Iraq is not Vietnam." "Iraq is not Algeria." "Iraq is not Afghanistan and the Soviets." Etc. And of course, Iraq turned out to be all of those things, and more--it could no more be otherwise than I could fly by flapping my arms. The people supporting invasion and occupation were the ones proposing that Iraq would be a single, vastly distant historical outlier, totally devoid of precedent or context, and those who thought that the mere incompetence of the administration, or its untrustworthiness, were substantive arguments against invading were almost universally people who wished (and still wish) to hedge their bets just in case it all worked out. McArdle says that only this latter sort are worth talking to because they were right for the right reasons, but the opposite is true.
But that's topic drift.
Even though I supported the war until Megan decided not to, I think IOZ is sort of right here:
"People who supported the invasion of Iraq were fatuous, bloodthirsty, ahistorical, immoral, politically naive, callous, unthinking, reprehensible morons--to the man. The proper attitude is contrition, silence, and contemplation. Making a gaudy spectacle of having "supported" something so awful, even if only to show how smart you were to change your mind when you noticed things going south, is disgusting."
But what do I know? I'm just a girl.
John,
Spagat's view of "main street bias" is easily debunked. Unlike Megan, I hate to denegrate economists, but Spagat is out of his discipline on this study. It certainly is true, that the majority of critics of the lancet study are non-scientists, with little understanding of sampling deaths in war time.
http://scienceblogs.com/deltoid/2006/10/science_on_lancet_study.php
Truth Seeker,
I always hate to "unload" on people like you who I don't know, but facts are facts and I can't have erroneous comments like yours circulating on the nets without a (truthful) retort. So howsah:
First, Spragat was a fine guy and a great speaker--I heard him lecture many times at the "Tute"--but you can't hitch your waggon to only one "star." Me, I look to others--you must, too.
Second, I don't care what Megan says, she's wrong about the Iraq death count. I know more about it because, as you might know, I am a professional econoblogger. The total No. of deaths, i.e., stiffs, is four million, give or take a million.
Does that answer your questions? It sure as heck should.;
Megan,
As you you know I am one of your most loyal posters. Believe me, I agree with you that the "death counts" in Iraq are completely exaggerated. My guess is that around 2,500 actually Iraqis have died--which is less than our American troops--for shame!
What we need to do--and you'll agree-is to increase the death counts--not decrease them. Question: How do you win a war? Answer: By killing the enemy.
Megan, the question shouldn't be, How many Iraqis have we killed?, but how many Iraqis DO WE NEED TO KILL to win the war.
Let's all just get back to basics, okay?
Truth Seeker, that is an unfortunate misconception from the way the case has been reported in the media. Spagat and Hicks are not the only critics of Burnham et. al.; there's concern in the conflict epidemiology community as well. I'll be getting into that soon, I hope.
"Conversely, few of the study’s supporters expressed much pleasure at the news that an extra 450,000 people might be walking around in Iraq."
I wasn't exactly thrilled to learn that there are 450,000 zombie-Iraqis on the loose either. What if the virus spreads? Has anyone considered a quarantine?
Manual trackback
Actually Rex, when the original study came out she blogged about it clearly without reading it. Her point of view was skewered in commments in her old blog and on blogs like Crooked Timber.
Indeed, her original comment about the study was that the number "seemed too big" to her, or some such utter nonsense.
If she finally has read it (which I still doubt), she has no clue about the methodologies used, and whether the methodology in the NEJM study is any better than the Lancet study or vice versa. She has no intellectual basis for her opinions.
I don't agree or disagree with any of the studies. I just think to comment on something you actually have had to have read it. That really is not too much to ask.
Anony_mouse appears with an equally ridiculous and air-filled statement ("My guess is that around 2,500 actually Iraqis have died.") Her ***guess***. Well I could guess that water is 2 parts Nitrogen and one part Fluffernutter, and I would be just as wrong as mouse. There are ways to learn things, and it usually begins by reading, not by guessing, and not by deciding out of a thin air that a number "seems" too big.
Funny how Megan missed something that was obvious to anyone who read it--the New England Journal study which gives a violent death rate four times lower than Lancet2 also gives a violent death rate which is essentially identical to that of Lancet1 for the first 18 months if the Fallujah outlier is excluded. That is, roughly 120 deaths per day.
That figure was about 4 times higher than the IBC number given at the time (60,000 vs 15,000, though IBC later added a few thousand to its total) and as I recall, that was why Lancet critics refused to believe it.
A competent honest reporter would have noticed this and informed the reader. That does not seem to have happened here.
Megan -
You refer to the WHO study of Iraq mortality, preferring it to others conducted by an American university. I'm sure you have a very good reason for not mentioning that the data for the WHO survey was "checked and edited" by a Health Ministry controlled by Shia cleric Moqtada al-Sadr. (You can check the newspapers if you're not sure who he is.)
Unfortunately you forgot to give us the reason for this oversight. So what was it?