Who is responsible for the inflated "100,000+" dead Iraqis myth?

[brentspeak]

According to this site, the figure is more like 26,000:

http://www.iraqbodycount.net/

Parading an inflated number of civilian casualties is propaganda, as revolting as one of Bush's Town Hall meetings. Citing such a figure only leads to easy discreditation. It also offends my sense of intellectual honesty.

[alarimer]

1. It was the Lancet, a respected medical journal

I have no idea how they came up with it though. My guess is that 100,000 is on the low side.

[GCP]

2. It was "The Lancet" - an extremely respectable and important British

Medical journal.

http://www.guardian.co.uk/Iraq/Story/0,2763,1338749,00.html

I'd be careful with that "myth" crack if I were you.

[brentspeak]

10. Tell me why I should be careful.

All you have is a link to an article that talks about the Lancet article. The website I linked to actually has an accounting for each of the Iraqi civilian deaths. It's not based on some crackpot methodology that the Lancet apparently used: interpolation of three randomly-chosen Iraqi neighborhoods.

[geek tragedy]

11. It's not the methodology that's so bad--it's the interpretation of the

results. A 95% confidence range of 8,000-200,000 most certainly does not mean that 100,000 died.

[brentspeak]

18. That's what bothers me

Broad interpolation of randomly-gathered data might be sufficient when measuring, say, literacy in a large population. But something like ascertaining the number of civilian fatalities in a war -- especially in today's age -- shouldn't require a whole lot of speculation. All one has to do is assemble a list of names, verify the circumstances, and achieve a decent ballpark figure.

[DFWdem]

29. Here's a link:

http://www.cnn.com/2004/WORLD/meast/10/29/iraq.deaths/

The researchers surveyed nearly 1000 Iraqi households in September, asking how many people lived in the home and how many births and deaths there had been since January 2002.

They then compared the death rate among those households during the 15 months before the invasion with the 18 months after it, getting death certificates where they could.

An expert on study methods who was not involved with the research, said the approach the scientists took was a reasonable one to investigate the Iraq death toll.

But Richard Peto, who is professor of medical statistics at Oxford University, cautioned AP the researchers may have zoned in on hotspots that might not be representative of the death toll across Iraq.

[Frederik]

44. They did the opposite of that

"Richard Peto, who is professor of medical statistics at Oxford University, cautioned AP the researchers may have zoned in on hotspots that might not be representative of the death toll across Iraq."

Au contraire, they excluded the most violent areas for this very reason.

[malachibk]

28. No, it doesn't.

But it could. It could also mean 195,000 died. Or 8,001.

A 95% Confidence Interval means that if the statistics were run an infinite number of times, 95% of those times the point estimate would fall in that range. That's the "frequentist" perspective. A biostatician might say s/he is "95% confident the true value falls in this range" but as an epidemiologist I bristle at that interpretation.

Unfortunately, this CI is quite broad and imprecise, probably due to large standard errors.

The original poster, though, I think is misguided in thinking it should be simple to get a verifiable number of Iraqi civilian deaths. The US surely isn't counting, and I can't imagine they'll let anyone else either.

And for what it's worth, I second the post that averred the Lancet is a top journal. It truly is.

[Frederik]

42. You misunderstand

the meaning of confidence interval, as do everyone who (without knowledge of population studies) criticize the Lancet study. It doesn't mean "the figure could be anywhere between 8,000 and 200,000 with 95% certainty", as many seem to think. The probability curve is bell-shaped and so the probability is highest, assuming the data are correct, that the number is 98,000. A confidence level of 95% is the norm in medical research, for instance - are all medical studies equally "flawed"?

The Lancet people said 100,000 was a "conservative" estimate, as they had excluded the most violent areas from their research (Falluja, for example).

[papau]

55. Well said :-) 100,000 is the BEST ESTIMATE - and 8000 would be a

lousy estimate.

[geek tragedy]

57. I agree in general about the bell curve. But when the confidence interval

dwarfs the mean like that, it suggests a woefully small sample size.

[Frederik]

Yes, it was small

The sample should have been larger, but then there was a war and all. It's certainly not an accurate estimate considering the sample size and the nature of the data collection, but it's not just guesswork either.

[geek tragedy]

87. I agree. eom

[seabeyond]

37. it is unforunate that our government and military say they arent

in the business of counting the innocent lives lost, so we dont have an accurate figure to go on, isnt it brentspeak. that IS the OUTRAGE isnt it brentspeak that we cannot even know how many INNOCENT people we have murder because our president doesnt think it is important.

[Frederik]

38. The people who carried out the Lancet study

are experts in their field. They know what they are doing, and it's not using "crackpot" methodology".

Iraqbodycount only count deaths that are reported, a spokesman for the site said he found the Lancet number to be credible and probable.

The number is certainly higher than 100,000 now, as the data collection was completed mid-04 and excluded the Falluja area because of the high level of violence there.

[brentspeak]

52. OK, maybe "faulty" would have been a better way to describe

how I feel their methodology. It might be a perfectly acceptable method to use when measuring hard-to-quantify things like human behavior in populations. "Casualties", though, can much more exactly be measured.

I've been trained to be sort-of ruthless when it comes to evaluating estimated statistics; always accept the confirmed data, and question with extreme scrutiny anything else.

[bloom]

91. Re: the Lancet study from .iraqbodycount.net

"The Lancet study's headline figure of "100,000" excess deaths is a probabilistic projection from a small number of reported deaths - most of them from aerial weaponry - in a sample of 988 households to the entire Iraqi population. Only those actual, war-related deaths could be included in our count. Because the researchers did not ask relatives whether the male deaths were military or civilian the civilian proportion in the sample is unknown (despite the Lancet website's front-page headline "100,000 excess civilian deaths after Iraq invasion", the authors clearly state that "many" of the dead in their sample may have been combatants ). Iraq Body Count only includes reports where there are feasible methods of distinguishing military from civilian deaths (most of the uncertainty that remains in our own count - the difference between our reported Minimum and Maximum - arises from this issue). Our count is purely a civilian count.

... Nonetheless, the Lancet's estimate of 100,000 deaths - which is on the scale of the death toll from Hiroshima - has, if it is accurate, such serious implications that we may return to the subject in greater detail in the near future. As of this writing we are more concerned with renewed air and ground attacks on Falluja, which last April left over 800 Iraqis dead, some 600 of them civilians (see previous IBC press release below).

It may already be noted, however, that Iraq Body Count, like the Lancet study, doesn't simply report all deaths in Iraq (people obviously die from various causes all the time) but excess deaths that can be associated directly with the military intervention and occupation of the country. In doing this, and via different paths, both studies have arrived at one conclusion which is not up for serious debate: the number of deaths from violence has skyrocketed since the war was launched (see IBC Press Release September 23rd 2003 ; also AP 24th May 2004 ).

We also recognise the bravery of the investigators who carried out the Lancet survey on the ground, and support the call for larger and more authoritative investigations with the full support of the coalition and other official bodies."

http://www.iraqbodycount.net/press/
----

I think it does not make sense to get hung up on whether people are "civilians" or not. It sounds like many in Iraq are getting involved and they wouldn't be in this war at all if it wasn't for us.

[leveymg]

3. That's CIVILIAN deaths.

Does anyone know a reliable source estimate for the number of Iraqi combatants we've killed?

[CitrusLib]

16. I was just wondering this yesterday. I haven't found anything.

[manic expression]

25. 80,000

Mario Cuomo said it in a speech. I also saw it somewhere else.

[leveymg]

36. 100,000 seems in the ballpark, then. Myth?

Propaganda? I don't think so. Guess you don't have to be so embarrassed the next time you hear that figure. Now, you can even cite some sources.

Good work, DUers. :thumbsup: :kick:

[manic expression]

56. Well, that is military deaths, which is separate from the 100,000

This would actually bolster the number to 180,000, using the Lancet's reliable estimate.

At any rate, if one includes ONLY the iraqbodycount's 20,000 and the known military deaths of 80,000, it equals 100,000, even though the civilian death number only includes deaths confirmed by multiple sources, which makes it woefully low. So yes, 100,000 killed because of the US actions can be correct even if one uses the iraqbodycount number of civilians (again, an extremely low count, not nearly the reality).

[MisterP]

47. without Falluja, and in mid-late '04, and only civilians (excluding

14-year-olds they assigned to carry rifles and guard a bridge as the old government collapsed: the World Emperor will have a million human souls on his account in Hell in just a coupla years!

[Sanity Claws]

4. It was a health or medical organization

I think from Britain.
The number was the increase in the death rate since the invasion. So it includes not only loss of life due to IEDs, etc., but also due to lack of medical care, violence in general, decrease in nutrition, increase in child mortality, etc.
I'm not sure the number should be considered propaganda. I think you have to understand what the author was trying to calculate.

[K-W]

5. You are confusing documented with estimated deaths. EOM

[SmokingJacket]

6. Well, here's an article.

http://www2.chinadaily.com.cn/english/doc/2004-10/29/content_386739.htm

Iraqbodycount counts reported deaths, while this study went around to families and asked how many people they lost.

I'm not sure if this study is 100% reliable, either, but it in no way compares to one of B's town hall meetings. It is a number that was arrived at in good faith, imho.

[SeveneightyWhoa]

7. How do you know its a "myth"?

Are you sure it's "inflated"? "Propaganda"?

I don't think either source has the correct number. I'd say this is more of an art than a science.

[geek tragedy]

8. It was Lancet, and much of the 'myth' has to do with a misapprehension

of statistics.

The survey found that there was a 90 or 95% chance that the death toll was between 8000 and 200,000. They picked the number in the middle--100,000 and ran with it.

[FogerRox]

13. WAsnt that 100k number from like about a year ago ?

Typically it might be safe to assume in a geurilla war the the insurgency might lose 10 to 1 the number of casuallities.

USA =1700-- this doesnt count the medivaced soldiers, that might have died in Germany --
Times 10= 17,000

But I would think the total Iraq listof dead, might closer to 200k at this point.

[geek tragedy]

20. The problem was that the Lancet study didn't really produce a number

like 100,000. It produced a VERY wide range--8,000-200,000.

One can't simply split the difference and say that's the number of dead. Stats don't work that way.

[mike_c]

31. that's a bit misleading....

They didn't just "pick the number in the middle and run with it" arbitrarily, which implies that they might just as easily have picked any number within that range with equal justification-- that confidence interval is based upon the estimated mean of 100,000 dead civilians, given the observed sample variance. 100,000 dead civilians WAS the estimated mean, regardless. The wide confidence interval suggests only that there was a great deal of variation within the data, which most likely means that the sample size was simply too small to eliminate the noise. That, in turn, IN NO WAY undermines the estimated mean, because the ONLY DATA IN HAND provide support for that particular estimate of the mean. The real mean is undoubtedly not 100,000, but there's NO support for it being very far from that estimate, at least within the data available.

[geek tragedy]

39. A mean with that gigantic of a variance simply isn't a solid number.

When the confidence interval is that big, it's just inaccurate to say that the 100,000 figure has any kind of certainty attached to it.

[Frederik]

48. Of course there's no certainty

attached to it. It's an estimate based on extrapolation of data collected in a war zone. But you seem to think they just "picked the number in the middle" and that it could equally well have been 8,000 or 200,000, which is not, to quote yourself, how statistics works.

It's only an estimate, but it's the only credible estimate we have.

[geek tragedy]

50. I guess I don't find that figure credible, given the confidence interval

It very well could be right. The actual number could even be much higher. But Iraq is a black hole of misinformation and confusion--the only thing we know over there is that we really don't know anything.

[Frederik]

81. Agreed

[mike_c]

49. only in a limited sense....

Making some necessary assumptions about the data distribution, of course, there is a 95 percent likelihood that the true mean falls within the interval, but the only number supported by the data is the estimated mean of 100,000. Simply picking any other number from within the interval is bogus because any other number has zero support-- NOT 95 percent support (that's a common misunderstanding), but NO support. One could could use that argument to "say" that there is a 95 percent chance that the true mean is 10,000, but (A) by the same logic it could be 190,000, and (B) that argument is spurious anyway, because there is only one estimated mean SUPPORTED BY THE DATA.

As for that being a "solid" number-- no estimated mean is a "solid" number, only the true mean is an actual parameter of the data. Likewise, there is no general rule regarding the relationship between the estimated mean, the true mean, AND the single sample variance. You can prove this intuitively by envisioning a *noisy* data distribution in which the estimated mean equals the true mean-- something the observer cannot actually know, of course. The noise does not necessarily decrease with sample size, although the squared deviation usually does.

[geek tragedy]

63. My point is that the confidence interval indicates that the sample size

was insufficient for this to provide the basis for any conclusions.

[mike_c]

68. I'm sorry-- I don't mean to keep harping on this...

...but data analysis is something near and dear to my heart. There is simply no genuine support for your last statement, as intuitive as it might sound. Think of it this way-- suppose the true mean of some set of widgets is 50-- which you do not know, of course-- and that they vary in widgetness from 0 to 100. Now suppose you select ONE widget at random, you measure it, and your result is 50. You happened to randomly select one widget whose widgetness EXACTLY corresponded to the true mean. There is a finite probability of this happening, so the example is NOT contrived. With a sample size of n=1, your estimated mean is exactly equal to the true mean.

OK, that example ignores the fact that you can't estimate a mean with only one sample, but you see my point-- it is entirely possible to accurately estimate the true mean with a small sample size. Imagine that you selected two widgets, where y₁ = 1 and y₂ = 99; again, the estimated mean would be exactly equal to the true mean, despite the sample size being exceedingly small AND the variance exceedingly high. (on edit-- that would yield precisely the large confidence interval you're using as justification for the estimated mean NOT being an accurate estimate of the true mean.)

If the data are normally distributed, values will be more likely clustered near the mean than distant outliers, further increasing the likelihood of obtaining unbiased, accurate estimates of the true mean with relatively small samples sizes, PROVIDED that the samples are random and independent.

[geek tragedy]

73. Extrapolating from 1000 households is risky stuff, imo.

If they were off by a total of 20 total deaths among 1000 households, that could tank the entire survey.

[Vladimir]

95. Its a standard sample size for a national opinion poll

in the UK, which has a bigger population than Iraq - not saying opinion polls are perfect, but its rare to see one go more than 10% out.

[geek tragedy]

97. But opinion polls involve relatively simple questions--yes or no or

multiple choice.

The interviews/surveys for this study were no doubt much more complex and involved. The greater the complexity, the greater the chance for error to creep in.

[Vladimir]

103. Well, you could say the same of most consumer surveys,

again done with similar samples. I don't think the sample size is such a huge problem here, although bigger is always better with samples. There are plenty of other sources of error, which is why I think the probability interval is so wide.

[geek tragedy]

107. It's one thing to estimate percentages--it's another to extrapolate

totals of deaths.

To put things in perspective, each death as measured in that survey counts for about 4,000 in extrapolation.

[Vladimir]

109. No, I don't think so

extrapolation is implicit in most surveys; as one example, with marketing surveys you are often surveying precisely so you can extrapolate to a predicted demand for your product.

[geek tragedy]

110. But the goal there isn't as precise as it is here--to measure the total

loss of life. There's a razor-thin margin of error--being off by 25 deaths out of 1000 households means a misestimation of 100,000 deaths.

[Vladimir]

112. Yeah, and if you make 100,000 units too many

your company may well go bust. Extrapolation is an everyday bit of statistics, on samples precisely this big if not smaller, and there are well established procedures for handling the errors, including the sort of inevitable systematics that come up with any study of this kind. To give you an example from my own research - I estimate the number of background events in a particular particle physics collision, based on a sample many many times smaller than the population, and a sample in which I do not find a single event of the type being extrapolated at that!

I think the Lancet estimate was, at the time, the best estimate out there.

[geek tragedy]

113. It's the best estimate out there, for sure.

In the land of the blind, the man with one eye is king.

[bloom]

118. How many people in your household have been killed in the war?

How many people were in your household in 2003?

And - How many people who were in your household in 2003 have been killed for any reason attributable to the war?


Seems like it could be fairly simple. The questions, that is.

[rinsd]

45. Strange basis for the data...

Maybe I misread but they were basing this on surveys of 30 families each for 5 researchers bringing the total to 150 families.

Seems like a small sample size to extrapolate data.

[mike_c]

62. as long as the data were random and independent, 150 samples...

...is quite capable of providing a good estimate of the true mean death rate. The biggest problem becomes one of unintended sample bias, which is why the researchers excluded the effects of Falluja-- averaged into 100 times that many samples, the Falluja effect might have been negligible, but it could have seriously biased the results at the smaller sample size. One common way to deal with this-- rather than the conservative approach the authors apparently took-- is to exclude any samples that fall outside of 2 standard deviations from the mean, because their strength as outliers is disproportionate to their frequency within the data. Increasing the sample size tends to "reel in" those outliers to within the 2 SD interval.

[rinsd]

89. Still doesn't make sense.

I found this slate article.

http://slate.msn.com/id/2108887/

Apparently the authors think there is a 95% certainty that between 8000 and 192,000 deaths happened.

On Edit: It also appears that they simply made up the pre-invasion mortality.

Also while reading the Lancet it appears to be more of a progressive journal showing the dangers of partisanship when coupled with science.

[mike_c]

101. I'm not going to insult your intelligence...

...by giving you a lecture, but with all due respect, your understanding of statistics-- and of interpreting statistical analyses-- is flawed (or was that misinterpretation made by the authors of the article you cited?). In particular, that is NOT the correct interpretation of the 95 percent confidence interval.

[Vladimir]

106. The Lancet now has a political bias?

We are talking about a peer-reviewed medical journal, with the 3rd highest worldwide impact factor as of 1999, here. What next, Nature is a mouthpiece for Marxism?

[ConsAreLiars]

111. You say "It also appears that they simply made up the pre-invasion

mortality." I have to assume that you never even looked at the study, and/or the author you cite really was only looking to criticize it, not to explain it, and so misled you into believing what you said.

Making such a blatantly false and easily checked claim is a good way to discredit your viewpoint and cast suspicion on your conclusions, but not a very effective way to argue against the estimate they came up with.

If the facts matter to you at all, you can find the study at: http://www.countthecasualties.org.uk/docs/robertsetal.pdf

(edit to add missing word)

[Vladimir]

69. 33 neighbourhoods of 30 families each, actually

brings us to circa 990 families, probably minus a dozen non-respondant ones or so.

See the post further down the thread giving the full (ish) methodology.

[Maple]

9. Lancet AND

Johns Hopkins

[Capn Sunshine]

12. There's a reason for the confusion

Every journalist working on actual casualty statistics has been killed or wounded by Coalition forces accidentally. Crossfire is a bitch.

More journalists have been killed in Iraq thus far than all the journalists killed in the entire 12 years of the Vietnam conflict.

[blue northern]

14. From the same site you linked:

We are not a news organization ourselves and like everyone else can only base our information on what has been reported so far. What we are attempting to provide is a credible compilation of civilian deaths that have been reported by recognized sources. Our maximum therefore refers to reported deaths - which can only be a sample of true deaths unless one assumes that every civilian death has been reported. It is likely that many if not most civilian casualties will go unreported by the media.That is the sad nature of war.

[Career Prole]

15. The Lancet article and estimate was based on a survey

followed by an extrapolation See Here.

Iraqbodycount.net counts only published reports from reliable sources.

Two vastly different methods with understandably vastly different results.

I don't find it at all hard to believe that the actual numbers may be larger than those shown by Iraqbodycount.net because not all deaths would have been reported in the media and thus wouldn't have made their count. Whether the total would reach the 100K the Lancet published I can't say...possibly no one will ever know for sure.

With Iraqbodycount.net, as they advertise, every digit represents an actual documented death.

[BurtWorm]

23. Iraqbodycount.com cites the Lancet article as a source.

I think IBC's count is based on confirmed or confirmable deaths, but that leaves out deaths that are unconfirmable for whatever reason, which means a LOT of potential deaths, seeing as how uncertain communication and records keeping is in the country.

[Career Prole]

100. That's why their minimum is the *bare* minimum.

Anyone claiming less is full of hot air.

http://www.iraqbodycount.net/background.htm#methods

[BurtWorm]

17. It was a Johns Hopkins team that did the count. Here are their methods.

http://www.jhsph.edu/PublicHealthNews/Press_Releases/PR_2004/Burnham_Iraq.html

The researchers compared the mortality rate among civilians in Iraq during the 14.6 months prior to the March 2003 invasion with the 17.8 month period following the invasion. The sample group reported 46 deaths prior to the March 2003 and 142 deaths following the invasion. The results were calculated twice, both with and without information from the city of Falluja. The researchers felt the excessive violence from combat in Falluja could skew the overall mortality rates. Excluding information from Falluja, they estimate that 100,000 more Iraqis died than would have been expected had the invasion not occurred. Eighty-four percent of the violent deaths were reported to be caused by the actions of Coalition forces and 95 percent of those deaths were due to air strikes and artillery.

“There is a real necessity for accurate monitoring of civilian deaths during combat situations. Otherwise it is impossible to know the extent of the problems civilians may be facing or how to protect them,” explained study co-author Gilbert Burnham, MD, associate professor of International Health at the Bloomberg School of Public Health and director of the Center for International, Disaster and Refugee Studies.

“Mortality before and after the 2003 invasion of Iraq: cluster sample survey” was written by Les Roberts, Riyadh Lafta, Richard Garfield, Jamal Khudhairi and Gilbert Burnham. Roberts and Burham are with the Johns Hopkins Bloomberg School of Public Health. Lafta and Khudhairi are with the College of Medicine at Al-Mustansiriya University in Baghdad. Garfield is with the Columbia University School of Nursing.

The study was funded by the Center for International Emergency, Disaster and Refugee Studies at the Johns Hopkins Bloomberg School of Public Health and the Small Arms Survey in Geneva, Switzerland.

[Name removed]

19. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Ms. Clio]

85. "Your sense of intellectual honesty my ass"

Mine, too.

[Ysolde]

21. As others have stated...

it was a study done by Lancet and Johns Hopkins researchers. I listened to the researcher detail how they came up with their numbers. It's not a "myth". It was a very well done and reasoned research project. That, I might add, completely left out any casualties from Fallujah. So, their numbers are going to be on the low side as they didn't take into account any of that carnage. If you didn't know where the number came from, nor how the researchers got to those numbers, why did you trash the data?

And, why do you find it so hard to believe that over 100,000 Iraqis have died? It's been over 2 years and they still don't have clean water. Many of the poorest are/were using barrels to collect rain water that originally contained radioactive chemicals. So, since you don't like their numbers nor research, how would you prefer we count the true human toll of this war?

[brentspeak]

27. While iraqibodycount.com can't be a completely accurate #

because it only reports media-obtained figures, by its very nature it's more credible because it can account for each one of its data. We can say, without any problem, and with a whole lot of certainty, that the number of Iraqi civilians killed in the war is at LEAST 26,000.

But we can't say with a whole of certainty that the figure is at least 100,000. Why? Simply because JH/Lancet is relying on extrapolated samples of data -- data that doesn't have names of victims, or anything else that can more firmly corraborate the study's conclusions.

Perhaps there really are at least 100,000 dead Iraqis. You would need names, circumstances -- to verify that number.

[iconoclastNYC]

34. OK then fund a study

Why don't you donate the millions of dollars it would take to do a study?

For the rest of us the Lancet's estimate is fine.

Anybody who thinks that 28,000 verified deaths is acceptable isn't going to be swayed by having the 100,000 VERIFIED.

People who support this war do so because it's a REPUBLICAN war. If this was Clinton's war all these flag waving war mongers would oppose it.

We oppose the war it because it was illegal, unnecesary, amoral, sold on a pack of lies, is damaging our military, and has made our country and the world less safe. If that's not enough reason to oppose it then statistics aint going to do a bit of good for you.

[brentspeak]

41. "For the rest of us"?

You mean 'for you'. Individually. And maybe some other people here.

I'm not sure why you're lecturing me about the war, as my opposition to it should be pretty obvious to anyone who's read my posts here would know.

This thread I started because I'm interested in the issue of accuracy with regards to Iraqi civilian deaths.

[Kraklen]

53. It's not obvious to me.

In fact, you seem to be going out of your way to pretend it's not all that bad.

[brentspeak]

61. Yeah? How do you figure that?

Because I insist on relying on confirmed data?

[Kraklen]

67. No.

By being intellectually dishonest and slandering peer-reviewed scientific data.

[BurtWorm]

46. Not all Iraqi deaths as a result of the war are reported in the media.

What needs to be figured out is how many the media miss. And that's not an easy task.

[mike_c]

78. this is only partly accurate....

Unless one invalidates the measurements of increased mortality reported by the Lancet and JH, I'd argue that their estimate is at least as credible as that provided by iraqibodycount.com, and probably much more so. First, IBC relies upon media accounts, which MUST be a subset of the actual number of dead-- there's no argument about that, the only question is how big or small a subset. So the actual number can only be higher than 26,000 if we accept the validity of IBC's source data (which is a separate issue, and a sword that cuts both ways). Your statement that the number killed must be AT LEAST 26K is correct in that limited sense.

But unless the original Lancet data are called into question, the increased mortality rate during the invasion and occupation is equally well documented, and since the IBC data represent only a subset of the actual mortality, one might argue that the Lancet data is the ONLY credible estimate of the true number-- the full set-- of civilian deaths.

As far as I am aware, no other estmate of the actual number of civilian deaths attempts to account for the catastrophic rise in post-invasion mortality among Iraqi civilians.

[DFWdem]

43. Are you sure?

The researchers surveyed nearly 1000 Iraqi households in September, asking how many people lived in the home and how many births and deaths there had been since January 2002.
They then compared the death rate among those households during the 15 months before the invasion with the 18 months after it, getting death certificates where they could.

If they did this survey at my house the answer would be zero in both cases. This isn't to say that there haven't been an abhorrent number of deaths in Iraq, but to simply say 100,000 as gospel is questionable at best.

[rinsd]

54. They're basing this on 150 families surveyed...

...and analyzed a total of 73 deaths.

This study is bullshit.

[dummy-du1]

72. Which study?

About the Lancet study:

The researchers used a traditional epidemiologic technique called a clustered sample survey. Without getting into technical details, the country was divided into a number (33 in this instance) of subgroups and a community was randomly selected from each cluster. In each community, Global Positioning System (GPS) devices combined with random numbers were used to select a particular point in the community. Then the nearest 30 household were surveyed; these 30 households are referred to as a cluster.

The important points in the previous paragraph are that the procedure was systematically designed to represent the population of the entire country and that random numbers were used whenever choices had to be made. The combination of a systematic procedure along with random numbers when choices must be made are techniques used by survey researchers to avoid any systematic bias due to the researchers intentionally or unconsciously selecting certain communities or households. If carried out successfully, these procedures allow a fairly small sample to accurately represent the population of an entire country. Thus, while we may have questions about the accuracy of election polls, they are usually at least in the ballpark of representing the vote-getting potential of a candidate. Seldom does a candidate get 10 percentage points more or less than an immediate pre-election poll.

The researchers ended up with information on 7,438 people pre-invasion and 7,868 people post-invasion. These people represented 988 households in 33 clusters.

http://www.zmag.org/content/showarticle.cfm?SectionID=15&ItemID=6565%20

[blogslut]

22. Nobody knows the number...

...because the people in charge of controlling the counting refuse to count.

Now, you tell me who are the real mythologists in this fucking illegal war?

[FogerRox]

24. Right on BLog slut

[iconoclastic cat]

26. Cool blog, slut!

:evilgrin:

[manic expression]

30. Umm....its not a myth at all

It's a fairly conservative estimate from a highly respected medical journal.

Iraq Body Count only counts deaths that are confirmed by MULTIPLE sources, which is not easy to get. This means that the vast majority of deaths are unreported and not counted.

[brentspeak]

32. We don't know what a conservative estimate of fatalities is

All we know is confirmed deaths. That's it. Are there more? Of course there are. How many more, there's no to know. We would have to wait until the war is over for there to be a proper accounting. You tally names, circumstances of death, etc.

For each victim, there should be a name.

[manic expression]

51. It's a conservative estimate in relation to what they found

Estimates are used all the time for just about everything. You can't get a name for each victim because it's unrealistic to do so. Using your logic, there wouldn't be a tally for the deaths in London for a long time, since they can't get clear confirmation on the names of the victims right away, and this is in a completely peaceful country, with one attack; compare that to Iraq.

How many more than the confirmed deaths? Well, the number is very surely 100,000 (total), while it is most likely that there are more. To say that there is 28,000 (or whatever that number was) deaths is to be negligent of the truth, especially when so much evidence points to much more.

The 100,000 number was found using reliable evidence and again, it was a conservative number based on what they found. What more do you want?

[brentspeak]

58. Allright, then number of bodies, then

Just anything that can more firmly put physical evidence to an estimated number.

For myself, what I say when discussing the issue with other people: There's at least 26,000 civilians killed in the course of this war, probably a lot more. How many more, we don't really know.

[manic expression]

65. OK, as long as you say:

The VERY LEAST number is 26,000. However, if you were to say that AROUND 100,000 civilians have been killed, that would be perfectly accurate. Whichever you prefer.

Another thing you could say: We KNOW that there's at the very least 26,000 civilians killed, probably a lot more, an estimate from a respected source using reliable methods puts the number at 100,000.

[OmmmSweetOmmm]

117. Orders were given to stop counting in December 2003. See my post 114.

Don't forget DU either. That is a gift that keeps giving and can add up to countless number of future deaths.

[G_j]

33. questions concerning the survey

note: a link to the Lancet report can also be found on this page.

http://www.casi.org.uk/briefing/041101lancetpmos.html

the Prime Minister's Official Spokesperson (PMOS) dismissed the study since its methodology was, he claimed, inappropriate:

<snip>
1) "Firstly, the survey appeared to be based on an extrapolation technique rather than a detailed body count."

-The fact that the survey uses an extrapolation technique does not automatically mean that it is less likely to be accurate than a body count. In fact, it is more likely to be accurate than existing attempts at body counts of Iraqi civilians. This is because in Iraq, where there are so many no-go areas, it would be impossible to count every casualty.

This is also the problem with Iraqi Ministry of Health figures obtained by counting bodies arriving at hospitals, cited by the PMOS in response to the Lancet report on 1st November: many of those who have died or been killed will never arrive at hospital in conditions of war, when access to roads and health facilities are severely disrupted, and when it might seem pointless to risk the journey for the sake of someone who is already dead.

Finally, attempts to do body counts through deaths reported in the press, like Iraq Body Count, are also necessarily underestimates, since press reports of casualties will be incomplete, not least because the areas where people are being killed are the same areas into which journalists don't dare go. As Iraq Body Count states on its website:

"Our maximum therefore refers to reported deaths - which can only be a sample of true deaths unless one assumes that every civilian death has been reported. It is likely that many if not most civilian casualties will go unreported by the media."

In these circumstances, population-based research like the Lancet study is thus arguably a much better source.

A detailed, comprehensive body count on the ground by military and medical personnel would perhaps be more accurate, but it is a task which the US and UK forces in Iraq refuse to carry out. For Downing Street to question the best available study on the grounds of preferring a methodology which they themselves refuse to operate is surprising, to say the least.

2) "Our worries centred on the fact that the technique in question appeared to treat Iraq as if every area was one and the same."

-The technique does not treat Iraq as if every area was the same. The survey takes a sample in each of 33 areas throughout Iraq’s 18 governorates, precisely so that regional variation is taken into account as much as possible. All regions of the country were sampled.

The PMOS might be referring to the fact that due to the deteriorating security situation in Iraq, the survey 'paired' 12 of the 18 governorates, and took samples from only 1 of the pair in order to reduce danger to the investigators by reducing the necessary travel. In these cases, however, the process of assigning samples to one of the two governorates was randomised, and weighted according to the population of the governorates, thus 'still sampling from all regions of the country' and ensuring that 'the sample remained a random national sample' (Roberts et al p.2)

3) "Secondly, the survey appeared to assume that bombing had taken place throughout Iraq. Again, that was not true. It had been focussed primarily on areas such as Fallujah."

-This is also untrue - see above. The survey was designed precisely to take variations throughout Iraq into account. Moreover, since the mortality rate due to violent deaths in the sample collected in Fallujah was so abnormally high, the Fallujah sample was excluded from the study’s final estimate. Thus far from overestimating the deaths caused elsewhere by extrapolating them from those clustered in Fallujah, as the PMOS suggests, in fact the figure of 100,000 deaths does not take into account the exceptionally high mortality probably caused in the area of the most intensive bombing, namely Fallujah. Had the Fallujah sample been included, the survey's estimate would have been of an excess of about 298,000 deaths, with 200,000 concentrated in the 3% of Iraq around Fallujah (Roberts et al p.5).

In addition, contrary to the PMOS’s claims about the tight focus of bombing, the Lancet study in fact found that in areas well away from Fallujah the most significant cause of the deaths it recorded was still violence, most attributed to coalition bombing (Roberts et al p.4).

4) "Consequently, we did not believe that extrapolation was an appropriate technique to use."

-Almost all surveys except censuses involve 'extrapolation techniques' (using a sample rather than the entire population), since it is simply impossible to survey every member of a country's population. Extrapolation is a universally recognised method that governments and academics use to obtain data, and it is remarkable to see a government spokesperson rejecting it outright.

In addition, the specific sampling technique used in the Iraq mortality survey, a clustered sample survey, is widely used by government and other surveys to gather data from geographical clusters of households. See, for example, the recent report from the Home Office Research, Development and Statistics Directorate, 'Asylum seekers' experiences of the voucher scheme in the UK, March 2002', which used cluster samples in order to reduce the cost of the survey (see pp.2-3 of this report).

<snip>

[Karenina]

59. G_j, I need a hug from you now

before I collapse in convulsions of nausea.

50% of Iraq'as population are CHILDREN.
No clean drinking water, power, food supply interrupted, hospitals bombed, birth defects from illegal weapons, the Health Ministry has GIVEN UP trying to count the bodies, occupation forces shoot first, cities flattened, refugee camps with no supply lines and some jerkoff challenges an INSULTING lowball number? Please hold me close and allow me to cry my OUTRAGE onto your shoulder rather than to confront such hubris in kind, as we both know violence begets violence...

[G_j]

71. I know what you mean

so... a BIG :hug:

it is truly heartbreaking and even if I tried I could not forget the children slaughtered by the war criminal and "chief".
Before calling the numbers a myth and writing them off as inflated the poster might have done a bit of research, me thinks...

:cry: :hug: :cry:

[Karenina]

79. Thanx SO MUCH, Dear One!

It's late here, and with the recent deaths of Jimmy Woode, Luther Vandross, Andy Stephenson and those in London and Iraq, ALL of which have affected me personally, I find myself unable to cope with SUCH BRUTAL INSENSITIVITY. I will go now and pull the covers over my head for some hours.

[G_j]

82. probably a wise idea

to add to it all, I cannot help but feel the London bombings also robbed us of us a singular chance to have global warming and poverty as the leading stories in the news because of the G8 meetings. Will that window be open again? I don't know, but I am afraid for our world,
and I am in grief. :cry:

[Tinoire]

102. Hug my friend

:hug:

I wasn't even going to post but then I saw this. Too revolted to put it mildly.

Take care :hug:

[Vladimir]

76. Well nailed n/t

[pocket]

35. I've seen the number 300,000 in a variety of places in the past week

The 100,000 estimate is pretty old now, so this would make sense.

[bloom]

88. I expect that 300,000 is more accurate at this point.... eom

[Chicago Democrat]

40. Its not an exaggeration...

The 26000 is for people who are counted.



Most deaths arent counted. Read the Lancet article. Could be 200,000. 80% women and children. Includes dysentary and other easily preventable deaths.


Consequential deaths vs Combat Casualties. All real.

[Jack Rabbit]

60. Johns Hopkins University

From the Bloomberg School of Public Health, Johns Hopkins Univerity
Dated October 28, 2004

Iraqi Civilian Deaths Increase Dramatically After Invasion

Civilian deaths have risen dramatically in Iraq since the country was invaded in March 2003, according to a survey conducted by researchers from the Johns Hopkins Bloomberg School of Public Health, Columbia University School of Nursing and Al-Mustansiriya University in Baghdad. The researchers found that the majority of deaths were attributed to violence, which were primarily the result of military actions by Coalition forces. Most of those killed by Coalition forces were women and children. However, the researchers stressed that they found no evidence of improper conduct by the Coalition soldiers.

The survey is the first countrywide attempt to calculate the number of civilian deaths in Iraq since the war began. The United States military does not keep records on civilian deaths and record keeping by the Iraq Ministry of Health is limited. The study is published in the October 29, 2004, online edition of The Lancet . . . .

The researchers conducted their survey in September 2004. They randomly selected 33 neighborhoods of 30 homes from across Iraq and interviewed the residents about the number and ages of the people living in each home. Over 7,800 Iraqis were included. Residents were questioned about the number of births and deaths that occurred in the household since January 2002. Information was also collected about the causes and circumstances of each death. When possible, the deaths were verified with a death certificate or other documentation.

The researchers compared the mortality rate among civilians in Iraq during the 14.6 months prior to the March 2003 invasion with the 17.8 month period following the invasion. The sample group reported 46 deaths prior to the March 2003 and 142 deaths following the invasion. The results were calculated twice, both with and without information from the city of Falluja. The researchers felt the excessive violence from combat in Falluja could skew the overall mortality rates. Excluding information from Falluja, they estimate that 100,000 more Iraqis died than would have been expected had the invasion not occurred. Eighty-four percent of the violent deaths were reported to be caused by the actions of Coalition forces and 95 percent of those deaths were due to air strikes and artillery.

Read more.

[Vladimir]

64. The Lancet study and the IBC study are completely incomparable

Lancet tried to estimate all Iraqi casualties, civilian or otherwise - IBC only counts civilian dead reported by the media.

[ConsAreLiars]

66. Your "intellectual honesty" should have alerted you to the obvious fact

that the IBC number only a fraction of the total, including only doubly-sourced newspaper reports. It is in no way a "true" number of the civilian casualties caused by the terrorists who control your government. Claiming that the deaths number 26,000 would be a lie, since we know that that is only a subset of the total number.

Your wish to characterize the Lancet estimate as "propaganda" is interesting. It is an estimate of deaths in the first 18 months only, and if you look at the graphs the death rate has been rising ever since the invaders first began the occupation. Perhaps you want to downplay the full extent of the carnage. Or maybe you are just reluctant to face the true horrors of the war crimes being committed. These numbers only tell a small part of the truth of what is being done there. If you have the "intellectual honesty" to see what is being done by your leaders, spend a few hours looking into the faces of some of the victims: http://www.robert-fisk.com/iraqwarvictims_mar2003.htm

Read the Lancet article for yourself: http://www.countthecasualties.org.uk/docs/robertsetal.pdf

[brentspeak]

74. Or maybe,

I just prefer speaking from a foundation of confirmed data. I'm not interested in minimizing anything; my academic background has kind of trained me to scrutinize things as sternly as possible, be as exact as possible.

[libnnc]

80. Well bully for you, then!

Heaven forbid your academic credentials be besmirched by a miscalculation of fatalities and a error in data. Oh the humanity of a mistake like that!

The humanity indeed.

[stepnw1f]

83. Thank You For That! (nt)

[OmmmSweetOmmm]

114. Is that the same "humanity" that stopped the count of Iraqi deaths? Read.

http://www.usatoday.com/news/world/iraq/2003-12-10-iraq-civilians_x.htm

Posted 12/10/2003 1:16 PM
Iraq's Health Ministry ordered to stop counting civilian dead from war


BAGHDAD, Iraq (AP) — Iraq's Health Ministry has ordered a halt to a count of civilians killed during the war and told its statistics department not to release figures compiled so far, the official who oversaw the count told The Associated Press on Wednesday.
The health minister, Dr. Khodeir Abbas, denied in an email that he had anything to do with the order, saying he didn't even know about the study.

Dr. Nagham Mohsen, the head of the ministry's statistics department, said the order was relayed to her by the ministry's director of planning, Dr. Nazar Shabandar, who said it came on behalf of Abbas. She said the U.S.-led Coalition Provisional Authority, which oversees the ministry, also wanted the counting to stop.

"We have stopped the collection of this information because our minister didn't agree with it," she said, adding: "The CPA doesn't want this to be done."

Abbas, whose secretary said he was out of the country

continued above

[ConsAreLiars]

84. Excluding large numbers of victims from your "count"

is hardly being "as exact as possible." Quite the contrary. Using the smaller number in place of the estimate is both deceptive and dishonest.

My "academic background" trained me to be as honest and true to the facts as possible, and not to systematically exclude data. Excluding civilian deaths from the total unless they were accompanied by two or more newspaper reports is irrational in the extreme when you are talking about total casualties.

If you really want to be "as exact as possible" you could use both. Use the IBC count for the limited purpose it claims for itself, and DO NOT falsely present that number as an estimate of the total. It is not, and using that number that way would be a lie. And add at least 50% to the Lancet estimate to include the last 9 months if you want an honest estimate of the total.

[brentspeak]

90. No

I'm not sure that the Lancet paper constitutes the kind of "data" that can be mixed in the IBC, so I'm excluding nothing. But it doesn't matter as far as I'm concerned, because I always talk about the civilian death total along the lines of "There's at least 26,000 civilians dead, probably a lot more." I've received too many responses on these boards blasting me for even saying that -- and seen others being blasted for saying it -- and insisting that I instead say "100,000", which is part of my reason for starting this thread.

It can hardly be called "dishonest" to say "at least X-number of dead civilians" when that X-number is a confirmed number.

[ConsAreLiars]

93. Actually, your original statement was completely dishonest..

You said, "According to this site, the figure is more like 26,000." That is not what that site says. When you misrepresent both what that site says AND what you yourself have said, well, I think the issue of your self-declared "intellectual honesty" is pretty well settled.

[brentspeak]

94. No, it just opens up the issue of your reading ability

Site: (MAX) - 25814

Me: "according to this site, the figure is more like 26,000"

[seabeyond]

98. isnt the outrage brentspeak that our government doesnt feel

it is necessary or warranted to keep these numbers. that they feel they are insignificant, the numbers of innocent that they kill. isnt the outrage that we cannot find a reliable number. that it can be anywhere from 20k to 300k

arent you outraged on that

because for however long this war has been going on, and i have been seeing bloody and dead babies, i have been outraged

[ConsAreLiars]

104. When you refuse to acknowledge that the number at IBC is only a fraction

of the total, you misrepresent the facts. It is a either deliberate misrepresentation of the meaning of that number, or a failure to comprehend what that site is doing. The fact that the site reports only some casualties (those documented in two or more "approved" news accounts) has been pointed out several times, so it is not simple ignorance on your part, which would be a treatable condition. Something else must be at the root of your attempt to discredit that number. Care to share?

http://www.iraqbodycount.net/
"Our maximum therefore refers to reported deaths - which can only be a sample of true deaths unless one assumes that every civilian death has been reported. It is likely that many if not most civilian casualties will go unreported by the media. That is the sad nature of war."

[mike_c]

105. well, speaking as an academician, I also find it distressing...

...that you dismissed the Lancet pub as "propoganda," and I say that from the perspective of a scientist who uses analyses like these daily.

[Walt Starr]

70. There is a simple fact here that everybody seems to overlook on the number

of civilians killed by the U.S. and U.K. in Iraq.

Nobody knows how many and nobody ever will.

That, my friend, is the saddest fact of all in this illegal invasion.

[JackRiddler]

75. Unbelievable

The people of Iraq never did anything to you or your country.

Your taxes went toward murdering thousands of them from the air, even before the invasion.

Your government told a series of big lies to justify attacking them.

Your government flattened their cities and occupied their country.

Then it ordered its loyal apparatchiks to stop counting civilian casualties.

Iraq has been turned into a death zone for non-embedded journalists, dozens of whom have been killed by your government's guns.

No outsider dares to venture into the most dangerous areas.

Now you want to say a few good scientists who try to come up with an estimate of the carnage (in the absence of comprehensive statistics thanks to the US suppression order) are as "revolting" to you as a staged rally for the figurehead of the regime that committed these crimes?

I want the real numbers too, sir, but your hyperbole offends my sense of moral honesty.

[G_j]

77. three completely unsupportable claims

"myth" "inflated" "propaganda" can you provide a link?

There is absolutely nothing to back up any one of these claims.
Didn't you know that many of us here like to see accusations backed up with facts?

[Serial Mom]

86. The US has not counted Iraqi civilian deaths ...

at the top of the site you linked to is this:

“We don’t do body counts”
General Tommy Franks, US Central Command

And I heard the Iraqi's themselves started doing ESTIMATES, it was well into the 2nd year of the invasion ... possibly just since January? I will be trying to find link to this ...

on edit: sp

[Ms. Clio]

92. Actually, Marla Ruzicka learned that the U.S. military does count

civilian deaths, right before she was killed.

The Mysterious Death of Marla Ruzicka:
The US Military has Detailed Statistics on Civilian Casualties
by Michel Chossudovsky

(snip)

The official position within the US Military, which is part of the media consensus, is that "nobody really knows exactly what the civilian toll is".

Marla Ruzicka had refuted the official statements regarding civilian deaths in Iraq. She revealed that 1,995 civilians died and 4,959 were injured in the first 50 days of the invasion ( http://www.uruknet.info/?p=m11278 )

(snip)

Marla Ruzicka had discovered, through careful investigation, that the US military authorities were involved in a cover-up. There was a policy of compiling precise statistics on civilian casualties. These figures, however, were classified and were not intended to be made public.

"These statistics demonstrate that the US military can and does track civilian casualties... Troops on the ground keep these records..."

Shortly before her death, she acknowledged having received (classified) statistics on civilian casualties from an unnamed high-ranking US military official:

"The numbers were for Baghdad only, for a short period, during a relatively quiet time. Other hot spots, such as the Ramadi and Mosul areas, could prove worse."

http://www.globalresearch.ca/articles/CHO504C.html


In an essay sent to Human Rights Watch shortly before she died, Marla wrote: "A number is important not only to quantify the cost of the war, but, to me, each number is also a story of someone whose hopes, dreams and potential will never be realised, and who left behind a family."

[Serial Mom]

99. I forgot about that! Thanks

I knew there was something I was searching for in the brain but the dots didn't connect!

The military has only claimed not to keep stats on Iraqi civilian deaths!

[Ms. Clio]

108. I will always have deep dark suspicions about her death

More from the same article:

Marla Ruzicka had also documented over a two year period the use of illegal weapons by the US military including cluster bombs, as well "parachute bomblets", and bombs filled with tiny nails dropped in densely populated urban areas, etc.

In other words, Marla Ruzicka was in possession of sensitive information on US sponsored war crimes, which extended beyond the count of civilian casualties. She was fully aware of the historical significance of this data; there are indications that she planned to release this information.

[OmmmSweetOmmm]

116. See link to USA article in my post #114. They were ordered to stop

counting back in December 2003.

[OmmmSweetOmmm]

96. Almost a year ago the Guardian published an article that had it over

100,000. I think you're numbers are completely off. Do the research......

http://www.guardian.co.uk/Iraq/Story/0,2763,1338749,00.html

[WillyT]

115. My Experience With Folks Whose Academic Training And Intellectual Honesty

have been sullied by lack of 'hard data', and feel the need to tell everyone within ear-shot of it, almost always turn out to be folks fresh out of high-school or college. And while I admire their idealism, I long to chat with them in their cynical years when their idealism has been seasoned with skepticism and (hopefully) humor.

:shrug:

[Zenlitened]

119. Your sense of intellectual honesty? Now THAT'S funny!

:rofl: