Talk:Lancet surveys of Iraq War casualties/Archive 2

From Wikipedia, the free encyclopedia

This page is an archive. Please do not edit the contents of this page. Direct any additional comments to the current talk page.

2006 Survey

Mozzie: References have been provided for the authors' defense of their methodology. Lapakhin:You incorrectly state that Lancet 2 attributes all excess deaths to violence. Lancet 2 reports an increase of .6 in the death rate for the post-invasion era due to non-violent causes. This means roughly 92.3% of the post-invasion increase is attributed to violence. Lapakhin is also in error regarding the original survey. Lancet 1 did not explicitly provide a percentage for portion of excess deaths attributed to violence, it can be computed from the available data as 87.8%. Thus, the percentage of excess deaths due to violence is much less than that stated by Lapakhin. Lapakhin's substitute language unfairly diminishes the results of the Lancet 2 study in verifying the results of Lancet 1, although I will admit that my original language may be deemed as biased in favor of the study by indicating that the confirmation of the original is a fact rather than an interpretation. As a compromise, I am replacing the statement with a quote from Lancet 2, thus indicating that the conclusion is the author's interpretation of the data. I also question the appropriateness of the Steven Moore quote, which lacks any evidence of his authority in interpreting statistical surveys and would like to see more discussion on this point. Finally, I am including the 95% level chosen for the confidence intervals in reporting those intervals - an oversight from my original entry. -- NeoRedpill 4:33, 22 October 2006 (UTC)

- The language now looks fine. FWIW, my comment about Lancet 2 attributing all the excess deaths to violence was meant to apply to the first 18 months post-invasion, i.e. that period also covered by Lancet 1. I did actually include a link to a blog article which set this out in more detail, but that disappeared, in conformity with the 'no blogs' rule which is being enforced on this site. :o)--Lopakhin 16:24, 30 October 2006 (UTC)

I need a reference for the authors defence of the quality of their work.--Mozzie 00:32, 13 October 2006 (UTC)

I have changed the assertion that Lancet 2 'confirms' Lancet 1's excess death estimate to 'claims to confirm', because of the disjuncture, set out in the article to which I linked, between the causes of death claimed in the first survey - roughly 40% nonviolent, from memory - and those claimed for the same period in the second survey, namely, all due to violence. I am aware that choosing subgroups of a statistical set in this way increases the confidence interval; but the claim that Lancet 2 'confirms' Lancet 1 is a strong one (which was also made by Lancet 2's authors), and I feel it needs to be tempered by providing Wikipedia's readers with access to differing opinions on the matter.--Lopakhin 14:25, 17 October 2006 (UTC)

I am also puzzled by this line: 'Indeed the sources of bias in the study push the figure down.' Could the person who wrote it possibly explain it? Is the word 'bias' being used in its statistical sense, or is it the more general lay usage? And in either case, could you possibly explain to what you are referring?--Lopakhin 15:27, 17 October 2006 (UTC)

I assume: "The miscommunication that resulted in no clusters being interviewed in Duhuk and Muthanna resulted in our assuming that no excess deaths occurred in those provinces (with 5% of the population), which probably resulted in an underestimate of total deaths" http://www.thelancet.com/webfiles/images/journals/lancet/s0140673606694919.pdf page 7 Gzuckier 17:20, 17 October 2006 (UTC)

Okay. And how about this line: 'The second The Lancet study reported a death rate about 35% greater than that reported by the Iraq Body Count project over the past year, and slightly lower than the US Department of Defense has reported during the same period.'

I'm pretty sure this is wrong. The Lancet 2 death rate was much higher than both IBC and the DoD figures - that's the whole point, isn't it? I think the person who wrote this has misunderstood the report, and I shall delete it unless an explanation is provided.--Lopakhin 15:59, 19 October 2006 (UTC) - in fact, looking at it again, I'm pretty sure he/she has based this on a misunderstanding of Figure 4, which is reproduced at the top of this main article. So I shall delete it - please call me out on this if i'm wrong.--Lopakhin 16:24, 30 October 2006 (UTC)

Criticism section

In an article of this kind there is a natural and understandable tendency for editors to include their own pet theories or those they have read on some blog or other. This tendency if unchecked leads to the dwarfing of well founded principal concerns by less important matters, and a resulting imbalance in coverage. Accordingly I've removed or reworded a lot of the criticism section, in some cases trimming back perhaps legitimate criticism simply because it hasn't been adequately cited (this is in accord with our Wikipedia:Verifiability policy). We really shouldn't be citing blogs here, either, unless they are corroborated by mainstream media or the criticism has been acknowledged by the researchers themselves.

Tbere were unfortunately also some instances of original research in the criticism section. I have removed or altered wording that appeared to me to be the unsupported personal opinion of a Wikipedia editor. I have also removed a similar amount of weaseling ("pointing out", versus "insisting", "some say", etc). These edits also are in accordance with Wikipedia's policies.

Care should be exercised in restoring any of the removed sections, to ensure that the restorations where undertaken are properly sourced, accurate and neutral, and comply with all other Wikipedia policies. --Tony Sidaway 10:50, 16 October 2006 (UTC)

Included the full name of the ILCS study. Added quote from study stating their belief that the study underestimates war-related deaths. While this admission might somewhat blunt the IBC use of the study to criticize Lancet, in the interest of fairness I also included the mention by ILCS that Lancet I (Lancet II had not yet been completed) also fails to correct for the same condition that may cause an undercount, thereby maintaining the discrepancy between ILCS and Lancet I. It is also unlikely that the cause of undercounting cited here would begin to explain the discrepancy, and in fact, neither ILCS nor IBQ attempts to explain the discrepancy. --NeoRedpill 02:28, 21 November 2006 (UTC)

Conflicts of interest

I've removed this paragraph because, in its current form, it looks too close to investigative journalism or original research:

The authors of the 2004 study state in their 2006 paper, "We declare that we have no conflict of interest." According to The Lancet's conflict policy,^[1] "all authors must disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work." The Lancet also states that "conflicts can...occur because of personal relationships, academic competition, or intellectual passion. A conflict can be actual or potential..." Some have argued that at least two of the authors, as well as the editor of Lancet, Richard Horton, have clear conflicts of interest in the publication of this survey. Author Les Roberts has run for the US Congress on the Democrat ticket on a largely anti-war platform,^[2] and author Gilbert Burnham contributed financially to his campaign.^[3] Editor Horton, too, is an outspoken and fervent critic of the Iraq war, and has charged that the British government "prefers... the killing of children instead of the building of hospitals and schools."^[4]

If true this is legitimate criticism of the standards used by the authors for declaration of potential conflicts. However we need to report whether they have actually been criticised for this, rather than present our own research on the matter. This may seem odd, but it's a consequence of being an encyclopedia rather than a commentary website.

Again we see weaseling here. "Some" have argued, If they have done so (and not just on some blog or other) then they should be cited. --Tony Sidaway 11:05, 16 October 2006 (UTC)

Pure leftist propaganda

This "study" is pure pre-election propaganda, a/k/a The Big Lie, a staple strategy of the Left. Roberts et al. inflated the death toll by about a factor of twenty (20).

That 655,000 supposed death count works out to be an average of over 500 deaths per day, most of them civilians, during the 3.5 years since Operation Iraqi Freedom commenced. If anyone believes that nonsense, then I would suggest that they look at the endless news reports which give daily and weekly casualty counts, and compare them to Roberts' numbers. Even during the Iraqi elections the death counts did not spike anywhere near that high[1].

if you want to make your criticism stronger you should attack the low end of the 95% confidence interval (i.e. ~400k deaths). Any criticism of the low end will obviously apply to greater estimates. Of course you probably think the holocaust never happened or that only a few thousand jews were killed ... Funkyj 18:45, 28 June 2007 (UTC)

"Iraq body count" (IBC) is an anti-war web site that tries to estimate civilian deaths in Iraq. They estimate the numbers of civilian dead with the goal of driving home the high numbers in an effort to end the war. Their current claimed death count is between 43,546 and 48,343, with a midpoint number of 45,945 -- which is just 7% of what Roberts claims.

Even IBC ridicules the Roberts number[2]. What's more, IBC is frequently criticized for overstating the death count. For instance, they admit that they include in their numbers crime victims whose violent deaths were unrelated to the war and the ongoing political and religious struggles. They estimate that such crime victims represent at least 36% of the total. If you subtract 36% from their midpoint number, the result is less than 30,000 -- which is less than 5% of what the Johns Hopkins / Lancet article claims.

On a bad day in Iraq, a few dozen people die from terrorist attacks, sectarian violence, etc. -- not five hundred. E.g., here's a January article in the UK's left-wing Guardian newspaper, which reports that on "the worst day of violence since last month's elections" there were "at least 55 people killed." http://www.guardian.co.uk/Iraq/Story/0,2763,1677843,00.html

If 55 died on an unusually bad day, the 'average' death toll cannot possibly be 500.

The Big Lie is that Operation Iraqi Freedom has resulted in a net loss of life. The truth is just the opposite.

During the 20 years that Saddam was in power, he was directly responsible for, according to most estimates, approximately two million violent deaths, most of them Iraqis. (Estimates vary, but none are under one million.) The largest number were war dead, in wars of conquest that Saddam launched against neighboring countries. Others contributors were his mass murders of Kurds and Shiites, genocide against Marsh Arabs, etc.: [3] [4] [5] [6] [www.freerepublic.com/focus/f-news/1480609/posts] [7] [8] [9] [10] [11] [12] [13] [14] [15]

Two million violent deaths works out to be an average of about 275 violent deaths per day, for twenty years -- which is about ten times the average daily violent death rate since Saddam was overthrown.

That probably explains why Roberts et al. thought it necessary to inflate their numbers by a factor of twenty. No lie is too big if it helps to discredit the hated Republicans on the eve of an election.

Saddam's genocide of the Marsh Arabs is described in this January 2003 Human Rights Watch briefing paper: http://www.hrw.org/backgrounder/mena/marsharabs1.htm

Here are some quotes from that article:

  "Numbering some 250,000 people as recently as 1991, the Marsh
  Arabs today are believed to number fewer than 40,000 in their
  ancestral homeland.  Many have been arrested, 'disappeared,' or
  executed; most have become refugees abroad or are internally
  displaced in Iraq as a result of Iraqi oppression.  The
  population and culture of the Marsh Arabs, who have resided
  continuously in the marshlands for more than 5,000 years, are
  being eradicated."

  "The waves of arrests were soon followed by reports of mass
  summary executions.  Among the reports received by Human Rights
  Watch at the time was one incident involving the execution of
  some 2,500 villagers in August 1992.  The victims, among them
  women and children, were rounded up in the marshes of
  al-Chibayish (west of al-Qurna) together with captured fighters
  of the opposition SCIRI.  According to testimony obtained by
  Human Rights Watch, including that of a survivor, they were
  taken to an army camp in northern Iraq, where they were executed
  over a period of about two weeks.34 The UN special rapporteur on
  Iraq detailed similar reports he had received."

Nor were all Saddam's victims human. Here are some quotes from the HRQ briefing paper, about Saddam's intentional environmental destruction:

  "Unlike other Shi'a, Marsh Arabs were also subjected to a
  government-engineered environmental catastrophe.  ...  Starting
  in 1991, in part to facilitate entry into the area by the armed
  forces, authorities built a series of dams, dikes, and canals
  aimed at preventing the waters of the Tigris and the Euphrates
  from flowing into the marshes.  The result, less than a decade
  later, was the destruction of the Middle East's largest wetland
  ecosystem."

  "An environmental study carried out by the U.N. Environment Program
  (UNEP) in 2001, based on previously unseen satellite images,
  revealed the extent of the devastation of the marshland region.
  The satellite images, taken in 1992 and 2000 by the U.S.'s
  National Aeronautics and Space Administration (NASA), confirmed
  the destruction of around 90 percent of the marshlands, qualifying
  it as 'one of the world's greatest environmental disasters.' ...
  UNEP concluded that...  'analysis of satellite imagery has shown
  that the marshland ecosystem had collapsed by 2000.'"

And don't forget Saddam's previous eco-terrorism, in 1991, when he ordered the intentional torching of hundreds of Kuwaiti oil wells, and intentionally dumped 20 times as much raw crude oil into the Persian Gulf as had spilled in the 1989 Exxon Valdez disaster in Prince William Sound. NCdave 12:48, 20 October 2006 (UTC)

This study was published in Lancet. What Lancet publishes is not always the truth but it is very rarely complete crap. The journal is quite possibly the most prestigious in the world and has quite a reputation to uphold, it doesn't let a lot slip through. Additionally they were extra careful with this study because of the potential it had to put a lot of egg on their faces. To your claim that the study was published at the time it was published to influence the results of the election I say even if it's true so what? It shouldn't impact on our evaluation of the the paper either way which should stand or fall on it's merit's, not on the political beliefs of it's authors. Your citation of the Iraq Body Count study reveals your ignorance, the methodology used by this study is counting deaths reported in the media and only a fraction of deaths in conflict's like this are reported in the media, all the IBC can do is show the bare minimum. You have accused those who conducted the study of outright academic fraud, something that would require an unfeasibly large conspiracy considering the number of people who conducted the study and which would put the authors careers at risk. Your reading of media reports reveals that you don't understand that just because a death isn't reported doesn't mean it didn't happen.

(The above anonymous comment was from IP address 220.236.55.37; here's my reply... -NCdave)

Lancet is a medical journal, prestigious in the field of medicine, but in this case they have stretched their scope to include a bogus political "study" which purports to use a variant of epidemiological methodology to estimate numbers which can much better simply be counted. What's more, the result of their "estimate" is about 20 times what simple counting gives, which proves that the study is wrong, yet they published it anyhow. You bet they have egg on their faces!

It is very strange to read you argue that for the study to be fraudulant would require an "unfeasibly large conspiracy." Actually, Just a small number of people are responsible for perpetrating this fraud. The Roberts study and supplement had just five principle authors, and conducted surveys at just 47 locations. Adding the very small number of Lancet editors and reviewers who approved it certainly does not make it an "unfeasably large conspiracy."

However, for the Roberts study to be accurate would require a truly massive conspiracy, with (quoting from the IBC) (1) "incompetence and/or fraud on a truly massive scale by Iraqi officials in hospitals and ministries, on a local, regional and national level, perfectly coordinated from the moment the occupation began," plus (2) "bizarre and self-destructive behaviour on the part of all but a small minority of 800,000 injured, mostly non-combatant, Iraqis," plus (3) "the utter failure of local or external agencies to notice and respond to a decimation of the adult male population in key urban areas," plus (4) "an abject failure of the media, Iraqi as well as international, to observe that Coalition-caused events of the scale they reported during the three-week invasion in 2003 have been occurring every month for over a year."

Now that's what I call an "unfeasibly large conspiracy!"

As for the IBC, their POV biases them to produce the largest possible numbers which they can justify, and their methodology for "documenting" deaths supports that bias. The IBC counts not just deaths due to the conflict itself, but all violent deaths reported in any two or more press outlets, taken from a very long list[16] of "acceptable" sources (most of them press outlets). Their criteria[17] for acceptability of sources do not include accuracy and reliability, and the press outlets that IBC deems acceptable are mostly liberal or leftist (even including Al Jazeera!). Many of these press outlets have no newsgathering presence in Iraq, but just report information obtained from other sources, so the IBC requirement that a death be reported in two or more places is no assurance that it actually occurred, since the information might have originated from a single source of unknown veracity.

What's more, the IBC admits that about 36% of the deaths that it includes in its figures are not due to the conflict, but are actually just common crime. The excuse that the IBC uses for including those deaths is that crime in Iraq must be due to the breakdown in law and order resulting from the conflict, making the crime victims (to use Roberts' terminology) "excess deaths." That is a flimsy excuse, considering that the so-called "law and order" which preceded the conflict included numberous examples of government-initiated mass murder and genocide, plus a couple of very bloody wars of aggression, and the IBC takes no account of those excess deaths.

Additionally, for another 11% of the deaths which IBC includes, the details are too sketchy to guess as to whether the cause was the conflict or simple crime.

The fact is that even though the IBC's POV and methodology are strongly biased to over-estimate the death toll, even the IBC can't stomach statistical fraud on the scale of the Roberts study.

NCdave 16:47, 23 October 2006 (UTC)

This page isn't for discussing the merits of the study, but rather for discussing the article about the study. Accordingly, I'll be removing this section from the talk page in a few days' time. --Tony Sidaway 12:28, 21 October 2006 (UTC)

Don't you dare delete it, Tony! Obviously the merits of the study are central to discussion of the article about the study, and they should be central to the article, itself. The article is POV-biased to the extent that it reflects misplaced credulity in a transparently bogus, politically-motivated study. Here's a good example about how such material should be handled: http://en.wikipedia.org/wiki/Holocaust_denial.

NCdave 16:47, 23 October 2006 (UTC)

Note that, through the magic of the mathematical integral over time, the "vast difference" between the IBC and Hopkins estimates in deaths amounts to a fairly constant factor of 1.25 in the Hopkins death rates vs. the IBC death rates.[18] Who here thinks it would require a "a truly massive conspiracy" to have 20% of the deaths in Iraq in its current state go officially unnoted? Hands? Note also from that graph that at present, the DOD estimate of the Iraqi death rate is in fact greater than the Hopkins estimate. Why isn't anybody accusing the DOD of "pure pre-election propaganda, a/k/a The Big Lie, a staple strategy of the Left"? Get on the stick people, or the Commies will win!! (see below)Gzuckier 19:06, 23 October 2006 (UTC)

Actually, the Roberts et al. (Johns Hopkins / Lancet) estimate of the death rate is 1326% (not 20%) higher than IBC's count, and that's using the IBC number which includes deaths that are not directly related to the conflict. If you subtract from the IBC count the 36% which IBC admits are due to simple criminal activity, and a like portion of the 11% for which they can't guess a cause, then the Roberts estimate of the death rate is 2294% (not 20%) higher than IBC's count. NCdave 09:24, 24 October 2006 (UTC)

You're right of course, I spent so much time adding up the IBC numbers by hand to see if they checked with those in the graph (they do) i was in too much of a hurry to notice that the two Y axis scales were not equivalent, and that the Hopkins study used one scale and the others the other scale. A note to that effect in the legend of the figure would have been nice...... Anyway, much as I'd like to erase my stupid comment, I instead strikeouted it and leave it as a reminder to Wikipedians, myself included, think more than twice before posting once. Gzuckier 18:48, 25 October 2006 (UTC)

NCDave: I have checked out some of these websites you have referenced as "sources" for your arguments. Like the arguments themselves, they simply do not measure up. Your 2 million number is completely unverifiable and includes, for example, deaths on both sides of the Iran-Iraq War, whereas the Lancet studies are confined to Iraqi deaths. You mention that some (anonymous critics) accuse IBC of overcounting, but I have heard (and find much more credible) arguments that IBC undercounts since it only counts deaths reported in specific publications. Even in their own FAQ, IBC states "It is likely that many if not most civilian casualties will go unreported by the media." [19] You also state that the Lancet studies would represent a reduction in death rates compared to the rate over the course of Hussein's reign, but by your own debatable tallies, the 275 deaths/day is well short of the Lancet rate which you state at the beginning of your rant is well over 500 per day. Later you seem to be comparing the presumed 275 per day prewar with the IBC count, which is irrelevant to any discussion of the Lancet study.

You reference an article written by John H Burns in the New York Times. The article gives a figure for deaths attributable to Hussein around 1/2 that you suggest. He states that "Doing the arithmetic is an imprecise venture" and includes Iraqis killed in the war against Iran and deaths attributed to US bombings. The Nico Price AP article providing an estimate of Baghdad residents executed used a similar -- although apparently less stringent -- methodology to the Lancet article that you are criticizing, yet you use it as support. The HRW report states that most of the Marsh Arabs became refugees, not execution victims, and the comments on environmental damage are also irrelvant to the discussion of the Lancet studies. In fact, none of your referenced articles that I looked at even mentioned the Lancet studies.

If the point of all your argument is that Saddam Hussein is a wretch of a human being who committed monstrous attrocities, I doubt that you will find many here that will argue against that point, but that argument simply does not have a place in the discussion of the accuracy of the Lancet study, for which you have provided no negative information. NeoRedpill 03:04, 16 November 2006 (UTC)

IBC / BanyonTree's edit

re: [20]

BayonTree asks, "why does IBC get a long, and incorrect, description? -Move it to their article. Don't call op-eds articles."

1) My description is neither long nor incorrect. Do you dispute that the IBC site has the purpose of driving home the high numbers to bolster anti-war sentiment? Or what?

2) They need a brief description here because their purpose leads to a bias toward reporting high numbers, which means that when they criticize the Roberts/Lancet numbers as being redicuously high then their criticism obviously is not due to bias. Is that not obvious?

If anyone is silly enough to believe the rediculous Roberts/Lancet propaganda, the fact that IBC (on the Left) and Bush (on the Right) are in substantial agreement about the death toll should give the silly people pause.

NCdave 14:29, 20 October 2006 (UTC)

NCdave, if I may say so, you appear to be editing this article, and using this talk page, with the intention of promoting your personal point of view. Please don't do that. It is absolutely forbidden by Wikipedia policies. --Tony Sidaway 12:29, 21 October 2006 (UTC)

Tony, that's rubbish. My edits are entirely for accuracy and balance. Please try to keep your political POV from coloring your opinions of other contributors, and at least please try to refrain from personal attacks. NCdave 15:21, 23 October 2006 (UTC)

You can't seriously be postulating Bush's "estimate" as having any sort of correlation with reality? Unless you know more about how he got it than anybody else does. Gzuckier 18:55, 25 October 2006 (UTC)

President Bush is an honest and honorable man, and his estimate is an honest one, actually probably a bit on the pesimistic side, because it is similar to the IBC numbers, which count many deaths that they admit are not directly due to either war or civil unrest.

To believe the Roberts numbers, however, requires that one accept (for example) the absurd proposition that a conspiracy of the vast majority of coroners and morgue workers routinely issue death certificates and then discard all record of having done so; that, in fact, that they discard all record of the vast majority of the death certificates that they issue. You can't seriously be suggesting that, can you? NCdave 00:15, 7 November 2006 (UTC)

"President Bush is an honest and honorable man, and his estimate is an honest one". Well there's no need for you to get sarcastic. That aside, just how much epidemiological training has Bush had, anyway? How do you handle the situation where honest and honorable Bush not only stood by Roberts' estimate of deaths in the Congo war, from the same methodology, but used it as evidence for sanctions?

Speaking of which, is there any other recorded civil war which has resulted in an annual fatality rate of only 0.04%? (30,000 out of 26,000,000, 0.12% total, over 3 years) In a similar three years, Bosnia had 100,000 fatalities, with 1/5 the total population of Iraq. That's a total of 2% over three years (0.7% annually), compared to Bush's "estimate" of 0.12% over three years, or the Roberts paper's estimate of 625,000/26,000,000 or 2.3% over 3 years. Where were all the people with their "gut level incredulity" over the "impossibly huge" 2% death rate in Bosnia, if they now ridicule the "impossibly huge" 2.3% death rate in Iraq? How about Roberts' estimate of the Congo, which the Bush administration embraced as mentioned above: 3,800,000 deaths out of 56,000,000 people over 5 years, or 1.3% per year. Couldn't be that there wasn't any self-interest in minimizing the death toll in Bosnia or Congo, but there is in Iraq, could it? Shouldn't the stunningly low death rate in Iraq, less than 10% that of other civil wars, be trumpeted to the skies if it's true? Lucky Iraq, with its peaceful war; due to the brotherly tolerance between Sunni and Shiite, the order maintained by the Coalition Authority, the well-functioning health and medical infrastructure, the pinpoint precision of our aerial warfare, and the general humanitarian restraint shown by all parties, I suppose. Gzuckier 20:52, 7 November 2006 (UTC)

Gzuckier: 'Speaking of which, is there any other recorded civil war which has resulted in an annual fatality rate of only 0.04%?'

Is it worth at all mentioning here that, according to a recent opinion poll carried out for the BBC and other organisations, the majority of Iraqis don't believe there is a civil war taking place there? See p.22. --Lopakhin 10:13, 21 April 2007 (UTC)

That said, how is it the IBC numbers that are so all of a sudden reliable (unlike their being hugely inflated before Roberts 2004) have minimum and maximum estimates? Couldn't they determine whether something was one body or two? I know, they say where there are differing reports they give max and min, but that just pushes the problem downstream. If this is supposed to be a cold hard actual count of deaths, how do two reports give different numbers? Is it possible that they are just ..... guesses by people on the scene? Who are not actually counting the number of bodies in the collapsed apartment building after all? And all this poop about these being hard, reliable numbers is just the emperor's same old clothes? Gzuckier 18:55, 25 October 2006 (UTC)

As I explained above, the IBC approach is designed to overstate the death toll. They don't examine death certificates, they don't try to identify the victims or verify the deaths, they just look at (mostly media) reports from other sources. If they find a claimed violent death mentioned in two or more media outlets, they go ahead and count it, even if the supposed victims are unidentified, even if there is no official record of the death, and even if neither of the media outlets has any newsgathering presence in Iraq. IBC even counts deaths which they can tell were NOT caused by the war and civil unrest; they justify counting such deaths by assuming that all violent crime is caused by a breakdown in law and order because of the war, which is an absurd assumption. NCdave 00:15, 7 November 2006 (UTC)

Can you demonstrate how it is impossible that there would be any deaths that were not reported to the IBC? Or even unlikely? I find it unlikely that there wouldn't be a LOT of deaths that were not reported to the IBC, and vaguely inflating the deaths they do see can hardly be expected to bring the estimate back to accuracy. Gzuckier 20:52, 7 November 2006 (UTC)

Trying to sort out criticisms to the Oct 2006 report

Hi, i had a first go at trying to sort out the different types of criticisms against the Oct 2006 report. Wikipedia is about external knowledge, and the type of knowledge represented by a politician criticising a scientist's work is different to that of another scientist criticising a scientist's work, so i think it's useful to separate them out.

The section of "lancet authors' responses" needs a bit of cleaning up. i didn't want to interfere too much in what other people wrote. However, i did remove the weasel word "however,", which didn't make sense after the reorganisation anyway.

And i think what would be really useful is if the lancet authors respond directly to the IBC criticisms - and then we can write a summary of that here. Boud 22:39, 27 October 2006 (UTC)

They have now published a second article in the Lancet, which responds to the IBC criticisms

Johnbibby 16:04, 29 October 2006 (UTC)

Sampling section

A group of us are planning to develop this section - sp please feel free to improve it, but do not go too mad about removing stuff yet please. (It's NOT original research as suggested by an anonymous editor)

Johnbibby 16:04, 29 October 2006 (UTC)

Please log in as a named contributor so we can have a reasonable discussion - I do not agree with what you say.

Johnbibby 17:59, 29 October 2006 (UTC)

Hi, not trying to discourage you, I felt the same when I had stuff removed for the NOR rule, but it makes sense when you read the policy page.

http://en.wikipedia.org/wiki/Wikipedia:No_original_research

"This policy in a nutshell: Articles may not contain any unpublished arguments, ideas, data, or theories; or any unpublished analysis or synthesis of published arguments, ideas, data, or theories that serves to advance a position."

As far as I'm aware none of the bullet points that I removed have been published, though I had seen the 'main street bias' criticism which is why i added the [citation needed] tag.

Specifically, I don't believe that 'political bias' is a valid criticism of this study or any other in fact. The validity of a study has nothing to do with the political views of the authors, left-wing, right-wing etc. etc. If a study is flawed then it should be able to be shown to be flawed without reference to the beliefs of the authors.

Also I don't think that comments such as "the availability of death certificates for 92% of the cases seems suspiciously high" are appropriate in the article unless sourced.

I'm not going to remove it again, but I do think if it's being developed then that should be done off the main article, which is why I suggested you develop the stub on this discussion page. Not personally sure on the policy on this, perhaps other editors could advise?

Mark M

Morgue stats the year before Iraq War - compared to following years.

Civilian deaths attributable to insurgent or military action in Iraq, and also to increased criminal violence. For the period between January 1, 2003 and July 20, 2006 as recorded by the Iraq Body Count project. Many of these type of civilian deaths are not reported. Other methods of estimating civilian deaths come up with much higher numbers. See also: Casualties of the conflict in Iraq since 2003.

--Timeshifter 23:30, 31 October 2006 (UTC). Several articles and websites refer to morgue stats in reference to Lancet and IBC casualty counts, excess deaths due to the Iraq War, etc.. Some use these references:

See these articles from the Associated Press (AP) and Iraq Body Count project (IBC):

• AP
• IBC

Google search of IBC site for the word "morgue":

• http://www.google.com/search?q=site%3Awww.iraqbodycount.net+morgue

IBC only records civilian death statistics from morgues if they are reported by English-language media.

May 23, 2004 AP article. Various quotes:

The death toll recorded by the Baghdad morgue was an average of 357 violent deaths each month from May [2003] through April [2004]. That contrasts with an average of 14 a month for 2002, Hassan's documents showed. ...

The figure does not include most people killed in big terrorist bombings, Hassan said. The cause of death in such cases is obvious so bodies are usually not taken to the morgue, but given directly to victims' families. Also, the bodies of killed fighters from groups like the al-Mahdi Army are rarely taken to morgues. ...

The death toll recorded by the Baghdad morgue was an average of 357 violent deaths each month from May through April. That contrasts with an average of 14 a month for 2002, Hassan's documents showed.

The toll translates into an annual homicide rate of about 76 killings for every 100,000 people.

By comparison, Bogota, Colombia, reported 39 homicides per 100,000 people in 2002, while New York City had about 7.5 per 100,000 last year. Iraq's neighbor Jordan, a country with a population a little less than Baghdad's, recorded about 2.4 homicides per 100,000 in 2003.

Other Iraqi morgues visited by AP reporters also reported big increases in violent deaths.

---

The May 23, 2004 AP article also reports: "Morgue records do not document the circumstances surrounding the 4,279 [Baghdad] deaths - whether killed by insurgents, occupation forces, criminals or others. The records list only the cause of a death, such as gunshot or explosion, Hassan said. It is the police's responsibility to determine why a person dies. But al-Nouri, the official at the Interior Ministry, which oversees police, said the agency lacks the resources to investigate all killings or keep track of causes of death."

---

The AP stats are for VIOLENT deaths only.

"...the morgue figures, which exclude trauma deaths from accidents like car wrecks and falls,..."

"But the AP survey of morgues in Baghdad and the provinces of Karbala, Kirkuk and Tikrit found 5,558 violent deaths recorded from May 1, 2003, when President Bush declared an end to major combat operations, to April 30 [2004]."

---

The Lancet survey also counted excess deaths in general since the Iraq War. The Lancet study counts ALL war-related Iraqi deaths in the broadest sense of the term. Both civilian and non-civilian. Including deaths due to the infrastructure degradation. Including deaths due to the increase in disease and lack of healthcare. Deaths due to lack of food, water, heat, airconditioning, shelter, sewage, electricity, you-name-it. Deaths due to inadequate, barely-functioning hospitals, medical clinics, etc.. Or not functioning at all in many cases.--Timeshifter 23:30, 31 October 2006 (UTC)