Bogus Lancet Study

Via The Command Post comes this study published in Lancet (free reg) which purports that 100,000 Iraqi have died from violence, most of it caused by Coalition air strikes, since the invasion of Iraq. Needless to say, this study will become an article of faith in certain circles but the study is obviously bogus on its face.

First, even without reading the study, alarm bells should go off. The study purports to show civilian casualties 5 to 6 times higher than any other reputable source. Most other sources put total combined civilian and military deaths from all causes at between 15,000 to 20,000. The Lancet study is a degree of magnitude higher. Why the difference?

Moreover, just rough calculations should call the figure into doubt. 100,000 deaths over roughly a year and a half equates to 183 deaths per day. Seen anything like that on the news? With that many people dying from air strikes every day we would expect to have at least one or two incidents where several hundred or even thousands of people died. Heard of anything like that? In fact, heard of any air strikes at all where more than a couple of dozen people died total?

Where did this suspicious number come from? Bad methodology.

From the summary:

Mistake One:

“A cluster sample survey was undertaken throughout Iraq during September, 2004”

It is bad practice to use a cluster sample for a distribution known to be highly asymmetrical. Since all sources agree that violence in Iraq is highly geographically concentrated, this means a cluster sample has a very high chance of exaggerating the number of deaths. If one or two of your clusters just happen to fall in a contended area it will skew everything. In fact, the study inadvertently suggests that this happened when it points out later that:

“Violent deaths were widespread, reported in 15 of 33 clusters…”

In fact, this suggest that violent deaths were not “widespread” as 18 of the 33 clusters reported zero deaths. if 54% of the clusters had no deaths then all the other deaths occurred in 46% of the clusters. If the deaths in those clusters followed a standard distribution most of the deaths would have occurred in less than 15% of the total clusters.

And bingo we see that:

“Two-thirds of all violent deaths were reported in one cluster in the city of Falluja”

(They also used a secondary grouping system (page 2, paragraph 3) that would cause further skewing.)

Mistake Two:

“33 clusters of 30 households each were interviewed about household composition, births, and deaths since January, 2002.”

Self-reporting in third-world countries is notoriously unreliable. In the guts of the paper (page 3, paragraph 2) they say they tried to get death certificates for at least two deaths for each cluster but they never say how many of the deaths, if any, they actually verified. It is probable that many of the deaths, especially the oddly high number of a deaths of children by violence, never actually occurred.

So we have a sampling method that fails for diverse distributions, at least one tremendously skewed cluster and unverified reports of deaths.

Looking at the raw data they provide doesn’t inspire any confidence whatsoever. Table 2 (page 4) shows the actual number of deaths reported. The study recorded 142 post-invasion deaths total with with 73 (51%) due to violence. Of those 73 deaths from violence, 52 occurred in Falluja. That means that all the other 21 deaths occurred in one of the 14 clusters where somebody died, or 1.5 deaths per cluster. Given what we know of the actual combat I am betting that most of the deaths occurred in three or four clusters and the rest had 1 death each. Given the low numbers of samples, one or two fabricated reports of deaths could seriously warp the entire study.

At the very end of the paper (page 7, paragraph 1) they concede that:

“We suspect that a random sample of 33 Iraqi locations is likely to encounter one or a couple of particularly devastated areas. Nonetheless, since 52 of 73 (71%) violent deaths and 53 of 142 (37%) deaths during the conflict occurred in one cluster, it is possible that by extraordinary chance, the survey mortality estimate has been skewed upward. “

Gee, you think? It’s almost as if military violence is not randomly distributed across the population of Iraq but is instead intelligently directed at specific areas, rendering a statistical extrapolation of deaths totally useless.

In the next paragraph they admit:

“Removing half the increase in infant deaths and the Falluja data still produces a 37% increase in estimated mortality.”

That puts their final numbers just above the high end of the range reported by other sources.

This “peer reviewed study” is a piece of polemical garbage. Everybody is supposed to take away the bumper sticker summary, “Coalition kills 100,000 Iraqi civilians, half of them children,” without reading the details. It tries to use crude epidemiological models like those used to study disease and applies them to the conscious infliction of violence by human beings. The result is statistical static.

(Update: Commentator Clashman below points out that the studies “conservative” estimate is actually around 66,000 instead of the 30,000 I had done in my head so the study is actually at least twice of what other sources place as the upper range at around 25,000)

(Update: Further related thoughts in )

(Update: I have posted more thoughts on evaluating the methodology of the study)

244 thoughts on “Bogus Lancet Study”

  1. Adding to the skepticism are two facts:

    The lead author was an opponent of the war, and

    The lead author submitted it to the Lancet on the express condition that it be published before the election.

    Do you suppose the guy might have a bit of bias of his own?

  2. Why does the medical profession do this to itself? People just love to trade on the good name of their profession (accumulated by someone else, I might add) to score cheap political points. Just like good leftists, always spending someone else’s money.

    I just read a trade paper (I’m a shrink) where the APA is actually considering a proposal to classify “racism” as a mental disorder! It’s truly infuriating.

  3. Does no one remember the UN said 4000 Iraqi children five and under were dying every month before the war?
    They were dying because of malnutrition and lack of medical care.
    That was before the liberation.
    That was because of the trade restrictions.
    Right.
    One might think it was because certain UN folk were doing a great job of feathering their nests.

  4. Good Stuff!

    Thanks for providing this information for my brain to process and make a more informed opinion of what I think about this.

    I already had my doubts.

  5. I’m sorry, but it sounds like you are contradicting yourself.

    You say that deaths shouldn’t be widespread, and that “most of the deaths would have occured in less than 15% of the clusters” (less than 4 clusters).

    And then you say they report that violent deaths occured in just one cluster (less than 4, certainly).

    Are you saying that their use of the word widespread, when they clearly indicate the total distribution of deaths, was wrong-headed?

    I take most objection to your comment that we haven’t heard reports of more deaths.

    The facts are that many, many stories aren’t getting reported. Journalists aren’t told by the military to stroll around. They are told that if they leave certain areas, their lives are in their own hands.

    Your wild guess of Iraqi military dead is entirely unsubstantiated.

  6. When I read the news account, I had problems, many of which you stated quite well. You also raised some other issues which I should have thought of, but thank you for making them. You have hopefully made the electorate a little smarter.

  7. godfodder,

    It is great news that the APA is considering clasifying racism as a mental disorder. This will prevent racists from being fired from their jobs as they will now have ADA protection. Further, given the notorious lack of success in dealing with such previous mental disorders as homosexuality, racists will have no incentive to change. Szasz was right.

  8. after taking classes in stats, conducting medical research, and learning about how you cannot draw conclusions on ridiculous stats, i find this report in the Lancet abominable.

    if the Lancet held the same statiscal standard to drugs currently in testing, or drugs that you and i will be taking, then 50 percent of all people who take the drugs would either die or be incapacitated.

    this publication has just denigrated itself by infusing such hilarious mystifying conclusions from statisitics that cannot even pass a simple logic test.

    what a shame.

  9. They are told that if they leave certain areas, their lives are in their own hands.

    Well, duh. I here that from police when I go to North Minneapolis.

  10. This story surfaced before and then and now seems tied to a web site (http://www.iraqbodycount.net/) that is part of a circuitous linkage of resources that reference each other in one large daisy chain, with each supporting the previous one’s data.

    Try tracking the related sources yourself and see if its still the case.

  11. Have a look at this report from UNICEF

    http://www.unicef.org/newsline/99pr29.htm

    To take some figures:

    85% of the population in the survey
    24000 households surveyed
    In 1999, there were 131/000 under 5s mortality
    In 1999, infant mortality (under 1 yr) was 108/000
    In 1999, maternal mortality was 29/000

    In the Lancet publication:

    Only 800 odd households surveyed
    Before war, infant mortality was 29/000
    After war, infant mortality 57/000

    Clearly bombing is good for the Iraqis, it halves their their infant
    mortality rate.

    Or you might like to think that this was a partisan puff piece put out as an
    “October Surprise” for the US election. Certainly, it would be sensible to
    treat it with sceptisism.

    JC

  12. Pingback: Who Knew?
  13. Speaking as an epidemiologist, I wouldn’t even spit on this study, let alone cite it. What idiots! Not just the authors, but anyone who takes the study seriously.

  14. Brice Tebbs,

    15 to 20 thousand people in 16 months is less than were being killed violently by Saddam Hussein. There is no way to get rid of an established totalitarian regime by force without killing 20 to 50 thousand people. The militant core of people that benefit from the regime are not going to give up until they see enough of their comrades killed to convince them to give up.

    During the civil war General Sherman reached this conclusion about the confederacy. He decided that the Union needed to kill, I believe, 200,000 confederates to get enough of the militant core of confederate leaders to end the war with a victory for the North.

    Because we have invested in military technolgy such as precision bombs and air survellience. We can do this a lot more cleanly now than in past wars. But there is no way the Baathists are going without tens of thousands of deaths. It is still better for Iraq than having them in power though.

  15. A few other notes about the study:

    -On author bias – when discussing what could have gone wrong, they routinely mention how they could be under-estimating the rate, and rarely mention how they could be over-estimating the rate.

    -Did they check their model assumptions? It looks like they dumped their data into some canned software – tends to be pretty unreliable.

    -They seem to be focusing on urban areas. What about rural death rates?

    -They assume an increase in death rate from the invasion. While this can also fit no increase, this does tend to make one question parts of the study – especially the confidence intervals

    -The people conducting the study were scared to travel around the country. How does this affect stuff?

    -Look at the chart 2/3rds of the way through with monthly death figures. Just think about it for a bit……

    Plus what the original post said.

    Mike

  16. Gads, that study wouldn’t even rate a D if it were submitted to me as a Grad school prof. The Lancet is following Scientific American into the ranks of pseudo-science. They better learn some math — fast!!

  17. Interesting analysis. Thanks for doing the work. Cliche warning – cliche warning: Statistics lie, and liars use statistics. War is hell. People are stupid.

  18. The real “health problem” is the “European disease” that has been endemic over there for 2,000 years, The soul-sickness to which I refer is, of course, anti-Semiticism. Charles at LGF has a number of excellent references exposing the political agenda infecting Lancet and many other elitist publications. Sometimes the rabidly-ill disguise their sick thinking with a cosmetic application of “concern for the suffering of the Palestinians” (in the case of anti-Israel propanganda). In this case, they are using “concern for Iraq” as cover for the fact that a primary reason they hate U.S. is that it is Israel’s greatest ally.
    Oh well, we had to go over and administer a strong dose of military medicine in order to cure their last outbreak of the European scourge (also called WW II). Now the disease is exacerbated by Islamofascism. Come Tuesday, we will find out if our current generation is up to the task. I hope and pray that we are.

  19. You can’t make this stuff up. Only they did.

    It is really getting bad here in Western Civilization Land. First we lost religion. Then art falls to pieces. And now even science is biting the dust.

    I’m with Ann Coulter on this one. Reasonable people have been playing nice for too long. They want a street fight, it’s time for a street fight. Only civilization itself depends upon it.

  20. “The number of 100,000 dead so far in Iraq in this peer reviewed study is the best estimate we’re likely to get since the Bush administration refuses even to engage in a count.” –Alterman

  21. “”Removing half the increase in infant deaths and the Falluja data still produces a 37% increase in estimated mortality.”

    That puts their final numbers just above the high end of the range reported by other sources.”

    Ummm, no it doesn’t. That puts their numbers at 66,600 dead.

    37% higher than an annualized mortality rate of 5/1,000 is 6.85/1,000.

    Now, if we take 24 million Iraqis as a given here, it should look something like this.

    Pre-War
    24,000,000 Iraqis/1,000=24,000×5=120,000×1.5 years=180,000 total deaths.

    Post-War
    24,000,000 Iraqis/1,000=24,000×6.85=164,400×1.5 years=246,600 total deaths.

    Difference=66,600

    Since when is that “just above the high end of the range reported by other sources”?

  22. Clashman–

    I couldn’t follow your math or your point. But this study is absolute crap. It shows a relative risk sans Fallujah of 1.5, which is one half the level accepted by leading medical journals for publication. Even if their numbers aren’t wildly skewed by sampling bias and the kazillion confounders they can’t control for, they are not statistically relevant. An RR of 1.5 means their data only “explains” 1/3 of the deaths. The other two-thirds are explained by something else.

    I call B.S. The Lancet “researchers” should be ashamed of themselves.

  23. Mortality before and after the 2003 invasion of Iraq: cluster sample survey

    Summary Background
    In March, 2003, military forces, mainly from the USA and the UK, invaded Iraq. We did a survey to compare mortality during the period of 14·6 months before the invasion with the 17·8 months after it.

    Methods
    A cluster sample survey was undertaken throughout Iraq during September, 2004. 33 clusters of 30 households each were interviewed about household composition, births, and deaths since January, 2002. In those households reporting deaths, the date, cause, and circumstances of violent deaths were recorded. We assessed the relative risk of death associated with the 2003 invasion and occupation by comparing mortality in the 17·8 months after the invasion with the 14·6-month period preceding it.

    FindingsThe risk of death was estimated to be 2·5-fold (95% CI 1·6–4·2) higher after the invasion when compared with the preinvasion period. Two-thirds of all violent deaths were reported in one cluster in the city of Falluja. If we exclude the Falluja data, the risk of death is 1·5-fold (1·1–2·3) higher after the invasion. We estimate that 98,000 more deaths than expected (8000–194000) happened after the invasion outside of Falluja and far more if the outlier Falluja cluster is included.

    The major causes of death before the invasion were myocardial infarction, cerebrovascular accidents, and other chronic disorders whereas after the invasion violence was the primary cause of death. Violent deaths were widespread, reported in 15 of 33 clusters, and were mainly attributed to coalition forces. Most individuals reportedly killed by coalition forces were women and children. The risk of death from violence in the period after the invasion was 58 times higher (95% CI 8·1–419) than in the period before the war.

    Interpretation
    Making conservative assumptions, we think that about 100,000 excess deaths, or more have happened since the 2003 invasion of Iraq. Violence accounted for most of the excess deaths and air strikes from coalition forces accounted for most violent deaths. We have shown that collection of public-health information is possible even during periods of extreme violence. Our results need further verification and should lead to changes to reduce noncombatant deaths from air strikes.

    Published online October 29, 2004 http://image.thelancet.com/ extras/04art10342web.pdf

    Center for International Emergency Disaster and Refugee Studies, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA (L Roberts PhD, G Burnham MD); Department of Community Medicine, College of Medicine, Al-Mustansiriya University, Baghdad, Iraq(R Lafta MD, J Khudhairi MD); and School of Nursing, Columbia University, New York, NY, USA (ProfRGarfieldDrPH)

    Correspondence to: Dr Les Roberts les@a-znet.com Mortality before and after the 2003 invasion of Iraq: cluster sample survey Les Roberts, Riyadh Lafta, Richard Garfield, Jamal Khudhairi, Gilbert Burnham

  24. “Our findings need to be independently verified with a larger sample group. However, I think our survey demonstrates the importance of collecting civilian casualty information during a war and that it can be done,”

    said lead author Les Roberts, PhD, an associate with the Bloomberg School of Public Health’s Center for International Emergency, Disaster and Refugee Studies.

  25. Before you all congratulate yourself too much, I should point that Ms. Love apparently didn’t read the Lancet paper in question. Or, more likely, she read it but didn’t understand what she was reading.

    Case in point: Ms. Love makes the following grand pronouncement (and with such flourish!):

    “And bingo we see that:

    ‘Two-thirds of all violent deaths were reported in one cluster in the city of Falluja'”

    What Ms. Love failed to comprehend (or maybe did comprehend but decided to leave out), is that the Fallujah data is an outlier, and is not included in the estimated 100,000 dead. To quote from the abstract:

    “Two-thirds of all violent deaths were reported in one cluster in the city of Falluja. If we
    exclude the Falluja data, the risk of death is 1·5-fold (1·1–2·3) higher after the invasion. We estimate that 98 000 more deaths than expected (8000–194 000) happened after the invasion outside of Falluja and far more if the
    outlier Falluja cluster is included.”

    Ms. Love’s faulty analysis wouldn’t make it into any refereed journal. It’s easy to post it on a blog, and fool some people into thinking that she actually has something to say.

  26. The Shorter Mortality before and after the 2003 invasion of Iraq: cluster sample survey
    The researchers conducted their survey in September 2004. They randomly selected 33 neighborhoods of 30 homes from across Iraq and interviewed the residents about the number and ages of the people living in each home. Over 7,800 Iraqis were included. Residents were questioned about the number of births and deaths that occurred in the household since January 2002. Information was also collected about the causes and circumstances of each death. When possible, the deaths were verified with a death certificate or other documentation.

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

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

  27. Of course liberations spill blood. But given the record of the Husseins and the UN’s oil-for-starvation-and-embezzlement program, the liberation of Iraq drastically reduced the number of needless innocent deaths over the next thirty years. The first 30 years of the Husseins cost a million lives. The likelihood that the next 30 years will come anywhere near that is extremely small. Quibble over 20K or 100K all you want. The bottom line is that life was preserved by the liberation. And this leaves aside the important point that it’s life of greater liberty.

  28. Just wanted to add that in the study, it says that many of the male violent deaths may be combatant deaths…
    That means that any number of the males who were recorded as having died violently, could have been killed as a military person or insurgent, aka not civilian.

    Their 100K figure includes these “maybe combatants”, yet they still say on the title “civilians”.

    Why? How can they say that 100K civilians have died, when in fact 50K of them could have been non-civilian combatants?

    Anyone notice their range for that number of possible excess deaths caused by the war?

    8K-194K? Isn’t that a bit of a large range??

    This survey is a sham, and the timing of it only goes further to prove it.

  29. William Trippe,

    I didn’t mean to include the Falujah to downgrade the 100,000 figure but to show that cluster-sampling with an asymmetrical distribution results in garbage. Violence in Iraq, especially air strikes, occur in very geographically focused areas down to specific neighborhoods in specific cities. Cluster sampling is not the kind of sampling you want to use in this situation. One or two clusters could throw off everything.

    Excluding Faluja, Table 2 in the paper shows that their “conservative estimate” of 51,000 was extrapolated from the unverified reports of 21 deaths in 14 clusters. That’s bullshit.

    This kind of methodology is not what we want to base public policy on.

  30. Pingback: Say Anything
  31. William Trippe wrote:

    Ms. Love’s faulty analysis wouldn’t make it into any refereed journal. It’s easy to post it on a blog, and fool some people into thinking that she actually has something to say.

    I would turn that statement around: the neat thing about blogs is that a reader who finds problems with a post can provide instant feedback in the comments section, and the author can then either refute the commenter’s argument or correct his own. How many months would such a feedback process require in a refereed journal? Some of the most able people who post on blogs couldn’t even get an article considered for publication in a refereed journal.

    It seems to me that blogs are an extremely valuable supplement to traditional means of vetting and disseminating information.

  32. Isn’t anyone else bother by the incredibly wide 95% confidence intervals? Come on now, they are 95% certain that the real war-related death increase lies somewhere between 8,000 and 98,000? That’s awful, right?

  33. I just heard the man who made this “100’000” report.
    He contradicted himself the whole interview (from memory):
    “journalist cnn”: You took your sample in bagdhad & Fallhoujah ?
    “Pr. 100’000”: No, it was too high numbers. We didn’t used them
    “Journalist”: How do you explain the difference between your report and the witness/journalists/humanitarians reports ?
    “Pr. 100’000”: Witness like journalist cannot go anywhere in Falloujah & Bagdad beccause it’s too dangerous.

    First he said that Bagdad/Falloujah & other dangerous places of the sunni triangle were not used, but after he contradict himself and suggest that the whole report is based on sunni triangle…

    I almost saw Michael Moore’s arm moving the puppet ;-)

  34. Brice,

    How many lives could you have saved if you had bombed Hitler in Berlin in 1935? Suppose it had meant 10,000 dead and massive destruction to the capital of Germany? Would it have been worth it?

    Now who in their right mind in 1935 would have given the OK for such an operation?

    Just as 7 Dec 1941 changed the weightings in the calculations of war and peace for some, so did 11 Sept 2001. For some of us, risks that we once dismissed now loom larger.

    Sorry to say 10 Sept is not coming back for many decades. The rule now is: shoot first and ask questions afterwards. If this scares our enemies it is a good thing. If it only kills them it is a good thing. If 10,000 innocents die to prevent the start of a nuclear war it is a good thing. Even if the knowns are indeterminant.

    9/11 changed the weightings of the risks for some of us. For others it will always be 1935.

    Did I mention that war sucks?

    Tell it to the jumpers of 11 Sept. See if you get an answer.

  35. I did an analysis here, some points basically the same as yours, plus one more… you can get a very rough estimate of the number of combatant deaths in their sample from the sex ratio… it matches other known estimates.

    http://obsidianorder.blogspot.com/2004/10/pick-number-any-number.html

    After doing that, the non-combatant death count directly due to coalition forces is also similar to other estimates, 5-10,000. Aside from that, obviously the sample is too small and the clustering adds a lot of variability and all the other problems you mention. Yeah, the study is politically motivated trash.

  36. Ms. Love,

    As an author of papers published in peer-reviewed journals, I was struck by the extraordinarily compressed time-line of this publication. Readers outside the biomedical fields might consider what the peer-review process involved:

    1. Data were collected in September 2004, and the authors had completed compilation, statistical analysis, drafting of text, artwork, and proofreading in order to submit their work in the form of a for-publication draft manuscript (MS) to the Lancet Editor.

    2. The Editor read the MS, chose peer-reviewers, had the reviewers comment on the MS, evaluated these comments, passed his/her favorable judgement on the MS to the authors, with any suggestions for necessary or advisable revisions.

    3. The authors revised the MS and resubmitted it.

    4. The Editor and perhaps the peer-reviewers reviewed and approved the revised text and figures. The MS files were sent to the Lancet’s copy editors for proofreading and digital typesetting. Author queries were generated and sent to the lead author, and the responses incorporated into the typeset version. Finally, the complete manuscript, ready for printing, was published on the Lancet’s website.

    Four to eight weeks is an unusually short time for a high-impact journal such as the Lancet to bring such an article into print. I would doubt that Lancet, JAMA, Nature, BMJ, Science, or similar high-prestige journals have ever compressed their review and publication schedule in such a drastic manner.

    Given the irregularity of this procedure, the serious methodological shortcomings that have already been identified, and the misleading and politicized interpretation of the study’s results in its summary, it strikes me that readers would be justified in demanding to take a closer look at what the “peer-review” process actually involved in this case. The authors of the study and the Lancet’s Editor should agree to release the correspondence and email between the lead author and editior. In particular, readers are entitled to see the peer-review evaluations of the submitted MS. This would enable readers to see whether Lancet maintained its usual standards in these highly irregular circumstances, or whether the reviews were cursory, or the authors’ responses inadequate.

  37. Most of what I’ve read on this page is reminicent of Orwellian newspeak. It’s frightening. We are so used to spin we’ve learned it off pat, meanwhile the truth wanders off in an unconcerned daze. We are creating differences that are widening year after year, getting more and more spitefull. Humans never learn do they. I feel a nasty world war coming on.

  38. >>Seen anything like that on the news?

    No…. if you happen to live in the U.S. that is – where the news is filtered and sanitized. I don’t think this figure will be any big surprise to European or Japanese or Middle Eastern viewers who get an unvarnished account (as they should) of American atrocities.

    The figure may be indeed be high but there’s no doubt civilians are paying a high price in this conflict, in large part because the american public -high on nationalism and self-righteousness and self-cnteredness- just doesn’t give a damn.

  39. Contact information for THE LANCET Publishing Group:

    Customer Services

    North America

    Toll Free Tel Phone: +1 (800) 462 6198
    Direct Tel Number: +1 (407) 345 4033
    Toll Free Fax: +1 (800) 327 9021
    Direct Fax: +1 (407) 363 1354
    E-mail:USLancetCS@elsevier.com

    Europe and Rest of the World

    The Lancet Customer Services
    Elsevier Ltd.
    The Boulevard
    Langford Lane
    Kidlington
    Oxford
    OX5 1GB
    UK

    T) +44 (0) 1865 843077
    F) +44 (0) 1865 843970

    E-mail:custserv@lancet.com

  40. right off the bat this argument fails: you state that the news would have covered that many deaths. I believe 1000+ US soldiers have died in Iraq, yet I wouldn’t know that by the photos of the caskets.

    geez, try something new…try thinking.

  41. And yet, blames (last name america) knows the number of casualties that have occurred. The facts have still been reported, just without the sensationalized photos that the military has NEVER released, in prior administrations.

  42. Scary revisionist math. Maybe you can apply that same methodology to the Holocaust and say that the 6 million dead were a myth (or just one-third, as if that makes it any better). Denial and obfuscation are the hallmarks of a scary time that seems to descend upon us. I oftern wondered how it could have been that people put their heads in the sand while death and destruction was obvious during WWII. Now I know.

    Being a Chicago Boy myself, I wonder how it would be any better that 300 people a year were murdered in the City instead of 900. Does that make it any better? And lest we forget, does the phrase “Lies, damn lies, and statistics” mean anything anymore?

  43. James:
    I think you don’t get the point of these critiques. The Lancet article is purporting to present SCIENCE, not opinion. Scientific studies have an accepted structure and use accepted methods of data gathering and statistical analysis. Any scientific study can be (must be!) examined against this framework of validity.

    Talking about whether or not 8,000 or 1,000,000 Iraqis died does not show callousness. People here are merely subjecting the material in the Lancet article to an analysis of the “quality” of the data/conclusions. The article IS about numbers of dead people, after all.

    A “junk science” article with illegitimate conclusions does NOTHING to further reasonable public policy.

  44. Indeed, it is scary that a supposedly “respectable” publication like the Lancet could engage in such blatant numerical revisionism in an attempt to influence an election.

    Whatever happened to honesty? Why is a medical journal using statistical lies to play politics?

    The only bright spot is that blogs like this have emerged to combat the lies.

  45. So the US media is ‘sanitized’ but the European media is objective about American ‘atrocities’, the undoubtedly sophisticated anonymous reader says, using the name of the Goddess Of War as a pseudonym.

    So many levels of irony, so little time…

    Enough games. Back to watching those unfiltered terrorist atrocities and other beheadings the Athenas of the world will never see.

  46. I am voting for Bush because he is demonstrably the better of the candidates who have any chance of winning. Kerry was for the war in Viet Nam (highly decorated patriot as announced to a smarmy degree at the Dem’s convention) and against the war in Viet Nam (leader of Vets against that war), and nothing has changed since. We need Churchill, not Chamberlin.

  47. 100,000 is such a shocking number and would be such a black mark on our national moral pride that it’s understandable that all of you should be working so hard to deny it.

    Carry on.

  48. Archie, 100,000 did not seem such a shocking number when the victims were gassed Kurds, Marsh Arabs or Shias dying under Baathist repression.

    Given the shallow hypocrisy of your fashionably self-loathing moral posturing, it is no wonder you want to believe in this number, is it ?

    “Carry on”.

  49. The study’s pre-invasion crude mortality rate also seems off. The study places this number at 5/1,000; but according to the WHO, the figure is 8/1,000 (in 2,003). Since the study checks how many more people die now than did then, that would seriously skew the data.

    In addition, the pre-invasion sampling period is 14 months, while the post-invasion period was 18 months (28% more), which would also cause an increase in the calculated extra deaths.

  50. No one died in my family in the last year. Using Lancet’s methodology I don’t think that anyone in Pittsburgh died in the last year.

    Funny, I don’t hear about that in the news. At least I should hear about funeral homes and cemetaries going out of business.

  51. Thanks, Shannon.
    I found the site of John Moore (who posts above) useful for context of publication.
    Your argument was interesting; I was left with some naive questions:
    A) Is my impression that Lancet is the major medical journal of England correct? It is 180 years old and Victorians refer to it with regularity and respect. Is it still esteemed?
    B) Do refereed journals generally agree with an author’s demand for immediate publication? Wouldn’t an editor be hesitant when such publication will hinders the usual lengthy process of vetting?

    Wouldn’t this time frame and its comments indicate that Lancet is more concerned with politics than science, speed than accuracy? Does that make a difference in how we as readess should interpret the findings?

    C) Are there any generally accepted and scientific guidelines for conducting such a study against which this science was held?

    D) Don’t we need more context for a statement like, “In the Iraqi culture it was unlikely for respondents to fabricate deaths”? How were the “field researchers” chosen? They were fluent in both Arabic and English and were doctors; these seem important qualifications. In a region like this, wouldn’t their previous and current positions be important in determining the nature of the responses they were likely to get?

    This is a little (and significantly unimportant) point – but how many deaths are likely to occur in a brutalized country in which all the criminals were let out of prison as the Coalitiion army advanced and the police fell apart? The release of political prisoners should not have made much difference in a peaceful society (Iraq? before? after?) but surely everyone in those jails were not political prisoners.

    Some may argue that is just my impression, living in Texas with all those guns – but it isn’t like Iraqis are gunless. And my husband, who edits a cultural journal, would love to know how they did the vetting & questioning & editing so quickly – he’d have fewer headaches with such efficient methodology.

  52. Most of the comments here seem to fall into the theme of “statistics suck!” Some of them fall into the theme of “so what if 100,000 people died, I bet the overthrow of Sadam has saved bundles more people.” But I’d like to shift the conversation to something Les Roberts, the lead author of the article said:

    Taken from the Sydney Morning Herald, October 29:

    “I emailed it in on Sept. 30 under the condition that it came out
    before the election,” Roberts told The Associated Press. “My motive in doing that was not to skew the election. My motive was that if this came out during the campaign, both candidates would be forced to pledge to protect civilian lives in Iraq”

    I agree 100%. Whoever wins this election should make an honest pledge to do this. Whether the true number is 15,000 deaths or 500,000 deaths I know we could be doing a better job at protecting civilian life. Right now that the situation is so bad we have a hard time counting deaths. This is horrible.

  53. Did they include in their pre-war-account also the people that were murderer by Saddams henchmen and the people that just “vanished”?
    They also died through violence.

  54. I’m going to make one more point here and then excuse myself from further “debate.”

    The Lancet article is a refereed journal. I challenge anyone here critical of the article to submit their analysis to another refereed journal. That is how research–and the refutation of research–is published and evaluated. Qualified peers weigh in, examine the approach, and conduct their own investigation if they wish.

    They do not simply pull some half-baked idea out of their ass and post it to a blog. Any asshole can do that, and at the end of the day, such commentary is worthless.

    If you feel you have a valid criticism of the research, attempt to publish it the way science is published.

    Blogs have their limits. This blog has clearly hit its limit, and this debate has essentially no value.

  55. Did they include in their pre-war-samples also the people that were murdered by Saddams henchmen and the people that just “vanished”?
    They also died through violence…

    To chel:
    Don´t you think that everything is done by the US military to avoid casualties?
    Didn´t they know that -to avoid killing civilians-
    without the wise advice of Lancet?

    Greetings from Germany

  56. Whoever wins this election should make an honest pledge to [protect civilian lives].

    This is absurd. You think that Bush (or Kerry) wouldn’t try to protect civilian lives without taking some “pledge”? And how many “civilians” would die before you declared that they’d broken their pledge? Two? Three?

    Right now that the situation is so bad we have a hard time counting deaths.

    No, we have a hard time counting deaths because: Iraq’s a big and dangerous country and it’s hard to get everywhere. Because communication systems aren’t what they are in the US, for crying out loud. Because the definition of “civilian” is kind of fluid, and the survivors often have nothing to gain in admitting that their relatives died while under arms. Because people lie. Because biased researchers will try to cook their data to fit their own agendas. I’m astonished that anyone could be the tiniest bit aware of the situation in Iraq, could be reading this blog, hell, could be aware of the existence of this blog, and not know these facts.

    Roberts allegedly said:
    I emailed it in on Sept. 30 under the condition that it came out before the election.

    This is a breathtaking statement. Where does a researcher get off insisting that a journal publish his work by a certain date? I’ll have to try that one on my next paper. Maybe if I argue it will get Bush out of office, the editor will rush it through.

  57. I think a lot of people here are having trouble distinguishing between their moral opinion of the war and objective criticism of this article. I just finished reading this article. It is, to put it mildly, junk.

    (1) The 95% confidence interval for the number of deaths is (8000, 194000). The researchers say the death toll is “probably about 100,000 people, and may be much higher.” It may be much lower too, but they neglect to mention that. They also neglect to mention that the confidence intervals for their pre- and post-war estimates of mortality rates overlap, which is usually taken as a warning sign that there may be no statistically significant difference.

    (2) No interviews were conducted in Basra, the second largest city in Iraq (Table 1). (Maybe it was too peaceful for their purposes?)

    (3) If you exclude Falluja the entire estimate is based on a grand total of 21 violent deaths. The assertion that many women and children are being killed is based on the deaths of 4 children and 2 women (Table 2).

    (4) Researchers attempted to verify the reported deaths in 78 of the households, and were only able to do so in 63 of these cases. This could mean that up to 20% of the deaths reported didn’t actually happen.

    As far as William Trippe’s suggestion that these criticisms and the other valid criticisms here be published in a peer-reviewed journal, he should know that these kinds of short criticisms are rarely published in any field — the ostensable purpose of peer-review is to screen out articles with such obvious flaws. Further, I counter that (1) given the obvious partisan leanings of this article and thus The Lancet, any criticism perceived as “conservative” would probably be rejected by The Lancet on its face, and other journals won’t be interested in publishing criticism of a half-baked article in another journal, and (2) if this article in The Lancet is an example of peer-reviewed work in this field, you’re better off reading a blog.

    Actually, what’s happening right here is a form of peer-review. If anyone thinks my comments are wrong, please speak up and I’ll do my best to clarify.

  58. Trippe:
    The readers (and posters) of this blog are not unfamiliar with the nature of refereed journals. That has prompted more rather than fewer questions.

    Refereed scholarly journals are seldom so explicitly aimed at politicizing research. Bloggers don’t research; they are closer to journalists and Kaplan (in the essay cited below) does some actual primary source interviewing. Frankly, the timing of this publication makes bloggers (or journalists) more important; the public electorate, much more than doctors, were the audience for this essay. Do you think if it came out in ten years anyone here would care? Or anyone reading us would care?

    See Fred Kaplan on Slate. As Instapundit points out, Kaplan takes a sledgehammer; the result (a pretty demolished study) is the same as Shannon’s. He begins by making this point:

    The report’s authors derive this figure by estimating how many Iraqis died in a 14-month period before the U.S. invasion, conducting surveys on how many died in a similar period after the invasion began (more on those surveys later), and subtracting the difference. That difference—the number of “extra” deaths in the post-invasion period—signifies the war’s toll. That number is 98,000. But read the passage that cites the calculation more fully:

    We estimate there were 98,000 extra deaths (95% CI 8000-194 000) during the post-war period.

    Readers who are accustomed to perusing statistical documents know what the set of numbers in the parentheses means. For the other 99.9 percent of you, I’ll spell it out in plain English—which, disturbingly, the study never does. It means that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000. (The number cited in plain language—98,000—is roughly at the halfway point in this absurdly vast range.)

    I suspect the major given is that any study comparing deaths in an authoritarian and secretive regime with those while Coalition forces are still battling remnants of the tyrant’s followers as well as outside terrorists is not likely to be all that accurate. The truth may come out, but it comes out slowly. Certainly, the results must be questionable enough (as are the motives) to make it an interesting observation, not one published in a respected scientific journal.

  59. William Trippe,

    Refereed journals are valuable, but they are not the only method of disseminating and vetting new information. Nor are they perfect. Sometimes peer-review fails and journals publish flawed research. The correct course of action in such cases is to challenge the research rather than to denigrate the medium (in this case blogs) in which the challenge appears.

    In the current case, a major journal published, apparently without serious review, a study whose author intended it to have a political effect. Indeed it’s conceivable that the principal goal of the study was political. Are you seriously suggesting that the only legitimate response to such hackery is to submit critiques to another publication committee and then to wait several more months for them to (maybe) publish it? Are you saying that the political hit doesn’t matter, what matters is to follow the proper publication protocol?

    Sorry, but you and the Lancet don’t get to frame the issue here. Politicized research deserves a political response as much as a scientific one, and this politicized research is getting both. Obviously you don’t like that. Too bad.

    And surely, someone who understood peer review would have a better come-back to someone he was arguing with than a statement like this one:

    They do not simply pull some half-baked idea out of their ass and post it to a blog. Any asshole can do that, and at the end of the day, such commentary is worthless.

    You point out an apparent flaw in Shannon’s argument. Shannon responds. Then, instead of responding to Shannon’s response, you spout a bunch of insults and MSM-type ranting about the imperfections of blogs. Wouldn’t it be more decent and mature of you to admit that Shannon is right or say that you need more time to think about it?

  60. To follow up on Skip Smith’s (1:56 pm) and Ginny’s (1:57 pm) comments, the cited 95% confidence interval could be interpreted this way:

    If 20 studies were conducted in September 2004 using Roberts’ methodology, we calculate that 19 of them would give excess-death-due-to-violence estimates of greater than 8,000 and fewer than 194,000 Iraqis. One identical copy of the Roberts study would give a number that is higher or lower than this range.

    As a thought experiment, if twenty groups each conducted such a study in September 2004, would the editors of Lancet-like journals rush the entire set to publication by Nov. 2? I strongly suspect that the only such studies to be ‘important’ enough to merit expedited review and publication would be those with extraordinary findings. Extraordinary findings in this case can only mean very high casualty estimates.

    To the extent this ‘thought experiment’ conclusion is valid, it invalidates Roberts’ confidence interval, as that calculation can only account for random error, and not for systematic errors such as bias in the whether-to-publish decision.

  61. Sigh… I don’t even know why I am bothering, but to respond to Mr. Gerwitz’s last post, specifically:

    “You point out an apparent flaw in Shannon’s argument. Shannon responds. Then, instead of responding to Shannon’s response, you spout a bunch of insults and MSM-type ranting about the imperfections of blogs. Wouldn’t it be more decent and mature of you to admit that Shannon is right or say that you need more time to think about it?”

    Actually, her rebuttal didn’t answer my point at all. I decided on further reflection that the debate simply wasn’t worthwhile. I stand by my prior post. If Ms. Love or anyone else feels they have a valid criticism of the Lancet research, he or she should attempt to publish it through the normal channels.

    As to my “MSM-type” ranting, please, leave the pedantic lectures for someone else. I write for commercial publications and I write for blogs (I have three of my own). Blogs have their purposes. This is not one of them. Now if you wish to discuss this further, feel free to email me or comment on my blog, but I am not contributing any more to an angry and ultimately pointless discussion.

  62. >>”I am not contributing any more to an angry and ultimately pointless discussion.”

    William, I think you are the only person here that is angry. Further, I think a lot of valid criticisms of this study have been raised here, so the discussion is hardly “pointless.”

    As far as your frustration at Ms. Love’s response to your criticism, I thought her response was perfectly fine. She was criticizing the use of cluster sampling in the study, and pointed out the outlier of Falluja as an example of the problems cluster sampling can cause in a heterogeneous population. To which you responded … Falluja is an outlier. If anything she should be frustrated with your response, not vice-versa.

  63. William Trippe (2:43pm):

    Manuscripts I have authored have never been published on a schedule as rapid as that achieved by Roberts et al. in the Lancet. But my work has never been of any political interest.

    On the other hand, I have never asked for the sort of deference to my work that you wish to accord to Roberts’ work, and that you argue that others should cede to them, as well.

    Your arguments to this point have been unpersuasive. If, on hearing such sentiments politely expressed, your response is to “not [contribute] any more” to a discussion that you find to be “angry and ultimately pointless”: that is certainly your prerogative.

  64. Amongst kinds of character, one has a choice between the equanimity and decency of Skip Smith, Amac, or Jonathan G ewirtz, on the one hand, and the insulting buffoonery of Trippe, on the other. The choice is easy, of course. (Moreover, I’ve found that excellence in character is more conducive to discovery of factual truth, other things being equal. But that’s an old story.)

    William Trippe, will you apologize to Shannon Love for calling her an asshole? It’s never too late to man up, Trippe.

  65. Actually, her rebuttal didn’t answer my point at all. I decided on further reflection that the debate simply wasn’t worthwhile. I stand by my prior post. If Ms. Love or anyone else feels they have a valid criticism of the Lancet research, he or she should attempt to publish it through the normal channels.

    I think you protest too much. If you find Shannon’s rebuttal inadequate, why don’t you simply point out why? You’re not willing to do that, yet you are willing to engage in arguments about other things. I find it difficult to take your protestations of offended professionalism seriously.

    BTW, WRT “normal channels,” this blog is a normal channel. At least, everyone here thinks so — including you, it seems, since you keep coming back. I don’t see where we ever held ourselves out to be a substitute for the Lancet. What we are is a small part of an alternative information-delivery network that functions as a check on other media, including nominally peer-reviewed ones. The reason we’re discussing these topics is that the Lancet publication committee’s vaunted review process failed, either out of sloppiness or because they decided to prostitute themselves for some cheap political thrills. Now they’re going to get the scrutiny they deserve, and not just from a bunch of part-time amateurs like us. But if you want to blame us for it that’s OK by me.

  66. William Trippe,

    “If Ms. Love or anyone else feels they have a valid criticism of the Lancet research, he or she should attempt to publish it through the normal channels.”

    You appear to be arguing that since this article appeared in A PEER REVIEWED JOURNAL that we should all just check our brains at the door and bang our heads three times on the floor in intellectual obsequiousness.

    Screw that.

    Arguments from authority are only necessary when one has nothing else to go on but unlike yourself, I and others on thread actually have at least a passing understanding of the science and math in this study and we conclude that this study has serious flaws on its face. Had this been a study of the incidence of a disease for example, it would have never made it into print much less been a rush job intended for political effect.

    As I argue in this post, this study reveals an obvious breakdown in the peer review system. It is represent corruption and scientific malpractice.

    This is matter is now far more important than the its immediate political impact. It is now about corruption within our scientific institutions.

  67. Hello Franz and Angie,

    I’m so confused by how negatively you guys reacted to my post. All I was trying to say was that I was hoping the candidates would really do all they can to minimize civilian casualties in Iraq. It’s not something I’ve been hearing them discuss much in their debates or stump speeches. If the candidtates feel like the electorate doesn’t care, it won’t be a priority in the administration.

    Any way you slice or dice it, a lot of civilians have died and this should be examined. Maybe we can’t do any better, but we can only know that be getting some facts. I hope that researchers will continue the data collection effort.

  68. Trippe’s assumption is that expensive specialized magazines such as the Lancet do not make mistakes. This naive assumption of purity -or innocence – is essentially the only argument he has to support his supercilious appeal to authority. Never mind the number of published studies who were peer-reviewed and have since been retracted, or the number of invalid statistical inferences recently found in a survey of such specialized magazines in the UK.

    Of course people who write in them and the publishers will protest of their superior quality by dismissing any critic as below their level. I mean, duh. But a peer review process can only be effective if it is itself open to scrutiny: who are the reviewers ? Who chose them ? Based what criteria ? What is their past record judging submissions ? In the absence of any such information, claims of inherent quality deriving from peer-reviewing are simply self-serving.

    Also, what of the papers that were rejected ? On what basis were they put aside ? We expect pharmaceutical companies to tell us about failed trials but the supposedly most reliable publications do not stand up to the most basic rules of transparency and accountability.

    this debate has essentially no value.Those who have chosen or appointed Mr Trippe as a judge of “debate value”, raise your hands. Self-appointed judges handing down summary judgments have no value in any debate, William. Such comments do reveal a lot more about you than they do about the debate you are, I’m afraid, a part of at this stage.

    As for those who comment on the situation being so bad that we cannot count casualties accurately, remind me how many weeks it took to have a final tally of 9/11 victims ? How many days were spent in Florida counting and recounting stupid ‘chads’ ? And you expect similar tallies in Iraq to be accurate to the minute ? I dare hope you are either being intentionally simple-minded or that your irrational expectations are only there to justify your conclusions and beliefs on this complex topic.

  69. I have a question I posed when this figure was put out as the number of deaths. Haven’t seen anyone answer it anywhere. Perhaps someone here has the background to address it:

    A Logistical Query: How in heavens do you bury, incinerate, or otherwise dispose of 98,000 people in this time-frame? Figuring five or six family members per death, this is roughly a half million mourners.

    So where was the wailing and tears on video? Where are the remains of all these human beings?

    You guys have reduced this to numbers. It’s not about numbers, it’s about practical reality. The magnitude claimed could not have just flown under the radar, just waiting for some survey to find it.

  70. There are some unavoidable truths about the War in Iraq. The first is that over a thousand young Americans have already perished. The second is that war kills mostly civilians, particularly woman and children and elderly civilians. Those that do not die in the bombs and the gunfire, perish for lack of sanitation, clean water, or wholesome food and electricity.

    There is another unavoidable truth about this President. George Bush is a religious zealot and fanatic that matches Osama Bin Laudin in every salient respect, a fanatical Crusader vs a fanatical Muslim whose underlings are bringing great suffering, death and destruction to the people of Iraq.

    It happened before in Vietnam, and America has learned nothing from the previous experience, which is why Americans are going to suffer its lessons a second time.

    By the way, most scientifically literate people around the world tend to give the Lancelot a great deal of respect as a publication and while a particular paper may be contraverial, it never the less has focused public attention upon the suffering civilians of Iraq.

  71. From an article in the July 2004 New England Journal of Medicine:
    http://content.nejm.org/cgi/reprint/351/1/13.pdf

    On page 6:
    48% of Army and 65% of Marines were “Responsible for the death of an enemy combatant” in Iraq
    14% of Army and 28% of Marines were “Responsible for the death of a noncombatant” in Iraq

    From what I can tell, these apply to “combat units.” Let’s guess that amounts to 80,000 of the troops there. That gives us a minumum of 40,000 Iraqi combatants and 12,000 noncombatants killed. Since some soldiers certainly killed more than one person, the actual number must be significantly higher. Also, the numbers were obtained in late 2003, so the current numbers would be higher yet.

  72. Joe Green (10:09pm):

    Your comment presents your opinions on whether the invasion of Iraq was a just war, on whether Bush is a religous fanatic, and on how medical professionals view The Lancet. I could offer my thoughts on these subjects too; in some cases they might not be far from yours.

    What neither of us would have done is to abstract our political and ethical judgements into a technical document, and then present that information and analysis to a sympathetic editorial board of a scientific journal for expedited publication for avowedly political ends. At this writing, this appears to be the course of action that Roberts et al. chose to take. They are complict with The Lancet’s editors and perhaps reviewers in corrupting an important information-transmission channel.

    This is reprehensible, whether or not many members of the Academy can see past their politics to understand what has taken place.

  73. This saga bears all the hallmarks of the way the authoritarian progressives attempt to control the political debate. Faced with a self evident truth (to them) which the populace stubbornly refuses to endorse, imposingly scientific and newscatching evidence is adduced to prove this truth, then when the evidence is subjected to real scrutiny and its scientific basis shown to be shonky (I’m being charitable and giving them the benefit of the doubt on the question of deliberate manufacture), the response is that the critics are just arguing about numbers not facts (ie degree not kind) so their point has still been proved.
    Keith Windschuttle in Australia has already revealed how this has worked in Aboriginal History, where grossly inflated and unsubstantiated figures of Aboriginal deaths were used to generate an impression of callous and ongoing murder. Historians here then defended themselves by claiming that the numbers were immaterial beside the fact that deaths did occur. Of course numbers do matter when they of the magnitude of 120 vs 1000. So also in this case the numbers do matter. The shock value of saying the death rate had increased minimally (or not at all) after the invasion would have been nil. We must stand up to these tactics and keep hammering back at the progressives and not let them establish the moral high ground. A lie is a lie, especially when, as in this case, it is deliberately promoted to achieve a political end. The real fascists here are those who abuse science to prove their lie. The likes of Trippe won’t admit this ever so there is no point in trying to debate with him rationally, but no amount of twisting and turning will change the facts and that is the best way to deal with him. For the first time since the “progressive era” began we, the people, have a medium in which we can force this unrepresentative elite to prove what they say. It is not a pretty sight watching them shrivel in a bonfire of their own making but it does warm the cockles of the heart.

    P.S It should be remembered how the progressives reacted (and still react) to claims of the millions slaughtered by the Stalinist regime. Clearly that was a difference in kind, not degree.

  74. I am not sure if anyone else pointed this out. If they have, I apologise for any duplication.

    It seems to me that the study presupposes the 14 month period prior to the invasion was some sort of a norm against which the post invasion mortalities could be compared.

    From what I have seen, the previous regime may have been somewhat preoccupied with burying things, like planes and leaders, in the desert to sustain its usual level of killing. Sort of a calm before the storm.

    Jeff

  75. Fred Kaplan has this to say.

    Readers who are accustomed to perusing statistical documents know what the set of numbers in the parentheses means. For the other 99.9 percent of you, I’ll spell it out in plain English—which, disturbingly, the study never does. It means that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000. (The number cited in plain language—98,000—is roughly at the halfway point in this absurdly vast range.)

    This isn’t an estimate. It’s a dart board.

    To borrow a phrase from Monty Python, “Now this is getting silly…”

    syn I’m with you except that I’m also getting tired of people like the authors of this propaganda piece thinking were the ones that are stupid.

  76. I am truly amazed at the scientific illiteracy being displayed in some of these posts. Everyone is attempting to show how the authors in this paper were somehow following a “political agenda” but those critics on these pages are hardly being “scientific” themselves.

    The correct way to refute points made in a scientific paper is to write a peer reviewed paper based upon sound research that finds other answers and explanations.

    The critics here sound more like mere right wing propagandists. Whatever the raw data did or did not say, and whatever the conclusions, the authors followed a scientific procedure that was open for other scientists to either accept, or devise additional experiments that would either refute or confirm the findings.

    It is a huge mistake to consider the Bush Administration as being supporters of science because that is patently not the case and obvious by merely examining their position in regard to medical stem cell research.

    The only kind word one can apply to these right wing spin and propaganda efforts is to call it what it actually is, quackery.

    On the subject of civilian casualties in the War on Iraq, many sources using other methods have reported numbers around 15,000 to 25,000. This paper uses methodology that sets the figure higher at around 100,000 and could have any number of explanations as to what is actually being impounded in the measurements.

    Afterall, its not as simple as merely and correctly applying Gaussian statistics to molecules in a gas where there are very large numbers and simple observables. Even when one considers Fermi particles, the nature of the particles being studied all by themselves produce very noticable differences in such population questions.

    I would close by simply stating that “political scientists” are nothing of the sort, and they would garner more respectability if they called their discipline “political studies”.

  77. Cool! Another argument from authority to put us in our place, while you avoid dealing with the substance of the arguments here. Meanwhile you just happen to have some opinions about the ignorant Bush administration, etc., that you’ll share with us. Such a deal.

    I take it you are unable to refute any of the criticisms of the Lancet piece.

  78. “The correct way to refute points made in a scientific paper is to write a peer reviewed paper based upon sound research that finds other answers and explanations.”

    Wrong. There is no one “correct way” to debunk such tripe as we find in this study. If you can not see past your partisan, political viewpoint to how truly outrageous the claims of this “study” are, you’re blind. One need know nothing more about the study than that it’s 95% Confidence Interval is 8,000 to 194,000 to know that it’s bunk. That would be like having a 95% CI that Pres. Bush was going to receive between 40% and 60% of the vote. This “study” is worthless on its face, and the Lancet is less for publishing it.

  79. To Kevin Bailey and Jonathon G ewirtz:

    Jonathon is acting like a spoiled and undisciplined adolescent while Kevin is not prepared to do the work of science. This is a display of immaturity and intellectual laziness.

    In Jonathon’s case he sees me as an “authority figure” because indeed I am a person of science, not ideology. I have explained how he can proceed and prove the authors are mistaken or misguided in a credible scientific manner that will impress the readers of the Lancelot.

    And for poor Kevin, he seems to be stuck in the elementary mathematics of Gaussian statistical probability distributions. But his inflamed rhetoric does not pass for knowledge, and in science, a million political barbarians can scream at the top of their lungs, yet its the quiet voice of a single scientist with a clear postulate backed up with a actual experiment that establishes the truth.

    Incidently, unlike the new Barbarians in Washington, science is a matter of informed consensus, not declaratory judgements from a politically soiled Supreme Court.

    One other thing. In science the standard of proof is much higher than Mr. Bush’s assertions about weapons of mass destruction in Iraq. In science, a SINGLE exception to a theory is sufficient to discredit the theorem or proposition under discussion.

    I am not sure that Kevin or Jonathon can stand the heat of this scientific kitchen, given all the frothing they have demonstrated above.

    Neither should quit their day jobs for positions in science.

  80. There’s been a lot of talk on the blogs and in the media about The Lancet’s recently published study by Dr. Les Roberts et al. that attempted to make an estimate of excess deaths in Iraq since the invasion in 2003. Something that I’ve been seeing a lot is people claiming that the study is either totally good or totally bad. That’s not the way studies work. Even the best published studies have flaws. Conversely, even some studies with extensive limitations can offer new information. Nothing can be absolutely proven in a single study; what a study offers is a piece of evidence. Each piece of evidence contributes our understanding of the phenomenon.

    I want to focus in on the study’s confidence intervals (CI). Fred Kaplan at The Slate stated, “It [the confidence interval] means that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000.” First of all, this not an accurate description of what a confidence interval actually is. The authors are not saying they are 95% sure that the true number of deaths is between these numbers. For a formal definition see The American College of Physician’s Primer on Confidence Intervals: . Kaplan though does correctly identify this to be a rather wide CI — just check out some of the abstracts from this month’s American Journal of Epidemiology to see some other CIs in published papers: . The wide CI means is that the estimate is imprecise. However, Kaplan begins his piece by stating that the estimate of 98,000 excess deaths is “so loose as to be meaningless.” Loose yes, meaningless, no. Confidence intervals tell us nothing about whether what we are seeing something meaningful. (Also, CI’s also cannot tell us if what we’ve observed is correct. You could have a very narrow CI and still be incorrect if, for instance, you had a fatal flaw in your study design.) Additionally I’ve been reading on blogs and in comments a feeling that the authors were somehow deceitful or purposefully trying to misguide people somehow in relation to the confidence intervals. This is not the case, if you read the actual study, the confidence intervals are reported clearly and according to convention.

    This is the first scientific effort to count civilian deaths. I respect the research team for volunteering to face danger while gathering data that may further the understanding of what is going on in Iraq. This study has flaws and no one, not even the researchers, would say that their findings are even close to being the final word or the end of the story. More research is necessary. Though the number 98,000 may not be a precise estimate, this research strongly suggests that civilian mortality has risen since the war began. The study also assists in our knowledge of the causes of these excess deaths, another very important and not obvious piece of the puzzle.

  81. To Joe Green

    A person of science???? Firstly, if this were a credible, albeit mistaken or misguided piece of research, I might agree with you. However, any piece of research that uses a 95%CI of 8000 to 194000 just isn’t credible and to follow your advice wold be to give it more credence than it deserves. As a graduate of statisical method, I wouldn’t even give this paper a grade it is so obviously lacking in any scientific understanding.

    Secondly, if this was a genuine scientific report, why was it so important to rush the peer review process and get it printed before the election. It seems to me that calls to respond in the “approved scientific manner” are just calls to avoid any questioning of the article until after the election is over. The authors insisted on putting it out there BEFORE the election so we have a right to respond BEFORE the election.

    Thirdly, I have no faith in peer review if this article is a sample of what is passed for printing in the Lancet. Despite your attempts to appear above the fray, you are obviously more concerned with ensuring that the truth is kept hidden from the electorate till after they have voted.

    In fact it is you who are either naive or immature. The truth is the Lancet, the authors of the report and their allies like you have been called up on their pathetic attempt to influence the election and they are now desperately trying to suppress any discussion that might show the people of the world how they are being conned by “scientists” in the hope that they will get the election result they want first. When the Lancet reveals 1. its peer review process and 2. its reason for fast tracking this article I will be happy to do the peer review thing. Until then I will be happy to say “GOTCHA”.

    BTW, the once scientist quote was cute and would have worked except for one thing .. there was a clear postulate in this case but not rigorous experiment. Lysenko was the real Godfather of this piece of nonsense.

  82. you guys need to wake up and realise insurgents are actually rebellion fighters… its easy to label them something more sinister to brainwash you all into accepting the events unfolding… these are people of Iraqi origin fighting against the US and its allies because they truly believe that the US is wrong and whether that is a fact or not is irrelevant.. point is they are honestly fighting for a cause they believe in and we’re the arrogant ones for not considering that.. who is to say that we are the ones in the “right”? at which point of Westerner’s evolution did such blind faith and arrogance become a regular trait of all?

    i aint supporting any side just saying that objectivity is a myth these days…

  83. Knowing how you guys appreciate a presentation that is ‘fair and balanced’, it seemed pertinent to add a few points.

    From an article from the Guardian [1] we have:

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

    and

    Even though the sample size appears small, this type of survey is considered accurate and acceptable by scientists and was used to calculate war deaths in Kosovo in the late 1990s.

    Futhermore, when interviewed [2] Gilbert Burnham (of the Johns Hopkins University), co-author of the study, makes no claim that the study is the last word on the subject and urges that further studies be undertaken “by an independent body such as the International Committee of the Red Cross, or the World Health Organization” – hardly the actions of someone trying to pull a fast one on us all! He says:

    “Now, I don’t want people to be thinking that this is the definitive study on mortality in Iraq. (…) I think that from the results we can say with confidence that yes, there is definitely an increased mortality rate. (…) Our best guess, on a conservative side, is 100,000. But it could be less and it could be more.”

    and to the question:

    “Was this study peer reviewed?”

    he answers:

    “Oh, my goodness, was it ever. [Laughs] First off, nothing, nothing ever gets in the Lancet without a vigorous peer review. It’s heavily peer-reviewed. And in the case of this article, it went through the full editorial review board several times and they sent it out for multiple reviews. I’ve written a few papers for the Lancet over the years and I’ve never had anything like the scrutiny that this one had.”

    As to his reason for publishing it when he did:

    “My motive in doing that was not to skew the election. My motive was that if this came out during the campaign, both candidates would be
    forced to pledge to protect civilian lives in Iraq.”

    In any case, when you think about it: with 150,000 plus troops stationed in Iraq, the number of deaths attributed directly to coalition forces (84 percent of the violent deaths) works out at one killing for every two soldiers in the 18 months that they’ve been out there. Does that seem so unlikely after all? Given the oft-repeated charges of heavy-handedness of US troops (by coalition partners, amongst others), the daily provocations they face and the sheer numbers (tens of thousands?) of bombing raids since the start of the conflict, I’d guess that the 100,000 figure may indeed be a conservative estimate.

    As Burnham himself said last week in an interview for Australian Radio National, most wars start out with the intention of causing minimal damage to the civilian population. Just about none succeed.

    Tony Cox

    [1]http://www.guardian.co.uk/worldlatest/story/0,1280,-4581202,00.html

    [2] Do a Google search for “did the insurgents
    cause this” and then go to the cache page.

  84. indeed I am a person of science, not ideologyWell, that settles it then. (Actual translation: “I claim scientific authority – based on what ? screw you, I’m a person of science – to support my own little prejudice”).

  85. OK Chel, I really have to reply to this.

    “The first scientific effort to count civilian deaths” .. Well no actually. It was the first attempt to apply a particular survey methodology and regression analysis to extrapolate an ESTIMATE of post invasion deaths. An effort to count civilian deaths would have focused on actual bodycounts not survey interviews.

    Secondly, I suspect that few people will read the report but focus on the executive summary as the media have done. The problem is that the summary bears only a tenuous relationship to the report data.

    The following discrepancies are major causes for concern.

    1. In the findings it is stated that “Most individuals reportedly killed by coalition forces were women and children”. However the report data in Table 2 shows that only 6 of 21(28%) non-Falluja deaths by violence were of women and children and 33 of 73(45%)of all violent deaths were women and children. Later in the report it states that 12 violent deaths were not attributed to coalition forces (11 men and 1 female). This reduces the total deaths attributed to coalition violence 61 of whom women constituted 4(6.5%) and children 28(45%) of which non-Fallujah deaths constituted just 4(6.5%). The report itself draws attention to the uniqeness of Fallujah where, according to the survey 24(85%) of child violent deaths occured. Only by including the Fallujah total (specifically the Fallujah child death total) can the statement quoted above be made. If a small community in say the US showed such a huge discrepancy in its death rates (7 times the rate elsewhere according to the survey), researchers would be looking for site specific causes to explain it. Not our authors .. they are happy to rely on skewed data in their statement of findings no matter how thet hedge it in the actual report.

    2. In the summary interpretation the authors state “Making conservative assumptions, we THINK that about 100,000 excess deaths or MORE may have happened”. What these conservative assumptions are is never clearly spelled out. The figure which they THINK (not assess) is just a mid point between the lowest and highest figures. On reading the evidence however their assumptions appear less than conservative. No attempt is made to verify child deaths (the most significant statistic in the Falluja results) and only minimal efforts to verify adult deaths (63 out of 983 households). Despite this the authors assume that all reported deaths and causes are genuine. It is interesting that on page 6 of the report they discuss the possibility of underreporting and dismiss it but do no even consider the possibility of overreporting. The one possibility they seem to accept as a possibility is that people may overstate the number of household residents to obtain greater rations. Of course assuming this would lower the estimate of the death toll so that in effect this possibility only increases the REAL death toll. Perhaps that is why it was included as a possibility.

    3. IN the Summary Interpretation it also states that “air strikes from coalition forces accounted for the majority of civilan deaths”. This is a bit like saying the sun causes daylight. In any conflict, even one with precision bombs, air strikes because they are more indiscriminant will inevitably cause greater casualties. However, the report offers no concrete evidence for this statement other than the “attributed” causes of death by the interviewees. If the authors had stated this as an axiom and then sought to determine whether these “collateral deaths” were greater or less than expected, that would have been a scientific study. Otherwise they are just stating the obvious while implying that it is a unique occurence in Iraq.

    4. Finally the very high number of child deaths in Falluja needs some more analysis. The figures show that 52 deaths by violence occurred in Falluja of which only 3(5.7% or less than the survey wide average) were women but 24(46%) were children. This discrepancy needs a more detailed study than is provided. It is unusual that such a large number of children should have died while so few women have died. The implication is that the children were detached from their mothers at the time of death or were targeted specifically. The latter is most unlikely, while the former points to the possibility of children being involved in the fighting. This latter fact was given some confirmation by a report from by an Australian reporter, Martinkus, showing young boys being trained as soldiers. It may also mean that a large number of children were killed in a single hit on a school or playground. Such an incident could not then be extrapolated to the whole country.

    All in all, the methodology of the report is shonky. The authors themselves are most careful to cover themselves in the body of the report. But their aim is clearly stated at the end of the report. “In view of the POLITICAL IMPORTANCE of this conflict, these results should be confirmed by an independant body such as the ICRC, Epicentre or WHO. In the interim, civility and enlightened self-interest DEMAND a re-evaluation of the consequences of weaponry used by coalition forces in populated areas.” This is a politically inspired report designed to push the coalition into self limiting its ability to fight back against those opposed not only to the coalition but to the interim Iraqi government. The executive summary replaces the weasel words of the report with confident assertions unsupported by their own evidence. It is time this debate was made clear. It is not about science it is about politics, but only the supporters of the report seem unable to admit this. If anyone doubts this look up the Lancet web-site (free registration) where you will find the executive summary printed. If you want to read the FULL report you have to download a PDF file. Guess what most will do!

  86. Tony Cox can’t refute Shannon Love’s criticism of the paper, so he gives us snark (Knowing how you guys appreciate a presentation that is ‘fair and balanced’), yet another appeal to authority/peer review, and his own armchair estimate of the number of dead.

    Meanwhile, Joe Green, also unable to refute Shannon Love’s argument, embarrasses himself by offering insulting buffoonery of the William Trippe kind (telling his opponents that neither should quit their day jobs for positions in science) and showing that he doesn’t even know what the fallacy of “appeal to authority” is (Jonathon is acting like a spoiled and undisciplined adolescent…. he sees me as an “authority figure”). The classic buffoon, ignoring his oppenent’s argument and making charges of illiteracy, speaking of himself as a “person of science” and referring repeatedly to something called “the Lancelot”. Will he apologize for his behavior? Did Trippe?

    Therefore:

    Are there any genuine inquirers (i.e., concerned with discovering the truth, whatever it happens to be) who take issue with Shannon Love’s argument? Would you please step forward and offer an argument? I myself would like to know whether Love is correct, not whether she can be bullied into submission. At the moment, it seems so, but I don’t know. Will any critic of Love please assist me? No buffoons, please, only genuine inquirers.

  87. In unmoderated comments on a hot-button topic, there are bound to be people on all sides of an issue who post emotionally. Unfortunately, the topic at hand is technical (extent to which the Lancet study is valid), and parts of the discussion require specialized experience (e.g. the extent to which standard peer-review processes were followed).

    The temptation is to leave Ms. Love’s example behind, and focus on the emotional and the polemical, perhaps mixing it in with the logical and the rational. This will predictably lead to attaboys from fellow-believers and brickbats from The Other Side. But, in most cases, not to much else.

    My vantage point is that, on the basis of what is currently known, the Roberts et al. study is so flawed in so many ways as to be useless as a technical document. Whether or not its conclusions of high casualties were pre-ordained from inception, its authors and the Lancet’s editors have colluded to rush it through peer-review procedures in order to meet a political timetable–publication in time to influence the American presidential election.

    This is not how good science is–or can be–done.

    Joe Green (1:03am), thanks for raising the points you did about peer review. You invite an essay explaining why I find your view of the near-sacredness of the process to be at odds with both its theory and its practice. I don’t have the time to write that essay today, and won’t ask readers to accept my conclusions ‘on authority.’ But readers can note, as one voice in this debate, that one practicing scientist finds the notions that ‘peer review trumps’ and ‘peer-reviewed results can only be contested via peer-reviewed critiques’ to be incorrect. If they were taken to heart by the scientific community, the overall quality of research and research reports would immediately suffer.

    Tony Cox (6:38am), thanks for the cited references to the authors’ quotes via Google and the experts’ opinions quoted in the Guardian report. Although I don’t think they mean what you think they mean–many of the quotes sound quite disingenuous to my ear–it moves the discussion forward to bring them up.

  88. On the issue of flawed peer-review it should be noted that the “Lancelot” has put its foot in the manure before, by publishing a politicized (and later retracted) paper on autism and the MMR vaccine.

  89. Joe Green writes, “I am not sure that Kevin or Jonathon can stand the heat of this scientific kitchen, given all the frothing they have demonstrated above.

    Neither should quit their day jobs for positions in science.”

    And this guy accuses ME of “frothing”? He also has apparently never heard of the logical fallacy *appeal to authority* which he and his friends use so freely. He did not bother to answer my contention that the astronomical spread of their 95%CI demonstrates the worthlessness of their study.

  90. As a reader, the posts that rely the most on logical presentation of relevant facts and contain little or no invective are the ones I find most thought-provoking and most persuasive.

    But I’m just a guest here, Shannon Love and the other Boyz surely have their own opinions on the matter.

  91. “Do you guys really believe we went into Iraq to save lives of Iraq’s?”

    Non sequitur. This has nothing to do with whether or not this “study” was flawed on its face.

  92. It’s interesting to compare the Lancet/Johns Hopkins figures with those of WWII casualty figures. The Harper Collins Atlas of the Second World War, edited by historian John Keegan cites 60,595 deaths in Britain, which suffered targetted bombing of civilian populations. Italy, which endured not only heavy bombing but a bitterly fought ground campaign involving lots of armor and artillery, lost 93,000).

    Japan suffered only three times the Lancet/Johns Hopkins estimate (300,000, according to the HC Atlas), despite “an incendiary campaign against Japan’s wooden-built cities that left sixty percent of the ground area of the sixty largest completely burnt out” (quote from Keegan’s A History of Warfare), as well as two atomic bomb attacks. Keegan says those two attacks alone accounted for 103,000 deaths.

  93. It seems to me, that it is fun for Joe Green to provocate a little bit (mixing in a little bit of politics here and there), but not that he is used to scientific work (political scientific work was at least done in the former Soviet Union, this was not meant, wasn´t it?).
    1. Lancet is a scientific medical journal.
    2. If you tried to get published your scientific work in the Lancet, urging these guys to publish it, you had a good chance to have it sent back.
    3. Everything went a little bit fast. Even if you had to offer the only 100% safe and side-effect free therapy to cure cancer (I assume this would be important enough, to publish it very quickly), those Lancet guys would review this veeery carefully.
    4. Remembering my own scientific work as a medical student and later as a doctor, the professor responsible for the statistical work would have thrown me out of his office, presenting him such a study and drawing the conclusions, like they were made by the Les Roberts et al..
    Simply said (with a medical example):
    If you had a count of 8000 white blood cells, you would be in good health.
    Having 98000 white blood cells in your WBC-count, your disease was leukemia.
    So, in a medical journal like Lancet they should be able to recognise the consequences, between saying “maybe there are 8000 death or maybe there are 98000 death” and “`there a 8000, not 98000´ respectively `there are 98000 not 8000´”.
    5. The comment of the publishers, why this article was published so fast, seems to me as pure hypocrisy but not as a scientific reason.

    To chel
    You said, you (and the Lancet guys) only meant, that both, Bush and Kerry should take care, not killing so many people in Iraq. Agreed.
    I´m sure they did that and will do that.
    A very bad person, P.J. O`Rourke, was asked in a chat of the British newspaper “The Independent” :
    “If you were President of the United States, how would you have handled Iraq? Would you have done a better job than G.W.Bush?”
    Answer: “As George Bush is a man of deeply average intelligence and so am I, I think I would have done just as badly. I wouldn´t put Iraq as high up the agenda as he did, although, having covered the first Iraq war and seeing what the Iraqis did to Kuwait, it certainly would have been on my mind to wipe them out”.
    Bush and Kerry don´t seem to have this on their minds, I suppose.
    Greetings from F.H.

  94. I have read no comments that lead me away from Shannon’s original conclusions.

    I generally pay attention to mainstream news sources, and have not heard a word of doubt expressed by the MSN. Has anyone else heard any refutation in MSN?

  95. Pingback: CounterPundit
  96. Hi Daemon,

    Thanks for your long reply to my comment from October 31, 2004 03:48 AM. First off I agree with you that most people won’t read the actual article in the Lancet. They’ll probably read an article that quotes it, scan a blog commentary on it, or see it mentioned on cable news. Maybe a few curious people will read the abstract but The Lancet is certainly no coffee table magazine and reading the whole darn thing is the only way to really know what the authors did and found. I often feel frustrated by the miscommunication of scientific information. It seems that almost daily an epidemiologic study is reported in the news in a way that’s confused or overly simplified. I wish there was something that could be done to communicate health and science better. I believe the problem stems from many things including a public that has minimal understanding and patience for methodology simply wants the bottom line, journalists who themselves have little background in science, and scientists who are poor communicators. But I digress.

    I stand by my statement that this study is the first published scientific study of civilian deaths. As far as I know, other attempts to quantify deaths have been based on press reporting. Press reporting is not bad or incorrect, it’s just a different method. It doesn’t involve the forming-and-testing-a-hypothesis element that a scientific study is characterized by. If I’m wrong and there have been other scientific studies of civilian deaths that have been published, please let me know.

    And now I’d like to address your points one at a time in the same order you presented them :

    1. Yes, I agree that it is confusing. The authors reported both results that included and excluded Fallujah (since Fallujah was such an outlier they reported both ways in their study so the reader could have maximum information.) Because of this many statements can easily be taken out of context. Some statements only apply when Falllujah is factored in, others when it’s taken out.
    2. The conservative assumptions are spelled out. They include (1.) Excluding Fallujah since it’s such an outlier. (2.) Requiring that the deceased have lived in the household for 2 months before they died. This probably undercounted some. It’s also possible that there was some overcounting too and the authors mention this. But the researchers do have reason to think that underreporting may be a problem with self report, especially with infants and younger children see:

    Document 1

    Document 2

    3. I (and many epidemiologists) would have to disagree with this point. Just because something may seem obvious to someone, that’s no reason not to get data and check it. After all, for years we thought peptic ulcers were primary caused by stress. Turns out it’s an infection and antibiotics help most people. It’s a good thing people looked into it.

    4. I totally agree (as do the authors of the study) that there is a great need for further research.

    And that’s really the general point that I was trying to make in my post. This study certainly has limitations, everyone can agree on that. It’s just a first step to understanding what’s going on and that’s how people should view it.

  97. Joe Green & others like him here have made a simple category error. They think that this is all about the development of scientific theory. It is instead about practical affairs.

    For instance, one does not need to publish a peer-reviewed paper in order to argue on publically available evidence that the proper peer-review process has been abridged under politically high-pressure circumstances. The scientific paper has entered the realm of practical affairs with practical ends in mind. It would be grossly inappropriate to slow down our practical judgment of the matter, slow it down to a careful scientific pace, when the paper itself has violated that rule & when time is, indeed, short.

    As complicated as those things are to say, so simple are they to see. I.e., Joe Green is simply being sophistical.

  98. Pingback: i cant think
  99. More articles of interest.

    100,000 Dead—or 8,000: How many Iraqi civilians have died as a result of the war?
    By Fred Kaplan, October 29, 2004
    http://www.slate.com/Default.aspx?id=2108887

    The Lancet: A Casualty of Politics
    By Tim Worstall, Tech Central Station, October 29, 2004
    http://www.techcentralstation.com/102904J.html

    A One-Sided Lancet
    By Sydney Smith, Tech Central Station, November 19, 2002
    http://www.techcentralstation.com/111902A.html

    100,000 Civilian Deaths Estimated in Iraq
    By Rob Stein, Washington Post Staff Writer, October 29, 2004
    http://www.washingtonpost.com/wp-dyn/articles/A7967-2004Oct28.html

    From the WaPo article:
    Previous independent estimates of civilian deaths in Iraq were far lower, never exceeding 16,000. Other experts immediately challenged the new estimate, saying the small number of documented deaths upon which it was based make the conclusions suspect.
    “The methods that they used are certainly prone to inflation due to overcounting,” said Marc E. Garlasco, senior military analyst for Human Rights Watch, which investigated the number of civilian deaths that occurred during the invasion. “These numbers seem to be inflated.”

    The Lancet reports 100,000 Civilian Deaths Post-War” Posted By Alan E Brain, The Command Post, October 29, 2004
    http://www.command-post.org/2_archives/016338.html

    From Alan E. Brain’s post: 2/3 of 100,000 is about 67,000. Given that the population of Fallujah was about 500,000, that many have fled the city, and given the typical 3:1 ratio of serious wounds to deaths from bombing, that means the Lancet’s peer-reviewed article claims that everyone remaining in Fallujah has already been killed or seriously wounded at least once. Nagasaki, a city 1/3 the size of Fallujah, took less than 1/3 of these figures, so Fallujah has been hit by the equivalent of at least 3 Atomic Bombs. Or so the Lancet implies.

  100. Now, if somebody wants, for the purposes of scientitific theory to treat the Lancet article as what is “current” in the field, that’s just fine—for the purposes of scientific theory, which proceeds necessarily at its own cautious pace.

    But it is a slimy & willfully obtuse piece of political sophistry to claim that the rest of us are bound, by the “scientific standards” supposedly embodied by this obviously timed last-minute October surprise, to hold our horses of response till election day is well past.

  101. It is useful to simply point out that various Iraqi mortality figures, as accumulated by news reports, vary from 15,000 to 125,000 depending upon who was collecting the information. That logically is the LOWER bound since the accumulated body counts are reported, however the figure is very likely higher than this figure because there had to have been more civilian casualties that were not reported.

    The study provides a very wide range the lower boundary of which EXCEEDS the actual news reports (8,000 compared to the other figures). The standard deviation about the mean is a statistical measure extracted from the data, but in this case, a high confidence bound obliges a lower figure, which is what the study actually produced.

    So “ballpark figures” all by themselves are in the right order of magnitude.

    The next thing that right wing propagandists question is the large standard deviations inherent in the study, which does not by itself in any way suggest that the study or analysis is actually flawed in any way, merely that the raw data contained a lot of “noise”. Given the small samples sizes, and the other related conditions of war, this is not a surprise.

    Perhaps the most helpful single feature of the paper that the authors point out was the ability of public health officials to gain rudimentary information on mortality under highly adverse security conditions, which is perhaps the single most useful item to emerge from this paper.

    Indeed, what I find impressive is the way the lower bound of the study is in the right ballpark with respect to the news reports accumulated from civilian deaths.

    One last thing. The application of Guassian statistical processes require that all “objects be idential but identifiable”, which is not a simple thing to assert in practice when you compare casualties in Fallujah to other locations in Iraq.

    In conclusion, the estimated figure of around 100,000 civilian casualties is a credible first estimate, and the estimates for casualties among women, children and elderly adults seem to be in line with what experienced military officers would observe in a combat zone.

    The spin doctors have simply spun out on this one, and once again, they should refute real science with better real science.

  102. Gee, isn’t it a bit soon (by your standards) to assert that this study is “real science”? Shouldn’t you, by your standards, wait till you get your claims & arguments vetted & peer-reviewed & published by a scientific journal?

    “Small samples,” “large standard deviations inherent in the study,” “the raw data contained a lot of noise,” a problem with achieving that all objects be “identical not identifiable,” “a credible first estimate,” “estimates…seem to be in line with what experienced military officers would observe in a combat zone.” (emphasis added)

    Oh and it’s all merely science, no political context, yeah, right. Though author insisted it be published before election day. Merely science puttering along, credible first estimate, lots of noise, etc. Sure, lots of credibility there. The author by his insistence on a politically based constraint for the study’s publication has chosen the form & context for the judgment of it. Can’t act politically & then insist on non-political standards. Straightforward intellectual dishonesty.

  103. ForNow states:

    “Joe Green & others like him here have made a simple category error. They think that this is all about the development of scientific theory. It is instead about practical affairs.”

    It is reasonable enough to make this assumption, given that the Lancet is one of the most respected medical science journals in the world.

    You just do not like the political consequences of the high estimated mortality rates for Iraqi civilians, which I suppose is a natural enough response from an Administration that is inept, incompetent and irrational as (R)Senator Richard Luger says it is.

    That situation however is not the fault of the Lancet or the authors of the paper.

    Let me put this into perspective. 100,000 souls is like 143 separate Rev Jones mass suicides in Guiana where some 700 plus Americans died at the hands of Christian religious zealots and fanatics.

  104. Lancet has now displayed a pattern of politically motivated publication. I remember when Scientific American was highly respected, before it became politicized & dumbed down. It’s absurd to claim that we are supposed to take for granted that Lancet is not subject to political motivations, & it is a category error—& a hard one to regard as sincere—to think that a paper published to exploit politically the brevity of time remaining before the election should itself be immune from political interpretation. Those who defend this on the grounds of scientific respectability are causing damage to science.

  105. ForNow,

    You sound more and more like a member of the Reformed Libertarian Church of Objectivism and its high priestess Ayn Rand, than you do as a scientist or even a lay person interested in science.

    This however is actually a religion at the opposite end of Communism and its high priest Karl Marx. A close study would show you that its an optical illusion to consider these separate theologies, rather they are two faces of a common coin.

    Obviously, neither Rand’s Objectivism or Marx’s Communism are even remotely connected to science. Rather they are both a form of fanatical religious zealotry of the type that motivate Mr. Bin Laudin and Mr. Bush.

  106. Also, I think that Joe Green thinks his argument rhetorically so strong that he is politically spared from responding to what I’ve saying—no grasp of it is built into his response. Politics has over-shaped his responses as surely as they over-shaped the faulty process of the paper’s preparation for publication.

    No Ayn Randianism here. I tend to go more with C.S. Peirce, without the triads.

  107. The paper has entered into a political domain, like William Shockley wearing a lab coat, for crying out loud, on the Morton Downey show. That lab coat won’t protect you from political judgment or from your acts’ disclosing their political meanings & intents.

    But I don’t think you understand what I’m talking about.

  108. The paper that has been published in the Lancet is a credible and useful paper. Its methodology and its results are not only sound, but also interesting in the field of public health.

    There does exist a valid and useful way to challenge the results using proper scientific studies and processes that are well known. But “bald statements” are not the same thing as valid and significant counterveiling scientific papers.

    The critics of the authors have criticized the large standard deviations in the paper citing that as reason why the paper is invalid. However, I have given one example of where the results of the study in fact overlap and tie into other valid and independent observations and reports, in the context of an examination of the LOWER BOUND by comparing the predicted lower bound of 8,000 to the news reports in the region of 15,000, taking into account the high confidence bound used by the authors.

    Now, this paper is not related to any political consideration, other than the desire of the authors to make its results available to interested American voters.

    I am quite sure that 100,000 civilian casualties or 10,000,000 civilian casualites in Iraq will not make much of a difference to blood thirsty neo-conservatives like Paul Wolfwitz and Richard Perle.

    However, for others that are concerned with “pro-life” moral issues in the election, 1,000 dead servicemen or 100,000 dead civilians is important.

    You can hardly blame the Lancet or the paper’s authors for these figures anymore than you can blame the Pentagon for anouncing the casualty figures that its servicemen have suffered.

  109. Joe Green still mixes the scientific issues with political issues. Even if it becomes boring…
    The study
    One of the results of the study is:
    There should be 8000 up to 194000 death in post-war Iraq.
    Conclusion: This means the death rate is higher than in pre-war Iraq OR NOT.
    The interpretation is up to you.
    In “Findings” you read: “We ESTIMATE that 98000 more death than expected…”
    In “Discussion” you read: ” The FACT that more than half the death reportedly caused…”
    Looking at “Figure 2” there seem to be nearly no violent deaths before the war. About 2 or 3 days before the war Saddams Special Forces gunned down 170 visitors of a mosque in Southern Iraq because they protested against his regime after after the FRiday prayer (source: Alliance Internationale pour la Justice, NGO, strictly anti-war). If one of their samples would have been made there, they would have completely different numbers.
    In “Discussion” the authors enter a completely different field: Military tactics:
    “In the interim, civilty and enlightened self interest demand a re-evaluation of the cosequences of weaponry now used by coalition forces in populated areas”.
    -Insertion: Medical studies that bring up a new aspect, that cant be approved 100% safe, usually end with the sentence: “Further studies should be done…”-
    Back to the last sentence of the study: If you want to compare, how carefully or not occupying forces are doing their job, compare pictures of Falluja with pictures of Grosny…
    And, are the authors used to military tactics?
    I think, after all, one can trust in the US military, doing their job as carefully as possible. Equally if the name of the next President is G.W. Bush or John Kerry.
    So we are medias res in the politics.
    I think the terms used and the opinion expressed in the sentence : “I am quite sure that 100,000 civilian casualties or 10,000,000 civilian casualites in Iraq will not make much of a difference to blood thirsty neo-conservatives like Paul Wolfwitz and Richard Perle” tell us, that Mr. Green is kind of a holy warrior. For these people science is an instrument of propaganda. This was a normal practice in the former Eastern Bloc. Nowadays we should stay away from such behavior…

  110. Shannon Love,

    Thanks for initiating and hosting this interesting discussion. I think that most sides in this debate can feel that their positions were adequately represented by commenters.

    The name-calling that took place was, for the most part, embedded in longer posts that attempted to deal with some of the substantive issues that Roberts et al. raised. So I think it’s fair for the reader to consider mockery, sarcasm, and ad hominem attacks as a part of the argument being advanced by a given post’s author.

    As is always the case, the technical and statistical and epidemiological issues will take time to resolve. Taking the decline of Scientific American as a sadly instructive example, The Lancet‘s Editors may never be sanctioned for downgrading the importance of pursuit of the scientific method. Too many physicians and scientists evidently think of its de facto replacement by political correctness as a feature, not a bug.

    I can only hope that the short-run political impact of this study’s publication will not be as its authors and editors so plainly intended.

  111. I am quite sure that 100,000 civilian casualties or 10,000,000 civilian casualites in Iraq will not make much of a difference to blood thirsty neo-conservatives like Paul Wolfwitz and Richard Perle. Hardly the kind of statement one should make if he is to pretend to credibility, unbiased scientific reasoning and an ability to state what is or isn’t political.

    Never mind…

  112. @AMAc
    Good point to refer to the decline of “Scientific American”. I can still remember the discussion about Björn Lomborgs “The Sceptical Environmentalist”…
    Greetings from Germany

  113. There is a Grima Wormtongue like reasoning in the posts by Joe Green et al. Keep everyone focused on the science and maybe they won’t notice the politics. The same goes for chel. I could point out that a methodology that may be appropriate for the study of epidemics and public health is not in fact appropriate for a study involving violent death which is man induced and specifically targetted and therefore not subject to extrapolation from a random sample no matter how broadly based. I could also point out the obvious problems of the assumptions of the study (e.g NO violent deaths pre-war in the whole sample .. this is Iraq isn’t it). But in the end this is all irrelevant (except maybe to some future researcher who will no doubt seek publication in a peer reviewed journal). The fact is that this was from start to finish a political exercise. The executive summary of the report makes bold assertions unsupported by its own research (was the report just to give verisimilitude to an otherwise bald and uninteresting tale?) aimed at creating an impression of sharply increased violent deaths of “innocent non-combatants” at the hands of coalition forces. The final paragraph of the report demands a political response to a problem that has not been identified, only suggested. And finally against all known Lancet procedures this report is carried out, written, peer reviewed and published all within 2 months in order to include it 4 days before the US election. ITS THE POLITICS STUPID!!! Once and for all let’s stop (and I plead guilty in this also) arguing the finer points of scientific protocol and state it bluntly. No critic of the anti report posts here has even attempted to answer the real questions. Why was this report fast tracked? Why was it published just 4 days before the election? Why, given that there have been questions about its methodology and conclusions, wasn’t it further reviewed before publication? The answer is obvious. To influence the US vote while minimising the time available for a response. Read Blithering Bunny for a link to the Lancet’s editor’s justification for publication and then tell me its about science and not politics. WE, the educated public, have every right to respond as we have. Its not about science, not about humanity, its about grubby partisan political advantage. Period. Thankyou for all those who have posted here and shown that they have a brain and a conscience and won’t be conned by politics masquerading as disinterested research.

  114. Sylvain Galineau wrote:

    quoting Joe Green:

    “I am quite sure that 100,000 civilian casualties or 10,000,000 civilian casualites in Iraq will not make much of a difference to blood thirsty neo-conservatives like Paul Wolfwitz and Richard Perle.”

    Sylvain opines:

    “Hardly the kind of statement one should make if he is to pretend to credibility, unbiased scientific reasoning and an ability to state what is or isn’t political.

    Never mind…”

    The science is separate from the politics. The paper published in the Lancet stands on its own merits. The political comments made, likewise stand on their own merits.

    The Lancet not only continued in the long tradition of meeting the highest standards of medical science by publishing this interesting and valid paper, but it also performed a public service for the American public.

    For many “pro-life” people that believe that killing is wrong, this paper provides useful new insights into the suffering of Iraqi civilians, women and children at the hands of American Occupation Forces.

    As for neo-conservatives like Mr. Wolfwitz and Mr. Perle in the Bush Administration, its understandable why they would seek to suppress medical science and its work in Iraq. Mr. Himmler did much the same kind of thing in the Third Reich.

    When “neo-conservatives” are confronted with the moral dimensions of their belief system, the usual response is the same one provided above by Sylvain — “never mind”.

  115. >>”The Lancet not only continued in the long tradition of meeting the highest standards of medical science by publishing this interesting and valid paper, but it also performed a public service for the American public.”

    Joe, what makes you view this as a “valid” paper? Do you disagree with the logical and statistical criticisms leveled here? If so, why?

  116. JoeGreen, you just don’t want to see do you ? This study does not have a secure scientific foundation. There maybe more (or less) “civilian” deaths than the estimates such as 16,000 odd produced by counting media reports, but this report adds nothing useful to the debate except as a political stick to beat Bush with. The Clustering technique is a not a particularly reliable tool with non-random distributions, the randomisation was compromised anyway, alhough some checking was carried out not all deaths were confirmed (and self reporting is notoriously unrelaiable), and probably most damning of all, the pre-invasion death rates are of an extremely doubtful nature and are almost certainly way too low. All of these lead to the conclusion that although the actual death rate is not particularly well known, this study is worthless in improving that situation.

    Were you also aware that the extrapolation of “100,000 deaths” is based on an actual body count of 21 ?

  117. Oh, Joe Green (is this your real name?):
    I am reading the Lancet since many years.
    And I believed this was a paper based on scientific knowledge.
    This all vanished with this article.
    It was published for political reasons, not for science. How mad!
    To get a bit more personal:
    About all the “pro-life” people: Why didn´t we hear from them anything when Saddam was killing hundreds of thousands of people (not to talk about, when Pol Pot was killing 2 million people in Kambodscha, 800000 were slaughtered in Ruanda, did we hear something from the “pro-life” people?). What a shame! What a hypocrisy!
    Sitting maybe in Cafe´s, talking about, that you should enjoy your life as good as possible, take what you can get, because you have only one life. One the other side telling the Iraqi population: Ok., is takes time, to free you from your dictator…, be patient…, tolerate some more thousands being killed or tortured or…, time will bring a solution…
    What a hypocrisy!!! I hate this!
    Friends of mine are still working in Iraq (WADI e.V., look at http://www.wadinet.de, working there in the kurdish non-flight-zone since 1993), they have different opinions like those of “Joe Green”. They came from the political far left, but they were cured through realities, they found in Iraq. They supported the war!
    Not: everything better but Bush, but: everything better than Saddam ruling Iraq!
    Greetings from Germany

  118. To Skip Smith:

    “Joe, what makes you view this as a “valid” paper? Do you disagree with the logical and statistical criticisms leveled here? If so, why?”

    I have read the paper from a direct perspective of mathematical statistics. The raw data is what it is, a given. The less than ideal conditions of collecting data is one aspect of this paper that gives it value in the sphere of public health, namely estimating mortality in a conflicted area or country in the absence of a functioning public health system that would otherwise provide accurate figures.

    The numbers fall out of the data directly. By picking a high confidence bound (the three sigma level at around 99%) means that the LOWER bound is necessarily driven lower. The paper provides the figure of 8,000. That compares to the INDEPENDENT number of 15,000 from body counts reported by news agencies. This therefore is consistent, since the lower bound of an estimate is LOWER than what has been provided by another independent “experiment”.

    So I find the paper interesting, controversial, and likely to prompt more work to refine the standard deviation to smaller levels, thereby sharpening the accuracy of the measurement.

    I think its useful to go back in the history of science and see what happens in science when a contraversial paper is published. I would refer you to Albert Einstein’s famous paper on “special relativity”. That also set off a frenzy and more papers.

    Eventually, a scientific consensus was reached. The postulate made by Einstein, namely that the speed of light is finite and constant, was eventually confirmed by other experiments and other papers.

    With respect to peer reviewed papers, I see this more as a safety net, than a censor. No scientist worth his keep is prepared to declare that he is “infallable”.

    One last point. With medical science in particular, the questions of ethics and morals are never far away.

    Most doctors swear an oath to “first do no harm”. The militarists in the current Washington administration, like those in Japan in World War Two, have very little regard for “ethics” and “morals” because what they call “collateral damage” is simply the killing of innocent civilians, woman and children that is the subject matter of this study.

    The paper is making the case that the war in Iraq has caused civilian mortality that lies between 8,000 and 200,000 souls. The “best estimate” is around 100,000. Given the increasing tempo of the insurgency, this could easily turn out to be on the low side.

    After all, this “Vietnam Two” is likely to produce similar results in terms of civilian casualties before this is all done. In Vietnam, over 2,000,000 civilians lost their lives in that “war of national liberation” and American casualties exceeded 55,000 young men.

    The military reasons why the United States is going to also lose this insurgency war is beyond the scope of this newsgroup. This news group has tried to address the question of civilian casualties caused by American “unilateral foreign policy” as practiced by the Bush Administration.

    I find the authors have produced a credible estimate worthy of further study, something they themselves suggested for the obvious reasons given in the paper.

    Finally, a word about “neo-conservatives”. Most of their ideological leaders are moral cowards. They play on religious prejudices, biases and ignorance. And when you blend that with religious fundamentalism, as has taken place in the Bush Administration, you have the moral equivalent of Osama Bin Laudin’s Al Quada organization.

    It takes integrity to recognize the moral equivalence of the two positions between Bush and Bin Laudin. Whether it is fuel laden aircraft falling upon civilians in New York City, or American bombs falling on Iraqi homes filled with civilian woman and children, the only correct description from the perspective of the civilian victims is that they were attacked by “terrorists”.

  119. To F. Hoffman

    My name really is is Joe Green.

    You mentioned Ruanda. The genocide in that African country happened because of the obstructionist behavior of the United States in the United Nations Security Council.

    The military forces with the United Nations were under the command of General Romeo Dellaire of the Canadian Armed Forces. They tried to prevent the ensuing mass murder, but were not permitted to take aggressive action against those doing the killing, and that in turn goes back to CIA interference with the Security Council of the United Nations. Read General Dellaire book, “Shake Hands with the Devil”.

    America had no interest in Ruanda because Ruanda does not have oil and gas reserves.

  120. Joe Green,

    I think you miss the primary point here. There are two different sets of numbers we are discussing here.

    The first set is the raw count produced by the survey itself. This is the 46 pre-invasion deaths versus the 182 post-invasion deaths.

    The second set is the statistical model produced by running the raw numbers through the studies demographic model. It is this second set that produces the broad spread.

    My primary objection is to the methodology used to create the raw count in the first place. Cluster sampling and self-reporting are not accurate measuring methods in this case.

    As evidence I offer up in this subsequent post the observation that the study found pre-invasion infant mortality rates substantially lower than any other study conducted before the invasion. This study’s methodology returned a pre-invasion infant mortality rate of 27/1000, lower than the mortality rate of the pre-gulf war one Iraq of 47/1000 and and nearly 1/4 of the 2002 unicef rate of 102/1000.

    The study’s methodology is garbage. That is the real debate.

  121. “My primary objection is to the methodology used to create the raw count in the first place. Cluster sampling and self-reporting are not accurate measuring methods in this case.”

    Actually, cluster sampling isn’t a _ridiculous_ method to use, at all. Most of the large scale population surveys that I can think of use some form of cluster sampling — in a lot of cases it’s the only approach that’s remotely affordable. Similarly, I’m hard pressed to figure out how it is that you get around self-reporting in some form in such a chaotic environment. Do you expect the central morgue registry to have recorded all deaths? Given the circumstances on the ground, I certainly wouldn’t. Lacking some central registry, what then is the alternative to self-reporting? I agree, there are some real substantive weaknesses in the study’s approach, but the authors have at least attempted, however imperfectly, to mitigate the issue to some extent.

    Where the survey falls down, and falls down big, is in the small number of clusters that it surveys. As you quite properly point out, this isn’t even close to enough sample to document a set of events as erratically distributed as deaths due to military action in a modern warzone. Frankly, I think the authors have little insight into the nature of death in wartime as a phenomenon, and are biased to think of it as being akin to a disease or even to death due to homicide, and this is skewing their thinking for the worse.

    All that said, I think it does become us somewhat to appreciate that this is survey work being undertaken in perhaps the most difficult environment that I have ever heard of. Typically the folks that collect the data for the surveys that I deal with only have to worry about getting cold and tired, rather than rendered violently dead.

  122. Dave,

    Cluster sampling is not bogus per se but it does amplify variations in asymmetrical samples. Violence in Iraq is highly asymmetrical. Some cities and even some specific neighborhoods get hammered while other see no violence at all. Air strikes in particular, which the study blames for most combat deaths, have occurred overwhelmingly within about 120 kilometers of Baghdad.

    Likewise, self-reporting isn’t a bad method for every phenomenon but in this case we have at least one sub-population, the Sunni’s who have a strong incentive to magnify the deaths. Moreover, self-reporting in many 3rd world cultures fares poorly for many cultural reasons. The actual number of deaths recorded in the survey is so small that even a handful of lies could significantly distort the study.

    Whether the researchers where forced to use cluster sampling and self-reporting isn’t the issue. I am sure that they chose these methods due to constraints of money, time and safety. The paper says as much. The question is, however, did the methodologies they were forced to use generate accurate results? I say it didn’t.

    The researchers and many defenders of the study seem to feel that any attempt to study the problem is worthwhile even if the necessary tools are not really available. I think they’re wrong in this. Bad science is far worse than no science at all. Not only does it provide bad information but it gives the bad info the aura of scientific authority.

    I think the honest, non-political researcher would simply admit that the conditions in the country and the resources available to the researchers did not permit a scientific evaluation of the change of mortality.

    Publishing this study was scientifically dishonest and unethical.

  123. For Shannon Love who wrote:

    “My primary objection is to the methodology used to create the raw count in the first place. Cluster sampling and self-reporting are not accurate measuring methods in this case.”

    Lets deal with this one step at a time. First of all with respect to the issues of “bias”, “self reporting” and so forth.

    This study was done in the context of a war zone and prior to that in the context of a totalitarian state, that never the less had some degree of external observability via the United Nations Food and Drug Program. Dr. McDermid who is a Congressman that visited Iraq prior to the war, did not report significant shortages of drugs in Iraqi hospitals before the war, in the sense that such shortages were the cause of infant mortality (and for that matter adult mortality).

    In the post invasion scenario, the combatants have strong motives to distort and pervert any numbers, the US Army had a strong incentive to deny “collateral damage” to civilians, particularly children; and would systematically produce under-reporting. As for the insurgents, systematic claims of American attrocities would over report casualties. Self reporting by eye witnesses in a war zone is perhaps the best available technique, and that is what the authors have done. Remember, the major value of the paper is to establish mortality estimates in the face of collapsed public health systems caused by the war itself.

    Now as for “sample sizes, clusters, etc”, you are trying to somehow shape a “representative sample” as perhaps a pollster might apply that idea in an election. But that is not the only method open to apply statistical analysis to a set of raw data.

    Now unless you are suggesting that somehow these authors are part of a “disinformation campaign” in the context of covert warfare, I think you are obliged to accept their raw data as a starting point that is valid.

    From what I can see, you have no quarrel with the mathematical analysis that has been applied to the raw data, unless I misread your post. Given that as the case, your only opening would be to attack the “imbedded information” that statistical analysis seeks to extract.

    There is something called a “chi test of statistical significance” but I have not looked very closely at how it might be applied, if at all. The test is used to measure “significance” of parameters that statistical analysis is trying to measure, and it is commonly used in wildlife ecology to measure population dynamics.

    On this foundation, I do not see your criticism as being valid, or alternatively sufficiently well developed (via an avenue such as a test for statistical significance).

    You have another parameter that you might also consider modelling, and that is a model that includes an “epidemic” kind of explosive growth in mortality caused by violence in Iraq. That is because, “terrorists” or “freedom fighters” are motivated and trained with processes that rapidly grow, sometimes even exponentially.

    There is no shortage of weaponry in Iraq, and there STILL is a shortage of American troops to guard and secure these weapon depots that were created before the war; so that insurgent’s only real limitation in inflicting troop mortality is related to their own number.

    The level of conflict will escalate and will inflict elevated “collateral damage” to civilian populations, bystanders, women and children.

    So far everything you have discussed is about the paper from a “public health perspective”.

    There does exist a different view, and that is a “military perspective” that would require a careful analysis of “bombing accuracy”, bomb “size” and “killing range” and “maiming range”, which are figures well known in the Armed Forces since these are the specifications used to design and build the bombs in the first place.

    Thus if you simply take the military bombing data, locations, “hits and misses”, you can arrive at mortality by a different route. I would be very surprised if the US Armed Forces did not have these numbers inside the Pentagon, although these would be closely guarded secrets.

    The point being, that both approaches are “valid” notions to arrive at mortality, although historically the antidotal stories of civilian victims of war tend to be the ones most often borne out after the facts.

    To conclude, I think that the paper is a useful effort under difficult circumstances, even dangerous circumstances for the interviewing teams. In a combat area of escallating violence, where each month, the US Armed Forces suffer greater casualties than the previous month, that the casualties on the other side is likewise mounting, and the civilians, woman and children caught in the violence is also mounting.

    The figure of 100,000 civilian casualties therefore could very easily be on the low side, and could be trending toward the UPPER LIMIT, not the LOWER LIMIT.

    Why is it so hard to believe that in a country of 24,000,000 souls (and that is decending into civil war with the loss of many troops and insurgents) that civilian mortality of 200,000 after a year and a half of warfare is not only possible and but highly likely???

    I think that these estimates are pretty useful for NGOs and other humanitarian organizations, including the Red Cross and the Red Crescent.

    And I would be very surprised if the Pentagon’s secret figures of civilian casualties are much under 240,000 level after a year and a half of daily combat operations in Iraq.

    There simply is no such thing as a “surgical strike” when it comes to civilians, and this study and past history bears this out in perfect hindsight.

  124. For F. Hoffmann who wrote:

    “What a hypocrisy!!! I hate this!
    Friends of mine are still working in Iraq (WADI e.V., look at http://www.wadinet.de, working there in the kurdish non-flight-zone since 1993), they have different opinions like those of “Joe Green”. They came from the political far left, but they were cured through realities, they found in Iraq. They supported the war.
    Not: everything better but Bush, but: everything better than Saddam ruling Iraq!”

    You are casting aspersions because you have no idea about my politics. Lets deal with the issues one step at a time.

    1. The paper that was published in Lancet was a pretty interesting, if controversial, paper related to the collapse of public health in a war zone. It has made an estimate of mortality due to the violence of the war, and estimated that the civilian casualty figure was 100,000 due to the violence caused by the American Invasion and Occupation of Iraq. That is not only credible, but fits rather well with the experience of war by civilians in other parts of the world.

    2. The criticism leveled at the Bush Administration in relation to these civilian casualties is not about the “far left”, or the “far right”, but rather from a “moral” and “ethical” perspective. As a signatory to the Geneva Conventions, the United States has a legal and a moral obligation to observe its imperatives. Indeed, the prison abuses at Abu Graub are directly traceable to “war crimes” committed by members of this administration, specifically Mr. Rumsfeld, Mr. Wolfwitz and Mr. Perle who specifically stated that they were not going to observe international law in the conduct of their “war on terror”. We tried Nazi war criminals in rather similar circumstances at Nurenburg.

    These in fact are “war criminals in the making”.

    3. With respect to indigenous Iraqi that have joined the insurgency to rid their country of invaders, these individuals should not be confused with “international terrorists” or “foreign fighters” that have been attracted to Iraq by the American Invasion because they seek to engage in the religious practice of Jihad against the Americans and the British.

    All people have a right to defend their homelands against foreign invaders, particularly in the absence of any United Nations mandated resolutions authorizing an invasion by member states.

    4. With respect to Al Quada that is now inside Iraq and growing, these fanatics have been bred in the failed states of Africa and the Middle East and are now exploding onto the world stage. They fundamentally cannot be stopped until you have stable nation states with the required security and police forces to either arrest them or destroy them.

    5. The American invasion and occupation of Iraq has given comfort to Russia to return to its old autocratic and anti-democratic ways, which is now playing out in the elections in Ukraine. “Anti-terrorism” is breathing new life into the politics of the old Russian KGB from which Mr. Putin hails. Like his father #41, this President has a strategic objective to reconstitute the Soviet Union.

    6. My final point is that right wing American fanatics and religious zealots are morally equivalent to the Muslim fanatics that attacked New York on September 11, 2001 and that they are killing Iraqi civilians in greater numbers then the Muslim fanatics were killing American civilians. There is nothing new here since the days of the Crusades in the Middle Ages. The only actual question is how to properly establish effective security arrangements inside all civilized countries, including Iraq, with minimum loss of life, and at minimum cost.

    Its important to point out that not all these actors actually desire peace and stability. Some relish the war, particularly when others do the bleeding and the dying.

    The lesson is that the only valid alternative to war is the “rule of law”, and that would require that we charge all those responsible for violating international law with war crimes, and punish them accordingly.

    Bringing Mr. Sharon and Mr. Bush Sr. to trial for international war crimes and bringing fanatics like Bin Laudin and Al Sawkawi to trial, would be the beginning of the end of violence if people could see “the rule of law” reassert itself.

    But that is not going to happen because warfare is profitable when the prize is the second largest petroleum reserves in the world, and corrupt CIA storefronts like the Carlyle Group are waiting to rip off the citizens of Iraq. And indeed, get the American taxpayer to foot the bill for the bloodshed both in terms of blood and treasure.

  125. The critique of cluster sampling as a method for assessing casualty rates is bogus. The problems with getting a representative sample through cluster analysis are incorporated in the standard errors and the design effects, as stated in the report. That’s why the confidence interval is so wide.

    If Shannon Love has some argument as to why the confidence intervals should be even wider than the ones reported, she should put it up (and, incidentally, publish a paper on the subject, as such an argument would certainly advance the state of knowledge on the subject).

    If not, then we take the calculated confidence interval as reflecting the uncertainty in the estimate as a result of cluster sampling, and we note that there is less than a 2.5% chance that one would have got a sampled death rate as high as this if the invasion had not made things worse.

  126. Joe Green (multiple posts):

    Thanks for raising technical points and writing about your politics in a clear way.

    The first task of a commenter is to present perspectives and arguments in a way that readers understand and appreciate. A second task, for many of us, is to persuade.

    In my case, while I find few of your rebuttals of Love (et al.)’s criticisms to be compelling, I do see that there are reasonable points to be made. I suspect Roberts’ and the Editors’ defenses would (will?) run along the lines you have sketched.

    If I had a political web-log, I would hope to have you as a frequent commenter.

  127. To AMac:

    Thank you for your comments. I originally posted here because its important to defend science, its methods, and the freedom of its practitioners from those that seek to silence researchers with propaganda, inuendo, manipulations and outright lies. Leonardo da Vinci suffered for these freedoms and we should not surrender them easily or lightly.

    Secondly with respect to the religious fundamentalism and fanaticism that is sweeping the world, you can find it in Christian, Jewish and Muslim circles. It is inevitably coupled with “conservativism” in all its various ugly forms and disguises, and it seeks to subdue and enslave mankind to a new dark age.

    Liberalism was and remains the fuel of the Enlightenment, and as the fanatical Christian right seek to destroy the fanatical Muslim forces of fascism, at a profit if possible, it is clear that there does not exist any peaceful end to this pathway. Rather, its only through the Liberalism of the Enlightenment with its values for knowledge, its love of learning, its compassion and its empathy for the poor, that a peaceful way forward might be found to what Churchill called “broad sunlight uplands”.

  128. “Whether the researchers where forced to use cluster sampling and self-reporting isn’t the issue. I am sure that they chose these methods due to constraints of money, time and safety. The paper says as much. The question is, however, did the methodologies they were forced to use generate accurate results? I say it didn’t.”

    Well, that’s not quite what you said in the initial comment that I quoted. You said that your primary objection was to the methodology used to generate the raw counts, and cited those limitations (clusters and self-reporting) as the source of your objections. This second quote sounds a great deal more like you are objecting because you don’t like the number produced, and using the methodological criticisms as a means of buttressing that objection. From what I can see here and elsewhere, you’re not alone in this regard and, in fact, you’re in what I not so humbly think is good company (i.e. myself). Personally, I tend to think that the discrepancy between this study and other sources is so large that a simple statement to the effect that passive means of observation are frequently insensitive isn’t sufficient.

    That said, it’s important that we be very careful in our criticism of our colleagues here. I have a very distinct feeling that what we’re looking at is a survey programme that got run over by a particularly harsh reality. There are hints throughout the paper pointing to the fact that a reasonably honest attempt at quantifying this phenomenon got overtaken by a rapidly deteriorating security situation, and there was a decision to salvage what they could of their research programme. To be perfectly blunt, this happens in one guise or another all the time — there’s a hell of a lot more imperfect or even downright “faith-based” science out there than even most practitioners of the art appreciate — salvage science of this type is not necessarily a bad thing, if the study is reasonably up front about it, which as near as I can determine they have been. All of us know what the structure of the research was, and have a pretty good idea of what its limitations are, as well as the potential implications of those limitations. Where we disagree is to what degree that matters in how we use the results, and given that so many of us are viewing these results through a political lens in the immediate run-up to an election, that disagreement is somewhat more vitriolic than it might otherwise be.

    Where I think their methodology might justly be criticised is in the fact that they are attempting to characterize a highly punctuated phenomenon using far too few clusters. I tend to think that this is due to their background in the study of disease, rather than war. That said, this is a different matter than criticizing the use of cluster methods — they do seem to have a decent methodology, they just don’t have enough resources to apply that methodology in order get a decent sample with any reasonable power.

  129. Joe Green,

    I don’t suppose it’s sporting on my part to point out that the greatest killers of the 20th century prided themselves on their atheism, reason and their devotion to science? Any randomly selected human in the 20 century was more likely to die at the hands of a mutant monster child of the Enlightenment than to die at the hands of a religious extremist.

    I my self am an agnostic scientific materialist but it is plainly evident to me that secular fanatics have proven far more dangerous than religious ones and even more irrational and dogmatic. Unlike the religious fanatic, they have no restraint of tradition or culture. Arrogantly convinced of their own rectitude and intellect, they sacrifice the innocent not to some god but to the hunger of their own egos.

    The 20th century welfare state is the result not a secular movement but springs from what might be best described as Christian socialism. The electorates that voted for these programs did so overwhelmingly because of their Christian beliefs, not their dedication to secular reason and science.

    Bad science kills. Lies decorated with lab coats have been used to oppress and murder more people in the last 100 years than all religions combined. This Lancet study is merely the latest in a long line of dubious “science” cranked out to advance a political agenda. It is a propaganda coup for the enemies of reason, peace and democracy.

  130. Shannon Love:

    “”A cluster sample survey was undertaken throughout Iraq during September, 2004″

    It is bad practice to use a cluster sample for a distribution known to be highly asymmetrical. Since all sources agree that violence in Iraq is highly geographically concentrated, this means a cluster sample has a very high chance of exaggerating the number of deaths.”


    I think you are wrong here. It should be the other way around. There is a greater chance of underestimate the number of deaths since it is likelier to pick places with no or little violence if the violence is concentrated to a few places. However, you are right in that if such a place is picked it will skew the extrapolation.

    An example: If you have a pot with 9 white and 1 black ball and want to extrapolate the number of black balls by picking one ball, then in 9/10 tries you will pick a white ball and extrapolate 0 black balls, wich is a slight under estimation. In 1/10 you will pick the black ball and extrapolate 10 black balls, a wild but uncommaon exageration. So I assume you are not correct when you claim there is a very high probability of an exagerated result due to the fact that the violence is highly concentrated. (It may be exagerated for other reasons though)

  131. incog,

    Cluster sampling doesn’t work as you describe. It works as follows:

    Suppose you want to determine the ratio of white to black balls in a sample of 100 balls. In this case, the balls are not all in one container but spread out over 10. Each container is a cluster.

    Now, as long as the balls are randomly mixed together you could safely choose just one container and get a fairly accurate measure of the ratio. In this case cluster sampling works because each container is safely statistically interchangeable with every other container. Measure one, you’ve measured them all.

    But suppose the person who filled the containers said, “Well I just opened a bag of black balls and filled containers until I ran out and then I switched to white.” Now you’ve got a problem. If you select just one container you know it will not be representative of the other containers. Cluster sampling will fail in this case because each container is not interchangeable with any other container.

    Cluster sampling is a pragmatic compromise used by real world researchers because of the constraints of time, money safety or other factors. There is no shame in using it but it has built-in errors the most major of which is that the more asymmetrical the distribution of the phenomenon being studied the more inaccurate cluster sampling is.

    The distribution of violence, medical care and other influences on mortality in Iraq is more like the second case than the first. The Falujah cluster just illustrates this point. With the inclusion of the Falujah cluster the casualty rate soared to something like 250,000 which is obviously nonsensical so they had to exclude it to get a number anyone would believe.

  132. This Lancet study is merely the latest in a long line of dubious “science” cranked out to advance a political agenda

    This is a pure and simple slur on the Johns Hopkins researchers and you should be ashamed of yourself for making it. You started with a few potentially interesting (in my view incorrect) criticisms of the use of the cluster sampling methodology. You are now making hysterical and baseless attacks on the honesty of people who have quite literally put their lives on the line in order to collect this data. This is not only a shameful way to behave in and of itself, it reduces the credibility of your substantive points.

  133. There is no shame in using it but it has built-in errors the most major of which is that the more asymmetrical the distribution of the phenomenon being studied the more inaccurate cluster sampling is.

    This is now the third or fourth time that I have asked you to explain why it is that you think this inaccuracy was not captured by the published and estimated standard errors, and you have ignored me every time. Please could you explain why it is that you think this study’s use of cluster sampling would lead to errors greater than those estimated by the software they used for analysing the data.

  134. Shannon,

    you are kind of making my point.

    “Now, as long as the balls are randomly mixed together you could safely choose just one container and get a fairly accurate measure of the ratio. In this case cluster sampling works because each container is safely statistically interchangeable with every other container. Measure one, you’ve measured them all.”

    Yes, in this case it works as it should. In my case above which is equivalently to stuffing 10 black balls in the first container and 10 white balls in each of the others it does not work properly. But, if you pick one container you will have 9/10 chance of picking a container with white balls, in which case your extrapolation of black balls is 0. This is quite close to the true value of 10 black balls. If you are unlucky you pick the container with the black balls, then you are force to extrapolate 100 black balls, wich is a big exageration. Now if the experiment is conducted many times the mean of all experiments will converge at the true value of 10 black balls, but a correct extrapolation from one single pick is impossible. The point is that the case when you under estimate the black balls is 10 times more common, thus if you extrapolate from one container the likliest result is a slight under estimation, but it could be a big over estimation as well. Thus, if the violence in Iraq is highly concentrated to small geographic areas then the case of an underestimation is liklier simply because it is liklier to pick a cluster with little or no violence. Of course the figure in the report could still be far to big, but I am trying to counter your point: “…a cluster sample has a very high chance of exaggerating the number of deaths”.

  135. I think the oft-quoted adage “a little knowledge is a dangerous thing” applies to the recent discussion here. There are a few people defending this study who know some statistical terms, but don’t really understand them or the statistical and logical issues raised by the critics.

    On the issue of cluster sampling: cluster sampling is done primarily to reduce the cost and effort of conducting a survey. Even the most basic references on cluster sampling warn that the clusters must be homogeneous. For example:

    “Elements within a cluster should be as heterogeneous as possible. But there should be homogeneity between clusters. Each cluster should be a small scale version of the total population.”

    http://encyclopedia.thefreedictionary.com/Cluster%20sampling

    Clearly this is not the case when one cluster is Basra and one is Falluja. The consequence of heterogeneous clusters is high sampling variability and thus unreliable estimates of population parameters (such as death rates). A thought experiment — how different would the study results look if we replaced the Falluja cluster with one of the Basra clusters (it easily could have happened during the random selection of clusters to sample)? That’s sampling variability — a different random draw changes the sample — and sampling variability is especially high when sampling heterogeneous clusters.

    Further, note that we can’t make this problem go away with some magical statistical black box. Dsquared states: “The critique of cluster sampling as a method for assessing casualty rates is bogus. The problems with getting a representative sample through cluster analysis are incorporated in the standard errors and the design effects, as stated in the report. That’s why the confidence interval is so wide.” The statistical corrections for cluster sampling assume the cluster sample meets all the usual assumptions (homogeneity across clusters), so the original criticism still stands.

    Further, a lot of people don’t seem to understand the basic concept of sampling variability. For instance, Joe Green says “The raw data is what it is, a given.” Well, yes and no. The *sample* is what it is, but that’s no guarantee that this is an accurate reflection of the *population* we are studying. If we were to conduct an identical study we might come up with completely different inferences due to sampling variability — and sampling variability will be high when sampling from heterogeneous clusters.

    Thus, conclusions such as: “The numbers fall out of the data directly.” are correct but meaningless. The researchers derived their numbers directly from the data they have, but if their sample isn’t representative of the population, their numbers are misleading.

    To summarize: for some defensible reasons (danger, distance, cost) the researchers chose to use cluster sampling on a heterogeneous population. The consequence of this is high sampling variability. The random selection of clusters (Falluja but not Basra) resulted in an unrepresentative number of deaths in the sample. The researchers then extrapolated these deaths to the population as a whole, without telling the readers of the flaws in their approach. The result is a study that estimates more civilian deaths in Iraq than during the London Blitz of WWII (the carpet bombing of London). I think that should strike everyone as implausible.

  136. incog,

    I probably phrased my original complaint poorly.

    Cluster sampling is more likely to produce an extreme result either way for any single study than more random sampling methods. In your last example, each individual study will always produce an incorrect result of either 100% white or 100% black. So, just out the gate, we should be suspicious of any extreme values returned by a cluster sample study. This study did produce values far different than other studies so we should be instantly suspicious.

    I can’t offer a proof off the top of my head but I am pretty sure there is also a promotional bias in this type of measurement i.e. a cluster with a lot of deaths (black balls) will have a stronger statistical effect than a cluster with zero deaths. I think that is why the Falujah cluster put their numbers through the roof. Don’t quote me on that.

  137. Shannon,

    “I probably phrased my original complaint poorly.”

    Well, yes.

    “Cluster sampling is more likely to produce an extreme result either way for any single study than more random sampling methods.”

    I generally agree.

  138. @ Joe Green
    1.) Lancet-article: Given the methods and numbers (through estimation) and looking at the interpretation of the results, the whole thing is so foggy, that everybody can make so different conclusions on the subject, that it becomes a matter of belief, not of science.
    2.) About the people WADI e.V., working in Iraq. It is surely not a matter of being politically left or right for them. Their minds are focussed on what is helpful for the people of Iraq. And in contrary to both of us, sitting on a desk, far from Iraq, they are still working there, as the only german NGO remaining in Iraq. Don´t tell me, they have no moral, risking their lives for other people. This distinguishes them from those been-7-days-in-Iraq-and-now-know-everything celebrities. Have a look at their website http://www.wadinet.de, its partially written in English!
    3.) “Insurgents”: On this website you can search for an article of the former official DDR-regime newspaper “Neues Deutschland”, written by an Iraqi professor (US critical, be sure): He lists up 4 categories of groups: Baathists including the former henchmen of Saddam, Al Quaida and friends, Shiites like Al Sadr, supported by Iran and ordinary criminals, bombing and hijacking for money. In each group he supposes a small number of people driven by patriotism. And in this article he benames terrorists clearly as terrorist, not “insurgents”. On the other hand: do you think that the normal Iraqui household has bazookas, Ak-47´, etc.. (“Give me the bazooka son, there´s a blackbird in the salads”).
    4.) About the justification of the war: Read Hillary Clintons congressional speech given in October 2002. It´s on her website.
    5.) How do you come to the point, that Al Quaida is growing? Numbers please, no opinions, etc.. And Al Quaida did exist long before the recent administration, ask Bill Clinton!
    6.) The Russians had problems with Cechnya long before 9/11. And Putin didn´t change the Russian constitution at the time the US marched into Iraq. He changed it every time, after a terrorist attack took place. The US criticized this, his good anti-war buddies Schröder and Chirac refused to criticise him.
    7.) Hair raising! Even if I don’t have much hair.. Muslim fanatics are morally equal to your government (I think that´s what you are trying to say)…
    Your mind derails here. Michael Moore is making money with that, but supposed you are not him…???
    8.) You citate Churchill in one of your answers. Read in Churchills “The Gathering
    Storm” page 1-30 and maybe you will understand, that he wouldn´t be on your side,
    but would be with Bush and Blair.
    So, have a good night!

  139. Skip; given that the Fallujah data point was thrown away, but that the Sadr City data point (with zero fatalities in a neighbourhood that saw heavy fighting), and given that the researchers chose not to make assumptions about the significant number of households where they came across empty bombed-out houses with nobody to interview, it is at least as likely that the survey figure has underestimated the mortality rate rather than overestimated it.

  140. Furthermore, Skip, have a look at the chart on page 3. It shows firstly, that every single governorate except Sulaymaniya saw a rise in the death rate post-invasion, and looking at the death rates, they seem to me to form a fairly normal histogram when the Fallujah cluster is excluded. In fact, the 95% CI for the post-invasion mortality rate is 5.6-10.2, which suggests a standard deviation roughly 25% of the mean. It seems to me that, once the Fallujah outlier is discarded, the data is telling us that the clusters are not all that heterogeneous after all.

    The intuition here is that heterogeneity works both ways; if it was really the case that things were going fine in Iraq apart from a few dodgy neighbourhoods, then it would be extremely improbable that pure mischance would have caused the Lancet to sample almost all bad areas and nearly no good ones.

  141. The result is a study that estimates more civilian deaths in Iraq than during the London Blitz of WWII (the carpet bombing of London). I think that should strike everyone as implausible

    It certainly strikes me as implausible that someone arguing in good faith would use that comparison.

    1. The wartime population of London was about 6 million; the population of Iraq is more like 22 million

    2. The population of London during the Blitz had adequate access to air-raid shelters; the population of Iraq did not.

    3. London was supplied adequately with electricity and clean water throughout the Blitz; Iraq was not.

    4. Children were evacuated from London during the blitz; they were not evacuated from bombing target in Iraq.

    Perhaps you would care to withdraw this argument?

  142. Erratum to my post above. Shannon has pointed out that some governates were not sampled. Therefore, after “every single governorate”, append “in the sample”, without, I think, loss of point.

  143. dsquared,

    “The intuition here is that heterogeneity works both ways;”

    I think this maybe where your failing in the appreciation of cluster sampling. There is no guarantee that cluster are in fact interchangeable for one another nor representative of the population as a whole. Again, statistic won’t tell you how representative your clusters are unless you already have a good idea of the distribution of the phenomenon your are looking at.

    In the case of this study, I think they based the clusters on the wrong factors. The controlling factor of life for Iraqi’s is not geography but ethnic group. An individuals chances of suffering harm either under Saddam or after the invasion was largely determined by which ethnic group Kurd, Shia or Sunni the individual belonged to. The geographic adjustments to the clusters had the practical effect of under representing Kurds and Shia while over representing Sunni. They also pulled clusters away from the relatively peaceful north and south and towards the violent center.

    I think you are also falling victim to a promotional bias in assuming that a cluster with zero deaths offsets any other cluster with an arbitrary number of deaths in it.

  144. They also pulled clusters away from the relatively peaceful north and south and towards the violent center.

    No; one cluster was moved from north to centre, and one from centre to south.

  145. The danger in explaining things like survey sampling is that people’s intuition often leads them far astray. I’m not sure how else I can explain the problems with this study without sounding condecending. Heterogeneous clusters don’t “balance out”, and a distribution of death rates across clusters that appears to be a normal distribution is not good news (ideally death rates should be identical across clusters, with all variance contained within each cluster).

    Rather than try to teach survey methodology to somone determined to disagree, I’ll just provide some useful cites. For anyone interested in learning more about cluster sampling (and sampling methodology in general), the classic reference on the subject is Leslie Kish’s “Survey Sampling” — a new edition came out in 1995. A slightly easier read is Scheaffer, Mendenhall, and Ott’s “Elementary Survey Sampling.”

    As far as the problems raised with comparing the London Blitz to Iraq — the Blitz was months of continuous bombing of a dense metropolitan area, while Iraq sees the scattered use of precision munitions. 100,000 deaths in Iraq is completely implausible.

  146. Skip, do you have any argument as to why the between-cluster variance cannot simply be taken into account by widening the confidence interval, as was actually done in the paper? Do you know what a “design effect” is, or why someone might wish to include one? Do you have any explanation of why it might be that all the governorates except one saw rising death rates, or do you believe that uneven cluster sampling is so evil that it can somehow bias results across time as well as space?

    Or are you, as I suspect, blowing methodological smoke by citing elementary textbooks on the subject of issues dealt with in the modern literature.

  147. Shannon Love wrote”

    “I don’t suppose it’s sporting on my part to point out that the greatest killers of the 20th century prided themselves on their atheism, reason and their devotion to science? Any randomly selected human in the 20 century was more likely to die at the hands of a mutant monster child of the Enlightenment than to die at the hands of a religious extremist.”

    That is pure rubbish. Karl Marx was the leading theologist of the communist religion. The leading Liturgical High Priest surely was Lenin followed by Stalin. All were atheists.

    Ayn Rand was the leading theologist of the capitalist religion. The leading Liturgical High Priest surely was Woodrow Wilson, the actual author of World War II, and the inspiration behind the War in Iraq. All were atheists.

    It is true that both of these “secular religions” in fact were created and animated by atheists.

    And it was not the fanatical religious zealots that pointed this out, but rather the iconoclast Christians (specifically the Catholics like Malcolm Muggeridge that called Stalin on the mass starvations in Ukraine for example, or other Catholics that led the labour union movement against capitalists like Henry Ford).

    In contrast, Osama Bin Laudin is a Muslim Fanatic that has perverted the Muslim Creed and has plotted the murder and death of countless thousands. Today we call him a “terrorist”, but in fact “secular atheists” also inspired “terrorism” with their “religion”.

    Similarly, Jerry Falwell is a christian fanatic that has perverted the Christian Creed and has plotted the murder and death of countless tens of thousands in Iraq, and elsewhere over his bloody career. Today we SHOULD be consistent and also call him a “terrorist”. George Bush and Jerry Falwell BOTH believe in the same perverted and twisted Christian Creed, and both have directly and indirectly caused the deaths of the 100,000 or more innocent civilians, woman and children that is the subject matter of this paper.

    Shannon Love goes on to state the following:

    “I myself am an agnostic scientific materialist but it is plainly evident to me that secular fanatics have proven far more dangerous than religious ones and even more irrational and dogmatic.”

    You would be mistaken. While it is true that communist fanatics under Stalin murdered in the millions, it is also true that religious “anti-communists” that were motivated by essentially the “capitalist” religion of Ayn Rand, also killed in roughly equal numbers.

    While there were pure Randian killers in the CIA like Howard Hunt in Guatamula for instance, there were far more that were motivated by the Pentacostal “christian” Churches in America. At one point, they went so far in Central America as to murder Catholic clergy, priests and nuns with American citizenship as the fanatics spread their “evangelizing message”.

    Shannon Love next writes:

    “Unlike the religious fanatic, they have no restraint of tradition or culture. Arrogantly convinced of their own rectitude and intellect, they sacrifice the innocent not to some god but to the hunger of their own egos.”

    For religious fanatics in Al Quada, not only is there no restraint but indeed there is a positive reward, i.e. Paradise.

    For Christian fanatics and zealots like Jerry Falwell and George Bush, there is the thrill of mass murder that far exceeds what even the 911 fanatics have done. You see, they actually do not know right from wrong, but believe in a perverse religious precept that they are doing “God’s Will” by committing crimes. This behaviour is very closely related to what CIA and KGB spies do when they break laws and commit “official crimes”.

    Shannon Love writes:

    “The 20th century welfare state is the result not a secular movement but springs from what might be best described as Christian socialism. The electorates that voted for these programs did so overwhelmingly because of their Christian beliefs, not their dedication to secular reason and science.”

    You are mistaken. Most of the social reforms of the 19th and 20th century stem from “enlightened self interest”. Your version suggests that the working class fought for and won concessions from their masters. The reality is that the wealthy capitalist classes “gave” concessions that in fact were “profitable” but not known as such to the workers at the time.

    You are about to see the Bush Administration, inspired by the Christian fanatics, do the same thing to Social Security. However, at the end of the day, the real measure of these “gifts” has always been “suffering”, and it is “suffering” and not “money” that is the measure of all things.

    Shannon Love writes:

    “Bad science kills.”

    No. Science is neutral, criminals kill and their instruments are varied and complex. In Iraq, technology has enabled American Christian fanatics like Jerry Falwell and George Bush to murder on a vast scale because in their spiritual world, their crimes against humanity are related to their perverse “beliefs” that they are following “God’s Will”.

    Shannon Love wrote:

    “Lies decorated with lab coats have been used to oppress and murder more people in the last 100 years than all religions combined.”

    You spoke too soon. If the Internet is still working after George Bush drops thermonuclear weapons on Iran, you can come here and correct your statement.

    Shannon concludes:

    “This Lancet study is merely the latest in a long line of dubious “science” cranked out to advance a political agenda. It is a propaganda coup for the enemies of reason, peace and democracy.”

    Balony. What you did Shannon was defend the indefensible. You were trying to discredit a report on the suffering of Iraq civilians, women and children, and to discredit their large numbers that magnified the crimes of the American Administration and the Christian fanatics behind it.

  148. Shannon Love wrote:

    “The 20th century welfare state is the result not a secular movement but springs from what might be best described as Christian socialism.”

    Charity is not “Christian socialism”. Socialism is a political system in which the government owns “the means of production”. Fundamentalist Christian Churches that practice “titheing” are engaged in a kind of emotional extortion racket, not charity and not socialism.

    Gerry Falwell and Jim Bakker are more like rackateers and swindlers along the lines of former CIA deputy director Carlucci that operates the Carlyle Group where the Bush and Bin Laudin Families are partners.

    CIA covert operatives (spies) named Steele and Baer have said as much on public television calling them a mafia crime family.

    To conclude, the mass killing of innocent civilians, women and children in Iraq is a crime against humanity ordered by American criminals in the White House. The purpose of the study, is to estimate the casualties in a country that has had its public health system destroyed by American Occupation Forces, bombs, ignorance and carelessness.

    After reading this paper, why would anyone STILL want to be an American citizen????? That is the only significant question that arises from the paper.

  149. Joe Green,

    Consider this thought experiment. Imagine yourself and your loved ones were ordinary Iraqis, not part of the Saddam’s Sunni 10%. Imagine that you know everything you know now. Imagine you get to choose whether America invades or not.

    Would you choose to live in Saddam’s police state or would you rather pass through the crucible of war just for the chance, not the guarantee, that it would lead to freedom and democracy?

    I would choose war and its risk for myself and my loved ones. I would hope that were the situation reversed, the Iraqi would do the same the for me.

    Of course, perhaps you are far more comfortable with the idea of living under tyranny. Perhaps, you would rather cower and wait for that knock on the door than to struggle for your freedom.

  150. On the other hand, Joe, you might also want to consider what you would say if you were offered the choice of

    a) no war
    b) a properly planned and resourced war and an orderly transition to democracy
    c) a badly planned and underresourced war and a chaotic occupation that might or might not end in democracy.

  151. >>”Skip, do you have any argument as to why the between-cluster variance cannot simply be taken into account by widening the confidence interval, as was actually done in the paper? Do you know what a “design effect” is, or why someone might wish to include one? Do you have any explanation of why it might be that all the governorates except one saw rising death rates, or do you believe that uneven cluster sampling is so evil that it can somehow bias results across time as well as space?

    Or are you, as I suspect, blowing methodological smoke by citing elementary textbooks on the subject of issues dealt with in the modern literature.”

    Ignorance I can handle. But ignorance combined with arrogance is really irritating.

    It is plain you don’t know the first thing about sampling methodology. To refer to Kish’s book, the “bible” of survey sampling (which you have plainly never heard of before) as “elementary” shows you’ve never read anything in this field. In fact, Kish talks extensively about calculating standard errors under heterogeneous clusters in the first edition of his book (1965), and the dangers of drawing inferences in this case. I suggest you pick up a copy and read it before portraiting yourself as an expert on survey methodology.

    Or perhaps you’re even more current on the subject than I am, and I’ve misjudged you. Could you please cite some of this “modern literature” you mentioned so I can see what methodological argument you are trying to make?

  152. I of course apologise; the other reference you gave was to an elementary textbook.

    I’ve provided a link on the other thread to a Gallup summary of the state of the art with respect to design effects; perhaps you’d care to explain why this doesn’t apply in the case of the Lancet study? Fourth time of asking.

  153. “c) a badly planned and underresourced war and a chaotic occupation that might or might not end in democracy.”

    First you guys tell us that we’re not using enough troops in Iraq. Then you tell us that we’re bogged down in Iraq and can’t do anything anywhere else in the world as a consequence. Make up your frigging minds. If we’re overstretched by what we invested, how can you criticise us for not investing even more?

    As for the choice: why don’t you ask the people that keep coming to the United States in leaky rafts, inner tubes, floating pickup trucks, and other similar means of transport? I think they’d take a ” badly planned and underresourced war and a chaotic occupation that might or might not end in democracy” (which the Iraq war was not, by the way – war ain’t like the movies, believe it or not, and by the standards of other real-life wars, Iraq went pretty well) over the certainty of a lifetime of tyranny.

  154. >>”I of course apologise; the other reference you gave was to an elementary textbook.”

    Was it the “Elementary” in the title that gave it away?

    >>”I’ve provided a link on the other thread to a Gallup summary of the state of the art with respect to design effects; perhaps you’d care to explain why this doesn’t apply in the case of the Lancet study? Fourth time of asking.”

    Is this website what you meant when you said “modern literature”?

    Look, this conversation about design effects is missing the point. Of course you can adjust the standard errors to account for estimated heterogeneity across clusters — but that does not change the inherent *sampling variability*. The study is based on 33 clusters, which is a pretty small sample size. As I said before, replace the Falluja cluster with a Basra cluster (which could have easily happened since clusters are selected randomly), and we’d get a point estimate of about 0 with large standard errors. Conversely, replace one peaceful cluster with a violent cluster, and we can get an estimate of 500,000 deaths with large standard errors. Would you still argue these numbers are plausible?
    In short, the tiny sample size means that if we change *one cluster* in this sample we get radically different results. That’s not a sample you should draw inferences from with any confidence. You seem to be arguing two contradictory things here — you first say that we can adjust the standard errors upwards to compensate for the fact that estimates with this sample are likely to be inaccurate, but then you accept the midpoint of this noisy estimate as an accurate estimate.
    There are other issues that have been raised as well. For instance, despite the claims of the researchers, overreporting of deaths is a valid concern. The unpopularity of the US in the areas of conflict and the fact that the researchers were unable to verify deaths in 20% of cases suggests at least some overreporting is likely.

  155. dsquared wrote:

    “On the other hand, Joe, you might also want to consider what you would say if you were offered the choice of

    a) no war
    b) a properly planned and resourced war and an orderly transition to democracy
    c) a badly planned and underresourced war and a chaotic occupation that might or might not end in democracy.”

    I think you Americans should have a look at the attitude taken by Canadians when they secretly extracted six Americans from the Iranian Embassy during the Carter Presidency.

    The only Canadian option ever considered was to “win”. But the hard question is “winning what”. And “winning when”.

    For the Iraqi, being in a country occupied by Saddam and his police apparatus had dangers and being occupied by American Forces has different, but just as deadly dangers.

    Second, as in politics, in warfare, you “never adopt a position you cannot defend”. The rule in ANY warfare is “first survival”.

    Saddam ruled using three interlocking sets of secret police. Taking down a Stalinist state that is still very active, is a highly dangerous thing to do, but the best starting point would have been to engage in passive covert warfare, spying in other words, and getting the information systems working so that it was possible to much better track and know what Saddam and his two lunatic sons were actually doing.

    The second major and very dangerous way to engage Saddam in a civil covert war would have been to get others involved in conditions of maximum cover and telling, not listening, to Americans and British forces operating the no-fly zones.

    It was clear from the very beginning that information was the key to taking down Saddam’s networks.

    Finally, the way to attack Saddam was to bring about and protect irrefutable evidence of human rights violations that could have been, and should have been used to bring international force to bear upon the regime. Had an army of “blue helmets” been assembled at the UN, there was a very large degree of probability that the Iraqi Army would have “stood down” while invading UN forces under a UN mandate would have collapsed the “criminal regime” of Saddam with very little bloodshed, chaos, or looting.

    Now that would have removed the threat, it would have taken down the Saddam Hussein Regime, arrested the top leadership on international warrents and put these on trial in The Hague. It would have resulted in a changed leadership in Iraq, but not likely the kind of “democratic wet dream” that Emperor George Bush II wanted, so that it could enrich the Houston Oil Barons that financed his political campaigns.

    Had this happened with the active co-operation of the Iraqi Army, and the complete detention of the regime and the Saddam Fedayeen, the military operations would have been complete in six months.

    But Bush, Rumsfeld, Wolfwitz and Perle screwed up big time because they utterly failed to provide the detailed and cautious leadership that the military forces are ENTITLED to receive from the civilian government.

    So now you have a quagmire, complete with casualties, dreadful costs, horrible publicity around the globe that is going to harm future American Diplomatic efforts for decades to come, and of course an entire generation of Iraqi that are going to remember who killed their families, relatives, friends, and children, loved ones and indeed the “flower of the country”.

    From a pure military perspective, I would have to say that Americans in the Pentagon were ALMOST acting as if they WANTED TO LOSE. That would prompt me to start looking for other explanations for such inept and incompetent leadership in the war, and what other hidden and undiscovered motives might still exist to explain American actions. Logic certainly was not one of them.

    Paul O’Neil resigned from the Bush Administration when be became convinced that George and his cohorts were actually inept and incompetent, and utterly clueless over fiscal and trade matters. In military matters as Senator Luger has observed, the problems that Mr. O’Neil observed were present in even greater quantities.

    Like one of the British papers ran for a headline that asked “How can 58 million Americans be so dumb?”

  156. My God, Skip, you really are a piece of work, aren’t you? The evening has thrown up *two* pieces of evidence that strongly suggest to me that you are purely and simply lying.

    First, your own response to my post. I have three issues with one single paragraph in it, starting with:

    “You seem to be arguing two contradictory things here — you first say that we can adjust the standard errors upwards to compensate for the fact that estimates with this sample are likely to be inaccurate, but then you accept the midpoint of this noisy estimate as an accurate estimate.”

    I have, in fact, never endorsed the 100,000 figure as a point estimate, and have on several occasions said that I don’t like it. I have, repeatedly, said that the important thing is the fact that the confidence interval on the relative risk ratio does not include 1.0 – that the invasion, according to this study, made things worse.

    This could be an innocent mistake, and I am prepared to give you the benefit of the doubt. However even if it was a mistake, the fact that you were prepared to put words in my mouth and did not read what I wrote with any care should cast some doubt on your credibility.

    Second:

    “In short, the tiny sample size means that if we change *one cluster* in this sample we get radically different results. That’s not a sample you should draw inferences from with any confidence.”

    This is not true. The Fallujah cluster was *one cluster*, and it was deleted, and we did not get radically different results. Since the Fallujah cluster was the biggest single positive outlier, it follows easily from the maths that there is no other single cluster which could be deleted to give radically different results. (In fact, if you deleted the Salamaniyah governorate clusters, you would get a much higher death rate, but this hardly helps your case).

    I don’t see how this could be an innocent mistake. The fact that the Fallujah cluster was deleted was mentioned very prominently in the study, was mentioned in Shannon’s original post and has been discussed ad nauseam in this thread.

    And third:

    “As I said before, replace the Falluja cluster with a Basra cluster (which could have easily happened since clusters are selected randomly), and we’d get a point estimate of about 0 with large standard errors. Conversely, replace one peaceful cluster with a violent cluster, and we can get an estimate of 500,000 deaths with large standard errors”

    Your guesstimates of point errors are clearly wrong from the data given in the study; it strikes me that your continued refusal to look at the data or to make specific comments about it cannot be considered to be innocent. In particular, you could not get such large changes by replacing one peacful cluster with a violent cluster, because there are 33 clusters; the Falluja cluster only caused a movement of 150k deaths in the estimate. Furthermore, you seem to be (either from ignorance or something worse) ignoring the fact that this argument is completely inconsistent with your other main claim. And here we move on to the other thing I learned last night which makes me believe that you are being thoroughly dishonest.

    Guess what, Skip? My library coughed up a copy of Kish’s book! I’ve now read what he has to say about cluster sampling.

    Guess what, Skip? When you were sermonising about the apocalyptic warnings Kish gives about the use of cluster sampling in heterogeneous populations …

    You neglected to mention that these warnings are all about the dangers of underestimating rare effects, not overestimating!

    Shame on you. Shame on me too, for not realising that this was obvious from the underlying stochastic process.

    For any observers still interested, a useful analogy to cluster sampling would be chucking stones into a minefield to see how dangerous it is. Imagine that you have been given the job of estimating the density of mines in a suspect field. You’re throwing stones into the bit you can reach, to see how many mines you set off.

    Obviously, this is not a great way to estimate how many mines there are in the field. But which is the greater danger?

    1) that, by chance, your stones all land in the spaces between the mines, causing you to believe that the field is much safer than it really is
    2) that, by extraordinary chance, all of your stones happen to land on mines, causing you to think the field is much more dangerous than it really is.

    Obviously, the second. And so it is with cluster sampling.

    You talk about “substitute a dangerous cluster for a safe cluster” as if the two were on an equal footing. In fact, if this is a heterogeneous population, there ought to be many, many more safe clusters than dangerous ones. Therefore, the survey methodology is much more likely in general to underestimate the change in the death rate than overestimate it.

    Now, you might have one argument left. It might be reasonable for you to claim that, through sheer bad luck, the survey carried out is a massive, multiple-sigma outlier; equivalent to having seven out of ten stones hit mines in a sparsely mined field. However in order to do this, you would have to claim one of two things.

    1) You would have to admit that such an outlier result could not possibly have been foreseen and the researchers should not be criticised for carrying out the survey and publishing their results. Shannon would have to withdraw her baseless and slanderous accusations about the survey’s motivation.

    or

    2) You would have to move to the claim that the sample was nonrandom, and that the violent clusters were picked on purpose. This would be an extremely serious claim of scientific malpractice, and I would advise you to be very sure indeed of your ground before making statements of this kind, which would certainly be considered defamatory.

    Unless I have missed some convoluted mathematical argument by which it can be shown that a cluster survey looking for rare effects in a heterogeneous population has some systematic tendency to oversample clusters in which that rare effect is present, it looks to me as if you have been thoroughly dishonest in your presentation of the subject so far, Skip. KIsh’s book certainly does not support your claim in the way you said it did.

  157. hahahahha. typo. Obviously, “obviously, the second” in my minefield example above should be “obviously, the first” [is much more likely]

  158. To Ken,

    “As for the choice: why don’t you ask the people that keep coming to the United States in leaky rafts, inner tubes, floating pickup trucks, and other similar means of transport? I think they’d take a “badly planned and underresourced war and a chaotic occupation that might or might not end in democracy” (which the Iraq war was not, by the way – war ain’t like the movies, believe it or not, and by the standards of other real-life wars, Iraq went pretty well) over the certainty of a lifetime of tyranny.

    Ken, I think it is long overdue for the United States to grow up and start acting like a mature member state in the International Community. Your Foreign Policy for example, might start at the point where you actually honour the treaties that you sign as a country.

    Do you need a few examples? OK let me give you one example.

    The USA signed the United Nations Charter and is a Founding Member State. When the US did that, along with other countries, you agreed to abide by the terms and conditions of that multilateral treaty. It OBLIGES your country to refrain from starting unilateral wars against other sovereign countries. If there are disputes between your country and another country, you have agreed to bring those disputes to the Security Council, and again to refrain from aggressive offensive warfare.

    And when you signed that treaty, you agreed to pay the dues of membership in that organization, for which you received many offsetting benefits, starting with the fact that the Head Office of the UN is on US soil in New York, where a great deal of funds collected from around the world is spent in that American city.

    But you also had an obligation to pay the fees that you as a country had agreed to pay. But you did not, and other countries that were paying their dues regularly had to pick up the slack for your national tardiness.

    Notice that I am not criticizing the United States for not signing any treaty, ONLY that you have failed to honour the actual treaties that you did sign.

    You also signed the Geneva Conventions on War by which you agreed to be bound in times of war. But in Abu Graub and other places, you have failed to meet the very things you agreed to do as a country.

    In Afghanistan for instance, there are many countries that are active and who have international treaty obligations but you do not read about the Germans, the French, the Canadians or the Australians violating the treaties that they have signed and agreed to support.

    The ABM treaty is another such broken American promise. So when people around the globe complain about “American Foreign Policy” what they are complaining about is the “lawlessness” that is practiced by some American Administrations.

    Make no mistake, Emperor George Bush II is an outlaw in the international community, every bit as much as Billy the Kid from your Ole West. In reality, Billy the Kid was a punk killer. Nothing else worthy of making myths.

    And frankly the rest of the world is getting real tired of it.

    Grow Up!

  159. Joe Green wrote:

    For the Iraqi, being in a country occupied by Saddam and his police apparatus had dangers and being occupied by American Forces has different, but just as deadly dangers.

    Where would you rather live?

    . . . Had an army of “blue helmets” been assembled at the UN, there was a very large degree of probability that the Iraqi Army would have “stood down” while invading UN forces under a UN mandate would have collapsed the “criminal regime” of Saddam with very little bloodshed, chaos, or looting.

    Now that would have removed the threat, it would have taken down the Saddam Hussein Regime, arrested the top leadership on international warrents and put these on trial in The Hague. . .

    The probability that this would have happened is zero. Saddam knew what he was doing when he bribed UN officials on a massive scale. Only the US was willing to do the dirty work of deposing the murderers.

    The fact that you put the term, “criminal regime” in quotes is interesting. I take it you believe the Hussein regime was legitimate. That’s about what I would expect from someone who expresses equanimity about Saddam Hussein, but takes the US to task for having the nerve to defend itself or (gasp) legally repudiate an obsolete treaty.

    To Ken,

    “As for the choice: why don’t you ask the people that keep coming to the United States in leaky rafts, inner tubes, floating pickup trucks, and other similar means of transport? I think they’d take a “badly planned and underresourced war and a chaotic occupation that might or might not end in democracy” (which the Iraq war was not, by the way – war ain’t like the movies, believe it or not, and by the standards of other real-life wars, Iraq went pretty well) over the certainty of a lifetime of tyranny.

    Ken, I think it is long overdue for the United States to grow up and start acting like a mature member state in the International Community. Your Foreign Policy for example, might start at the point where you actually honour the treaties that you sign as a country.

    You won’t answer Ken’s question, but no matter. We already know that you think the US government is morally equivalent to Hussein’s regime, and that the US should blindly obey your tendentious interpretations of international treaties rather than defend itself. You long ago became a broken record, and not a very convincing one at that.

  160. dsquared wrote:

    Obviously, this is not a great way to estimate how many mines there are in the field. But which is the greater danger?

    1) that, by chance, your stones all land in the spaces between the mines, causing you to believe that the field is much safer than it really is
    2) that, by extraordinary chance, all of your stones happen to land on mines, causing you to think the field is much more dangerous than it really is.

    Obviously, the [first]. And so it is with cluster sampling.

    It isn’t “obviously” the first. There is simply no way to know, given the inadequate data. You assume that the mines are randomly distributed, but there is no way to know if this is true, and overestimating the number of mines might be as or more costly than underestimating them, if doing so causes you to change your behavior in ways that affect how you fight the war. (And changing the way in which we fight the war was the study author’s stated reason for rushing publication before the election.)

  161. Jonathan,

    You said as an answer to dsquare, “It isn’t “obviously” the first”. Well actually it is. The odds of hitting a mine are low if the mines are rare, it is as simple as that (the intuitive argument is correct and easely proven mathematically). And if the mines are scatterd as Shannon claims “..highly geographically concentrated” then your odds are even worse becuse it makes it likely that mines are on top of each other (which reduces the area covered by the mines and therefore the likelyhood of hitting a mine with a rock).

    Shannon initially stated “..this means a cluster sample has a very high chance of exaggerating the number of deaths” which is (as proven in some posts) a very misleading statement, because the opposite is true.

    There is a chance of overestimation as a result of bad sampling, but the matter of the fact is that a sampling is MORE LIKELY to result in an under estimation. There are a wealth of reason why the study could be over the top, but the odds are against bad sampling.

  162. Jonathan, that would only be the case if you started with absolutely no information about the field; if you didn’t know whether you were dealing with a normal minefield (a field with a few mines in it) or a heap of mines with a few patches of grass between them.

    That’s not the situation with respect to Iraq. We know that deaths due to violence were the uncommon events, and that the majority of the population did not die. Therefore we know that we are dealing with small clusters of airstrike deaths in a general environment of survival, not small clusters of survival in a general environment of death. Therefore we know that the cluster sample is more likely to underestimate than to overestimate the change in the death rate.

  163. dsquared, I would love it if you kept your ad hominems to yourself, because they seriously degrade the signal-to-noise ratio of your posts which are otherwise worth reading. Characterizing your interlocutor Skip Smith as a “piece of work” or a liar is doubly pointless, given that all he is required to do here is support his arguments, not defend his character.

    Should we assume that you are tireless in your defense of the Lancet because you took out a bet on the number of Iraqi victims being 100,000, and are hoping to cash in? Unlikely as that is, it’s not impossible – however that doesn’t mean I would entertain the possibility even for a second. Neither you, I, Skip nor the Lancet authors have a monopoly on compassion for Iraqi victims of the war, so we might as well treat others who enter into detailed debate on this matter as honest participants, rather than utterly corrupt ones.

    I for one hope that Skip isn’t put off by the aspersions being cast on his character, and continues to engage in this debate in the same calm manner as before. You or I may not agree with all he has to say, but let’s let him raise his doubts without demonizing him for it, shall we? Perhaps you should save your ire for the folks who openly welcome these deaths, and hope to see more of the same. Hound them out of voluntary debating chambers like this one, if you wish, since they contribute nothing but an insight into their diseased minds.

    But well-informed skeptics and critics of the Lancet study need to have their doubts answered – it benefits no one if they are declared to be morally defective because their reasoning fails to accord with your own. Otherwise how will we – or you – know that it is their arguments which have been exhausted, and not just their willingness to be subjected to abuse?

  164. http://obsidianorder.blogspot.com/

    Hi dsquared,

    have another look at the study and notice the following numbers.

    Violent deaths in Fallujah:
    24 children
    25 men
    3 women

    Violent deaths in the rest of Iraq:
    4 children
    13 men
    2 women
    2 elderly

    The authors write:
    “Most individuals reportedly killed by coalition forces were women and children.” That is INCLUDING Fallujah.

    Basically, if you look at the actual numbers reported by the Lancet study (throwing out Fallujah as entirely unreliable for a number of reasons), they support my own estimates nicely. Look at the link provided and THEN get back to me.

  165. On calm reflection, I agree with GKL and apologise to Skip. Although I still think that he should have mentioned that the Kish book’s discussion of cluster sampling placed more emphasis on underestimation rather than overestimation, I probably went too far and withdraw the personal remarks.

  166. Heiko, you’re going to have to be a lot more specific because I am not sure what exact point you’re inviting me to comment on. As I read it, your remarks are about attribution of responsibility for the deaths, rather than questioning the death rate numbers themselves; I agree with you that more emphasis should be placed on the confidence interval than the extrapolated figure of 100K.

    Other than that I’d make the following comments on your post:

    1) The statement that because violent deaths were recorded in only 15 out of 33 clusters, the violent death rate in most of the country is zero is a howler; you can’t mean this literally.

    2 Your discussion of the by-governorate death rates looks partial to me; several very violent areas were not sampled (Samarra, most particularly). and the Al-Sulyamaniya province which showed a falling death rate is in the Kurdish-controlled region. In general, all the points I have made to Skip about the tendency of cluster samples to underestimate rather than overestimate rare effects, apply to your critique too.

    3) It is not obvious that the sample size is “too small”. Too small for what? Too small to give a narrow confidence interval? Yes, it is. Too small to support the conclusions drawn from it, in particular the very important conclusion that the overall death rate rose rather than falling? No, it is not too small to say this.

    4) Your numbers for the murder rate in Compton is too high; it is more like 55/100K than 80. In general, this comparison is meaningless, however; you are comparing urban areas in the US with a mixture of urban and non-urban in Iraq, and the murder rate in Iraq ought to be lower as there are tanks in the streets keeping order!

    5) You appear to arrive at your estimates of “noncombatant dead” by a series of arbitrary assumptions based on sex ratios and guesses. You cannot seriously be proposing that someone should have carried out these subjective and ad-hoc adjustments for a paper in the Lancet.

    6) I don’t understand where you get your estimated figure for combatant and non-combatant deaths from. It just appears for the first time in your conclusion with no explanation of how it relates to the arguments made above. I can’t comment on this figure without understanding how you arrived at it; in the meantime, I think you ought to have addressed the possibility of underestimation. As it stands, you only consider possible downward adjustments to the figures in the Lancet study; this is biased and reduces the credibility of your whole process.

    I’ve posted this comment on your own site and would be happy to continue the discussion there.

  167. Oh, I get it. We win as soon as will kill more people than Sadaam did. Or do we win as long as we kill one less person? Why doesn’t the US want the name of civilizan deaths tracked? Shouldn’t everyone understand the benefits of not attacking the US while having huge oil reserves?

  168. To Jonathon Gerwitz who wrote,

    “You won’t answer Ken’s question, but no matter. We already know that you think the US government is morally equivalent to Hussein’s regime, and that the US should blindly obey your tendentious interpretations of international treaties rather than defend itself. You long ago became a broken record, and not a very convincing one at that.”

    But I did answer the question. What part of “no” do you fail to understand? The US invaded Iraq to “defend itself???” Surely you jest!

    You can compare the actions of the American Government to the actions of the Iraqi government over any fixed interval of time. If you did that, a pretty good estimate of American dead would lie around 1200 counting civilians, women and children, compared to around 100,000 Iraqi civilians, women and children that is the result of American Government, not Iraqi Government, Decisions. Now you cannot properly ascribe ANY of the deaths from 911 to Iraq. Your own Commission on that very question concluded that Iraq and Saddam had nothing to do with the attacks. So the 3000 dead plus others lost in aircraft, properly belong to Al Quada and to Saudi Arabia, NOT Iraq.

    That is the US “score”. On the Iraqi side, Saddam and his secret police were responsible for tens of thousands of deaths, particularly in suppressing the Shi’ites following the first Gulf War. Estimates are that between 30,000 to 50,000 lost their lives in a fruitless uprising against Saddam after Kuwait was liberated. Now of this figure, it really is important to apportion blame for these deaths, because you cannot lay all this at the feet of the Iraqi fanatics surrounding Saddam. Some of it belongs to General Swartzkoff who signed the armistise that gave back Saddam the helicopter gun ships that he subsequently used against the Shi’ite to suppress their uprising, even while President George Bush I was encouraging them to overthrow Saddam. You should regard George Bush I’s treachery as at least being partly responsible for Shi’ite deaths. That fact alone should have been ringing the alarm bells when Mr. Rumsfeld said US troops would be welcome as “liberators”, because such actual American treachery at the highest levels is not soon forgotten by the Shi’ite that lost family members and friends in the revolt.

    You should also carefully assess the role played by the United States in the genocide conducted by Saddam against the Kurds. Saddam never had chemical weapons until the US provided these to him, and the facts are that Saddam received many supplies for biological weapons as well from the US. Remember that the US was allied with Iraq against Iran.

    So when you ask the question of “moral equivalents”, in terms of murder and mayhem, the answer is that US Administrations under both Bush Presidents were indeed moral equivalents as far as the murdered victims were concerned.

    I now turn to your point about “tendentious interpretations of international treaties”. International Treaties are agreements between Sovereign Nation States. They are written to be followed and enforced in ways that prevent open conflict. There are countries that consistently follow their treaty obligations, Germany and Japan, Canada and Australia being four rather good examples.

    The US and Israel are not good examples because the record of broken treaties exists for these countries that exceed even what Saddam violated when he too disregarded the Resolutions of the United Nations.

    Finally I make the point that your own hypocrisy is hardly an adequate explanation for American failures to honour their treaty obligations to other countries.

    When countries arrogantly and blatently breach treaties that they have entered with other countries, they acquire a reputation commensurate with such hypocrisy. And when others criticize American Foreign Policy, THIS is what they are zeroing in on as THE problem with the US in the world today.

    You are capable of better behavior, and its time for you as Americans to recognize it and meet the challenge.

    For you Jonathon, that would mean to start participating reasonably in the discussion of this paper and its merits, rather than regurgitating a lot of “neocon” garbage that passes for policies these days in Washington.

  169. Jonathan Gerwitz asked:

    “Where would you rather live?”

    There are many law abiding and peaceful countries that would be very nice places to live. Iraq is not one of them. The US is not one of them. Both are countries with very high rates of mortality due to violence and crime.

    Canada is fine, and so is Sweden or Norway. Britain is crowded but Australia is wonderful. If you like warm climates, Malaysia is fine as well. My all time favorite would likely be the Cook Islands in the Pacific because they enjoy a stable government, peace and tranquility are the norm, and political lunatics do not occupy office, but give their quaint speeches on soap boxes.

    Joe Green stated:

    “. . . Had an army of “blue helmets” been assembled at the UN, there was a very large degree of probability that the Iraqi Army would have “stood down” while invading UN forces under a UN mandate would have collapsed the “criminal regime” of Saddam with very little bloodshed, chaos, or looting.

    Now that would have removed the threat, it would have taken down the Saddam Hussein Regime, arrested the top leadership on international warrents and put these on trial in The Hague. . .”

    And Jonathan wrote:

    “The probability that this would have happened is zero. Saddam knew what he was doing when he bribed UN officials on a massive scale. Only the US was willing to do the dirty work of deposing the murderers.”

    You are misusing statistical ideas here just as carelessly as the rest of your critique of the paper. For starters, when you say “the probability”, in fact you are grabbing something straight out of the air without any scientific foundation whatsoever. On what basis do you say “probability … is zero”? This is a bald statement because you have made it in a vaccuum.

    Finally, you are completely WRONG that only the US would do the work of international peacekeeping and surpressing human rights violators.

    If you ever needed any evidence of the CORRECT way to deal with tyrants and mass murderers, you need look no further than what NATO (and NOT the US alone) did in Yugoslavia to suppress “ethnic cleansing” (mass murder) in Kosovo. It was precisely because of the overwhelming force applied by NATO with nothing more than its air power, that the Yugoslav Army folded and withdrew from the field, thereby minimizing casualties to both armies, and to civilians as well.

    The actions in Kosovo by NATO not only was consistent with the UN Charter (section 52) but the UN itself was involved at every step of the way, including the protection and support of NGOs.

    In Iraq, the Saddam Hussein regime would have collapsed faster than a 800 year old man if the same tactics had been applied. Unfortunately that did not happen, and you now have massive and growing casualties and the US finds itself alone in a quagmire with no plausible exit strategy.

    Well, the rest of the world told you not to go into Iraq in this fashion, but I guess you knew better.

    “You broke it, now you pay for it” is how Colin Powell would put it.

  170. dsquared wrote (in response to this comment):

    Jonathan, that would only be the case if you started with absolutely no information about the field; if you didn’t know whether you were dealing with a normal minefield (a field with a few mines in it) or a heap of mines with a few patches of grass between them.

    That’s not the situation with respect to Iraq. We know that deaths due to violence were the uncommon events, and that the majority of the population did not die. Therefore we know that we are dealing with small clusters of airstrike deaths in a general environment of survival, not small clusters of survival in a general environment of death. Therefore we know that the cluster sample is more likely to underestimate than to overestimate the change in the death rate.

    This is only true if we understand the data distribution and if the sampling methodology is unbiased. Neither condition appears to be true in this case. The minefield metaphor is misleading, because the mines in minefields tend to be distributed randomly or according to some regular pattern, while in the case of civilian war casualties we do not understand the distribution, and indeed it is precisely the distribution that we are trying to determine. We can’t just assume it is random or follows a particular pattern. There aren’t enough data to do good analysis. Better analytical methods cannot be effective in the absence of better (more extensive, better sampled) data.

  171. Joe Green quoted me:

    “You won’t answer Ken’s question, but no matter. We already know that you think the US government is morally equivalent to Hussein’s regime, and that the US should blindly obey your tendentious interpretations of international treaties rather than defend itself. You long ago became a broken record, and not a very convincing one at that.”

    . . . and then responded:

    But I did answer the question. What part of “no” do you fail to understand? The US invaded Iraq to “defend itself???” Surely you jest!

    To see if Joe really did answer my question, see the comment in which I asked it.

    Joe also wrote:

    Jonathan Gerwitz [sic] asked:
    “Where would you rather live?”
    There are many law abiding and peaceful countries that would be very nice places to live. Iraq is not one of them. The US is not one of them. Both are countries with very high rates of mortality due to violence and crime.
    Canada is fine, and so is Sweden or Norway. Britain is crowded but Australia is wonderful. If you like warm climates, Malaysia is fine as well. My all time favorite would likely be the Cook Islands in the Pacific because they enjoy a stable government, peace and tranquility are the norm, and political lunatics do not occupy office, but give their quaint speeches on soap boxes.

    Got that? The U.S. is comparable to Iraq in terms of violence and as an undesirable place to live, but Canada is “fine” — and so is Malaysia, a country whose recent prime minister was a notorious Jew hater and terror sympathizer.

    I leave it to readers to decide whether Joe or I has the better handle on reality. If anyone has questions, I suggest putting them to one of the numerous recent Canadian immigrants to the U.S., or to an immigrant from Cuba, Iran or the Arab world.

  172. Jonathan Gerwitz wrote:

    “Got that? The U.S. is comparable to Iraq in terms of violence and as an undesirable place to live, but Canada is “fine” — and so is Malaysia, a country whose recent prime minister was a notorious Jew hater and terror sympathizer.

    I leave it to readers to decide whether Joe or I has the better handle on reality. If anyone has questions, I suggest putting them to one of the numerous recent Canadian immigrants to the U.S., or to an immigrant from Cuba, Iran or the Arab world.”

    The traditional propaganda methods no longer work Jonathan. And it does nothing for science or mathematics either.

    You asked a question and I answered. The mortality rate in the US due to crimes and accidents with guns on a per capita basis exceed that of all European countries and Canada by an order of magnitude. Today, thanks to a one party state in the US, terrorists and para-militaries can buy unlimited amounts of automatic assault rifles and pistols, together with any quantities of ammunition. The American second amendment virtually guarentees that this high mortality will not be reduced in the foreseeable future.

    I did not raise the issue of Israel and Jews, but you did, so I will answer. If you compare again, mortality from gun use in a country like Malaysia to Israel, and include all forms of political violence in both countries, Malaysia is a much better choice for a violence free place to live. Your chances of being blown to bits in downtown Kuala Lumpur compared to downtown Tel Aviv asymtotically approaches zero. That has nothing to do with prejudice, but rather objective observations.

    IF you like guns Jonathan, you are welcome to live where there are plenty of guns, either in Houston or in Baghdad, or even Falujah in Iraq.

    I think it would be a very interesting study to compare prewar Iraq with the United States for mortality due to gun use.

    I am sure it would raise more than a few eyebrows.

    As for the rest of the snake oil that Jonathan is peddling here, it never was a replacement for sound science and mathematics.

    The study remains credible, and given the new levels of killing now underway throughout Iraq, the actual mortality numbers throughout Iraq for civilians, women and children will rise even further and past even the high numbers in the study.

    Use some common sense. If you beseige a city like Falujah, bomb it, mortar it, shell it, burn it, pound it into rubble, and then assault it yet again, shooting at anything that moves, you will get what John Kerry described before the American Congress as a young man who had just returned from Vietnam. Very high civilian casualties, many dead children, and “free fire zones” that constitute war crimes under the terms of the Geneva Conventions.

    Not that something like the Geneva Conventions would ever stop someone as determined as Jonathan. Or Emperor George II.

    Of course neither appreciates that these activities are actually building Al Quada because these conditions breed new recruits for the terrorists. So that what started out a couple of years ago as about 5,000 Al Quada fighters in Afghanistan now exceed 50,000 and is growing daily and is spreading like lymphatic cancer.

  173. Jonathan,

    “This is only true if we understand the data distribution and if the sampling methodology is unbiased.”

    No, this is not the case. As long as we know that the entities we are studying are rare (and we know they are) and the sampling is random, then the underlying distrubution is irrelevant to decide the fact that an under estimation is the most likely outcome. It is however interesting to note that the distrubution proposed by Shannon Love ande Fred Kaplan (http://slate.msn.com/id/2108887/) even further increases the chance of an under estimation.

    “There aren’t enough data to do good analysis. Better analytical methods cannot be effective in the absence of better (more extensive, better sampled) data.”

    This is true, but does not alter the fact that under estimation is still more likely.

  174. incog wrote:

    Jonathan,
    “This is only true if we understand the data distribution and if the sampling methodology is unbiased.”
    No, this is not the case. As long as we know that the entities we are studying are rare (and we know they are) and the sampling is random, . . .

    The sampling isn’t random. That’s the whole problem here.

  175. Joe Green:

    Why do you think so many people, including Canadians and people from Malaysia, immigrate or want to immigrate to the United States?

  176. The sampling isn’t random. That’s the whole problem here

    Yes they were. The samples were taken by selecting a random number and matching it up to a GPS grid. There’s detailed discussion of the sampling process in the paper, and it was random. In what way was the sampling nonrandom? What do you mean here?

  177. Jonathan Gerwitz wrote:

    “Why do you think so many people, including Canadians and people from Malaysia, immigrate or want to immigrate to the United States?”

    Like most of the points you tried to use in attacking the paper, you are also grossly mistaken about people wanting to come and live in the United States.

    Here are the facts. Of the 160,000 Canadians that live in the United States, 93% retain their Canadian Citizenship and are “expats” that have no intention of ever becoming Americans. Moving in the other direction is a larger number of Americans that live in Canada, something like 600,000. Of these 180,000 live in Calgary and are mostly Republicans from Texas that work in the oil industry, an industry in Canada that is owned and controlled by foreign interests for the most part.

    The largest group of Americans living outside the United States, live in Calgary making it the most American city in Canada.

    Most of the Canadians living in the US are retirees or “snowbirds”. These folks live down there because of the milder weather. Their presence in the US represents a net contribution to the American economy since most of them are spending their pensions in the US and do not have jobs in the economy.

    I do not have the figures for Malaysia, but from what I can see, people in Malaysia have much better and more lucrative opportunities closer to home. In particular, America is mostly a hostile place for the predominately Muslim population of Malaysia. And while the Chinese minority in Malaysia suffer from some racial discrimination, it is not nearly as great as the Chinese and Japanese suffer in the US.

    The only group of immigrants struggling to get into the US are from Mexico, and that is because their standard of living has been severely depressed because of NAFTA and the corruption of the previous Silanas Mexican Government that signed this bad trade deal with the US with the former Republicans under George Bush I.

    The Americans abused the Mexicans with this deal since today about half the national income of Mexico comes from the sale of oil and gas, and the other half comes from the meagre earnings of illegal farm workers that earn less than the minimum wage for their work in the US. Mexico is the second largest supplier of oil and gas after Canada.

    My view is that the US in the context of the current Administration is unworthy of good neighbours such as it has in Canada and Mexico.

    The NAFTA trade deal should be cancelled, and trade, particularly “free trade” should not be setting immigration policy.

    Indeed, the rest of the world should be considering trade embargoes against the United States for its illegal and aggressive war in Iraq to take over its oil supplies that belong to the people of Iraq.

  178. Mr. Green, I am interested in meeting some of the Chinese that feel comfortabe in Malaysia. And I’m interested in the discrimination against the Japanese and Chinese citizens of America. Do you have a wide knowledge of these communities and some statistics, polls, etc. that would support your contentions?

    I’m also interested in yours and Kofi Annan’s definition of “illegal.” Is what is being done presently in the Sudan legal? It clearly will not be considered so by some in the security council to be just that.

    Legal or illegal premises a set of rules and a governing authority. I guess the “rules” are the whim of the security council (the resolutions against Iraq don’t count) and the UN is the governing authority?

  179. Ginny wrote:

    “I’m also interested in yours and Kofi Annan’s definition of “illegal.” Is what is being done presently in the Sudan legal? It clearly will not be considered so by some in the security council to be just that.

    Legal or illegal premises a set of rules and a governing authority. I guess the “rules” are the whim of the security council (the resolutions against Iraq don’t count) and the UN is the governing authority?”

    When the Secretary General of the United Nations says that the invasion of Iraq by American and British military forces is “illegal”, he is saying that these activities are in contravention of the United Nations Charter, which is the result of a binding International Treaty signed by the sovereign nation states that created the United Nations in the first place. Both the United States of America and Britain as founding member states of the United Nations, have signed the UN Charter, and are bound by its terms and conditions, at least legally speaking.

    Law abiding countries like Canada, for example, are meticulous at properly limiting their activity to the strict confines of what they have agreed to by signing the Charter. There are many other countries like Canada that do likewise. But there are those that do not live up to the terms of their promises and commitments. The US and Britain, as well as Israel, do not.

    So that is the foundation of the Secretary General’s label that the war in Iraq is “illegal”. Only the United Nations Security Council has the authority to authorize offensive warfare by member states upon other member states, as agreed to by the signatory powers of the UN Charter.

    In other words, both the US and Britain violated their own solumn promises to uphold the UN Charter.

    I might add that Iraq also violated the UN Charter when it invaded Kuwait. It was an open and shut case of a breach of the UN Charter by Iraq. The Gulf War was correctly mounted by the UN under American leadership and many countries, including Canada, were involved in the military effort to free Kuwait from Iraqi military forces.

    Since the war in Iraq today is “illegal”, it also means that those Iraqi in Falujah that are fighting Americans and the American installed puppet government in Baghdad, have the status of “freedom fighters”.

    Iraqi have suffered the loss of over 100,000 civilians, women and children, and they are fighting an invader that covets their natural resources as well as to protect their own national identity and culture.

    With the Falujah operations, the war is about to enter a new phase where it will properly become a “war of national liberation”. This is precisely the same thing that happened in Vietnam where the Vietnamese fought against foreign colonial powers (first French, then American) in order to secure their national independence.

    There is still one other aspect and that is that Saddam’s Iraq, may have in fact fully disclosed what it knew about chemical, biological and nuclear weapons development in Iraq and that their actual reports were accurate in the first place. The US and Britain took the opposite position and that makes their position before the Security Council very bleak because they in effect lied to the Security Council.

    Technically the US and Britain could be censured, but because each is a permanent member of the Security Council with a veto, that is not going to happen.

    At the bottom of it all, is the fact that the US and Britain have thrown overboard mostly “defensive postures” and “collective security” in favour of a policy of “unilateralism” which in effect is a throwback to pre World War I behavior and diplomacy. No doubt there exist “secret treaties” between the US and Britain, and as we know from history, such “treaties” lead directly to global war.

    If the US and Britain do the honourable thing, they would quit the United Nations and stand upon their own position. For the same reason they should withdraw from the NATO collective security treaties.

    If they did that, Canadian, German, French and other NATO country troops could come home from Afghanistan.

    In a “unilateral world” the word of any country does not matter much or count for much. Its strictly “every man for himself”.

    And of course, the US election just past, has seen most Americans endorse this position and policy, however unwittingly.

    That is why I say that we are now living in very dangerous times. We have post Cold War nuclear weapons technology that is controlled by pre World War I diplomacy alone and where decisions on war and peace are in the hands of MBAs without a shred of military or diplomatic experience. Its something like getting into a 747 jumbo jet that is piloted by a eight year old that loves model airplanes.

    That cannot be very reassuring for rational people. A nuclear holocaust is not impossible, and given the politics and irrational role of religion in politics these days, if it does not destroy mankind, all the accumulating problems of global warming will.

    So light up a good cigar, pour yourself a strong glass of brandy, put your feet up on the desk, and try to enjoy the unfolding final act. There really is almost nothing that any sane and rational person can do at this late stage.

    In the end, there is a kind of grim justice because Republicans will perish alongside Democrats. And Russians and Iraqi will suffer the same fate that Americans and Brits will endure.

    In the end, Lord Mynard Keynes was right.

  180. Joe Green wrote:

    Here are the facts. Of the 160,000 Canadians that live in the United States, 93% retain their Canadian Citizenship and are “expats” that have no intention of ever becoming Americans. Moving in the other direction is a larger number of Americans that live in Canada, something like 600,000. [. . .]

    The population of the USA is approximately ten times that of Canada, yet on your numbers there are only 4 times as many Americans in Canadia as there are Canadians here. It seems to me that Canadians wouldn’t be in the USA in such proportion to Canada’s population if your bizarre fantasy of a violent, Iraq-like USA were anywhere close to being true. Not to mention that you are trying to convince someone who lives in the USA that he should believe you over his own experience.
    Joe Green wrote:

    When the Secretary General of the United Nations says that the invasion of Iraq by American and British military forces is “illegal”, he is saying that these activities are in contravention of the United Nations Charter, which is the result of a binding International Treaty signed by the sovereign nation states that created the United Nations in the first place. Both the United States of America and Britain as founding member states of the United Nations, have signed the UN Charter, and are bound by its terms and conditions, at least legally speaking.

    From Chapter 7, Article 51 of the UN Charter:

    Nothing in the present Charter shall impair the inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations, until the Security Council has taken measures necessary to maintain international peace and security.

    The US was attacked Sept. 11 and has an absolute right to defend itself, UN or no. It has the right to decide what measures constitute adequate self-defense. The UN Security Council has no legal or moral authority to impede our defensive efforts, and anyway took itself out of the game by refusing to respond in a timely way, if at all, to the threat against the US.
    Not to mention that members of the UN Security Council were taking bribes from Saddam Hussein and are therefore doubly disqualified from having any legitimacy to judge our self-defense efforts. There is even circumstantial evidence that the UN Secretary General, who like you lectures us on our national responsibilities, was on the take. You have a lot of nerve to criticize the US in this context. And for rooting for our enemies you should be ashamed of yourself. You are nothing more than a squalid little America-hater who uses verbose pseudo-legalisms to rationalize his bigotry. Go back to your civilized Canadian burrow and thank God for the USA, because you’d be speaking Russian by now if we weren’t here.
    IMO the main thing the US did wrong after Sept. 11 was to go to the UN. Bush should have repudiated the organization and kicked it out of NYC when it would have been politically easy to do so.

  181. Jonathan Gerwitz wrote:

    “The US was attacked Sept. 11 and has an absolute right to defend itself, UN or no. It has the right to decide what measures constitute adequate self-defense. The UN Security Council has no legal or moral authority to impede our defensive efforts, and anyway took itself out of the game by refusing to respond in a timely way, if at all, to the threat against the US.”

    According to the 911 Commission that wrote its report on the attack, they stated flatly that there was NO connection between Al Quada – Osama Bin Laudin, and Saddam Hussein in Iraq. In its unanimous report (both Democrats and Republicans), the Commission found no evidence that could even remotely link Iraq with the attack.

    The US Bush Administration launched a “pre-emptive” war upon a sovereign nation state (Iraq), that was not responsible for this attack on 911, which led directly to the American breach of its own Treaty obligations under the terms of the United Nations Charter.

    IN FACT, the ACTUAL attack on 911 was mounted PURELY WITHIN the United States, and NO FOREIGN countries were ACTUALLY involved.

    Now it is true that some Republicans lied to the American people about this, they lied to the United Nations about Iraq, and they lied to the American Congress.

    Some of these lies even date back to Richard Nixon who was a Republican that almost subverted the American Constitution and had maintained an “enemies list” where John Kerry’s name appeared, and where the Nixon White House hired a team to discredit the young naval officer freshly returned from Vietnam. That same Nixon team reappeared as the “swift boat veterans”.

    You also alluded to Canadians speaking Russian in your diatribe, so let me address that part of your attack.

    The greatest threat to Canadian National Sovereignty actually comes from the United States, not Russia. In terms of Canada’s territorial boundaries and limits, Russia recognizes Canada’s position vis a vis the Canada US border in Alaska (and they should know because they sold Alaska to the United States) and rejects the US position that the border is “shifting” as the banks of the North Slope are shifting (the Canadian position is that the boundary is unchanging along the 141 W longitude line)

    Furthermore, Canada under the Law of the Seas claims the North West Passage as domestic territorial waters, a position endorsed by the signatories to the Law of the Seas, but refuted by the US.

    America, bold as brass, sailed the USS Manhatten into the Northwest Passage without seeking the appropriate Canadian legal permissions and support, with the result that this oil tanker got stuck in the ice in some of the most ecologically sensitive waters on the planet. The Canadian Coast Guard Ice Breaker Sir John A MacDonald had to free the tanker, which just about completely underscores the world’s problem with the “American attitude”.

    Russia of course recognizes the Canadian position and rejects the American one with respect to the Law of the Seas and the domestic waters of the North West Passage.

    At the end Johnathan decends into abrassive name-calling because of the points of law that he lacks the capacity and ability to refute.

    What is actually at stake is not my personal preference, but rather the “rule of law”. The United Nations has no ability to “make laws” that can be imposed upon the US for the obvious and simple reason that the United Nations is not a “government”.

    For the Republicans to present it this way is just one more Big Lie added to a growing list of lies. What the US has done in Iraq is to contradict its own stated position vis a vis the United Nations Charter.

    Finally, unable to contain his mindless anger, Jonathan foams that the United Nations should be moved out of New York.

    I think that would actually be a good idea because the United Nations should have a better home than what the US has provided. Surely an organization that works for world peace deserves better than to have the US government constantly interfering with its peaceful effort with cladestine wiretaps, buggings, and other illegal activity.

    The old KGB Internal Security Apparat and the new US Homeland Security apparatus were all cut from the same cloth. All of them worked against open government, integrity of office, and the “rule of law” which is the only alternative to war.

    But Jonathan just does not get it. And that is really why his vitrionic attack upon the Lancet paper is so off the wall, irrational, unscientific and filled with religous zealotry that would make Jesus himself blush.

    You see what so infuriates him, is the realization that he is a fanatic no different from those that hijacked four airliners with nothing more than box cutters.

  182. We are so nice around here.

    “…he is a fanatic no different from those that hijacked four airliners …”

    People say things like this and we stay nice.

    Why do we do this?

    I don’t know.

    Why is it that the comment boxes on blogs inevitably attract people like this who say these kinds of things?

    I don’t know.

    Anyway, Joe has many strong opinions. I think Joe needs his own blog so he can voice those opinions there, in his own place, at whatever length he wishes. But this is not my post, so Joe is free for now to say anything he wishes, and take any tone he wants.

    Joe is a difficult case. Some of Joe’s points are interesting, other things are simply abusive or silly. How much is it worth sorting through the rag bin for the godd stuff?

    I don’t know.

    It is a matter of subjective opinion. I have low tolerance.

    Jonathan is very patient with people like Joe. That is his way. It seems to work for him. Jonathan is nicer than I am.

    But the world is big place, with room for all kinds of people. It is so big in fact that it includes Canada, for example, a very large place. Many people live in Canada and appear to like it there. Joe lives there. Joe lives in a far away place, yet here he is among us.

    The miracle of modern technology permits him to be here with us. Is this a good thing?

    I don’t know.

  183. I think a lot of the tolerance in this thread is due to dsquared. When asked to moderate his tone and give his arguments, he did, and made a rather crucial point about the likelier direction of bias in a cluster sample. I don’t like the point; my gut tells me that the median estimate is way too high. But it’s important to know the likelier direction of bias, and it gives me pause. At the very least it provides an honest out for the preponderance of statements that the estimate may be an under estimation.

    We also have posts like GKL’s reference to the MORI article, Chel’s references to the epidemiology articles. Not to mention Shannon’s initial analysis and AMac’s continuation. Even the occasional Guardian citation was worth following.

    Contrast this to similar streams on Iraq or other presidential policy, which generally degenerated to the Himmler, Bush as religious fanatic terrorist etc. comments without the information content.

    It has been a huge thread, but well worth the time in reading it, and maybe even worth the anger i had to swallow at times. Probably i got more out of it than anything other than the affirmative action / judge by intent vs. judge by results thread.

    And so we are nice because not being so would drive away folk like dsquared.

    Matya no baka

  184. Lex made the following comments:

    :Anyway, Joe has many strong opinions. I think Joe needs his own blog so he can voice those opinions there, in his own place, at whatever length he wishes. But this is not my post, so Joe is free for now to say anything he wishes, and take any tone he wants.

    Joe is a difficult case. Some of Joe’s points are interesting, other things are simply abusive or silly. How much is it worth sorting through the rag bin for the godd stuff?

    I don’t know.”

    Lex, this thread began as a discussion of science and mathematics in the problem of solving an “expectation value” for the number of casualites in Iraq as a result of the American invasion of that country. Usually in scientific research, the entire community of scientists, mathematicans and engineers arrive at a “decision” based upon “consensus” as validated by independent experimentation and research.

    When I write papers I try to follow the “golden rule” and try to consider everyone as my peers, no worse although quite possibly brighter or more insightful than me. But over the years I have also seen the growing sophistication of the art and science of “propaganda” as perhaps best typified by the phrase “plausible deniability” that somehow passed for the abdication of moral responsibility for henious acts of cruelty, inhumanity, futility and ultimately self defeating stupidity.

    Scientists, mathematicans and engineers are leaders in society and others look to us as models, particularly when we write papers, articles and opinions.

    I reacted vigorously to what Jonathan Gerwitz posted here, not because I have any particular disagreement with the authors of the original paper, or that I have disagreements with critics that have an interesting and dissenting position. Rather what I read between the lines was Jonathan’s words as a propagandist and “attack dog” for a jaded and perverted American political party that in this election “lied to most of the people, most of the time”. But they will not succeed in lying to “all of the world, all of the time” as Lincoln once shrewdly observed.

    This election was marred by a dreadful abuse of religion and theology in the United States when it was put in the service of unscrupulous politicans without any checks and balances on the truth, on science or on mathematics which is the language of science. The propaganda techniques now in play are hiding the truth from Americans about the many young people and civilians being killed in a war that has been completely mischaracterized by Mr. Bush.

    You are also right that Canada is a large country and its also true, even if you do not like my words, that Canada has been a long and steadfast friend of the United States and its people.

    Canadian peacekeepers died in Vietnam for instance, trying to keep a separation between the retreating American forces and the advancing North Vietnamese Army; which never the less saved lives on both sides of this war to which our country was never a party.

    Now make no mistake, 911 was a trauma for Americans and Canadians also perished on that day in New York. Canada declared War on Al Quada pursuant to our NATO obligations the following day and dispatched the Canadian Navy to patrol and interdict terrorist shipments and other illegal movements. We for example have performed fully one half of all boardings of such cladestine ship movements in the Gulf, and have participated with other NATO Navies that have intercepted illegal movements of terrorists, explosives, even missiles from North Korea.

    But there is also no mistaking the actual costs to America’s long term relationships for this ill advised invasion of Iraq. As the published study illustrates, civilian casualties leave deep scars for generations. Iraqis are not going to discuss THEIR trauma in academic terms, for most of them, like most Americans discussing THEIR trauma of 911, it will be personal and sad.

    Canadians know from first hand experience, that peacekeeping is long and hard work, and we have no illusion about what it takes to actually succeed. In Cyprus for example, Canadian troops kept Greek Cypriot and Turkish Cypriots separated for the better part of thirty years before tempers died down, and people there cooled down enough to talk and have a debate, rather than pick up a gun and start shooting and killing, thereby setting off another cycle of violence, reprisal and still more casualties and more victims. And Canadian peacekeepers know better than anyone that you cannot pick sides, and that if you can simply stop the shooting, you have succeeded, if only one hour and one day at a time.

    Much has been written around the world about American Foreign Policy. If you actually wish to foster the growth of democracy, you need to understand that you cannot achieve that by installing puppet governments as if they were plastic roses imitating a rose garden. You have to be prepared to tend real roses, water them, feed them and repot them from time to time. You cannot smell a plastic flower.

    Finally I wish to close with a word on “religion and science”. Scientists, mathematicans and engineers have to be especially careful in ensuring their work is free of “religious bias”, “assumptions” or “distortions”. That is not to say that scientists, mathematicians and engineers cannot be religous. But the duty to separate one from the other is paramount because when that does not happen, the public becomes very confused and that is when charlatans, witchdoctors, and lunatics confound the public and mislead many of them into even more serious errors. In the runup to the invasion, well over half the American public believed that Saddam Hussein was responsible for the 911 attacks although there was not a shred of evidence supporting this belief by Americans.

    When Jonathan wrote what he did about the paper, precious little that I could find in his criticism was either novel or technical, rather it had the flavor and odor of an attack ad, and an attempt to cover over the suffering of civilians in Iraq at the hands of American Arms.

    One of the harder things for a real friend to say is things we do not want to hear. Real friends are people who tell us that our fly is open, or that we need to brush our teeth, or that we need a fresh shirt or more deorderant. Cynical aquaintances usually will say nothing, except once you are gone, to repeat these things behind your back, often magnifying the slight or offence.

    My core concern over the Bush Doctrine of “pre-emption” is that it will destroy NATO and SEATO, and it will seriously harm critical co-operation between modern nation states to not only fight global terrorism, but its close relatives – narco-terrorism, and organized crime. Indeed, the Bush Doctrine could destroy the United Nations itself.

    I never intended to suggest that somehow Canada is free of these problems because the Hells Angels are alive and well in Canada unfortunately, and the Carlyle Group together with the Bin Laudin Group are still operating in the USA.

    One last point, and that is related to events of 911. The Canadian Armed Forces and the United States Air Force worked together on that day within an integrated command structure for North American Air Defence, or NORAD. Hundreds of civilian airliners were ordered to land in Canada as US airspace was closed, and tens of thousands of passengers that were stranded, including many Americans, were accomodated, housed and secured in the aftermath of that terrorist event.

    We will defeat terrorism but it will be with the “rule of law” and not with large invading armies that inflict large amounts of “collateral damage” upon the civilian populations of other countries.

    In any event, we shall see if Mr. Bush learned anything in this bitterly polarized election on the need to help Americans come together, on the need to work with America’s genuine friends, and to understand that you do not elevate your own stature by diminishing the standings of decorated war veterans, whether that is John Kerry or John McCain.

    The leadership of the democracies that America seeks is not a matter of an automatic right, it must be earned, like trust must be earned.

    We shall see if America fares any better in Mr. Bush’s second term, than it did in the first.

  185. Joe, I confess that I am scratching my head over some portions of the foregoing, but I like the tone of it a lot better. So that’s progress. We are starting to apply the Rodney King rule I think.

    BTW, SEATO was disbanded in 1977. Can’t blame that one on Bush.

  186. Lex wrote:

    “BTW, SEATO was disbanded in 1977. Can’t blame that one on Bush.”

    Well you are right that SEATO was created back in 1954 as a kind of organization that was rather like NATO and disbanded officially in 1977. At the time, the free world was greatly concerned about the expansion of communism from the Soviet Union and later from China.

    The SEATO organization functioned on the same principle as did NATO, namely it was a collective security arrangement that considered an attack on one as an attack on all.

    Unfortunately, the US acted “unilaterally” in Vietnam (the issue was also offshore oil and gas) since the civil war or if you like a “war of national liberation” was unfolding in a context that did not involve the necessity of a collective response, however much the US wanted to provoke it. As a result as you suggest, a very sound principle in Asia fell apart for lack of American foresight and planning, and the entire community became much more dangerous. It remains so to this very day as you can see in North Korea.

    Its a really good example of how unilateralism has actually failed American interests in Asia.

    And it grew out of the same kind of wrong headedness that you now see unfolding in Iraq. Far too much jingoism and not nearly enough study, planning and thought.

    Incidently I would blame part of the failure of SEATO on Mr. Bush Sr. He was the director of the CIA during much of this time span, and he fostered some of the bloodiest covert warfare in Asia which helped maintain ruthless and anti-democratic dictators such as Suharto and Marcos. Indeed, it brought with them the kind of corruption and crime that would undermine any credible collective security arrangement. I put this squarely on the shoulders of the United States, and its failure to properly lead. Instead, it got itself embroiled in a civil war that it could not win.

    Now you can see the same dynamic playing itself out in Iraq. The US has sided with the shi’ite against the sunni in Iraq, which is the principle political cleavage in Iraq. By supporting the shi’ite, the Bush Administration is virtually guaranteeing the installation of a Fundamentalist Muslim fascist state that mirrors the one in Iran. Its hard to believe that this is by accident, rather its a continuation of the long established policy of supporting dictatorships that are friendly to American administrations, and particularly to the Oil Barons of Houston.

    It would be interesting would it not to apply the methodology of the paper being discussed in this thread to the communist insurgencies in Indonesia, the Phillipines and Malaysia and then see how the numbers worked out because its well known that the civilian casualties were very high. Over half a million dead in Indonesia alone by some estimates.

    Incidently Christopher Hitchins, the conservative author and critic, regards Henry Kissinger as having committed “war crimes” during this period, something that Mr. Kissinger has not rigourously objected to in the Courts.

    Few Americans in hindsight would regard the period of the Vietnam War as being one that made them “safer”.

    Indeed, following the Vietnam War, there was a sharp runup in the inflation rate caused primarily by the waste and carnage of the war, and by the necessity to pay for it. This today is being mirrored by the vast sums being spent on the Iraq War.

    If in fact, young Sunni Iraqi are joining the insurgency to “fight the invader”, then no amount of American military resources can “win” anymore than it could in Vietnam.

    Finally Allawi is a figure rather like Nugen Kwo Kei. No one in Vietnam was fooled into believing that he was anything other than an America puppet and no one is being fooled in Iraq either.

  187. Indulge me please, because I’m no statistician, but it seems to me that:

    (a) 24.4 million (the population of Iraq) divided by 100,000 (the Lancet authors’ non-Fallujah estimate of excess deaths in the 18 months since the invasion) = 244; therefore 1 in 244 Iraqis has died, by this estimate, thanks to the war – about 4 in every thousand, 40 in every ten thousand;

    (b) 7868 (the number of living Iraqi individuals sampled) divided by 244 = 32, meaning that the figure of 100,000 deaths is extrapolated from 32 excess deaths discovered by the Lancet.

    This appears to accord neatly with the fairly detailed Economist article at: http://economist.com/science/displaystory.cfm?story_id=3352814, which says that

    Of the increase in deaths (omitting Fallujah) reported by the study, roughly 60% is due directly to violence, while the rest is due to a slight increase in accidents, disease and infant mortality.

    21 violent deaths is the figure for non-Fallujah violent deaths given in Table 2 of the paper, which is of course “roughly” 60 percent of 32.

    If so, then “100,000” is the number of deaths one would discover in Iraq if the Lancet’s sample of 7868 turned out to be perfectly representative of excess deaths in the country as a whole. How confident (in the normal usage of the term) can we be that this “perfect-world” number is, in fact, the “most likely” number of excess deaths country-wide?

    From what I understand of the issues involved, there’s nothing controversial about the survey’s falling short of perfection: this is what one would expect from the fact that it used cluster sampling – as would any other survey of its kind – and that clusters differed from each other to some degree (we don’t know how much, because the bar charts in Figure 1 relate to Governorates, some of which had several clusters each), and the survey design explains the width of the confidence interval (CI). But the CI deals with the purely random errors inherent to the mathematical design of the survey and (unless I am mistaken) the inter-cluster results obtained: it doesn’t address measurement or sampling errors due to poor survey design (if any).

    So I’m at a loss to explain the confidence (again, in the normal usage of the term) of the authors in the “100,000” number. Surely it should never have been promulgated as enthusiastically as it has been. I respect the position taken by dsquared (that the interval is more interesting, in that it shows an indisputably elevated relative risk of death post-war, especially from violence), but for good or ill it’s with “100,000 excess deaths” that the Lancet study is now associated, and this is down to no-one but the authors, who have promoted this number from the beginning.

  188. Wow. Some of you people are actually talking as if the U.S. military indiscriminately using air raids against residential targets somehow “defeats terrorists”, and “saves lives”. It’s always rather curious to refer to saving lives by taking many more. And yes, Saddam was brutal, we know this, but how many other brutal leaders, oh yeah, INCLUDING SADDAM, PINOCHET, THE SAUDI RULING FAMILY, THE SHEIK OF IRAN,and of course OSAMA BIN LADEN, have been supported by U.S. money, arms, and training throughout the world? There is your problem. As long as we have a bogeyman to pursue, in the name of stopping communism, terrorism, or what have you, we will be playing an awful role in shaping the lives of millions of people at gunpoint, whether it be our gun, or the gun we gave someone else to use.

  189. GKL,

    Yes, given that almost any small error in counting deaths would change the 100K figure by such a large measure as to make it useless, I too am at a loss to explain the vehemence of some of the study’s defenders. Some of them, like the clueless Desro above, casually latch on to the controversy as a vehicle for their existing agendas. That I can understand, though I don’t have much respect for such people. I am more puzzled by thoughtful commenters like dsquared, who continue to insist that it’s reasonable to make strong inferences from such limited and uncertain data as the investigators had in this case. Surely it would not be irresponsible to do more data gathering — i.e., not just analysis of the currently available data — before reaching a conclusion.

    BTW, is it just me or has interest in this topic dwindled since the election? I can’t imagine why that would happen.

  190. Wow, I didn’t realize this thread was still live.

    So Dreso, on what basis do you call U.S. air strikes in Iraq, “indiscriminate”?

    I think a lot of the problem that people like you have in thinking about the war is that you don’t have any understanding of the technical aspects of how it is fought. You seem to think it’s like WWII or something similar. You find the Les Roberts study creditable because you have no grasp of the scale of the conflict nor the methods employed.

    Perhaps you should educate yourself about the on the ground realities before you castigate others.

  191. What Shannon said.

    If the USA wanted to do indiscriminate bombing, we could, like when we burnt down Dresden. We don’t because it would be wrong as well as ineffective to do so.

  192. It seems to me that there are additional factors at work that may well show additional mortality, not less. The Lancet study uses geographically fixed clusters in its sampling technique but this will undersample actual loss of human life. In real civil wars, which is what is now unfolding in Iraq, civilian populations are not “static” but are driven from their homes, or are trapped in them between the forces actually doing the fighting.

    In places where there is a great deal of fighting, such as Falujah, these are not so much “outliers” as much as valid samples with very high mortality rates. One would expect that as the civil war spreads, that there will be other such clusters that experience the same effect, so that the exclusion of such clusters may well be reconsidered and included in a valid manner.

    The other aspect of the study is the inherent assumption of causality, namely that the violence of the war is not related to such other factors as malnutrition and lack of medical treatment. But in fact the war has a pretty direct causal link in that the lack of security leads directly to loss of food supplies and medical supplies.

    All that suggests to me that the Lancet study has likely UNDERSTATED the mortality in Iraq, not exaggerated it as the appologists for the American Government would have you believe.

    There have been other studies that have suggested that since the First Gulf War, that over 2,000,000 Iraq civilians have perished because of the sanctions, because of a lack of food and medicine, INSPITE of the Oil for Food Program, and we know that well over 200,000 Iraqi casualties were inflicted upon the Iraqi conscripts, but not the elite Republican Guards. And we know that perhaps something like 30,000 to 50,000 predominately Shi’ites have perished in an uprising encouraged by George Bush Sr. but not followed up with support on the ground once the uprising began.

    There is no doubt that this is a “covert war” and that in such conflicts a great deal is “hidden”. One such “hidden” factor is the very high mortality to civilian populations that the Lancet study only hints at.

    Finally there are additional mortality factors in play in Iraq that one would not normally think of and that the effects of spent depleted uranium munitions since the first Gulf War. Because of Iraq’s deserts, and clustered populations along the rivers, and the subsequent concentrated use of depleted uranium munitions in such urban areas, the incidence of radiation disease through the food chain leading to secondary mortality from cancer, lukemia and others cannot be easily dismissed as “normal” mortality, such as takes place in Western Countries.

    There are also secondary effects in the population life expectancy which one would expect would be falling. If this figure has fallen to 48 years, it already implies a mortality of 500,000 per year. In contrast, if this figure were closer to 96 years, the mortality figures would approach 250,000 from “natural causes” comparable to Western figures for a country with a similar population. The difference is around 250,000 and that is a figure that I would think is pretty reasonable and practical because it captures these secondary effects of the war on mortality.

    In the war in Falujah and other places in the Sunni triangle, there have been hundreds of deaths reported nearly on a daily basis. With reports of some 15,000 actual reported casualities, this alone pushes the figures much higher than even the Lancet study, in other words, something approaching the high side of their study estimate, something closer to 200,000.

    All this actually begs the question about the Pentagon’s statements about “collateral damage” because its clearly much larger than the military will admit.

    There is a good explanation. Most military studies of the effectiveness of bombs and weapons are done in the context of a reasonably well prepared and equipped enemy. However, civilians are not so protected and so equipped (flak jackets and body armour), so that “survivors” in the military calculations are actually “casualties” in real life. Anyone who has actually seen a 500 lb. bomb going off in real life, knows that it will kill unprotected individuals at much greater ranges then the military studies would admit for well protected troops.

    Even with a killing efficiency of 200 Iraqi civilians and insurgents combined for every US soldier killed, that still is a highly immoral calculus when you look at the final objective, which is to loot Iraq of its oil and gas resources so that the Houston Oil Barons can rule the world the way the “real axis of evil” — AKA “Cheney, Rumsfeld, Wolfwitz and Perle” have planned in a permanent American hegemony over the Middle East.

    Even an election cannot turn a Big Lie into Enlightened Truth. Shit is Shit and it can never be gold.

  193. Suffice it to say that you wrote your own epitaph, Joe “Even (the loss of) an election cannot turn a Big Lie into Enlightened Truth. Shit is Shit and it can never be gold.” This almost perfectly encapsulates the Lancet study in its entirety.

  194. To Ed Snack and Andy B:

    “There are none so blind as those that will not see.”

    You obviously are not interested in science, mathematics, politics or the war. Rather your utterances prove the point I started out with, that propagandists that attacked this paper in the Lancet (such as those in the US under Karl Rove) have pretty much achieved the same standards that Herr Gobbels introduced to the Third Reich.

    And just as obviously, you actually have nothing more to contribute to the factual scientific discussion.

    I rest my case.

  195. I am more puzzled by thoughtful commenters like dsquared, who continue to insist that it’s reasonable to make strong inferences from such limited and uncertain data as the investigators had in this case.

    Because I understand the underlying sampling theory. If the underlying death rate hadn’t risen, it would be very difficult to get a sample in which the sampled death rate had risen so markedly. It is the data which is supporting strong conclusions, not me.

    btw, Joe writes:

    In real civil wars, which is what is now unfolding in Iraq, civilian populations are not “static” but are driven from their homes, or are trapped in them between the forces actually doing the fighting.

    which I don’t think is quite right; although there is an effect here, it’s also the case that the people left behind in these clusters are going to have much clearer recall of recent deaths than of deaths before the war, which “recall bias” would tend to drive up the estimate of the change in the death rate. The Lancet team didn’t believe that this was a big factor either way, based on internal consistency of recalls in their dataset, and based on death certificate evidence where they asked for them.

    Marc Mulholland notes elsewhere that the estimated death rate for Iraq is twice as bad as the worst year in Northern Ireland if you take the low end of the range and ten times as bad as the worst year in NI if you take the high end of the range. Given that the British Army never used airstrikes in Belfast, I would suspect it is more reasonable than people think to assume that Iraq is nearer ten times worse than Northern Ireland than two times.

  196. “Its a really good example of how unilateralism has actually failed American interests in Asia.”

    Unilateralism and Vietnam did quite well in buying time for Asian countries such as South Korea and Taiwan to prosper and grow as democracies.

  197. It’s about the oil stupid. With regard to Iraqi deaths and the average American driver I am reminded of the quote by the Cliff Roberson character in “3 Days of the Condor” when questioned by an outraged R. Redford about a clandestine CIA plan to invade the Mideast (paraphrased of course) “When their (the American People) lights go out, their homes are cold in teh winter and there isn’t gas for their cars, they won’t care where the oil comes from, they’ll just want us to get it for them.” The daily deaths of gallant American soldiers doesn’t take the average American from his color TV, six pack, and dinner, why should the deaths of a couple of hundred Iraqis a week bother him? Only the brain dead or naive think our involvement is about freedom of the Iraqis, it is about world oil markets, and the founding of a friendly secular client state amidist a gaggle of threatening, backward theocracies with mindsets stuck in the 11th century. Just as the civil war morphed into a struggle for the freedom of the negro from a struggle over economic imbalance and states rights, it has become politically expedient to paint Iraq as another Crusade to set men free.

  198. No Joe, the fact is that I have been following the arguments closely. It is clear to see that no amount of logic or rationale is able to penetrate your skull. You are arguing a position, and when segments of your argument are attacked, you shift the argument. Anyone could pour an inordinate amount of time into telling you that black is not white, it is black, and you still would defend the indefensible. I don’t have that time to waste.

  199. dsquared quoted me, then replied:

    I am more puzzled by thoughtful commenters like dsquared, who continue to insist that it’s reasonable to make strong inferences from such limited and uncertain data as the investigators had in this case.
    Because I understand the underlying sampling theory. If the underlying death rate hadn’t risen, it would be very difficult to get a sample in which the sampled death rate had risen so markedly. It is the data which is supporting strong conclusions, not me.

    This is the fundamental point of our disagreement. I think there are too many plausible uncertainties about the data, dsquared disagrees. Perhaps someone will eventually try to replicate the study using less-controversial survey techniques. Until then, it looks to me that there’s enough reasonable doubt about the result to remove it from serious consideration as a policy tool.

  200. I think there are too many plausible uncertainties about the data

    Which ones, precisely? At present, my assessment is that the only really tenable critique is the “Lying Iraqis” theory. I might be open to very persuasive arguments that the sample was nonrandom in such a way as to systematically overestimate casualties, but I must say I can’t see any weaknesses in the construction, and Fred Kaplan’s caviling doesn’t convince.

  201. Shannon Love –

    I am calling these air strikes indiscriminate based on the fact that you can’t use air power in an urban firefight vs. these type of insurgents. You will inevitably end up bombing civilian targets, such as peoples’ houses in insurgent controlled areas, the “10%” of civilians our military reports have not fled the city. Is it realistic to expect an entire city to “flee” somewhere so that the big boys with guns can solve the problems? Has anyone looked at any pictures of the reality on the ground, heard any interviews with angry Iraqis about how their houses are being destroyed, their families killed, and most of this all by air strikes? I know I can be accused of being a flaming liberal, etc, but i don’t see any possibility for “victory” in Iraq. When you hear “militants”, or “insurgents”, say things such as “without fear, you cannot die”, what does that show you about the mentality of these people? Does anyone remember when Ho Chi Minh said that if ONE American dies for every TEN Viet Cong, his strategy was successful? The best we could hope for is free elections in which radical Islamists are elected as the leaders of the country, in a style much similar to Iran. You can debate the specifics of civilian casualties, humanity in military targeting all you want, but you ignore the basic humanity involved in the situation. When one family is accidentally killed by a coalition airstrike, even if totally unintended, how are all the other families living within miles of there going to be affected? They are going to hate the Americans that much more, and probably cooperate with us that much less, while possibly aiding insurgents. It is strikingly similar to Israeli military tactics in the Gaza Strip and West Bank, those tactics being “there are terrorists in here somewhere, so lets just bust down every door and scare the daylights outta every family on the block, while destroying half their house and putting giant holes in the walls everywhere”. This will find the terrorists, will kill the terrorists, and it will also result in public opinion of the people on that block going 90+% against us, guaranteed. The only way to really win in the end is to win the popular support of Iraqis for what we are doing, which the Bush Administration has miserably failed at doing. What message does it send when we allow the country’s historic and culturally significant sites to be looted and ransacked while Marines guard the Oil Ministry building? As long as we keep fighting, even if we feel the fight to be a noble one, civilian casualties WILL occur, and in urban fighting, will occur at a substantial level. I can only see Iraqi public sentiment turning further against us, and our battle against “terror” there may as well be, in the end, a battle against the Iraqi people, who have little hope left as unemployment there reaches all-time highs, while at the same time Kellogg Brown & Root workers are paid the salaries we should be giving Iraqis to rebuild the country, and most importantly, the oil infrastructure. I am just very sad in the end, because i see no reason for the reality of the situation there to be publicized, broadcast on TV, etc. The best way to validate claims of civilian housing being destroyed is to find Iraqis who are now homeless because of this, and who are screaming foul. They could be found, but who will find them in Fallujah, Sadr City, Samarrah? Our reporters wouldn’t walk the streets in a full metal jacket in places like that. No one but our military will, so I guess we’ll just have to take their word that civilian casualties are very minimal.

  202. Jonathan G ewirtz –

    I don’t feel this is about my agenda, or anyones. I know, as the study says, that this would need to be verified by a much larger independent sampling, which is something that will never happen, as not many are willing to risk their lives in conducting such a broad, large scale inquiry in a war zone. Your answer would be then, I suppose, not to accept the figures, and really, not to even worry about those deaths, because you are not being killed, nor is your neighbor, or anyone on your block. Maybe interest in this topic has lessened since the election because people realize whether we like it or not, we are in for at LEAST 4 more years of fighting a conflict that is unwinnable. So my question in end to you is, do we continue to try and make the best guesses we can about civilian casualties, using methods which are open to skewing and have large margins for error, or do we not not care at all? I guess we can’t accurately find the number of civilian casualties, so we may as well not try at all. As George is so fond of ignotantly and hatefully saying: “We’re fighting them over there, so we don’t have to fight them over here”. Beatiful. Now THAT’s humanity. I’ll bet most Iraqis love that.

  203. Desro, yes, it’s not about your agenda. I was mistaken. How could it possibly be about your agenda when you falsely attribute bad motives to me because I don’t share your defeatist assumptions. No, that has to be false consciousness on my part. You, OTOH, have pure motives, so you must be factually right as well. Keep at it and your ad hominem ravings will no doubt convince me. Maybe you will even convince George Bush, another person who is bad by virtue of his disagreement with you. Man, I wish I had it all figured out like you do.

    BTW, who is “THE SHEIK OF IRAN”??

  204. KPatton wrote: Just as the civil war morphed into a struggle for the freedom of the negro from a struggle over economic imbalance and states rights, it has become politically expedient to paint Iraq as another Crusade to set men free.

    You conveniently forget that the states rights issue and secession were driven by the slavery issue. You choose to believe this is about oil. Just like you choose to believe the Civil War was about everything except what it was about.

    Desro wrote: The best we could hope for is free elections in which radical Islamists are elected as the leaders of the country…

    That’s the best we can hope for? Why do you suddenly abandon all knowledge of the research and polling showing wide support for a secular, democratic government in Iraq? Because it doesn’t support your doomsday scenario?

    We all know the reason the numbers are being skewed is to make an antiwar political statement. If you’re a pacifist and opposed to war, fine. However, you shouldn’t need to hide behind poor statistical analysis to make your point. It’s that simple.

  205. dsquared, the “lying Iraqis theory” is as good as any, plus the fact that a small difference in the number of deaths would change the result significantly. I am no statistical maven but I know enough to be skeptical of research outcomes that are not robust WRT small changes in key variables, especially when such outcomes are inconsistent by a factor of at least 3 as compared to the previous estimates (which, I think, may themselves have been exaggerated for political purposes). These considerations, plus the obvious political enthusiasm of the investigator, plus the unusually quick review period, plus the other obvious data-gathering difficulties in a war zone, do not necessarily mean that the research outcome is invalid, but I think they strongly suggest that that outcome is not to be trusted unless and until it can be validated independently. I mean, what does it take for you to decide that investigators’ efforts here might be better spent in seeking additional data than in performing ever more involved analyses on existing data that are about as precise as a fly’s path over a pile of shit?

  206. To Andy B who wrote:

    “Couldn’t have said it better myself Joe. You really summed up the study with that line.”

    Thank you, but you need to focus better on the target. Your target SHOULD BE the critics of the paper published in Lancet.

    The author’s best estimate of Iraqi mortality due to the war is around 100,000 with the upper limit approaching 200,000 and they attributed most of it to air strikes from US Forces. That is the long and the short of it.

    Americans are killing a lot of Iraqi for their oil.

    The pychotic aspect of American Foreign Policy in Iraq is that Americans actually expect the Iraqi people to “love them” for all this loss of life.

    As Gordon Lightfoot once wrote “Freedoms just another word for nothing more to lose”. “Freedom ain’t worth nothin but its free”.

    The authors did a much needed job of putting realism into Emperor George II’s war in Iraq and paint it with a human face and in stark terms. To put this into perspective, America’s bombing campaign in Iraq has now claimed the lives of between 12 and 24 times as many people as have perished with Saddam’s chemical weapons attack on the Kurds.

  207. I would also like to point out that the argument Kristofferson advances is that a lack of commitment and meaning that comes from moving on–admittedly an American tendency with our emphasis upon individualism–causes regrets and perhaps the emptiness he so artfully captures in “Sunday Morning Coming Down.”

  208. I mean, what does it take for you to decide that investigators’ efforts here might be better spent in seeking additional data than in performing ever more involved analyses on existing data that are about as precise as a fly’s path over a pile of shit?

    Jonathan, I am close to losing my temper with you. You have no substantive arguments with the study (or rather, you have attempted to raise several and failed every time), but rather than apologise for impugning the integrity of good scientists and starting an intelligent discussion about the numbers, you are retreating to vague and unprovable innuendoes about bias, using an argument (trying to pretend that the IBC number is a comparable one to the Lancet number) which you must know to be fallacious, and now, to top it all, trying to sneer at the researchers for not taking even longer and exposing themselves to even more risk of death in doing the US Army’s job (as required by the Geneva Conventions, startlingly) for them. Have a word with yourself. Your own political bias is the issue here, not the Lancet’s.

  209. dsquared, your suggestion that my expression of cautious doubt about the validity of this study “impugn[s] the integrity of good scientists” is a stretch, to put it mildly. You also misunderstood my invocation of the Iraq Body Count’s civilian-death estimate: the point was that until the Lancet study the obviously-politically-motivated IBC estimates were the highest ones anyone was making, so absent better information one should be cautious of any study whose results include outcomes that are multiples of the IBC numbers.

    You keep trying to frame this argument as being about “the numbers.” It is not, or not principally. It is about the plausibility of the theory by which the numbers were established. There are too many possible reasons, including the “lying Iraqis” one that you acknowledge, for why the data may be unreliable. Since less than 80 reported civilian war deaths in 900 families were used to infer total deaths in a 25m population, almost any systematic errors in measuring war deaths, in attributing such deaths to civilians or in determining pre-war death rates would skew the results significantly. And because this research hasn’t been done before in this particular society under similar conditions, there are many conceivable reasons why the reporting might be inaccurate. That’s what it comes down to. One can’t prove that it’s accurate via analysis. One can only attempt to replicate the study so as to get more data and more experience with local conditions and peculiarities. The fact that doing that might be dangerous is not an argument against the necessity to do it if one wants to gain an accurate understanding of what is going on.

    WRT biases, I have never hidden mine. OTOH, you seem to have some issues. And so, I think, do the authors of this report. My impression is that their agenda is at least partly political — else why ask to be published before the election? They were not necessarily pulling for a particular candidate, but clearly they saw publication as an engaged act. And The Lancet went along with it. I am puzzled that you don’t find this behavior inappropriate, or at least suggestive of a need for more scrutiny. No study has such important results, or is so dangerous to conduct, as to be beyond the ordinary requirements of independent confirmation.

  210. Thanks for your responses in which you do nothing other than accuse me of having “pure motives”. Thank you for not responding to the humanitarian arguments I went over. One could look at it as a “defeatist” attitude I suppose, but I think it is a valid viewpoint that we have taken a heavy handed approach here, which has not won the hearts of many Iraqis. Sure, some polls show Iraqi support for Bush, the liberation, new secular gov’t led by Allawi etc etc….but who is being polled? Are these polls conducted in relatively safe areas, or in more heavily insurgent populated areas? Predominantly Shiite or Sunni areas? You use broad statements and simply dismiss my arguments as liberal rants I suppose. I am just looking at the reality of what happens when you fight a widespread insurgency composed of the same people we were supposed to liberate. These are not, in majority, foreign fighters as many would like to say, it appears these are heavily Sunni Muslim fighters who are now angry at their very small role in the upcoming gov’t, apparently thinking that the new governing arrangement will be cutting their due influence short, which may or may not be true. So I still say, when we fight and kill these “insurgents”, we are killing Iraqi citizens, at least a majority of the time. I hope that our military’s estimate of insurgent casualties is way off, because they are saying things like 1500 insurgents killed in Fallujah, as opposed to 38 GIs. This is rather disturbing to me. Probably not to those of you who are good honest god-fearing Americans, though. I imagine you revel in insurgent and terrorist death because they are the great enemies we face these days. I just happen to think that 1500 deaths, when none are necessary, and we should not have invaded a sovereign country, are not a good thing. I repeat my assertion, ignored by Jonathon I suppose, that mass numbers of Iraqis, “insurgents” or not, being killed by “coalition” forces, is not good P.R. for the occupation. In fact, it only encourages more insurgents to take up arms against us. Experts are already predicting a very difficult time having any type of elections on the schedule that we are sticking to now. I don’t see this as a motivated argument, just a practical one that plays off of some basic human emotions. To draw a paralell and maybe help you realize what I am saying….Imagine that at some point in the hypothetical future that some other nation invades and takes over the U.S. because they think we are threatening in some way. I know this is all hypothetical, bear with me. If Americans begin a civilian resistance at a point after which the bulk of our military has been defeated, I would see that as justifiable. Now imagine, Jonathon, that youre next door neighbor, your cousin, your aunt, your brother, someone who means something to you, dies when their house is flattened to the ground by a 500 lb. bomb. When you angrily ask what they ever did to deserve that, your occupiers tell you that she was providing safe harbor for insurgents. Now, after this, do you say “OK, I understand” to your occupiers? Or do you try to gather whoever you know and fight for what you have left. I know this isn’t numbers, or scientific fact here, or polls, or studies. This is just logical thought, which I hope you can follow through. I am sorry I come off as confrontational, I am merely passionate about this, and I believe that our country in the end will be much worse off for this prolonged occupation, which has divided our populace and the world’s, with many people like myself pointing out various drawbacks. I believe that at this point, the only thing to do is hold elections, and in that I agree with Bush and his crew. I just am pessimistic about the final resulting state of Iraq, and I believe my pessimism is justified. You say public opinion polls there show widespread support for a secular democratic state. Well….link me to these polls then! I do consider myself fairly well informed, and I don’t believe I have been arguing anything on basis of incorrect fact. I just want you to do more than say “Oh youre ranting, and you have an agenda, blabla” You are the one sounding like a broken record slamming everyone who does not agree with you, except for dsquared, because his tone is more moderate. I realize there are disagreements. I just don’t understand how I am strictly “pure motive”. Like I have said, this is not a study, just a point of view, that seems to fit in with logical thought. BTW, I apologize for the Sheik of Iran. I meant to write the Shah of Iran, who ruled until 1979 I believe when he was deposed by the Islamic Revolution there. The Shah had built Iran’s military to the 4th or 5th largest in the world at the time, buying weapons from primarily our gov’t. This caused a backlash, however, as a strong military did nothing to help the poor masses, and they revolted, using a broad based appeal to Islam as their rallying cry. Another link of interest pertaining to the deteriorating situation in Iraq: http://english.aljazeera.net/NR/exeres/A66151CB-2105-418B-BFAA-73211A631611.htm I believe has great relevance especially with elections approaching. I don’t know if any candidate will be able to really solve the problems faced there today, but it will be interesting to see who the people think is best capable.

  211. Desro, I almost deleted your last comment but decided not to because you made some serious points. However, your argumentative style is abusive, as you repeatedly impugn my motives for disagreeing with you, and I am tired of debating you. If you’re as passionate about these issues as you say you are, go and start your own blog instead of using this comment thread as a vehicle for your off-topic opinions.

  212. Insurgents? Not….

    Fallujah is Iraq’s murder capital — or more precisely, the outlaw town used as a staging area for murder committed by Iraq’s secular and religious reactionaries.

    And “reactionary” is a much more apt description for these thugs than “insurgent.” Words matter, and insistently describing the murderers in Iraq as insurgents distorts the aims and true nature of these enemies. Saddam’s old cronies (the secular reactionaries) and Musab al-Zarqawi’s suicide bombers (the religious reactionaries) don’t hold elections, they don’t dig sewers, and they don’t build hospitals. The secular reactionaries want to return Iraq to a Sunni-dominated dictatorship — the corrupt, murderous hellhole Iraq was in March 2003. The religious reactionaries have a grander target, with their “golden age” a bit deeper in time. They want to run the entire world along the lines of an 11th or 12th century Muslim caliphate.

  213. Does anyone any longer have any doubt about the accuracy and veracity of the Lancet Study?

    If anything, given recent past events, this figure will look conservatively small. Just look at how many American troops alone have been killed since this debate raged a month an a half ago.

    One sure thing you can observe. Karl Rove attended to the propaganda needs of his Government every bit as diligently as Herr Gobbels did to his.

  214. Americans, get out of Iraq!

    Stop attacking people and Nations

    Stop causing wars

    blah blah

    Thanks to the blog admin for deleting the rest of my boneheaded anti-American rant and saving me additional embarrassment! ChicagoBoyz rock!

    regards,
    Peter from Europe

Comments are closed.