Talk:Casualties of the Iraq War/Archive 3

PLOS Medicine Survey. 276,000 violent deaths out of 460,000 excess deaths
I corrected the article. It said 132,000 war-related violent deaths reported by the PLOS Medicine Survey. This is obviously wrong since it is less than the death tolls from body counts.


 * http://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1001533
 * http://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.1001533&type=printable (pdf).

"Methods and Findings" section in the abstract at the beginning of the above-linked PLOS Medicine Survey (emphasis added):

"approximately 405,000 (95% uncertainty interval 48,000–751,000) excess deaths attributable to the conflict. ... We estimate that more than 60% of excess deaths were directly attributable to violence, with the rest associated with the collapse of infrastructure and other indirect, but war-related, causes. ... We used secondary sources to estimate rates of death among emigrants. Those estimates suggest we missed at least 55,000 deaths that would have been reported by households had the households remained behind in Iraq, but which instead had migrated away."

405,000 plus 55,000 equals 460,000 excess deaths.

60% of 460,000 excess deaths equals 276,000 violent deaths.

Figure 2 on page 7 of the PDF makes it clear that the majority of excess deaths were due to violence (emphasis added):

"Figure 2. Estimates of numbers of deaths per week in Iraq for 2-y intervals, 2001–2011, by cause as reported by households in the University Collaborative Iraq Mortality Study. ... War-related, but not violent, deaths above the normal baseline are in the salmon-colored area. War-related violent deaths are portrayed in red." --Timeshifter (talk) 16:48, 23 March 2018 (UTC)

iraqbodycount.org home page says 288,000 total violent deaths including combatants
See:
 * https://www.iraqbodycount.org

"Documented civilian deaths from violence: 181,455 – 203,561"

"Total violent deaths including combatants: 288,000" --Timeshifter (talk) 16:14, 22 May 2018 (UTC)

Tag-team edit-warring again
Two users return to this page on a regular basis to edit-war in fringe content and text, and remove reliably sourced text. These editors are now back, and are removing academic studies ("usually the most reliable sources" per Wikipedia's RS policy) and are removing abundantly sourced text (see endless discussions above) that clearly describes some casualty estimates as being non-credible. Snooganssnoogans (talk) 22:56, 9 July 2018 (UTC)
 * The gall to say only "one author". It's been repeatedly pointed out to this editor that this is egregiously false. At this point, this user is just engaging in vandalism. Snooganssnoogans (talk) 00:44, 10 July 2018 (UTC)


 * You've single-handedly reverted more than anybody here, forcing through your preferred version last time this came up only because of your increased willingness to hit the "revert" button. What I object to in your preferred version is that you have used your own personal judgment to label certain academic studies "credible" and others "not credible." You have decided that one particular academic, Michael Spagat, is correct, but that the much more heavily cited authors of the Lancet papers are not. That's original research, and it's not allowed here. Michael Spagat's views are already amply discussed is this Wiki article (he's given way too much weight, in my opinion). You don't have to elevate his particular views to "the Truth" in the lede, in contrast to the other studies whose results conflict with his. -Thucydides411 (talk) 01:39, 10 July 2018 (UTC)
 * Your claim that this is only Spagat is patently false. This has been repeatedly pointed out to you, and even the most cursory skim of sources and text would make it clear. I cannot rehash this with you every 4 months, and it serves zero purpose to debate this with you given your past behavior. Snooganssnoogans (talk) 10:50, 10 July 2018 (UTC)
 * Actually, Thucydides411 is right. The Lancet survey (the Lancet has impact factor 47.8) is cited by 324 publications, while Daponte's work criticizing the Lancet survey is only cited 5 times (so it's hardly significant enough to even mention).  Spagat's 2010 article in Journal of Defence and Peace Economics (impact factor 1.308) is only cited 8 times according to Scopus, the same measure used for the 324 of the Lancet survey.  So Spagat is also hardly significant enough to even mention.  These critics of the Lancet survey are given far too much space in this article.  Jrheller1 (talk) 16:38, 10 July 2018 (UTC)
 * Thank you . I have twice now begun the project of reviewing the number of times the Lancet papers are cited (versus Spagat), and also reviewing the impact of the citing papers, and the nature of their references to Lancet or Spagat (positive, neutral, negative). The summary is that the Lancet papers are standard in the field, overwhelmingly positively cited in textbooks and reviews, and that's not surprising given the methodology. Your assessment of Spagat is dead on. The problem however is that it takes a monumental amount of time and work to review just how influential those Lancet papers have been, since it means going through dozens / hundreds of publications. By contrast looking up every Spagat reference is a piece of cake.
 * From a purely theoretical perspective I don't mind that Snoogans disagrees: I think that with good faith effort and patience on all sides, it can be clearly demonstrated that the critique of the Lancet conclusions is fringe, academically. However, the incredible certainty with which this fringe view has been advanced sometimes stands in comical contrast with the weakness of the argument. Until Snoogans is actually presented with a more comprehensive picture of the citation structure you reference however, Jrheller, it's not wholly their fault that they've come to this conclusion. -Darouet (talk) 17:07, 10 July 2018 (UTC)
 * "it can be clearly demonstrated that the critique of the Lancet conclusions is fringe, academically". Yes, please bring the thorough lit review that you claim to have been working on. I'm very excited to see what you've been cobbling together for months on end. So far you haven't brought a single RS to these discussions AFAIK despite coming here to edit-war on a regular basis. Snooganssnoogans (talk) 17:29, 10 July 2018 (UTC)
 * No, you're incorrect and you should familiarize yourself with the references and text in the article, and the previous discussions before you chip in. I cannot discuss this again and again every four months just because tag-team edit-warriors like Darouet and Thucydides put their heads in the sand and desperately want to bring their fringe nonsense into this article. You mention two papers that critique the Lancet study when they are two drops in a bucket of papers that do so. Thucydides is also blatantly lying when he says that only Spagat has criticized the Lancet study (and it's not the first that this editor has done so), so your assertion "Thucydides is right" reflects extremely poorly on yourself. Edit: I briefly thought "should I comment on the ridiculous citation number thing?" but then I remembered that I did ages ago when discussing this with one of the two tag-team edit-warriors. The short story is of course that controversial and shoddy studies get a lot of cites, as shown by Wakefield's autism-MMR study in the Lancet (2700 cites) and LaCour and Green's 2014 Science study which was debunked months after publication but still has 77 cites. Are the cites an indication of quality in the case of studies that are controversial and widely described as shoddy? Of course not. Snooganssnoogans (talk) 17:24, 10 July 2018 (UTC)
 * Aaaaaand Jrheller1 starts by removing Andrew Gelman's assessment of the Lancet study. The absurd comment by the user now makes sense. One of the most renowned statisticians in the world described the Lancet study as seriously flawed, so of course the content must be scrubbed from existence. Glad to know where Jrheller1 stands. Snooganssnoogans (talk) 18:40, 10 July 2018 (UTC)
 * The first Gelman source is literally an "interview with Michael Spagat". The second Gelman source is a newspaper article (Washington Post) written by Gelman that is heavily based on Spagat's work.  There is already far too much reliance on Spagat and his collaborators in this article. Jrheller1 (talk) 22:19, 10 July 2018 (UTC)
 * You can't argue that the Lancet study is credible while scrubbing expert content describing the Lancet study as seriously flawed. The gall! Snooganssnoogans (talk) 05:34, 11 July 2018 (UTC)
 * The only researchers seriously questioning the validity of the Lancet survey are Spagat (and his collaborators including Gelman) and Daponte. Their attacks on the Lancet survey are published in low impact journals and have very few citations, so they should not be given that much weight.  If the Lancet survey had really been "discredited", the Lancet would have retracted the article. Jrheller1 (talk) 22:19, 10 July 2018 (UTC)
 * There's a difference between retraction and being found to be seriously flawed and non-credible. If you're familiar with scientific publishing you'd know that, but apparently aren't. Your claims about Spagat's studies are false, and your claims mirror those made by Darouet and Thucydides in past discussions. I'm not gonna rehash them, but suffice it to say that Spagat's studies have won awards and have been described by experts as thoroughly discrediting the Lancet study. Snooganssnoogans (talk) 05:34, 11 July 2018 (UTC)
 * The award Spagat's paper in Journal of Peace Research (impact factor 2.153) won was for the best paper of the year in that journal. This is a paper with 13 citations showing in "Web of Science" (with many of these being later papers by Spagat and his collaborators).  The fact that a paper with so few citations was the best paper of the year for that journal only shows that this is not a very good quality journal. Jrheller1 (talk) 06:17, 11 July 2018 (UTC)
 * You're only discrediting yourself with these comments. First, you delete Andrew Gelman's assessment, and now you shit on the Journal of Peace Research. Snooganssnoogans (talk) 10:19, 11 July 2018 (UTC)

You're pushing through this text, despite it clearly not having consensus here. You're claiming that it's "long-standing," but the only reason it's actually in the article is because when you introduced it originally, in February, you reverted a dozen times, over the objections of multiple editors, to force it through. Please stop edit warring, and work with me and other editors here to achieve a consensus text. -Thucydides411 (talk) 00:32, 12 July 2018 (UTC)
 * I don't know the history of the "long-standing" text, but I agree that this is edit warring. I also agree with above comments during this discussion that this places too much weight on minor articles disputing the Lancet findings. Based on what I've seen in other similar articles about civilian casualties, Lancet is a very well regarded source for this. Please stop edit warring and try to reach a consensus with editors. Seraphim System  ( talk ) 05:54, 13 July 2018 (UTC)
 * The Lancet also published Wakefield's fraudulent autism-MMR study. Then when other academic publications described that study as non-credible, Wikipedia reflected those sources and proceeded to describe Wakefield's study as non-credible. That's how this works. We don't describe studies as credible that are widely seen as flawed and non-credible. We go by reliable sources. (And for what its worth, this has been discussed in depth on this talk page for almost a year, and I'm so far THE ONLY person to have actually brought academic assessments and reliable sources to this debate). Snooganssnoogans (talk) 06:00, 13 July 2018 (UTC)
 * If you think Lancet is not a RS, I recommend posting on RS/n. As far as I know Lancet is considered a reliable source. Seraphim System  ( talk ) 06:02, 13 July 2018 (UTC)
 * The dispute is not over whether the Lancet is RS. The dispute is over whether this particular Lancet study should be described as credible, and a large number of recognized experts and peer-reviewed studies say it isn't. If you don't know what this content dispute is about, you should read everything that's been posted on this talk page for the last year rather than shoot from the hip and talk nonsense. Snooganssnoogans (talk) 06:09, 13 July 2018 (UTC)
 * I don't see any evidence to back up what you're saying, only Gelman, and I agree that the current version of your edit puts too much weight on one primary study. What is the secondary source for Gelman? Seraphim System  ( talk ) 06:26, 13 July 2018 (UTC)
 * There are at least a dozen academic sources cited in the body to substantiate the text. I can't lead you along like a child on this when it's in the article for everyone to see. The fact that you're deleting the assessment by one of the foremost statisticians in the world does not exactly inspire much faith that you're actually serious about contributing to this Wikipedia article and that my time would be well-spent going through each source. Snooganssnoogans (talk) 10:56, 13 July 2018 (UTC)
 * Repeating my comment from above: the only reason I deleted the Gelman sources was because they are just more Spagat. The first source was Gelman interviewing Spagat, the second source was a newspaper article by Gelman based largely on the work of Spagat. Jrheller1 (talk) 15:58, 13 July 2018 (UTC)
 * Please clarify what point you're trying to make. You deleted the assessment by one of the world's most prominent statisticians because the statistician was deceived by Spagat? Gelman, one of the world's foremost statisticians, is incapable of providing an expert evaluation on this matter because he's an idiot and/or somehow easily swayed? Snooganssnoogans (talk) 16:20, 13 July 2018 (UTC)
 * The point is that somebody has turned this article into a promotional website for Michael Spagat. This is a serious problem with the article that needs to be fixed. Jrheller1 (talk) 17:12, 13 July 2018 (UTC)
 * Nonsense. Snooganssnoogans (talk) 18:48, 13 July 2018 (UTC)
 * If you are serious about reaching a consensus I would suggest trying to be WP:CIVIL with editors. I didn't check the sources in the article, instead I checked various databases before commenting, and what I saw does not corroborate your arguments on the talk page. This already seems to be adequately covered in the article, but the language you are trying to edit war into the lede does not seem to be summarizing the article text. Seraphim System  ( talk ) 16:15, 13 July 2018 (UTC)
 * The language absolutely summarizes the body of the article. And not only did you not check the sources in the article, but you took the egregious step to actually delete peer-reviewed academic research and delete the assessment by one of the world's foremost statisticians. Snooganssnoogans (talk) 16:20, 13 July 2018 (UTC)

I think it's very illuminating that almost a year ago, I was disputing with Darouet and Thucydides on this very subject, and during the course of those discussions I added more than a dozen peer-reviewed academic sources to the article (in addition to pre-existing content) to substantiate my arguments. That didn't change their mind for a second, and if I recall correctly, they tried to delete all or most of the sourcing. And now, almost a year later, I add a dozen more high-quality academic sources (in addition to the already abundant high-quality sources in the article) and the editor Jrheller1 proceeds to remove all of them. Snooganssnoogans (talk) 16:41, 14 July 2018 (UTC)

For the record, Snooganssnoogans is correct that his additions have been supported by several editors in previous discussions (including and myself), that counting citations is often a misleading metric to assess academic consensus, and that the lead author of the 2006 Lancet study (which failed replication by the IFHS) no longer stands by its estimate of excess deaths.TheTimesAreAChanging (talk) 17:32, 14 July 2018 (UTC)
 * That's fine, but it doesn't excuse edit warring. I don't support the changes to the lede because they seem to be pushing a POV. I've checked outside sources, and there are other studies besides the Lancet study for the high estimates. The lower estimates have been criticized as well, but the changes refer to them as "credible" - that is perhaps more of an issue, when all the estimates have been criticized. What makes them credible? Perhaps most importantly, the proposed wording in the lede has challenged and there is currently no consensus for it - it is difficult to discuss changes to the lede when an editor is convinced that his edit warring is justified to defend a previous consensus (that has now been challenged by multiple editors). Seraphim System  ( talk ) 17:43, 14 July 2018 (UTC)
 * Please tell us about the outside sources that you claim to have been reading. So far you haven't contributed a single sentence to the article or cited a single source on the talk page. Your contributions at this point to the article are the repeated mass-removals of peer-reviewed research and expert assessments. Snooganssnoogans (talk) 17:47, 14 July 2018 (UTC)
 * Seraphim, this tag-team edit wars on many articles, not just here. Snoog on the other hand is a principled and widely-read academic who is civil and articulate in his arguments. I suggest you take a broader look at who and what you're defending here.  SPECIFICO talk 18:32, 14 July 2018 (UTC)
 * Good timing, I just pinged you. I admit I don't know the history of the dispute, or the editors involved. But I would note that the article also discusses criticisms of lower estimates for various reasons so describing them as "credible" estimates does not seem to be supported by the current version of the article text or the reliable sources in the article. (for exmaple the section on IBC: The IBC project's director, John Sloboda, has stated, "We've always said our work is an undercount, you can't possibly expect that a media-based analysis will get all the deaths.") - This is one of those situations where I think the article text itself needs substantial work before the lede, but would show preference to a more neutral and simple lede until then. Seraphim System ( talk ) 18:42, 14 July 2018 (UTC)
 * Well, this isn't the place to get into detail, but just to confine myself to what I see on this page, D+T seem to be tag-teaming against consensus, logic, and evidence to support their POV and it would really be good if they'd stop doing that. They've convinced nobody of their concerns.  SPECIFICO talk 18:48, 14 July 2018 (UTC)
 * The lede specifically says that the IBC does undercount. I'm not sure what you're trying to say. Are you saying that the IBC estimate is credible and that the lede is wrong in currently separating the IBC from the credible estimates? Snooganssnoogans (talk) 18:53, 14 July 2018 (UTC)
 * It would help if readers didnt have to guess which credible estimates you are talking about. Other estimates around 176,000 (within the range given as credible) are also based on IBC and thus would seem to be included as credible estimates by the lede language as it is currently written. Seraphim System  ( talk ) 19:09, 14 July 2018 (UTC)
 * The 176,000 estimate is for a longer period (i.e. it includes the Iraqi Civil War 2014–). The body of the article needs a re-write (in particular a trim of non-scientific surveys) and the lede could have greater clarity in what precise time periods are covered (i.e. what counts as a casualty of the "Iraq War"). The lede could also definitely mention the name of the study behind the 151,000 estimate (IFHS, published in the NEJM) and the 460,000 estimates (published in PLOS Med) to make it super-clear for readers. I mentioned ways to improve the article at the start of the year but did not undertake the work because the article has effectively been held hostage by Darouet and Thucydides for as long as I've edited it. You would not believe how much effort it took just to remove the most rubbish of rubbish sources from the article (Darouet and Thucydides push a lot of fringe content across Wikipedia), clean up the article, and add actual peer-reviewed research. Or well, you might believe it, given how you yourself behaved on this article. Snooganssnoogans (talk) 19:30, 14 July 2018 (UTC)
 * Well it sounds like we have two separate estimates rather than a range. I think we can start by adding WP:RS to the new section that's been opened. Seraphim System  ( talk ) 21:51, 14 July 2018 (UTC)

Some history is relevant to 's accusations here. SPECIFICO has been stalking me through Wikipedia, looking at my contribution history and reverting my contributions at various pages. That's how they arrived at this page, and it's how they arrived at a page as random as Near-Earth Object Camera to revert me. They then "warn" other editors about my supposedly terrible editing history. This is a serious behavioral problem with SPECIFICO, which is unfortunately on display here. -Thucydides411 (talk) 00:28, 15 July 2018 (UTC)

The major problem with the lede, as it now stands, is that it arbitrarily defines "credible" and non-credible estimates, based on ' own judgment of the sources. IFHS, which has been heavily criticized for not sampling the regions with the highest levels of violence and for being run by government officials affiliated with a Shiite militia, is somehow "credible," while a highly cited study in the Lancet is excluded from the list of "credible" surveys because one researcher believes it suffers from "main-street bias." Never mind that it might also suffer from biases in the opposite direction that might be larger, such as "recall bias" and "household breakup bias," which are discussed by Hagopian et al. (2013), and which might lead the 2nd Lancet study to be an underestimate. All survey methods will suffer from various biases, tending both to increase or decrease the estimate. However, Snooganssnoogans has been arguing here for labeling one particular estimate as not credible, leaning heavily on low-citation-count criticisms by Michael Spagat.

This has all been argued over and over again, as one can see from the above threads. I think it would be best to go to a structured dispute resolution, where we could go through the issues one at a time, and actually build up a consensus point-by-point. -Thucydides411 (talk) 00:44, 15 July 2018 (UTC)
 * Lies from start to finish, some of which have been repeated multiple times. Snooganssnoogans (talk) 01:00, 15 July 2018 (UTC)

New lede text and removal of excessive Spagat references
The current lede text is not very good. What do you all think of this?

Estimates of Iraqi deaths resulting from the Iraq War (the 2003 US invasion of Iraq and the ensuing occupation and insurgency) range from 150,000 to over a million. Sources of these estimates include scholarly journals, the US military, the Iraqi government installed by the US, and polling organizations. Reports on Iraqis seriously injured due to the Iraq War are less common and even more prone to error. The only sources for Iraqi injury figures are the Iraqi government, US military, and the United Nations Assistance Mission for Iraq (for 2006 only).

One other serious problem with this article is the excessive Michael Spagat references (as many other editors have noted). If nobody else starts trimming these down, I will try to start doing it sometime this weekend. Jrheller1 (talk) 19:38, 13 July 2018 (UTC)


 * thanks for the suggestion. I think that a few things are important:


 * 1) We should note, in the lede, the difference between epidemiological survey results (there are a few) and the media counts made by the Iraq Body Count. I suspect many readers who come by this page would want to know what those two things are.


 * 2) I think we should also mention US and allied casualties.


 * 3) In reviewing articles and books related to the dispute above, I see there is a lot of commentary on other health effects of the Iraq War in the academic literature, often discussed in conjunction with deaths and physical trauma. These health effects include PTSD. It might be worthwhile to have a bit of information on this as well.


 * I plan to make a post on this topic soon, sorry that it's taken some time. -Darouet (talk) 19:59, 13 July 2018 (UTC)


 * The article should definitely discuss US and allied casualties. There is some discussion of this already.  For example, the "Additional statistics for the Iraq War" section mentions the astounding (to me) 675 wounded US soldiers per month by 2007.  The Coalition military casualties section also discusses this (including some discussion of PTSD).  Jrheller1 (talk) 17:24, 14 July 2018 (UTC)

This is very simple: On Wikipedia, we adhere to reliable sources, and we in particular privilege peer-reviewed academic research and assessments by recognized experts. Jrheller1's removal of peer-reviewed research is just straight-up vandalism. Snooganssnoogans (talk) 11:28, 14 July 2018 (UTC)


 * The focus of the article needs to be "Casualties of the Iraq War", not criticism of various studies on casualties. You only seem to be interested in adding sources critical of studies with higher casualty numbers than you would like.  In addition to these facts, the sources you are adding (which are mostly Spagat or based on Spagat) are all published in low-impact journals.  If these authors had anything really significant to say, they would publish in a high-impact journal like The Lancet, or New England Journal of Medicine, or Science.  Jrheller1 (talk) 17:24, 14 July 2018 (UTC)
 * I think both sides in this dispute should know better then to escalate by making multiple rapid fire edits and disputed changes while the talk page discussions are open - this includes removing sources without discussion as well as continuing to restore disputed language to the lede that has been opposed by most editors on the talk page. Thus I've restored the version that roughly seems to have consensus at this time (which includes the Spagat references, but does not include the changes being edit warred into the lede). I will also be requesting page protection. Seraphim System ( talk ) 17:31, 14 July 2018 (UTC)


 * There is no consensus to change a long-standing version of the article that fully reflects the body of the article. You just removed more than a dozen peer-reviewed academic sources, and restored a recent version of the article that you yourself have an affinity for. The gall to criticize editors for "removing sources without discussion" when the only editors who are at this moment removing sources without discussion are you and Jrheller1. Snooganssnoogans (talk) ;17:38, 14 July 2018 (UTC)


 * Snoogansnoogans, your version (and it is your version) of the article is not exactly "long-standing". You came along on October 17, 2017 and started flooding this article with links to low-quality journal publications attacking the Lancet survey and other surveys with higher casualty numbers than you would like and edit-warred your version into place. For years before October 17, 2017, nobody felt the need to attack the surveys with higher casualty numbers in the lede. Jrheller1 (talk) 18:13, 14 July 2018 (UTC)


 * I think we can certainly discuss further improvements and expansion of the lede (see above) but at this point, there is tag team edit warring on both sides. Also, your comments on this talk page have been more directed at the editors here then on content. Most likely this will end up at some kind of multi-party dispute resolution, which I would be willing to participate in. Seraphim System  ( talk ) 17:47, 14 July 2018 (UTC)
 * "your comments on this talk page have been more directed at the editors here then on content." This is an untruth. In the course of disputes on this page, I've added DOZENS of peer-reviewed academic studies and expert assessments to support my arguments. Sources that you removed without providing any substantive reason why. Snooganssnoogans (talk) 17:52, 14 July 2018 (UTC)
 * The issue that has been raised on this talk page, which you have not addressed, is whether this content is DUE for inclusion (See WP:SCHOLARSHIP) since you are adding more about something that is already adequately covered in the article (opinions and quotes about the high estimates). I also don't see anything supporting your characterization of other estimates as "credible" in the lede. Seraphim System  ( talk ) 18:11, 14 July 2018 (UTC)
 * The lower bounds of 151,000 is from the IFHS study, which has widely been described as highly credible (see many of the sources I added and which you deleted) and the upper bounds is the 2013 PLOS Medicine estimate of 461,000, which has been disputed but not comprehensively so. The 2013 PLOS Med study was generally described as an improvement on the Lancet study, even by Spagat. I'm not sure you realize this but the lead author of the 2006 Lancet study, the one that you're disputing over, is a co-author on the PLOS Med study. So, the lead author of the 650,000 estimate no longer stands by it. As for estimates over 650,000, none are seen as credible. In fact, one of peer-reviewed studies that I added earlier today described the ORB survey (which estimated more than 1 million casualties) as simply being "ignored" in serious debates. I didn't even bother to add it to the ORB section, as no one can seriously be arguing that the 1 million estimate is credible. Snooganssnoogans (talk) 18:25, 14 July 2018 (UTC)
 * You have not found any source where Burnham says that he does not stand by the results of 2006 survey. He was a co-author of a paper with a slightly lower estimate; this does not imply he does not stand by the results of 2006 survey. Jrheller1 (talk) 18:42, 14 July 2018 (UTC)


 * IFHS is a deeply flawed survey, which has met with severe criticism. It didn't sample the highest-violence areas of Iraq, and the interviews were conducted by people associated with the Iraqi Health Ministry, which was run by the leader of a Shiite militia at the time. If the 2nd Lancet survey is going to be described as not "credible" because it might suffer from "main-street bias," then surely a survey that suffers from the much more serious "not-sampling-the-regions-where-violence-was-occurring bias" and "Sunni-interviewees-are-afraid-to-tell-people-associated-with-Shiite-militia-their-family-members-were-killed bias" will be labeled not "credible." -Thucydides411 (talk) 00:49, 15 July 2018 (UTC)


 * Also, it's complete madness to characterize me and TheTimesAreAChanging as tag-team edit warriors. Me and TTAAC have repeatedly disputed with one another on different Wikipedia pages, and had nasty exchanges. However, both of us do have respect for Wikipedia's RS policy, and both of us add lots of academic content to Wikipedia. Snooganssnoogans (talk) 18:15, 14 July 2018 (UTC)

Protection and reversions
Let me explain what happened. I saw a request for full protection at WP:RFPP left by User:Seraphim System. I acted on that request fully protecting the article for three days. Subsequently, User:Snooganssnoogans complained at my talk page that the last version is not the pre-war consensual version. I checked the history of edit-warring and reverted to what I believe was the last pre-war version (9 July). I have no idea whether this is a consensus version, and it is not my business to determine this. Per WP:WRONGVERSION I did not have to revert anything at all, but I think it would be probably easier for all of you to discuss if none of the sides gets its way from the very beginning. Please use this time efficiently, and if you come to consensus leave an unprotection request at WP:RFPP. Let me note that (i) I am not involved with this article and I do not want to be involved, I merely protected it to prevent ongoing disruption; (ii) I will likely not be available (or very little available) starting from about three hours from now. Please discuss between yourselves, not with me, and seek mediation if needed. Thank you.--Ymblanter (talk) 18:12, 14 July 2018 (UTC)
 * I don't think so - The changes to the lede were opposed by 3 editors on the talk page, and User:Snooganssnoogans continued to edit war against the talk page consensus insisting he was restoring a long term consensus version - those changes should not have been protected. I don't think this makes discussion more likely, since he was edit warring against a recent consensus and will most likely continue to do so when that consensus is restored after the protection is lifted. Pinging  as a past participant - restoring "longstanding version" is not a justification for edit warring. We can't just grandfather SPECIFICO in as a default support for "longstanding version". Seraphim System  ( talk ) 18:35, 14 July 2018 (UTC)
 * Darouet and Thucydides411 have a long history, including Arbcom Enforcement and other ignominies, with POV tilts and edit-warring that I won't repeat or document here. I'm familiar with Snoog and with TTAC and I've never seen them agree or even collaborate on anything at all. Both, however, are widely read and tend to take sourcing and detail very seriously.  SPECIFICO talk 18:41, 14 July 2018 (UTC)

Additional sources for higher Iraqi death tolls
I am creating this section to list additional reliable sources for higher Iraqi death figures that should be included in the article. Jrheller1 (talk)


 * The report written by Physicians for Social Responsibility called Body Count: Casualty Figures after 10 Years of the "War on Terror" Jrheller1 (talk)


 * The portion of the PSR report related to Iraq casualty numbers was written by Joachim Guilliard, who frequently appears on Russian state propaganda networks and has promoted conspiracy theories about White Helmets in Syria. Snooganssnoogans (talk) 21:41, 14 July 2018 (UTC)


 * John Tirman estimated Iraqi deaths at between 800,000-1,300,000 in this article in the Nation. Jrheller1 (talk)


 * I don't mind adding Tirman's opinion. He is after all a recognized expert, and the Wikipedia article already mentions his defence of the 2006 Lancet study. Unlike other editors, I don't delete expert assessments just because I don't like them. Some editors can be principled like that. His estimate of 800,000-1,300,000 is however simply a guesstimate extrapolation and not derived from study or data. He basically mentions that two studies found deaths of 400,000 and 650,000 respectively for a certain period, and that the IBC saw a doubling of reported deaths from the end of those studies, and that therefore the death rate must be double what the studies found. Snooganssnoogans (talk) 21:41, 14 July 2018 (UTC)


 * NYT article discussing war widows; cites two different Iraqi government estimates of 900,000 and 1 million war widows respectively. Jrheller1 (talk)


 * Per the source: "The Ministry of Social Affairs pays widow’s benefits to 86,000 women, most of whom, it says, lost their husbands in the latest war. This figure corresponds with conservative estimates of 103,000 to 113,000 Iraqi deaths in the war, according to a nonprofit group that tallies casualties, Iraq Body Count." You should perhaps read more carefully. Snooganssnoogans (talk) 01:13, 15 July 2018 (UTC)
 * They were only paying benefits to 86,000 widows. But there were 900,000 to 1,000,000 total widows in 2011.  If most of the 86,000 they are paying benefits to lost their husbands in the war, probably most of the 900,000-1,000,00 total widows also lost their husbands in the war. Jrheller1 (talk) 01:22, 15 July 2018 (UTC)
 * That's not what the source says. The source clearly says that the widow population are the result of multiple horrors that have befallen the Iraqi people before and after the onset of the Iraq War. The source does not say that most or a large part of the widows are from the Iraq War. The only text in the NYT piece that ties widow population to the Iraq War is the quote I showed you, and that's the only time the NYT links the widow population to casualty figures. Snooganssnoogans (talk) 01:31, 15 July 2018 (UTC)
 * The source says the Iraqi government is paying benefits to 86,000 widows.  The source says that most of these 86,000 were widowed by the Iraq War.  So it is reasonable to conclude that most of the total number of widows (900,000 - 1,000,000) were also widowed by the Iraq War. Jrheller1 (talk) 01:38, 15 July 2018 (UTC)
 * No, it's not. That's WP:OR and dangerously reckless OR at that. For all you know, payments to widows may be specifically linked to widows who lost their husbands in the Iraq war, and not apply to widows who lost their husbands in Saddam's atrocities, the Iran-Iraq War etc. Snooganssnoogans (talk) 01:48, 15 July 2018 (UTC)
 * Your statement "payments to widows may be specifically linked to widows who lost their husbands in the Iraq war" is wrong. See this other NYT article that states that benefits to widows are linked to "having political connections or agreeing to temporary marriages with the powerful men who control the distribution of government funds". Jrheller1 (talk) 02:03, 15 July 2018 (UTC)
 * Exactly, I have no idea what the requirements for payments or the demographics of the widowers are. That's why I go by what RS say. I recommend you adopt that approach in your Wikipedia editing also. Snooganssnoogans (talk) 02:13, 15 July 2018 (UTC)
 * Since most academic sources agree that IBC is not a reliable estimate I don't think we can use this source - there seems to be academic consensus that IBC is not reliable based on its methodology - Lancet on the other hand is debated in detail and defended by many scholars. Seraphim System ( talk ) 02:27, 15 July 2018 (UTC)


 * One important point that we should make is the difference between body counts (of which IBC is the most prominent for Iraq) and statistical surveys as a means of measuring mortality during a conflict. As to labeling certain estimates reliable or unreliable - I don't think we should be in that business, unless the sources are absolutely clear. We should discuss the strengths and weaknesses of the different estimates, with the appropriate weight, based on reliable sources. That means, for example, not piling on every criticism that can be found of the Lancet surveys, but still mentioning the major criticisms (e.g., "main-street bias"). -Thucydides411 (talk) 03:07, 15 July 2018 (UTC)


 * The lede already notes that the IBC is an undercount. There is no dispute here over describing the undercount, so I don't understand what you're trying to get at. You're like a broken clock. The dispute is over describing the over-count because there are fringe POV pushers who for some reason want to inflate the number of casualties in the Iraq War and mislead readers. The Lancet study has widely been described as seriously flawed and/or less rigorous than the 150,000 and 460,000 estimates, and there are dozens of sources in the article that substantiate that language. The lead author of the damn Lancet study is a co-author on the study that downgrades the casualty numbers. You're defending an estimate that the lead author no longer stands by, the lead author was censured over by the American Association for Public Opinion Research, and which has been widely described as flawed, weak, discredited and less reliable and less rigorous than other surveys. Snooganssnoogans (talk) 09:19, 15 July 2018 (UTC)


 * Reuters and NPR sources for ORB poll estimating approximately 1 million deaths in September 2007. Jrheller1 (talk) 17:22, 15 July 2018 (UTC)
 * The ORB poll is already in the article. Snooganssnoogans (talk) 17:35, 15 July 2018 (UTC)

Extended Discussion

 * You wrote above which has widely been described as highly credible (see many of the sources I added and which you deleted) - can you post these sources on talk. I don't see many sources supporting this, all I see is one opinion taken from Small Arms, Crime and Conflict that says the IFHS study is more reliable. This is not cited or explained further, so it is only one expert opinion. Another expert opinion about IFHS:
 * This is discussed in multiple sources, in detail, including disputing IFHS - while both studies have similar figures of excess deaths, the disparity is over whether they are violent deaths. The criticism from the Lancet team, and discussed in secondary sources, is that violence-related deaths may have been underreported to the IFHS team - who were employees of the Iraqi health ministry. (John Tirman has additionally noted that the Minister of Health who supported the 150,000 figure was affiliated with Muqtada al Sadr - this was discussed in multiple sources, one of which I cited above, but I can keep looking). Attempting to describe this as "credible" and the other studies as "disputed" doesn't seem to be an accurate summary of the discussion in these sources). Seraphim System ( talk ) 23:07, 14 July 2018 (UTC)
 * What a disgraceful request. You DELETED multiple sources which described the IFHS study as rigorous and reliable, and now you come here requesting sources that substantiate it, saying you haven't seen any. You don't read the stuff that you delete??? That explains a lot, and it frankly merits sanctioning. I have not read the two links that you've cited, but I will. Unlike you, I actually read sources that other editors bring up in content disputes rather than dismiss them out of hand. Snooganssnoogans (talk) 23:22, 14 July 2018 (UTC)
 * No I did review it, most of the sources are not about IFHS and don't support the version of the lede you have been edit warring into the article. The other changes you should discuss on the talk page, since they are disputed (by other editors, not by me) and you guys have been edit warring over this for days. I didn't "delete" anything, I asked you several times to discuss the disputed content on the talk page. Nothing is ever "deleted" it's all in the revision history. I think most likely this will have to go to DR/n. If other editors file for dispute resolution, please ping me. I will participate but I'm not going to respond to accusations of bad faith. Seraphim System  ( talk ) 23:38, 14 July 2018 (UTC)
 * Wonderful. We're now in postmodernist territory where there is no such thing as "deletion" and sources that literally say "reliable", "rigorous" etc. don't really say it at all... Snooganssnoogans (talk) 00:17, 15 July 2018 (UTC)
 * I read the sources. (1) the POQ article is good and from a highly reputable journal (unlike Jrheller1, I'm not gonna spout nonsense and say that POQ is a "low-quality journal" because it "only" has an impact factor of ca 1.5). The source does not describe the IFHS study as not credible. It points out a weakness of the study, one it has in common with the Lancet study: namely that surveyors can't be sent into the most violent areas (which the POQ article describes as "understandable"). (2) The article in the DePaul Journal of Health Care Law is an OK overview, and it does not describe the IFHS as not credible. Furthermore, the DePaul Journal of Health Care Law is not peer-reviewed and it is run by students at DePaul (the editor-in-chief is a DePaul undergrad), so there are weight concerns and we should not add it to the Wikipedia page. Snooganssnoogans (talk) 00:13, 15 July 2018 (UTC)
 * What are the sources that support this version of the lede? Don't tell me they are all in the article and I've "deleted" them - it is not clear at all why you believe those sources, like Spagat citing Spagat, support the lede's characterization of a particular range as credible. It is actually a gross representation as there is no study that indicates a range of 150,000 to 460,000 as credible. You've arrived at this range based on the figures obtained in two separate studies - each of those studies represents a range that is calculated based on the figure they obtained. Representing them as a "credible range" is a careless but serious error. Seraphim System ( talk ) 00:43, 15 July 2018 (UTC)
 * The lede is supposed to summarize the body of the article. The body of the article characterizes the 151,000 and 460,000 estimates as reliable and credible, while the IBC, the 2006 Lancet study and ORB study are highly disputed and widely described as flawed. This version of the lede is primarily supported by this section but of course also the text in the ORB and IBC sections. Snooganssnoogans (talk) 00:55, 15 July 2018 (UTC)
 * But it doesn't, there is no "credible" range of 150,000- 460,000 range, and I haven't seen any secondary source that describes the casualty estimate this way. Every source I've seen discusses a number of disputed estimates, noting the criticisms that have been raised about each one. Additionally the section of the article you indicated is also disputed - there are 15 citations, at least 5 of which are directly to Spagat's primary work and another one like this . I'm having a hard time understanding why anyone would edit war over restoring these. Seraphim System ( talk ) 01:30, 15 July 2018 (UTC)
 * There are far more than 15 citations. Snooganssnoogans (talk) 01:35, 15 July 2018 (UTC)
 * There are far more than 15 citations. Snooganssnoogans (talk) 01:35, 15 July 2018 (UTC)



, "while both studies [Lancet 2006 and IFHS] have similar figures of excess deaths" is actually false. Although I have not had the time to look into this more carefully and many of Wikipedia's sources on this matter are currently dead links, based on [https://en.wikipedia.org/w/index.php?title=Iraq_Family_Health_Survey&oldid=839086452#400,000_excess_deaths? Wikipedia's article on this topic] it appears that Tirman and Roberts made a number of extravagant and misleading claims about the IFHS: For example, Tirman falsely asserted that the Iraqi Ministry of Health was "controlled by Moktada al Sadr" (who has never held any official position in Iraq, although he remains very influential in the Iraqi government) and therefore not credible; Tirman simultaneously argued that the IFHS vindicated the Lancet study as it "also found a sizable mortality figure—400,000 'excess deaths'" (emphasis added), even though the authors of the IFHS expressly stated that it did not measure excess deaths (Tirman then criticized the media for accurately reporting the IFHS's major findings); Roberts even had the gall to suggest that "The NEJM article found a doubling of mortality after the invasion, we found a 2.4-fold increase. Thus, we roughly agree on the number of excess deaths" (!) (emphasis added), as if a difference of 250,000 would be insignificant even if Roberts's underlying analysis was accurate (and it isn't—the IFHS actually found that "The non-violent mortality rate increased by about 60%," not more than 100%, concluding only that "Further analysis would be needed ... to assess how large the mortality increase due to nonviolent causes is, after taking into account that reporting of deaths longer ago is less complete" and making no blanket statement about excess deaths). Wikipedia should avoid repeating Tirman and Roberts's misleading comments as fact, although there is no doubt that the IFHS's 151,000 estimate is (as the current lead indicates) the absolute minimum.TheTimesAreAChanging (talk) 07:37, 17 July 2018 (UTC)


 * The various statements that you say are "extravagant and misleading" are actually reasonable and accurate.


 * For example, Tirman falsely asserted that the Iraqi Ministry of Health was "controlled by Moktada al Sadr" (who has never held any official position in Iraq, although he remains very influential in the Iraqi government) and therefore not credible. The Iraq Family Health Survey (IFHS) was conducted in late 2006 and early 2007:
 * The survey was done by Iraq Ministry of Health employees during late 2006 and early 2007 in all 18 provinces, divided to get a valid sample of each area. But Iraqis hold a deep distrust of central authority, given the tribal nature of their society and the years they lived under Saddam Hussein, whose grip on power was built partially on a web of informers. -Fox News, from an AP wire
 * At that time, the Iraqi Health Ministry was apparently infiltrated by Muqtada al-Sadr. US forces detained the deputy Minister of Health in early 2007 because he had allegedly been using Health Ministry resources to kidnap and murder Sunnis:
 * According to the military statement, the arrested man is suspected of providing significant employment to Shia militia members who used health ministry facilities and services for "sectarian kidnapping and murder". -The Guardian
 * Tirman's statement that the Iraqi Ministry of Health was "controlled by Moktada al Sadr" appears to actually have been true. "Controlled by Moktada al Sadr" does not mean "Muqtada al-Sadr was the Minister of Health." It means that the Deputy Minister of Health was funneling resources towards Muqtada al-Sadr's organization.
 * Tirman's point was that Sunnis might have been wary of telling people employed by the Ministry of Health that family members had been killed. As the AP notes in the passage I gave above, Iraqis already distrust the central authorities, and as the Guardian notes, the Iraqi Health Ministry had ties to a Shiite militia at the time the survey was conducted, and was using government resources to conduct kidnappings and murder.


 * Tirman simultaneously argued that the IFHS vindicated the Lancet study as it "also found a sizable mortality figure—400,000 'excess deaths'" (emphasis added), even though the authors of the IFHS expressly stated that it did not measure excess deaths (Tirman then criticized the media for accurately reporting the IFHS's major findings). The IFHS did not explicitly compute the number of excess deaths, but they collected enough information to allow others looking at the data to do so. The authors said that they themselves had not computed the number of excess deaths. That doesn't mean that Tirman could not look at the data and do so himself, to see if it was consistent with what the Lancet surveys found.


 * Roberts even had the gall to suggest that "The NEJM article found a doubling of mortality after the invasion, we found a 2.4-fold increase. Thus, we roughly agree on the number of excess deaths" (!) (emphasis added), as if a difference of 250,000 would be insignificant even if Roberts's underlying analysis was accurate (and it isn't—the IFHS actually found that "The non-violent mortality rate increased by about 60%," not more than 100%. Roberts says that mortality doubled after the invasion, but you cite the finding that the non-violent mortality rate increased by 60% as a refutation. There's a very big difference between non-violent mortality and mortality. You're saying that Roberts had "the gall" to suggest something which is factually accurate - IFHS figures showed overall mortality doubling after the invasion, driven by a huge increase in violent mortality.


 * These statements, which you said were misleading or false, were actually factually accurate. -Thucydides411 (talk) 10:47, 17 July 2018 (UTC)
 * Looking over the data in table 3 here, I see that reported mortality increased from 3.17 per 1,000 to 6.01 per 1,000—an increase by a factor of almost 1.9, so it is hardly much of an exageration to say that the mortality figures "doubled." However, it remains the case—as the authors of the IFHS noted—that "the level of underreporting is almost certainly higher for deaths in earlier time periods," which means that the actual increase is likely to be somewhat less than reported and that extrapolating 433,000 "excess deaths" from this increase is open to question. Moreover, despite Tirman/Roberts's claims of "rough agreement" between the sources, it remains strikingly incongruous that only 8.2% of the calculated "excess deaths" in the Lancet study were non-violent whereas almost 58% of the "excess deaths" that Tirman/Roberts extrapolated from the IFHS would have been non-violent.TheTimesAreAChanging (talk) 12:04, 17 July 2018 (UTC)


 * Keep in mind that the IFHS did not survey the most violent areas of Iraq, which could be one cause for it finding a smaller increase in violent deaths. It was also carried out by a government agency that some of the populations most affected by violence had good reason to fear. The fact that IFHS did not see a steep rise in violence in 2006 also raises questions about its results, given that 2006 was the most violent year in the Iraq War.
 * Nevertheless, I think it's clear that there are very serious issues with the IFHS, which have been pointed out by many sources. There's no justification for declaring IFHS as "credible," and the Lancet surveys as not credible. We should discuss the strengths and weaknesses of the different estimates, without inserting our own judgments (in Wikipedia's voice) on which studies are credible. -Thucydides411 (talk) 14:09, 17 July 2018 (UTC)

(unindent) As most of us know already IFHS survey only covered a 3+ year period: March 2003 through June 2006. IFHS estimated 151,000 deaths due to war violence.

PLOS survey covered March 2003 to June, 2011. PLOS estimated around 276,000 violent deaths in that period (by doing the math). Its confidence interval (for violent deaths) goes much higher than that (also by doing the math).

Logically, expanding the IFHS death totals out until the PLOS end date of June 2011 comes out with violent death totals higher than the PLOS midrange number. I am sure some reliable source points this out somewhere.

US troops withdrew on Dec. 18, 2011. So the PLOS survey period covers almost all of the US part of the Iraq War before these wars: Iraqi insurgency (2011–13) and Iraqi Civil War (2014–present).

Here are the current Iraq Body Count yearly civilian war death totals below. See IBC database chart. They can be used to guesstimate the IFHS death total out past 2006. For example there were around 55,000 civilian IBC deaths through June 2006 (end of IFHS survey). Another roughly 55,000 IBC civilian deaths occurred by the end of 2009. That leaves 2010 and 2011 where around 8000 IBC civilian deaths were counted. So, the IFHS number can easily be doubled to 300,000 violent deaths.

As Thucydides411 pointed out, the IFHS violent death total is considered to be an underestimate by some due to the various sources and reasons he elucidated. --Timeshifter (talk) 16:38, 17 July 2018 (UTC)
 * After further consideration, I believe that Thucydides411 and Timeshifter have made some valid points and have edited the article accordingly.TheTimesAreAChanging (talk) 10:44, 20 July 2018 (UTC)

Lead: "Credible estimates of Iraq War casualties range from 150,000 to 460,000"
150,000 is not a credible number for violent deaths due to the Iraq War. Body counts of violent civilian deaths alone approach that. And IBC has stated in several places that their body count is low due to many reasons.

Iraq Body Count: "IBC has documented 112,017 - 122,438 civilian deaths from violence between 20 March 2003 and 14 March 2013." From this IBC page.

Body counts have the least credibility of all for getting a total number of violent deaths according to nearly all serious researchers and experts in this field. Even IBC acknowledges this.

The lead should summarize what the range of survey estimates are for all violent deaths due to the Iraq War, and all excess non-violent deaths due to the Iraq War. And the difference should be made clear.

It should be pointed out that the survey estimates are all highly disputed. The word "credible" has no place in the lead. The PLOS survey is the most credible. And it has a huge confidence interval. Until there is a much broader survey across all parts of Iraq including the areas that have seen the most violence, the word "credible" should not be in the lead.

Body counts should not be in the lead at all, except to show why a survey result of 150,000 is not credible. --Timeshifter (talk) 20:55, 15 July 2018 (UTC)
 * "150,000 is not a credible number for violent deaths due to the Iraq War. Body counts of violent civilian deaths alone approach that." You're comparing apples and oranges: The IBC may have documented 112,017–122,438 violent deaths as of March 2013 whereas the IFHS estimated 151,000 violent deaths as of June 2006, but the IBC had documented only "47 668 civilian deaths due to violence during the period studied in the Iraq government/WHO survey." As Snooganssnoogans has noted, "the lede could have greater clarity in what precise time periods are covered (i.e. what counts as a casualty of the 'Iraq War')".TheTimesAreAChanging (talk) 07:51, 17 July 2018 (UTC)


 * What is the basis for calling IFHS' result of 151,000 deaths credible, and calling the 2006 Lancet study non-credible? There are serious problems with the IFHS study, given that it was conducted by one of the sectarian factions taking part in the war and did not even survey the most violent regions of the country. As it stands, the statement in the lede about "credible" and "other" estimates (implying that the latter are not credible) is simply a personal judgment of one or a couple of editors here, and is not based on actual sourcing. Rather than us arguing here about what we personally believe to be credible, we should be basing the article on the mainstream scientific view. That mainstream view isn't defined by one author (e.g., Michael Spagat), but rather by the consensus of many researchers. Given the varied statements by different researchers on the various studies (including criticism of IFHS and praise for the 2006 Lancet study), I don't see a scientific consensus to call IFHS "credible" and Lancet non-"credible." -Thucydides411 (talk) 04:02, 6 December 2018 (UTC)

Article is a mess
The "Tables" and "Total Iraqi casualties" sections are largely redundant and the long-term goal here should be to merge them together. If we retain the tables at all, it should be to provide a quick overview of the various estimates, but in some cases the tables as presently constituted actually feature more detailed statistical data than the main body of text, such that readers would have to actively switch between both sections to get all of the relevant information.TheTimesAreAChanging (talk) 11:23, 20 July 2018 (UTC)
 * Yes, the tables need to be fixed ASAP. They make the article unreadable. Snooganssnoogans (talk) 11:33, 20 July 2018 (UTC)

2006 Lancet paper
has added in several paragraphs to the section on the 2006 Lancet paper, giving a heavily unbalanced view of that work. The 2006 Lancet paper has received significant praise, and is one of the few rigorous studies on mortality in the Iraq War available, yet the section on the 2006 Lancet study presents almost only criticism, relying heavily on one particular critic - Michael Spagat. I and a few other editors tried to rectify this a few months ago, in order to present a more balanced and neutral view, but we were prevented from doing so by the persistent edit warring of (see the string of reverts from 9-14 July 2018, within the timeframe covered by this diff). As it is, this article's coverage of the 2006 Lancet survey is extremely biased, as is the declaration in the lede about "Credible estimates," which is a personal judgment of one particular editor, and which does not represent the mainstream scientific view on the subject. -Thucydides411 (talk) 03:52, 6 December 2018 (UTC)


 * Agreed. The article version from October 13, 2017 (just before Snooganssnoogans started editing the article) is more neutral.  Jrheller1 (talk) 01:07, 11 December 2018 (UTC)


 * Since has been adding ever more POV material into the article, without consensus, and has been edit-warring to keep it in the page, I suggest we all go to some sort of structured dispute resolution. That will be more productive than the current cycle we're in:
 * Snooganssnoogans adds yet more material from Spagat.
 * Someone else reverts it.
 * Snooganssnoogans reverts as many times as necessary to keep the new material in.
 * The other person gives up.
 * Snooganssnoogans claims the material is "long-standing" the next time around, simply because they edit-warred to keep it in.
 * I'd rather be in some sort of process where we can actually evaluate the publications about Iraq War casualties and determine how to give due weight to them. -Thucydides411 (talk) 14:46, 18 March 2019 (UTC)


 * I think the best possible solution would be to fork the article into two different articles: one for Snooganssnoogans to add Spagat material to his heart's content and another starting from the pre-Snooganssnoogans version for most of the other editors of this article. That would be the most useful solution for Wikipedia readers and also the best way of preventing future edit warring.


 * Of course, this isn't possible under the current Wikipedia system. It is possible under the StackExchange system, though: on StackExchange, multiple answers from different users are allowed for each question.  This is one of the strengths of the StackExchange system.  There is no edit warring on StackExchange, partly because a user can simply create a new answer to a question if he isn't satisfied with the existing answer.


 * I believe that adding the possibility to fork articles would be a major improvement to Wikipedia that would reduce edit warring and increase the number of good editors active on Wikipedia. Wikipedia needs to learn from the success of StackExchange in attracting skilled, knowledgeable editors.  I believe that adding the possibility to fork articles is one way of doing this. Jrheller1 (talk) 00:57, 19 March 2019 (UTC)


 * That's an interesting proposal, but I was thinking more along the lines of Wikipedia's existing Dispute Resolution mechanism. -Thucydides411 (talk) 02:26, 19 March 2019 (UTC)


 * I realized you were probably referring to DRN, but it would be more efficient and more useful to readers to fork into two articles. The Snooganssnoogans version would focus on debunking the higher casualty estimates and the other version would attempt to be neutral.  I don't think that deleting the material added by Snooganssnoogans would benefit readers, but I also don't think that pretending the current article is neutral is good for either readers or editors.  Perhaps it is even possible to fork the article within the current Wikipedia system.  The Snooganssnooogans article could have a different title (something like "Debunking high Iraq war casualty estimates").Jrheller1 (talk) 15:39, 19 March 2019 (UTC)

Where's the sourcing that describes the Lancet study or the ORB survey as over-inflated? Without that, I would lean towards the wording suggested by Thucydides411. I'm not impressed by Snoogans edit warring, they reverted 4 times in a 36 hour period, and should definitely know better. Mr Ernie (talk) 16:43, 19 March 2019 (UTC)
 * "Where's the sourcing that describes the Lancet study or the ORB survey as over-inflated?" Where? In the article. Snooganssnoogans (talk) 20:40, 23 March 2019 (UTC)

Systematic misrepresentation of sources to downplay importance of Lancet papers
For some perspective :


 *  Burnham et al., (2006) Mortality after the 2003 invasion of Iraq: a cross-sectional cluster sample survey. The Lancet, v368 pp.1421-1428.
 * Cited 331 times according to ScienceDirect
 * Cited 717 times according to Google Scholar

The Lancet et al paper is obviously very highly cited, and looking at those references, you can see they almost unanimously take the Lancet article seriously and cite other similar papers from the same authors. For instance of the ‘’most cited’’ papers that cite the Lancet 2006 paper,


 * Wang, Haidong, et al. (2016), cited 1031 times according to Google Scholar, “it has been challenging to accurately document the number of casualties from wars and deaths resulting from malnutrition, infections, or disruption in health services during wars.147–150” Where references 147–150 are all to articles by Gil Burnham, Les Roberts and Christine Tapp.
 * Adhikari, Neill KJ, et al. (2010), cited 653 times according to Google Scholar. “However, during times of war, we should remember that evidence from systematic household cluster sampling suggests that most excess deaths, and, by extension, most demands for intensive care, do not arise from violence but from medical disorders resulting from the breakdown of public health infrastructure (eg, cholera), or from the discontinuation of treatment of chronic diseases caused by interruption of pharmaceutical supplies.[88,89]” Where references 88 and 89 are to the Lancet paper and to a similar mortality survey from the Congo.
 * Kassebaum, Nicholas J., et al. (2015) cited 441 times according to Google Scholar. “The accelerated declines in cause-specific [years of life lost] rates for nearly all causes is occurring despite threats to human health such as climate change, antimicrobial resistance, obesity, emerging infectious diseases, and conflict.[6, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60]” Where references 57–60 are all to Burnham et al and to Christine Tapp.

Since (like Burnham, but unlike Spagat) Tapp is also cited regularly by authors when reviewing Iraq mortality, here is what Tapp's article, which is reviewing casualty estimates in Iraq, has to say.
 * Tapp, Christine, et al. "Iraq War mortality estimates: a systematic review." Conflict and health 2.1 (2008): 1, cited 49 times according to Google Scholar. “The studies used a wide range of methodologies, varying from sentinel-data collection to population-based surveys. Studies assessed as the highest quality, those using population-based methods, yielded the highest estimates... Our review indicates that, despite varying estimates, the mortality burden of the war and its sequelae on Iraq is large... Of the population-based studies, the Roberts and Burnham studies provided the most rigorous methodology as their primary outcome was mortality [16, 18]. Their methodology is similar to the consensus methods of the SMART initiative, a series of methodological recommendations for conducting research in humanitarian emergencies [33]. Another population-based study, the Iraq Living Conditions Survey, reported lower death estimates that we assume is due to the survey being conducted barely a year into the conflict, a higher baseline mortality expectation, and differing responses to mortality when houses were revisited [21]. However, not surprisingly their studies have been roundly criticized given the political consequences of their findings and the inherent security and political problems of conducting this type of research [36, 37]”

By contrast, here are some criticizing chapters / articles by Michael Spagat, who thanks to Snoogans is referenced seven times by name in our Wiki article section on the 2006 Lancet paper:
 * Spagat M (2010) Ethical and Data‐Integrity Problems in the Second Lancet Survey of Mortality in Iraq. Defence and Peace Economics, 21(1), 1-41, cited 17 times according to Google Scholar,, with 5 of those citations from Spagat himself.
 * Spagat, M (2012) Estimating the human costs of war: The sample survey approach. The Oxford handbook of the economics of peace and conflict, pp. 318-340, cited 6 times according to Google Scholar

Thanks to Snoogans, our wiki article cites some scholars who approvingly cite Spagat's otherwise largely unknown work (NP Jewell - cited by 4, MAP de Montclos - cited by 1).

However, the Lancet paper is cited as an authority by most scholarly work, and by work with far more academic attention (typically by a factor of 100) compared to Spagat or other articles that criticize the Lancet paper.

The impact of our not citing well-known papers that reference the Lancet paper, and instead cherry-picking a few critical papers that are relatively unknown, is that our article hugely misrepresents what is known about casualties of the Iraq War. -Darouet (talk) 20:20, 19 March 2019 (UTC)


 * It is a complete falsehood that papers that cite the Lancet study "almost unanimously take the Lancet article seriously". As the body of the article extensively documents and which has already been made clear to you multiple times, the study has been thoroughly rebutted in high-quality journals by multiple scholars (despite the repeated inaccuracy that only one scholar has rebutted the study) and those rebuttals have been lauded by recognized experts, leading many of them to say the Lancet study is deeply flawed. Snooganssnoogans (talk) 20:37, 23 March 2019 (UTC)

Academic sources overwhelmingly view the Lancet papers positively
As I noted above, the Burnham et al (2006) Lancet paper is cited 717 according to Google Scholar. The earlier Roberts et al (2004) Lancet paper is cited 475 times. I pointed out above that of the papers that reference these Lancet studies of Iraq War mortality, the top cited (1031 citations, 653 citations , 441 citations ) treat the Lancet articles as the authority on this topic and cite other papers by Burnham and Roberts (but do not cite Spagat or related criticisms). An often-cited review of Iraq War mortality estimates described the 2004 and 2006 Lancet papers as "the most rigorous" of "population-based studies", noting that it was not surprising that the studies were "criticized given the political consequences of their findings." The review's summary results stated, "Studies assessed as the highest quality, those using population-based methods, yielded the highest estimates." These figures and quotes leave no doubt that the Lancet articles by Burnham, Roberts and others are the gold standard in estimating casualties of the Iraq War.

A later high-quality estimate of Iraq War mortality authored by Hagopian et al (2013), cited 85 times according to Google Scholar, found lower mortality than that observed by the 2004 and 2006 Lancet papers, estimating 405,000 deaths between 2003–2011 attributable to the war. The Hagopian et al paper reviews earlier work:

Hapogian et al explain reasons why their later survey might report lower death rates. Noting improved sample size in their own work and better random sampling, they write,

I will go through more well-known and often-cited papers that reference the Lancet papers to further demonstrate their positive reception in the academic community. -Darouet (talk) 00:16, 20 March 2019 (UTC)

Here is the more recent assessment of professors Levy and Sidel, published in 2016 in the prestigious Annual Reviews series, Levy, B. S., & Sidel, V. W. (2016). Documenting the effects of armed conflict on population health. Annual review of public health, 37, 205-218:

Like Tapp et al., 2008 referenced above, Levy and Sidel 2016 find the two Lancet papers and the Hagopian et al., 2013 paper to be the most rigorous, describe these papers as widely accepted by the scientific community, and note that they may in fact underestimate mortality (a concern also discussed by the papers themselves, see quotes from Hagopian above). -Darouet (talk) 16:52, 21 March 2019 (UTC)


 * It took almost a year but finally some sources were added to the article or the talk page. The sources (which are not WP:SYNTH) belong in the body, and the lede should no longer suggest that the Lancet study is not considered credible. The lede should instead say that the Lancet study is disputed in the literature, which is what the body reflects. Snooganssnoogans (talk) 20:31, 23 March 2019 (UTC)
 * As for the number of citations, that is not a reflection of the quality of the paper. After all, rubbish studies get cited a lot when others critique those studies. Snooganssnoogans (talk) 20:31, 23 March 2019 (UTC)


 * As showed above, the articles that cite the Lancet papers, which have themselves have the most citations, view the Lancet papers positively. That would not be the case if the Lancet papers were "junk." I would be against saying that the 2006 Lancet paper is "disputed" in the lede, because that does not reflect the academic literature. It reflects the views of one particular person, Michael Spagat, whose papers are only lightly cited. The amount of room that his views have been given in this article are completely out of proportion to the citation counts of his papers. The section on the 2006 Lancet paper should be edited to reflect the most highly cited papers that discuss it - those that  has listed above. -Thucydides411 (talk) 21:35, 23 March 2019 (UTC)


 * "It reflects the views of one particular person, Michael Spagat" - This is completely and utterly false, as has been repeatedly pointed out to you. I have gone through this with you what must be at last a half a dozen times by now. Snooganssnoogans (talk) 21:50, 23 March 2019 (UTC)


 * It happens to be true. He's on almost every single paper you've cited to criticize the Lancet studies. All of those papers have much lower citation counts (usually by about two orders of magnitude) than the papers that lists above. Are you going to keep on edit warring to skew the article towards the views of this single, marginal commentator? Instead of forcing your views through against everyone else, please self-revert and respect the community process. You don't own this article. -Thucydides411 (talk) 22:13, 23 March 2019 (UTC)

Thanks for doing all the work to look through the literature. I've updated the section on the 2006 Lancet study to reflect the review papers, and removed the endless citations of Spagat et al. The article should follow what highly cited review papers say, rather than being a compendium of every low-citation-count criticism of Burnham et al. that's ever been published. I've also updated the lede to reflect the fact that we're going by the highly cited reviews now. As such, I've removed the claim that the studies the review papers consider to be the most rigorous - and which the reviews say are the most widely accepted in the scientific discipline - are "disputed." -Thucydides411 (talk) 02:50, 26 March 2019 (UTC)

Thucydides411, can you explain why you deleted the following sourced text? Burnham, Edward J. Mills, and Frederick M. Burkle noted that the IFHS's data indicated that Iraqi mortality increased by a factor of 1.9 following the invasion, compared to the factor of 2.4 found by Burnham et al., which translates to some 433,000 excess Iraqi deaths (violent and non-violent). Timothy R. Gulden considered it implausible that fewer than one-third of these excess deaths would have been violent in nature. Francisco J. Luquero and Rebecca F. Grais argued that the IFHS's lengthy survey and use of IBC data as a proxy for particularly dangerous areas likely resulted in an underestimate of violent mortality, while Gulden hypothesized that respondents may have been reluctant to report violent deaths to researchers working with the Iraqi government.[205] In a similar vein, Tirman observed that the Iraqi Health Ministry was affiliated with Shi'ite sectarians at the time, remarking that there was evidence that many violent deaths may have been recategorized as "non-violent" to avoid government retribution: "For example, the number of deaths by auto accidents rose by four times the pre-invasion rate; had this single figure been included in the violent deaths category, the overall estimate would have risen to 196,000."[206] Gulden even commented that "the IFHS results are easily in line with the finding of more than 600,000 violent deaths in the study by Burnham et al." However, the authors of the IFHS rejected such claims: "Because the level of underreporting is almost certainly higher for deaths in earlier time periods, we did not attempt to estimate excess deaths. The excess deaths reported by Burnham et al. included only 8.2% of deaths from nonviolent causes, so inclusion of these deaths will not increase the agreement between the estimates from the IFHS and Burnham et al."[205] I added that content in July 2018 in a good-faith effort to balance the article and address some of your stated concerns. Is it deficient in some respect, or was it merely "collateral damage" from your reverting to a pre-Snooganssnoogans version of this article? In truth, a lot of good sources were caught in the crossfire of your mass deletions.TheTimesAreAChanging (talk) 08:23, 26 March 2019 (UTC)
 * The most highly cited academic literature (including the Lancet studies) do not treat the higher casualty estimates of the Lancet studies and ORB poll as questionable. The views of Spagat and co-workers are a minority position among experts.  They do not belong in this article.  All of the minority-position-based material (including the criticism of the minority position by Tirman and others) should be moved to another article. Jrheller1 (talk) 15:15, 26 March 2019 (UTC)

Stop the edit war. If there is no consensus to add the material, then do not re-add the material until you can gain consensus. Quoting our policy here - The onus to achieve consensus for inclusion is on those seeking to include disputed content. and, you need to self-revert and gain consensus on the talk page. Admins would be well justified in blocking everyone participating in the edit war, and I hope it doesn't come to that. Mr Ernie (talk) 09:16, 26 March 2019 (UTC)


 * Please make your comments here, so that it's easier for people to follow the discussion. You asked in your edit summary what my rationale was for removing certain material. My rationale was simple: I took the two most highly cited review articles that specifically discuss Burnham et al. (i.e., the 2006 Lancet study), and I summarized what they say about the study. I removed the long discussion of the minority viewpoint (of Spagat et al.), which dominated the section on the 2006 Lancet study. The section on the 2006 Lancet study was much longer than the sections on the other studies, and weirdly focused on the views of a small minority of the scientific community, while almost completely ignoring the majority view. I remedied this by just paring back the section to what two major review articles have to say. -Thucydides411 (talk) 16:10, 26 March 2019 (UTC)
 * First of all, while Tapp et al. 2008 establishes that population-based studies are the gold standard for estimating civilian mortality in wartime, it is more than a decade old and should not necessarily be used to refute more recent sources or to summarize the current academic consensus on this topic, certainly not on the basis of citation-counting on Google Scholar. Notably, the Iraq Family Health Survey (IFHS) employed the same methodology as the Lancet but produced a far lower estimate of violent deaths in Iraq, leading several experts (notably Spagat) to contend that the Lancet study had failed replication, and others (such as Tirman) to argue that the IFHS results are actually consistent with the Lancet's findings. It's quite odd to remove dozens of sources in favor of just one or two review articles, paring back the section so much that this major line of criticism is not even addressed. Your statement that these deletions are justified because "the section on the 2006 Lancet study was much longer than the sections on the other studies" is rather curious; given how widely cited the Lancet study is in the literature (both in terms of praise and criticism), why would you expect it to receive no more coverage here than any other estimate of Iraq War-related mortality?TheTimesAreAChanging (talk) 17:54, 26 March 2019 (UTC)

Dispute Resolution Process initiated
I have launched a dispute resolution process at DRN here:. In requesting resolution I have pinged these editors:. This is an important enough article that we need to "get it right," and I think this is possible. Please divert further comment to that dispute resolution page. -Darouet (talk) 19:04, 26 March 2019 (UTC)

There is a dispute resolution process going on. Rather than continuing to revert, please engage there. -Thucydides411 (talk) 03:19, 28 March 2019 (UTC)

Please consider reverting Snooganssnoogans' continuation of the edit war while the issue is in dispute resolution. At the moment there is no consensus for inclusion. Snoogans has gone to 3RR twice now in the last few days at this article. Mr Ernie (talk) 12:43, 29 March 2019 (UTC)
 * Wtf? This version is the long-standing version of the article (i.e. the consensus version). My edits have sought to restore the long-standing version. Snooganssnoogans (talk) 22:07, 29 March 2019 (UTC)


 * The only reason why this content is long-standing is because every time the issue has come up, you've repeatedly reverted to keep it in. I think there's actually consistently been a majority of involved editors opposed to the inclusion of so much criticism from Spagat, but nobody has been willing to hit "revert" so many times as you. -Thucydides411 (talk) 23:10, 29 March 2019 (UTC)

It's unfortunate that dispute resolution went nowhere. If everyone involved had presented their analysis of the sources and proposed text for the lede, then it could have gone somewhere. In any case, the discussion there and above establishes clearly that the scientific consensus views the Lancet studies as among the most rigorous studies on the subject, but that all attempts to study mortality during the war face significant difficulties. I propose the following, which is supported by the sources, for the second paragraph of the lede:
 * Estimating war-related deaths poses many challenges (Wang et al. 2016, Adhikari et al. 2010). Experts distinguish between population-based studies, which extrapolate from random samples of the population, and body counts, which tally reported deaths and likely significantly underestimate casualties (Tapp et al. 2012). Population-based studies produce estimates of the number of Iraq War casualties ranging from 151,000 violent deaths as of June 2006 (per the Iraq Family Health Survey) to over a million (per the 2007 Opinion Research Business (ORB) survey). Other survey-based studies covering different time-spans find 461,000 total deaths (over 60% of them violent) as of June 2011 (per PLOS Medicine 2013), and 655,000 total deaths (over 90% of them violent) as of June 2006 (per the 2006 Lancet study). Body counts counted at least 110,600 violent deaths as of April 2009 (Associated Press). The Iraq Body Count project documents 183,348 - 205,908 violent civilian deaths through April 2019.

The sources in parentheses will of course be converted into proper references. The entire discussion about the sources can be found here, particularly in the 3rd round of statements by the editors.

I'm pinging all editors involved at DRN:. -Thucydides411 (talk) 21:09, 4 May 2019 (UTC)
 * Yeah, that seems like a fine summary to me.TheTimesAreAChanging (talk) 08:52, 5 May 2019 (UTC)
 * It's fine for me too, but I recommend that you add the sentence "All estimates of Iraq War casualties are disputed." at the end of the paragraph, citing Hagopian 2013 and Levy and Sidel 2016. -Darouet (talk) 01:02, 6 May 2019 (UTC)


 * Okay, done: . -Thucydides411 (talk) 22:36, 7 May 2019 (UTC)


 * It's completely bonkers to characterize body counts as obviously undercounters when laughable estimates such as the 1.2 million estimate by ORB is described as one of the most credible ones. This misleads readers into thinking the credible estimates range from 151K to 1.2 million, which is inane. The body of the article makes abundantly clear that body counts undercount (something I've never disputed and something which I've never sought to edit-war out - because there is not a single reason why I'd want this page to falsely suggest that the casualties were lower than credible estimates show) AND that the ORB and 2006 Lancet estimates are highly disputed. The other studies which have not been seriously disputed should not be lumped in with ORB and 2006 Lancet studies. It's absolutely beyond why you're so intent to inflate the Iraq War casualties... going so far as to inflate it to estimates that the authors of those studies don't stand by. Snooganssnoogans (talk) 22:53, 7 May 2019 (UTC)


 * Before reverting again, please seek consensus on the talk page for your proposed lede. I proposed the new text for the lede a few days ago, and pinged all the editors who were listed at DRN. Two of them commented, largely agreeing with the proposed text. You did not reply. Then, when I make the change to the lede, you suddenly appear and revert my edit. I would have preferred to have gone over all these issues at DRN with you, where we could have discussed them one-by-one, in a structured fashion. Instead, when the mediator asked for proposed wordings and sources, while I and another editor provided what was asked for, you instead attacked us. The lede is now very well sourced, and represents the views expressed in highly-cited papers and review articles. If you have issues with it, then please provide the sourcing that backs your concerns, and show how the lede does not represent the reliable sources. -Thucydides411 (talk) 23:59, 7 May 2019 (UTC)


 * I have already explained again and again in excruciating detail for what is almost a year at this point why the long-standing version of the lede should be kept, and why the inflated casualty estimates (which you for some inexplicable reason want to present as the most credible estimates) should reflect how RS describe them: as disputed. I can not rehash the same arguments with you again and again, and it seems pointless given that you repeatedly and shamelessly lie about certain aspects (e.g. that only one academic has critiqued the Lancet study). And now you have bogglingly opted to describe the ORB estimate of 1.2 million casualties as one of the most credible estimates, which is just beyond belief. Under no circumstance whatsoever can the lede to this article state that there were 1.2 million deaths in the Iraq War. It's a laughable estimate and does not reflect credible estimates in the academic literature. It's feeding readers bullshit. Snooganssnoogans (talk) 02:00, 8 May 2019 (UTC)


 * Can you please explain why you are unwilling to do a RfC to resolve this? Why do you return to this page every 2-3 months to edit-war your fringe version into the lede and/or even worse, seek to scrub all academic sources from the body which criticize the inflated estimates? Snooganssnoogans (talk) 02:06, 8 May 2019 (UTC)


 * You're the only person arguing against the new lede. Stop trying to force through your preferred version. I suspect the reason why the other editors support the updated lede is that it represents the academic sources. You'll notice that the new lede describes all casualty estimates as "disputed," rather than the few you personally picked out - based on your original research - to denote as "disputed" (which are, on the contrary, precisely the estimates that the academic literature singles out as the most rigorous). Why didn't you comment when I asked for opinions on the new text? Instead of commenting, you just show up afterwards, see that three editors agree on the new lede, decide that your own opinion trumps that, and revert. -Thucydides411 (talk) 04:09, 8 May 2019 (UTC)


 * Saying every estimate is "disputed", and then listing the absurd ones (the inane 1.2 million estimate and the Lancet study estimate that the lead author no longer stands by) along with the credible ones, without clarifying which is which, is crazy (but in some ways consistent with your pattern of editing which for example has equated far-right conspiracy websites with reliable news sources such as the NYT and WaPo - as if they're all the same). The version of the lede that you keep edit-warring in now misleads readers that credible estimates of Iraq War casualties range up to 1.2 million. Furthermore, the lede literally says that one type of estimate definitely undercounts while failing to note that the ORB and 2006 Lancet study are widely considered overcounts. That's not reflecting the academic literature. Snooganssnoogans (talk) 12:35, 8 May 2019 (UTC)
 * Why didn't I comment immediately? Because I cannot spend the rest of my life dealing with this BS. There's nothing new here. Nothing that I haven't discussed with you before. What is it that prevents you from doing a RfC to resolve this once and for all? Why don't you want community-wide input? Snooganssnoogans (talk) 12:35, 8 May 2019 (UTC)


 * Please self-revert. You don't own this article, and you can't dictate to three other editors what will and will not be included in it. The consensus is clearly against you, and repeatedly reverting to force the article to say what you want it to say is a serious breach of Wikipedia's rules on edit warring and consensus. -Thucydides411 (talk) 18:13, 8 May 2019 (UTC)

The first two sentences in the proposed version of the lede are fine. The rest of the proposed version of the lede is not. Snooganssnoogans (talk) 22:19, 8 May 2019 (UTC)


 * I take it that you're not going to self-revert then, and that you're now in breach of WP:EW. You seem to think that the only thing that can overrule your personal opinion on what should be in the article is an RfC, and are ignoring the consensus of other editors here. In the above sections and over at dispute resolution, I and other editors have gone back to the academic literature to show what the scientific view is on the Lancet studies - that they are widely viewed by peers to be among the most rigorous studies on mortality in the Iraq War. You haven't presented sources to dispute that. Instead, you've dug out about a dozen papers by the same person, and claim that that represents the scientific consensus - instead of the review papers on the subject, which indicate otherwise. -Thucydides411 (talk) 22:43, 8 May 2019 (UTC)

RfC on lede paragraph
Which of these two paragraphs should be in the lede (note that the sources are from the body and will not be in the lede). The non-bolded parts are identical in both versions whereas the bolded parts clarify the differences between the two versions: Snooganssnoogans (talk) 22:19, 8 May 2019 (UTC)

Version A:
 * Estimating war-related deaths poses many challenges. Experts distinguish between population-based studies, which extrapolate from random samples of the population, and body counts, which tally reported deaths and likely significantly underestimate casualties. Population-based research estimates of Iraq War casualties range from 151,000 violent deaths as of June 2006 (per the Iraq Family Health Survey) to 461,000 total deaths as of June 2011 (per PLOS Medicine 2013), over 60% of them violent. Other population-based research estimates, which are disputed in the scientific community,                    such as the 2006 Lancet study and the 2007 Opinion Research Business (ORB) survey, put the numbers as high as 655,000 total deaths as of June 2006 (over 90% of them violent) and 1.2 million violent deaths as of August 2007 respectively. Body counts — which underestimate mortality — counted at least 110,600 violent deaths as of April 2009 (Associated Press). The Iraq Body Count project documents 183,249 – 205,785 violent civilian deaths through Feb. 2019.

Version B:
 * Estimating war-related deaths poses many challenges. Experts distinguish between population-based studies, which extrapolate from random samples of the population, and body counts, which tally reported deaths and likely significantly underestimate casualties. Population-based studies produce estimates of the number of Iraq War casualties ranging from 151,000 violent deaths as of June 2006 (per the Iraq Family Health Survey) to over a million (per the 2007 Opinion Research Business (ORB) survey). Other survey-based studies covering different time-spans find 461,000 total deaths (over 60% of them violent) as of June 2011 (per PLOS Medicine 2013), and 655,000 total deaths (over 90% of them violent) as of June 2006 (per the 2006 Lancet study). Body counts counted at least 110,600 violent deaths as of April 2009 (Associated Press). The Iraq Body Count project documents 183,348 - 205,908 violent civilian deaths through April 2019. All estimates of Iraq War casualties are disputed. 

Please indicate whether you support version A, B or some other version, along with your reasoning. Snooganssnoogans (talk) 22:19, 8 May 2019 (UTC)

Survey

 * Version B: Version A's contention that some estimates are uncontroversial, while others are not, is pure original research on the part of . In fact, the studies that Snooganssnoogans has decided to single out as "disputed" are precisely the studies which the academic literature indicates are the most rigorous: "Of the population-based studies, the Roberts and Burnham studies provided the most rigorous methodology as their primary outcome was mortality" (Tapp et al. 2008, "Iraq War mortality estimates: A systematic review". Conflict and Health. 2(1). doi:10.1186/1752-1505-2-1), "Although the Roberts and Burnham studies faced some criticism in the news media and elsewhere, part of which may have been politically motivated, these studies have been widely viewed among peers as the most rigorous investigations of Iraq War–related mortality among Iraqi civilians; we agree with this assessment and believe that the Hagopian study is also scientifically rigorous" (Levy & Sidel 2016, "Documenting the Effects of Armed Conflict on Population Health". Annual Review of Public Health. 37:205–218. doi:10.1146/annurev-publhealth-032315-021913). Version A incorrectly describes the Lancet studies as particularly disputed, despite them being described in the academic literature as among the most rigorous studies on Iraq War mortality. Version B is neutral, does not describe any one survey-based estimate as more reliable than another, and notes that all casualty estimates are difficult and disputed. Version A's assessment that particular studies are more reliable than others is based only on one editor's personal opinion, and contradicts the sources. I encourage editors commenting here to take the time to look over the discussion above, where an attempt has been made to look through the academic literature in a systematic fashion. In contrast, Version A has been loaded up with a smattering of cherry-picked articles that: 1. Do not support the proposed text in Version A, or 2. Rely heavily on the work of one critic of the Lancet studies, Michael Spagat, whose work has not been widely cited in the academic literature. Version B is based on the most widely cited academic literature that deals with the Lancet studies, and on review articles in prestigious journals (e.g., Annual Reviews). -Thucydides411 (talk) 22:34, 8 May 2019 (UTC)
 * "Rely heavily on the work of one critic of the Lancet studies, Michael Spagat" - By my count (just from the sources above), there are at least 40 recognized experts who describe the Lancet study as widely disputed and/or point to serious flaws in it. Snooganssnoogans (talk) 23:10, 8 May 2019 (UTC)
 * "Version A's contention that some estimates are uncontroversial" - this version does not do that. It simply reflects what reliable sources say about the estimates. The body of the article does not substantiate that all the estimates are widely disputed. The body makes clear that the body counts are too low (per countless RS) and that the high 2006 Lancet and 2007 ORB estimates are widely disputed (per countless RS). If other estimates are widely disputed, then the lede can be updated to take account of that. Snooganssnoogans (talk) 23:16, 8 May 2019 (UTC)


 * Most of the sources that you cited don't actually say what you're claiming they say. We've gone over this before. You've thrown in references that don't even comment at all on the reliability of the Lancet estimates, simply because you believe that those references somehow undermine the Lancet studies. Almost all the articles that directly criticize the Lancet studies are by Michael Spagat, and they're hardly cited in the academic literature. This is original research on a grand scale. By contrast, Version B's sources are only review articles that directly assess the reliability of the Lancet studies, and the most highly-cited academic sources that deal with the Lancet studies.
 * Version A does absolutely call some estimates controversial and others uncontroversial. When you state that "estimates range from X to Y, but other, disputed estimates say Z," you're singling out Z as particularly disputed or controversial. "the high 2006 Lancet and 2007 ORB estimates are widely disputed" - the academic literature actually says that the 2006 Lancet study is among the most rigorous studies. Your proposed lede contradicts that academic literature, in favor of your own personal view. -Thucydides411 (talk) 23:28, 8 May 2019 (UTC)


 * The cited sources all characterize the Lancet study as widely disputed and/or point to serious flaws. If any of them don't, point them out so we can remove them. Snooganssnoogans (talk) 23:47, 8 May 2019 (UTC)


 * I've already gone through them before with you. Marker (2008) is a detailed analysis of the strengths and weaknesses of several different estimates of Iraq War mortality, and it explicitly states that it is not forming its own judgment on whether or not those estimates are accurate. Its heaviest criticism is actually directed against the IFHS estimate, which was conducted by the Iraqi Health Ministry, which had links to a Shiite militia leader at the time. Yet you're using this source to back up your claim that the Lancet studies are particularly disputed. Rosenblum and van der Laan (2009) is a highly technical paper about a new statistical method, and doesn't give a judgment on the 2006 Lancet study. Then there is the entire mass of different articles by Michael Spagat, which I won't go into - it's clear that he criticizes the Lancet studies. It looks like your methodology for selecting studies to cite was not to look at review articles on the subject, or the most highly cited articles that discuss the Lancet studies. Rather, it looks like you went to the articles by Michael Spagat, which are rarely cited, and then looked for the few articles that cite his work. That's not a rigorous way to evaluate what the academic community thinks about the Lancet studies. In contrast, the above sections "" and "" conduct a systematic overview of the literature, and that is what Version B is based on. -Thucydides411 (talk) 00:07, 9 May 2019 (UTC)
 * (1) The Marker paper did criticize methodological choices made in the Lancet study. RS literally say the Marker paper criticized the Lancet study. The Marker paper does not direct heavy criticism against the IFHS estimate. The strongest criticism against the IFHS estimate is one it has in common with the Lancet study: namely that surveyors can't be sent into the most violent areas (which Marker describes as "understandable"). (2) Rosenblum and van der Laan show that the confidence intervals in the Lancet study are inappropriate, and lead to flawed results. As a result, both studies should be kept. And this is my final response (so as to not WP:BLUDGEON the process). Snooganssnoogans (talk) 00:41, 9 May 2019 (UTC)


 * Marker (2008) discusses both strengths and weaknesses of the methodology used by several different estimates of Iraq War mortality. That includes both critiques of specific choices made in the 2006 Lancet paper and other, as well as praise for different aspects of their methodology. The point of Marker (2008) is not to "discredit" or "criticize" the Lancet study, but rather to propose ways in which future studies of mortality can be improved. It specifically discusses how the 2006 Lancet paper's methods were considered state-of-the-art at the time. Using this paper to support your idea that the 2006 Lancet study is more disputed than any of the other studies is just original research - such a claim is not found in the paper, and is in fact flatly contradicted by the review articles on the subject. Every study ever published has flaws, but you're arguing, in flat contradiction to the review articles on Iraq War mortality, that the 2006 Lancet study is particularly flawed and disputed. The review articles say exactly the opposite - that the 2006 Lancet study is among the most rigorous on the subject. -Thucydides411 (talk) 01:40, 9 May 2019 (UTC)


 * Version A. The difference between the two versions is whether or not to characterize the 2006 Lancet study and 2007 ORB survey as "disputed" or to simply lump them in with the credible estimates without further clarification. By failing to clarify that these two estimates are widely disputed in the scientific literature, we mislead readers into thinking credible estimates put the number of casualties as high as 1.2 million. At least 25 reliable sources (either peer-reviewed studies or expert assessments by recognized experts in the field), #4-#32 characterize the 2006 Lancet study as widely disputed and/or highlight serious flaws in the study. This includes giants in the field of quantitative study of conflict, such as Andrew Gelman (who describes the study as "discredited"), Nils Petter Gleditsch (who said the study had "major biases" which resulted in an overcount) and Lars-Erik Cederman. The main study critiquing the Lancet study won "Article of the Year" by the Journal of Peace Research, which is the leading journal for quantitative research on conflict and hosted by PRIO, with the jury of leading scholars in the field declaring that the study had "convincingly" showed the Lancet estimate to have significantly overestimated the number of casualties. The lead author of the Lancet study, Gilbert Burnham, no longer stands by the estimate, as he published a separate study in 2013, which produced a far lower casualty estimate. Since the Lancet study has been widely disputed in the academic literature, we should describe it as "disputed" to reflect what RS are saying. As for the 2007 ORB survey estimate, the 2007 ORB survey is so bad that a Oxford University Press textbook on the science of counting war casualties says it shouldn't be taken seriously and that the estimate has largely been ignored. Snooganssnoogans (talk) 23:03, 8 May 2019 (UTC)
 * There are a whole number of problems with your above statement.
 * Andrew Gelman admits in the interview you're citing that he is not an expert in the issue. He hasn't published on Iraq War mortality. He's a well known statistician, but in this context, he's commenting on something outside his area of expertise, and his comments are at odds with what the scientific literature says.
 * Winning an "article of the year" award from some organization or another does not change the fact that the article in question has hardly been cited in the academic literature, and that review articles on the subject take a very different view on the Lancet studies.
 * The claim that "Gilbert Burnham no longer stands by the estimate" is an outright falsehood. Hagopian et al. (2013), which you're referencing, cites the Lancet study positively, and quite contrary from how you're portraying things, Hagopian et al. (2013) explores reasons why their new estimate might have missed a significant number of deaths (e.g., households that suffer casualties breaking up). The imputation that Burnham no longer stands by the estimate is entirely your original research, and directly contradicts what he himself has said.
 * "Since the Lancet study has been widely disputed in the academic literature, we should describe it as "disputed" to reflect what RS are saying". The academic literature explicitly contradicts this statement. Levy & Sidel (2016), in the prestigious Annual Reviews series, says that the Lancet studies are widely viewed among academic peers as the "most rigorous" studies on the subject. The highly cited Tapp et al. (2008) also calls the Lancet studies the most rigorous studies on the subject. You've cherry-picked work largely by one person, Michael Spagat, which is rarely cited by researchers working on the subject, and decided that that overrules what the rest of the academic field says.
 * -Thucydides411 (talk) 23:20, 8 May 2019 (UTC)
 * The "some organization or another" is in this case the Peace Research Institute Oslo, a premier research institution in the field of quantitative research on conflict, and the jury was comprised of heavyweights in the field of quantitative social social science such as Lars-Erik Cederman and Sara McLaughlin Mitchell. The Hagopian et al. (2013) study (which the lead author of the Lancet study signed onto and which estimated far lower casualties) does not cite the Lancet study positively or negatively - it just cites it. It outright states "Our methods may have avoided biases that served to overrepresent deaths in the [Lancet study]." As for your final comment, when 2-3 studies and experts say the Lancet study is rigorous but 25+ say it's widely disputed and/or seriously flawed, then the study is "disputed". Note that "disputed" can include both praise and criticism. Snooganssnoogans (talk) 23:39, 8 May 2019 (UTC)


 * You're talking about one random prize a paper you like received because, as you know, that paper has not been widely cited, and it takes a very different view from papers that are much more heavily cited, and which appeared in more prestigious journals (e.g., Annual Reviews, which publishes papers that summarize the state of research in various fields).
 * Hagopian et al. (2013) specifically discusses reasons why their new estimate might be biased low. As to biases in other methods that it mentions, it does not single out the Lancet studies - in fact, that very same sentence references IFHS, which you are trying to single out as supposedly being more credible than the 2006 Lancet study.
 * The key point in this is that, as shown above, in the section, the Lancet studies are overwhelmingly viewed positively in the academic literature. You weren't able to reply in that section to the systematic source analysis that did, but are instead grasping at individual, low-citation-count papers to justify the text that you want. -Thucydides411 (talk) 23:48, 8 May 2019 (UTC)


 * Studies that solely point out methodological flaws in a single study do not tend to get cited a lot (even in cases when they just straight-up murder the original study, e.g. see these two recent prominent examples that come to mind). If you think they do, you seem to misunderstand why scholars usually cite things and how citations in science work. And I did respond to Darouet. And this will likely be my last response, so as to not WP:BLUDGEON the process. Snooganssnoogans (talk) 00:10, 9 May 2019 (UTC)


 * You've been rationalizing this entire time why we should ignore citation counts. The 2006 Lancet study is the most highly cited study on the subject. You say we should ignore that, because even flawed studies can be highly cited. One editor then went out and looked up how the 2006 Lancet study is referenced in the literature, and found that the most highly cited papers that reference the 2006 Lancet study do not treat it as discredited - but rather use it to positively reference points they are making. That editor then also went and looked at how review articles on the subject of Iraq War casualties refer to the 2006 Lancet study, and found that they explicitly say that the 2006 Lancet study is among the most rigorous, and that it is widely viewed as such among academic peers. Against all that, you're referencing low-citation-count articles largely written by one person, Michael Spagat, and saying that both citation counts and review articles don't matter. I understand fairly well how citations are used in scientific literature, and your contention that they're irrelevant is just not true. -Thucydides411 (talk) 01:46, 9 May 2019 (UTC)

Note: I advertised the RfC on the NPOV noticeboard to get more input. Snooganssnoogans (talk) 01:36, 9 May 2019 (UTC)

Version B: The overall text is already unwieldy for a lead, and unless there is overwhelming agreement among sources that some estimates are unreliable, the proper place for expressing doubts about ANY sets of figures, is in the body. Casualty estimates in almost all conflicts are controversial and will often vary for many reasons, including legitimate variations in methodology, and/or flawed methodology. The purpose of the lead should be to simply communicate clearly what the range of estimates is, and by who/when they were created. Version A seems to be trying to put too much info into an already unwieldy mass of figures. Pincrete (talk) 18:30, 10 May 2019 (UTC)

Version B Based on the previous discussion it appears that calling out certain sources as unreliable is cherry-picking at best. Thucydides has, IMO, argued convincingly that the Lancet study is not widely viewed as unreliable relative to other studies on the subject. LokiTheLiar (talk) 06:59, 11 May 2019 (UTC)

Version B In the discussion above, Darouet and Thucydides convincingly demonstrate above that the Lancet paper is considered a high quality paper by experts in the field. Its citation count and the citation counts of other papers relying on it are very high. The most influential discussions in the literature consider it authoritative, and so we shouldn't be singling it out as disputed, especially not in the lead. If there is significant criticism, that can be covered in the article body (it currently appears to be over-covered in the body, but that's a separate issue). Red Rock Canyon (talk) 19:58, 11 May 2019 (UTC)

Version B, agreeing with, , , and. Two reviews of Iraq War mortality estimates describe the Lancet papers as the gold standard in the field, supporting their extraordinarily high citation counts (over 1200 together): Levy and Sidel (2016) Annual Reviews Tapp (2008) Conflict and Health Lastly, the much lower IFHS mortality estimate that version A would promote is considered less credible by academic sources because it substituted body count data — known to be too low — in the most dangerous and therefore highest mortality areas of Iraq, an issue reviewed by many experts for instance here. Because that's behind a paywall, I'll quote Dr. Luquero and Dr. Grais from Epicentre in Paris, a part of Médecins Sans Frontières: IFHS "imputed missing data for [high mortality governorates, (HMGs)] using the Iraq Body Count, which probably underweights the HMGs as a result of publication bias..." Dr. Gulden at the University of Maryland also notes this point and adds that because the IFHS study was run by Iraqi "government-affiliated survey takers," mortality estimates will further be depressed. -Darouet (talk) 21:16, 12 May 2019 (UTC)


 * NEITHER - do not want to add either of these, they both seem like OR writings and not LEAD summaries and the current lead doesn’t need this. Cheers Markbassett (talk) 02:18, 14 May 2019 (UTC)


 * Version B Feels more clear and concise (atleast for an article this controversial) --NikkeKatski &#91;Elite&#93; (talk) 13:07, 16 May 2019 (UTC)


 * (summoned by RFC bot) Version B as more organized and neutral (it says all estimates are disputed, which does make sense, whereas v. A singles out some). Staszek Lem (talk) 19:53, 17 May 2019 (UTC)
 * Like Markbassett I don't like either option, and I cannot claim to have properly digested either option, both being practically unreadable, so what either might be doing in the lede in particular, is beyond me; I would find either one hard to read even in the body. And anyway, whether lede or not, that string of refs in A looks like satire. If the refs really are relevant, they should go into some sort of table of refs, and not into running text. But if one option rather than the other is to be chosen, B seems a little better than A, I think &mdash; I doubt that anyone would read either paragraph properly anyway. JonRichfield (talk) 07:59, 25 May 2019 (UTC)