I had thought this question was pretty much settled in favor of the Lancet Report (655,000 dead because of the war as of July 2006) but in this thread ManiacMan insists on calling it grossly inflated, so here we go.
N.B.: “Dead because of the war” is not limited to “dead from war-related violence.” The latters is a subset of the former.
The only reason the the question was ‘pretty much settled’ here on the Lancet report is because it provided the biggest numbers, and therefore the most ammunition, to the anti-war crowd. It diverged heavily from pretty much all other estimates.
Here’s a good article regarding the problems with the Lancet report:
The data collection methodology appears to have used ‘clustered’ data samples taken from areas of intensive casualties, then using those as representative of the entire country to come up with the grossly inflated figures. The sample sizes also seem very small from which extrapolate a nationwide figure. From the article:
Fraud, biased samples, slanted extrapolations, all of which result in a casualty figure almost an order of magnitude greater than any other study had found.
Sighhhh…
Sigh…
Geeze. I don’t care if some of you feel that the issue is “settled” in your minds. I am pointing out that the methodology used to estimate the number is dead is flawed. I mean c’mon they conducted “in-person interviews with a random national sample of 2,212 Iraqi adults, including oversamples in Anbar province, Basra city, Kirkuk and the Sadr City section of Baghdad”.
And I then I quoted how **"The Bosnian Book of the Dead - a database that reveals 97,207 names of Bosnia and Herzegovina’s citizens killed and missing during the 1992-1995 war. An international team of experts evaluated the findings before they were released. More than 240,000 pieces of data have been collected, processed, checked, compared and evaluated by international team of experts in order to get the final number of more than 97,000 of names of victims, belonging to all nationalities. " **
Doesn’t the Bosnian estimate sound more credible?
So, are you saying that you don’t believe in statistical methods? Or, are you saying the statistical methods were dramatically misapplied and the reviewers of the Lancet paper didn’t catch this?
There’s one problem with comparing Iraq and Bosnia. Bosnia in 1991 had a population of about 4 1/2 million people. Iraq in 2003 had a population of roughly 25 million people. Even IF we take the Bosnian numbers as a guideline (say we round up the Bosnia numbers to 5 million total and 100,000 dead for convenience), that means that the EXPECTED number for Iraq under that guideline would be 500,000 dead.
If people are arguing about something very well they can argue ad infinitum, right? Statistics is used to settle an argument just so long as both sides can agree on the method and the approach. Once a study is conducted where everyone can agree on the method and the approach I will give it more credibility. When people scream these huge numbers with the aforementioned criteria it is just rhetoric.
I wasn’t comparing the two countries, but I was comparing the studies that estimate the deaths from both conflicts and how the Bosnian study seems to be more robust compared to the recent Iraq study.
When the same techniques were applied to Fallujah, the result was an estimate so obviously out of whack with reality that it was dropped from the study. Then those same techniques were used elsewhere.
And you can do all the statistical analysis you want, but if the original data was collected in a fraudulent manner, it doesn’t really matter, does it? The Lancet study used extremely biased and untrustworthy data collectors, and applied no methods to check for fraud.
Let’s see… if a tobacco-related illness study was funded by RJR Reynolds, and used surveyors who were employees of tobacco stores to collect data, and the results were an order of magnitude off of every other study of tobacco-related illnesses, would you call the study valid so long as the statistical analysis of the received data was scientifically valid? What would you say to someone who ignored all other studies and used this one as the definitive piece of evidence to show tobacco wasn’t dangerous?
Seriously BG…what gave you the impression this was ‘settled’?? Because YOU thought it was? I’ve seen plenty of threads where the figures were disputed…in fact I was under just the opposite impression to you. My take is that only the rabidly anti-war crowd think Lancett is completely credible and are wont to swallow it whole without skepticism.
sigh I thought this was all settled…
-XT
Perhaps we have only a paltry hundred thousand innocent lives snuffed out. Such comfort.
Why - have there been other instances where Soros-funded research came up with biased results?
Way you guys talk about him, you’d think his name was Blofield or something.
Didn’t they check death certificates or something?
The U.S. Census Bureau and the Centers for Disease Control use cluster sampling. The article provides no evidence to back up the assertion that the clusters were chosen in a nonrepresentative manner.
The proper place to dismember an article published in a refereed journal is, of course, another refereed journal. (Not that others aren’t free to chime in, but it’s an indication that the critics actually know something about statistics, and aren’t just BSing.) When come back, bring such an article, or some evidence that every reputable journal that would be an appropriate place to criticize the Lancet study’s methodology is politically biased.
Perhaps when the dust starts to settle we will find out a more accurate number. Until then we should practice restraint and remember that any study must be validated by independants. People need to stop throwing around basically mere opinion as fact…myself included.
Whoa, steady up there, big hoss!
Its one thing to criticize the choice of statistical methods, but intepreting the intent of those decisions is another thing altogether. Do you suggest that the methods were chosen with an eye to producing exaggerated numbers? And then they furthered that nefarious effort by gathering “fraudulent” data?
When did you first catch on, what gave the game away? What reliable statistics have you, gained by an honest eye? How is it that you are sure that the methods chosen were chosen for reliability of result, rather than practical necessity? I haven’t the slightest clue how I might go about gathering reliable statistics. Have you? So much so that you can tell us with authority that their methods were bogus?
And **Atry’**s question is sound: if, as you say, these things are so clearly bogus, why is the scandal not more widely exposed? Do these people fear to offer comfort to Bushco and annoy Ernst Stavro Soros? Something this big, if the bogusity was so apparent, I would expect most, if not all, of the refereed statistical journals to jump on it with both feet and kick it do pieces. Did they?
Have you any substantiation for your insinuation of fraud? If you haven’t, would you mind terribly if the rest of us assume that it was done in good faith?
And if you have such proof, would you bring it forth, toot sweet, I am quite in the mood to be astonished.
That’s my sense about this, too. I have to say I don’t understand this insistent on going with any one particular number. No matter how you look at it, it’s A WHOLE LOT OF PEOPLE. And with the chaos that is Iraq, how is anyone really to calculate this? Would we feel any better if it were “only” 100,000 instead of 500,000? Those numbers of dead are just too high to imagine. Do we determine how moral we are by comparing the numbers dead in the war with the numbers dead while SH was in power?
The sad truth is that I don’t expect we’ll ever know the correct number. Can we all just agree that it’s too many?
The point is, it’s not just in the past. The numbers in which Iraqis have died since 2003 is indicative of the numbers in which they are dying now, which is relevant to any discussion of whether and when and how we should pull out our troops.
Actually, the numbers probably HAVE come down since then, but that’s not a result of “the surge” or improved security. It’s a result of the ethnic cleansing of neighborhoods already being mostly effected.
Don’t you think it’s some combination of the two? It would seem an odd coincidence that just at the point when the added troops were finally in place (sometime in late summer last year), there was a dramatic drop in violence against Iraqis and against US troops.
Which I think is all the Lancet study really was ment to demonstrate. It’s at best an order of magnitude estimate. It’s not like its authors are pretending that their methodology is anywhere close to ideal, they give huge margins of error. But the fact is that Iraq remains too chaotic and dangerous (and I doubt the money or interest exists) to do the sort of comprehensive census that Maniac mentioned was done in Bosnia. So the only real option was to pick a a few random samplings and extrapolate.
Wiki discusses several estimates. Lancet is not even the highest.
That is the problem, it is apparent people like **ManiacMan ** and Chouan believe the ~100,000 figure is acceptable to achieve the victory they believe is inevitable.