After accusing the researchers and the Lancet of fraud and treason, Shannon Love is back with another accusation. The latest crime he accuses them of is rounding things off:
One easily graspable example in the Lancet study's dishonesty is the key sentence in the Summary, the one repeated in the media world wide, that pegs the "conservative" estimate at 100,000 excess deaths. The actual given estimate is 98,000. What pure scientific purpose is served by rounding the number up to 100,000? There is no technical reason for doing so. They chose that number because a big, round numbers stick in people's minds. Its a number chosen only for its marketing value.
The nice thing about this criticism is that if they hadn't rounded things off, Love could have accused them of presenting the results with spurious precision: "Look saying 98,000 implies that the result is known to the nearest 1,000, but it isn't!"
Then, apparently in all seriousness, Love demands to know how the information in the study could be used to save lives in Iraq. The answer is pretty obvious, of course, if you want to save lives you have to know how they are being lost, as explained here. Love follows this up with a full-on "shoot the messenger" post. Hey, if science gives results that he doesn't like, that proves that science has been subverted and can no longer be trusted.
Another of the "ChicagoBoyz" has joined the fray with some blatantly cherry-picked quotes. Jonathan Gewirtz attempts to mislead his readers by contrasting a quote from dsquared in response to Love's charges of treasonous and fraudulent conduct by the Lancet with a completely unrelated quote from Love. In comments, Gewirtz went on to accuse dsquared of being "essentially dishonest".
The criticisms of the Lancet study by the Chicago Boyz have mostly been entirely without merit (like the rounding argument above). Even when there has been some basis to the criticism (the study could have made it clearer which estimates were based on data including the Falluja cluster), they have proceeded to make a mountain out of a molehill (Love claimed that the treatment of Falluja proved that the researchers had conspired with the Iraqi insurgents to falsify the results).
Tim,
I agree with you to the extent that this debate has been extraordinarily and unnecessarily acrimonious, though, in coming-late-to-the-party, I see this on both sides of the divide.
That some criticisms of the Lancet study have mostly been entirely without merit does not, of itself, make Roberts' paper better, or worse.
I recently addressed my four (or five) top concerns about this article as a comment timestamped 28/3 13:48 on the Deltiod 'Lancet Links' post. Some of your readers may be interested.
Re: acrimony at ChicagoBoyz, I hope you will have the time at some point to show your work, or retract assertions you made in response to a comment I left there. (timestamp 3/26 11:23am).
Thank you, Tim, for doing the unpleasant work of reading Duh Chicago Boyz so we don't have to.D
Is this the best way to deal with Shannon Love? He reminds me of a two-year-old I know who has discovered that the adjective "f__cking" is terrific for grabbing the attention of adults. (Hopefully she hasn't yet grasped the fact that it can also be a verb.) If she and Shannon are ignored, will they grow out of it, or is a good kick up the arse required?
I agree, Shannon's off the mark on the above comments. The survey has enough mountainous problems without those mentioned above, like only proving 8,000 excess deaths and not sampling the most heavily Shiite and Kurdish areas.
So, anyone want to do an estimate on how many Saddam would have killed/tortured/maimed since April 2003 had he been left in power? When are we going to get a quantitative analysis of the value of freedom and democracy vs. tyranny and oppression? Maybe we can ask the N Koreans and S Koreans to help out on that one (I think the death toll in N Korea is a few million now). Also, some on the left might disagree, but most people would probably say a GDP per capita ten times higher is worth something too. http://www.cia.gov/cia/publications/factbook/rankorder/2004rank.html
Since the ten times higher GDP is from pumping and selling oil as fast as possible, and since the rate limiting step before 02 was the embargo, I find the issue about GDP, shall we say misleading. Of equal interest is GDP-oil revenue.
No, I was referring to S Korea vs. N Korea on the ten times higher GDP per capita.
TallDave wrote, No, I was referring to S Korea vs. N Korea on the ten times higher GDP per capita.
Oh, OK---you mean you weren't making any sensible point.
tallDave,
The Lancet study did not, in any sense whatsoever, "prove" that there were 8,000 excess deaths.
Rather, it described a probability distribution -- roughly (tho not quite correctly) speaking, it found that the most likely number of excess deaths is 98,000 and that there is only a 5 percent chance that the number was below 8,000.
The 5 percent cutoff point, tho conventional, is entirely arbitrary. if we picked a 1 percent cutoff for significance, we would get a smaller number; if we picked a 10 percent cutof,w e would get a higher number. Either way, it's just shorthand for describing an entire curve of probabilities.
I am guessing that you have neither read the Lancet study nor ever taken a statistics course. Correct?
Oh I don't know liberal, I think that the fact that North Korea's GDP is lower than South Korea's is a very telling, even a decisive, critique of the Lancet study.
Lot of talk but you never did explain why it was rounded up to 100,000.
"there is only a 5 percent chance that the number was below 8,000."
Not quite. There is a 5% probability that the true value is outside the confidence interval. That is, there is a 2.5% probability that it is below 8,000, and a 2.5% probability that it is above the high point of the range.
Simon:
If you're confused about why it was rounded up to 100,000, there's a pretty good website here; it's colourful, and it even has a little quiz at the end so you can see how well you've understood the concepts.
Feel free to ask more questions if you're still confused.
Via Juan Cole's blog a statement by general George Casey that there is 50-60 attacks by insurgents per day in Iraq. How many do we read about in the media? 2? 3? That suggests the fraction of the violence that is reported, giving further credit to how the casualty figures could be as high as 100,000 without most of them being reported.
http://www.abc.net.au/news/newsitems/200503/s1332344.htm
Thanks Jonathon loved the colours. Do you have a lot of other links like that up your sleeve ? Like "How babies are made" and "How to count to ten without fingers or toes".
Doesn't help me with my question.
You really are a half wit, or maybe I should round that up to full idiot.
Thomas,
It is not only the majority of attacks by so-called "insurgents" the media do not tell you about, it is also the majority of attacks by U.S. forces, most of which are on cities, and therefore will inevitably result in civilian casualties - you simply cannot bomb population centers without killing anywhere from a few to a great many non-combatants. And every day Iraqis are killed at checkpoints, or for not showing the proper deference to U.S. convoys. There are also enough credible reports about "security contractors" (aka mercenaries) killing civilians to indicate it is a significant addition to the dangers of living in Iraq while Iraqi.
I'll bet you hear a lot from the media about "victories against insurgents", both bogus and real, and I'll bet are either "joint U.S.-Iraqi" operations or "Iraqi forces" doing thus and so".
Amac, you wrote:
Violence accounted for most of the excess deaths whether or not Falluja is included. You are correct on the second part of your statement -- coalition air strikes are are only the major part of violent deaths if the Falluja cluster is included. I noted this flaw here.
Shirin, I live in Sweden so I get slightly less biased reporting than what you get in US mainstream media. I'm quite aware that this underreporting of violence is going to be at least as severe when it comes to violence from coalition troops, or mercenaries, but every example helps in showing those who seem to assume that people always have the courtesy of dying in front of a camera so that reports like Iraqbodycount will give an accurate estimate.
I read one embedded journalist who claimed that while he was in Iraq he witnessed on average one war crime per day in the US unit he followed. Not all may have been lethal, but nevertheless it adds up. The brutality of US troops reportedly also was a contributing factor to Spain pulling out. The Spanish people were against the invasion all the time, but after a while the military became disillusioned too.
Tim Lambert (29/3 16:37):
Thanks for responding. In my first contention, I misinterpreted excess deaths as deaths. So we agree. I retract that contention, and hold to the second one.
Nice summary Tim!
I've been following this story on Chicagoboyz ever since it started way, way back. The attitudes expressed there are so interesting to me. Even a suggestion that a lot more civilians are dieing than previously thought causes a lot discomfort. I wouldn't have guessed this because I thought the strong pro-war opinionated people would feel the war is so good and so right that it doesn't matter if 1,000, 100,000 or 500,000 civilians die, it's still worthwhile.
Also their standards in science interest me. If you look back on the blog they really went to bat for Lawrence Summers and his comments that were based on flimsy conflicting science that President Summers didn't even do. When it comes to the Lancet study they start to get all upset when the Summary rounds to 100,000 and question the heterogeneity of the clusters.
I am a bit puzzled by the latest exchange between AMac and Tim. If I understand them correctly they have concluded that less than 11 of the 21 violent deaths (excluding Falluja) are due to US airpower. How do we know? Is there an implicit assumption as to how the violence attributed to Iraqis is distributed between Fallujah and other clusters?
It may well be a silly question since I am a little bleary-eyed trying to make sense of the various niggles; but any clarification would be appreciated.
The Lancet paper reminds me of the technical marketing materials one is forced to wade through in the computer industry. The information must be true enough to pass legal scrutiny or the company can be sued for deceptive advertising yet everybody knows that the companies spin, slant and obscure the way the information is presented to such a degree that it is easy for all but the most diligent reader to come to erroneous conclusions.
I suppose if your standards for scientific papers can be met by a Microsoft white-paper, then this paper does pass muster. If you have higher standards however, you should be concerned.
liberal,
No, I had a point -- freedom matters, and S Korea is the proof. Sorry if that was too hard for you to decpiher.
lemuel,
I'm trying to control my laughter over your post, but it's just too hard. Oh, so 95% is totally arbitrary and just as good as any other percentage, say 1% or 99%. Have you ever seen a math book before, or heard of math? Did you ever stop and wonder why 95% is conventional, or were you under the impression it was just a randomly chosen number between 1 and 100?
The convention is that 95% confidence constitutes proof, and therefore this study (if one accepts its methodological flaws) proves 8,000 excess deaths.
Kevin Donoghoue,
The breakdown was revealed in an email sent to Mike Hardwood. Tim has a post on it somewhere. The email says that all 12 of the deaths not attributed to the coalition occurred outside of Falluja. This means only 9 of the deaths measured outside of Falluja are attributed to Coalition actions. Since the study also says that 3 deaths were caused by Coalition small arms fire outside Falluja that means 6 of the measured deaths occurred in Coalition airstrikes.
Excluding the Falluja cluster changes not only the magnitude of the number deaths by violence but also the key findings of the actors causing them and the nature of the victims.
The study literally reaches opposite conclusions based on whether the Falluja cluster is included or not.
Tim:
In your reply to AMac, you link to an earlier post where you acknowledge the study's conclusion that coalition air strikes caused the majority of violent deaths (Interpretation statement on page 1) is only true if Falluja is included. In the linked post, you indicate this is a " pretty minor flaw."
I have to disagree. For the benefit of those commenting here who did not read the recent Chicago Boyz threads, Tim revealed that he had corresponded by e-mail with Richard Garfield, one of the study authors. The purpose of this correspondence was to clarify an interview Garfield had given to " EPIC" (Education for Peace in Iraq) in November 2004. In the EPIC interview, Garfield had made this statement:
"In areas of Iraq, with the exception of the North, all had a rise in the mortality rate and most were due to violence. Real change was in deaths due to violence.[The majority of the 57,600 deaths due to violence was attributed to air assaults.]"
In the interview, Garfield attributed 30,000 of the 100,000 excess death estimate (excluding Falluja) to coalition bombing. This figure is consistent with the study's own key Interpretation statement which reads:
"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."
However, in his correspondence with Tim, Garfield acknowledges that he was in error, that the 30,000 bombing figure is completely incorrect, and was the result of him inadvertently including Falluja into the mix. This admission leaves us with an ex-Falluja death extrapolation for coalition bombing of approximately 17,000, using the study's actual ex-Falluja data.
How did one of the study's leading figures make such a key mistake? How does his explanation that he inadvertently included Falluja account for this 30,000 figure? Including the Falluja numbers does not give us an extrapolated toll from bombing of 30,000, in fact this toll would be past 150,000. It's clear that the 30,000 figure comes from the Interpretation statement's conclusions concerning attribution of violent deaths, not the Falluja data. As I noted at Chicago Boyz, study authors other than Garfield also reinforced in interviews the fact that the 100,000 excess death estimate from the Interpretation statement completely excluded the Falluja cluster.
Given that the attribution of causes for violent deaths has been one of the most prominent battlegrounds between defenders and opponents of this study, I can't see how this is a minor flaw, Tim. Moreover, when one of the study's own authors nearly doubles the death estimate for what the study purports to be the number one cause of death for its most controversial and quoted conclusion (the 100,000 excess death estimate), well that isn't minor either.
However, what really sets Garfield's mistake apart is the fact that his error doesn't misrepresent the study's key Interpretation statement concerning attribution of deaths, it actually agrees with the statement.
Thomas:
I don't think linking numbers of insurgent attacks to an underreporting of casualties is a sound argument. The volume of attacks themselves is not a good indicator of insurgent success in terms of casualties inflicted, as evidenced by U.S. military fatalities in particular. For many months now, the coalition has reported daily attacks on its forces, numbering as low as the teens per day, to as high as the 50's.
Compare that number of daily attacks to the overall U.S. death toll in combat, which is 1,160 as of today, and I think you'll see that there are days in Iraq where we have several dozen attacks on U.S. forces, with no U.S soldiers killed, or perhaps one or two soldiers killed. I realize that insurgent attacks also directly target civilians on occasion, but I think the vast majority of insurgent attacks are clearly directed at the coalition forces and Iraqi security forces. The daily Associated Press roundup of violence in Iraq seems to be a fairly comprehensive summary of violent incidents that result in deaths.
Shirin:
"you simply cannot bomb population centers without killing anywhere from a few to a great many non-combatants."
I agree with you to some extent, however I don't see that the Lancet study, in and of itself, is a reliable tool for giving us the answer to the above. Has coalition bombing killed " a few" non-combatants, or has it killed " a great many non-combatants?"
As I mentioned in my last post, the study data extrapolates to approximately 17,000 bombing deaths without Falluja's inclusion. We are relying on 6 deaths from 3 bombs to project this 17,000 figure. As I've mentioned before, if we do have 17,000 bombing deaths (actually, prior to the Garfield correction, I was prefacing this argument with " if we do have 30,000 bombing deaths"), then we must have multiple neighbourhoods in Iraq that have suffered double digit deaths, even triple digit deaths from bombing, even though the Lancet study didn't pick up any such neighbourhoods outside of Falluja.
As a result, a second identical study with the same methodology and sample size could give us 2 such clusters, which the authors couldn't exclude on the basis of being outliers (if they count say 10-15 deaths each). This would result in an excess death extrapolation from bombing alone of 50-70,000, without Falluja. If they were identical surveys, which one would you adopt as accurate, and why? As Heiko pointed out in an earlier thread, cluster studies in a combat environment may be capable of providing more reliable data and extrapolations on some death subsets (such as homicides) than they can on death subsets like bombing.
Further evidence of this comes from a point I made at Chicago Boyz concerning the civilian death toll from the NATO bombing of Serbia in 1999.
Estimates of the number of bombs dropped on Serbia range from 20,000 to 25,000, during a campaign that saw bombing occur for 78 straight days. Human Rights Watch accused NATO of killing approximately 500 citizens during the campaign. Canadian law professor and ant-war activist Michael Mandel claimed 500 to 1,800 citizens killed when making his attempted indictment of NATO political and military leaders before the International Court of Justice. The Serbian government claimed as many as 2,000 civilian deaths (which would clearly be the highest likely toll, given that there would be absolutely nothing in it for Serbia to underestimate the number in their condemnation of the bombing).
As I said at Chicago Boyz, I'm not suggesting that the bombing campaign in Serbia is a duplicate of the bombing activity in Iraq. However, the bombing of Serbia cannot be written off as inconsequential or light in scale. It illustrates, in my view, the danger in automatically assuming the Lancet extrapolations for bombing are accurate, or even underestimates as some claim.
I believe an actual count of the violent deaths resulting from regime change is the method that will result in the most accurate and reliable figure for at least the violent death toll component of the excess death figure. Obviously the security situation in Iraq does not allow for such a count at this time, but one is nearing completion in Bosnia today, by Bosnian war crimes investigator Mirsad Torkaca. Torkaca is seeking to verify the identity of each victim and circumstances of their death..
While a count will never uncover every victim, it can provide a definitive minimum casualty figure that cannot be reasonably refuted, and, if properly undertaken, can also lay claim to being close to the actual overall number of violent deaths. The situation in Iraq will some day allow a similar exercise. Whether one will occur or not is anyone's guess.
.
TallDave,
You have no conception of maths as applied to statistics. Go read some books on the subject matter. Its the level of ignorance as exemplified by your comments that really upsets me. A 95% CI indicates that there is a 5% chance that the real figure lies outside the CI. A %5 chance is very small in terms of probabilty. Wilkling to bet your life on it?
The Lancet Study: A Closer Look
Well, with all of the argument surrounding it, I decided to have a look at the Lancet Study myself. Here is what I concluded after reading it:
The Lancet Study is based on a survey conducted in 33 "clusters" throughout Iraq in which 988 households containing 7868 residents served as the basis for the study's findings. Each cluster contained 30 households and an average of 238.4 residents. One of the clusters just happened to end up in Falluja, where an overwhelmingly disproportionate number of deaths had occurred. For that reason it was not considered in the calculations that lead to the famous 100,000 figure because, according to the authors, "the Falluja cluster is an obvious outlier and might not belong with the others."
So for its findings the study relied on 32 clusters with (if we take the averages) 960 households and around 7630 residents. In these 32 clusters, 46 deaths were reported in a 14.6 month (or 442 day) period from January 1, 2002 to March 18, 2003 leading up to the war, which yields an average of 3.15 deaths per month. This period is referred to as the "pre-invasion" period by the study's authors. In the same clusters, 89 deaths (21 of which were violent) were reported in the 17.8 month (or 548 day) period from March 19, 2003 to September 16, 2004, which yields an average of 5 deaths per month. This period, which includes the major combat operations phase of the Iraq conflict, is referred to as the "post-invasion" period by the study's authors.
The numbers above serve as the fundamental basis for the Lancet Study's estimate of "100,000 excess deaths" in Iraq, which the study claims were mainly attributable "to coalition forces." Here is essentially how they used the numbers they provided to get that result:
First, because the two periods under consideration are different (14.6 months or "110,538 person-months of residency" versus 17.8 months or "138,439 person months of residency"), one has to adjust for time. One of the ways this can be done is by simply taking the average number of recorded deaths per month "pre-invasion" (3.15) and multiplying it by 17.8 months. That gives us right about 56 deaths. Remember for the "post-invasion" period, we had 89 deaths in 17.8 months. So now, after adjusting for time, we still have 33 "additional" deaths (89 minus 56) in the "post-invasion" period in our 32 clusters.
You may be scratching your head at this point and wondering...how did they get 100,000+ "additional" Iraqi deaths from 33? Well, this is how:
First you take the estimated population of Iraq: The survey uses 24.4 million. You divide the total population of Iraq by the overall number of people included in the 32 clusters surveyed (7630) for a result of 3198. Then you take your 33 "additional" deaths and multiply them by 3198...voila...there you have your 100,000+ "additional" Iraqi deaths. Of course there are numerous different ways one can plug in the numbers that will lead to slight variations, but the end result will always be right around 100,000.
For the record, I attempted to get a hold of the exact methods and computations carried out to reach the 98,000 figure cited by the study for the 32 clusters that served as the basis for the study's 100,000 estimate from the study's leader, L. Roberts. His reply:
The study used then another equation to calculate its "confidence interval" of 8,000 to 194,000 "additional" deaths.
That's right ladies and gentleman, this entire study is based on a few dozen deaths. To be more precise the entire study and its 100,000 "additional" death estimate is based on 89 deaths in a 17.8 month period during and after the war versus 46 deaths in a 14.6 month period before the war in a 32 cluster sample group. In other words, if even one cluster was disproportionately affected by death (or a lack thereof) either before, during or after the war, the results of the study would be dramatically off base. If even just a few deaths were incorrectly recorded or invented, it also would have dramatically changed the study's results. If even just one or two of the surveyed families was the unfortunate victim of a particularly violent bombing incident or terror attack, the entire survey could be way off base and fatally skewed. In fact, a change of just 10 deaths for any reason would have the effect of throwing this survey off by around plus or minus 30,000 deaths. The conclusions built on this data are comparable to houses built on quicksand because all of the results and assumptions rely on a sample group that is simply too small to give us any reliable or relevant data whatsoever.
How could anyone claim that this sample group was large enough to be taken seriously when debating an issue of such enormous gravity? How could anyone rely on a survey that makes such heavy claims of 100,000 "additional" dead in Iraq when it is based on a difference of around 33 deaths (time adjusted)? This is explainable only if we consider the worldview and ideology of those who have exhibited a burning desire to believe the results, whether true or not.
IBC vs. Lancet
The Iraqi Body Count is another left-leaning project aimed at recording the number of deaths in Iraq during and after major combat operations. Unlike the Lancet study, which relied on a mathematical estimate derived from a small sample group, the Iraqi Body Count is dependent primarily on the media for its casualty tally, but has also included information from the Iraqi government and in a few cases from NGOs.
Even assuming that all of the media sources and all of the other data sources the IBC relies upon are trustworthy and reliable (they include Al-Jazeera), the maximum casualty count according to the site is just over 19,500. That would mean, if we believe the Lancet study's estimate of 100,000 "post invasion" deaths to be true, that over 80,000 Iraqi deaths, caused "mainly by the coalition" have gone completely unnoticed by the international media, Iraqi authorities and NGOs. That is over four-fifths or more than 80% of the deaths the Lancet study estimates have occurred.
Infant Mortality and the Lancet Study
The study's "pre-invasion" infant mortality rate was also derived by combining a tiny sample group with a flawed assumption. Here again, the study relies on a minute number of actual recorded infant deaths to reach its reported "pre-invasion" infant mortality rate of 29 deaths per every 1000 livebirths for all Iraq. To be more precise, 8 actual infant deaths serve as the basis for this statistic that is purported to accurately reflect the infant mortality rate in a nation of over 24 million.
Here is how the authors came up with the figure: The study recorded 275 births and 8 infant deaths in the 14.6 months before the war. Using those figures, the authors came up with the 29 infant deaths per 1000 live births infant mortality figure which they then proceeded to accept as an accurate estimate for the infant mortality rate for the entire nation.
The study defends the 29 figure by noting that "the preconflict infant mortality rate (29 deaths per 1000 livebirths) we recorded is similar to estimates from neighbouring countries." In so doing it blatantly ignores a number of key facts. First, other "neighbouring countries" were not subject to sanctions or Saddam Hussein's reign of tyranny in the same period. Secondly, a UNICEF study conducted in Iraq for the period from 1994 to 1999 came up with an infant mortality figure of 108 per 1000 livebirths. That would mean if the UNICEF numbers are accurate, the infant mortality rate would have had to drop by over three-and-a-half times within 3 years in an Iraq under sanctions and Hussein's rule.
The "58-Fold" Canard
Some particularly outspoken critics of the Iraq war have pointed with horror and outrage to the Lancet study's finding that: "Violence-specific mortality rate went up 58-fold during the period after invasion."
One problem with the 58 fold figure is that it is derived using all 33 clusters including the data gathered in Falluja. The authors came to the figure using this data: "After the invasion, 142 deaths (73 of them violent) were reported in 138,439 person-months of residency. Before the invasion, respondent households reported 46 deaths (1 of them violent) during 110,538 person-months of residency."
So again, you simply adjust for time by dividing 110,538 by 138,439. That gives you .79846. Then you multiply that by the ratio of the number of violent deaths "post-invasion" to the number of violent deaths "pre-invasion." That figure is 73. 73 multiplied by .79846 gives us 58.3.
But get this, if even one more death had been reported for the "pre-invasion" period, it would have cut the 58-fold" figure in half to 29! In fact, the "58-fold" figure depends on just one reported death in the "pre-invasion" period.
Here again, the data is fatally flawed and rendered useless by the fact that the sample group is simply far, far too small to produce meaningful results.
Falluja Not a Risk???
In the "Methods" section of the Lancet study, the authors changed the locations of some of the clusters they planned to visit. The reason, according to them was the following:
In other words, some of the clusters were moved in order to (in the words of the authors) "lessen risks to investigators." Yet the same study authors who were so worried about risks had no problem allowing their "investigators" to drive right into Falluja in September 2004 and conduct their survey while the city was still under the control of violent insurgents and subject to a Coalition siege featuring almost daily bombardment from the air and ground. Let's not forget about the kidnapping victims who ended up at Al-Qaeda headquarters in Falluja during this same period.
Now, how, exactly, is that consistent? If the study leaders were as concerned about safety as they claim they were, why would they allow their investigators to drive right into Falluja at a time like that? What could have possibly been their motivation? Does that not seem just the slightest bit contradictory? All things considered, one has to wonder about the logic that went into selecting the other 32 clusters chosen as well.
Saddam's Long Rule of Violence Not Accurately Reflected by Study
And how can we say that the 14.6 months preceding the war are an accurate reflection/representation of Saddam Hussein's 24 year regime and the violence wrought by that regime over the years? The year and a half before the war happened to be one of the less violent periods of Hussein's regime. But is this period an accurate reflection of the average number of deaths suffered at the hands of the Baathist tyranny over the years? Of course not! And there is no guarantee that Saddam and his vicious sons would not have carried out future massacres had they been left in power. There history (they are responsible for killing hundreds of thousands) certainly would not have made such an event unlikely.
Again, for the reasons given above, the Lancet study is fatally flawed. Above all, the data derived from the study is useless because the group sampled is too small and too subject to radical change through small anomalies.
One last point: Tim has mentioned the death certificate issue. He has pointed out that in the 78 households where documentation of death was requested (or in the words of the authors, "where confirmations were attempted", in 61 cases (81% of the time) it was provided. However, let's keep in mind that the study recorded a total of 231 deaths. That would indicate that, unless all 231 deaths in fact occurred in just 78 households, that the study did not even request proof of death in many cases of reported death. Why don't you deal with these issues Tim, dsguarded and the other Lancet defenders???
---Ray D.
TallDave
Your north Korea/south Korea analysis omits a couple of important points.
1. The area that is now North Korea has ALWAYS been poorer and more backward than the South. At least some of the current disparity in income is due to that.
2. South Koreans have only been enjoying something approaching western cncept of "democracy" for about the past decade.
People who want to use South Korea as an example of how a foreign invasion can lead to democracy and prosperity need to take a look at the history of Sotuh Korea from 1952 to around 1980.
They can start by googling for "Kwangju Massacre".
Syngman Rhee's principal point of differentiation with Kim Il Sung was that he had better economic advisers.
"The convention is that 95% confidence constitutes proof"
No, TallDave, that is NOT the "convention", and no matter how many times you repeat it it will never be the "convention".
Ray, all your objections have been dealt with here in the posts and comments again and again and again. The most egregious example is your repeated claims that the IBC contradicts the Lancet study. It does not. They measure different things. This has been explained to you several times. Why do you keep ignoring this?
Ray--
When you write sentences like :"That's right ladies and gentleman, this entire study is based on a few dozen deaths. To be more precise the entire study and its 100,000 "additional" death estimate is based on 89 deaths in a 17.8 month period during and after the war versus 46 deaths in a 14.6 month period before the war in a 32 cluster sample group."
People are either going to assume you don't know the underlying CLT assumptions or you do and are just blowing smoke to discredit the survey in people's mind. Either way there would be no reason to respond since either ignorance or hackdom wouldn't allow you to see the answer.
Hi Tim,
I never said the IBC contradicts the Lancet study. Did you read my post? I clearly state that the two measure different things. My main point is that the IBC's data casts serious doubts and raises serious questions on the Lancet's estimate of 98,000 "additional" deaths. (If, in fact you consider the 98,000 figure to be an accurate, reliable estimate) Here, again, is what I wrote:
You can see that I clearly draw a distinction between Lancet and IBC and what they measure. Lancet also counts military deaths whereas the IBC only counts civilians. It is clear that they arrive at their figures using different methods entirely. However, the massive difference in their estimates of the total casualty count, again, raises very serious questions about Lancet that you simply can not answer and that would indicate the study is seriously flawed.
@ Rob
So Rob, is the study's 98,000 figure based on something other than "89 deaths in a 17.8 month period during and after the war versus 46 deaths in a 14.6 month period before the war in a 32 cluster sample group." Really? Please fill me in.
Please point out the error here or with any single point in my argument.
---Ray D.
Ray you really going to flip when you find out how they determine whether a TV show is a hit or not.
Ray, the Lancet study includes the deaths from increases in disease, accidents and murder. This is clear if you read the study and I have pointed this to you before.
Actually, I think the IBC and Lancet studies are quite compatible, despite measuring different things and both being horribly biased and using unreliable methods. The Lancet study found at least 8,000 excess deaths (at 95% confidence)vs. pre-war, the IBC found around 20,000 civilian deaths from military action. Sounds pretty close to me!
Ian: 2. South Koreans have only been enjoying something approaching western cncept of "democracy" for about the past decade
And the N Koreans haven't. Freedom and democracy matter, however one arrives at them.
Shirin,
From a very basic statistics book:
When a study reaches "statistical significance," it means the researchers are very confident (usually 95% confident, a widely used convention) that the change observed in that study was not a fluke of chance.
So if you want to play stupid semantic games about whether 95% is proof or "statistical significance," fine. It just shows the weakness of your arguments, and the point stands either way.
This study's only statistically significant finding is that there were at least 8,000 excess deaths.
For the record I would like to correct an error in my long post: There were 188 total deaths (46 "pre-invasion" and 142 "post-invasion" recorded if we include Falluja, not 231. I mistakenly reached 231 by added 142 and 89 when 89 actually made up part of the 142 and the 46 is the correct pre-war number. Just want to be clear.
---Ray D.
@ Talldave,
Actually, a great number of the civilian deaths recorded by IBC are victims of suicide bombers and terrorists. Just look at the database. They are basically saying the US is responsible every time some whacko straps on a bomb and detonates it next to innocents.
---Ray D.
Ray,
Oh yes, I'm quite aware. I've been following the IBC for some time. I love their disclaimer for blaming the U.S. for casualties from enemy suicide bombers, which is something like: "We place responsibility for these actions solely on the shoulders of those who acted without UN authorization."
@ Rob:
That's certainly a cynical way of viewing it especially since American television audiences have not been subject to concentrated genocide or bombings.
---Ray D.
I just wanted to point out that the infant mortality rate recorded by UNICEF in autonomous Kurdish areas not impacted by sanctions for 1994 to 1999 was 59 per 1000. So much for the argument that oil-for-food improvements account for the huge improvement you claim occurred (108 to 29 in three or four years.)
Actually, a great number of the civilian deaths recorded by IBC are victims of suicide bombers and terrorists.
Quite correct. Tallying deaths attributable to one source or the other yields some interesting results (the insurgency has killed a bit more than the coalition where it is possible to tell.) The IBC also includes a few undifferentiated block entries in the 1000's, via morgue and hospital data.
Defenders of the CLT assumptions seem not to care very much about the treatment of outliers or the implications of the very wide confidence interval. Or reporting the CI appropriate to the Fallujah cluster (alone or study-included.) Is the consensus that the CI width isn't important because it doesn't include zero?
@ Tim,
The study itself says that violence "not accidents or disease" accounted for most of the excess deaths and air strikes from coalition forces accounted for most violent deaths. So again Tim, you want us to believe that tens of thousands of deaths caused mainly by violence mainly attributable to air strikes have gone completely unnoticed by the world press? Let me point out that IBC does not even claim to count "additional" deaths as does Lancet.
---Ray D.
Oops let me address 1) as well
1. The area that is now North Korea has ALWAYS been poorer and more backward than the South. At least some of the current disparity in income is due to that.
Not true.
"While the economic advantage led the North Korean leadership to feel confident enough to invade the South in 1950, it could not sustain the lead: North Korea started to lag behind the fast growing South from the late 1960s, and then suffered a tragic decline in living standards in the 1990s."
http://www.answers.com/topic/history-of-north-korea
S Korea may not have had true democracy until recently, but post-war they always had more economic freedom than the North, the consequences of which are easy to see.
TallDave,
You appear to be confusing insisting on a reasonable level of clarity and precision in language with "playing stupid semantic games".
If one wants to communicate clearly and effectively, it is important to choose the appropriate words and terms. Proof and statistical significance are two very different concepts, and the two terms are not synonymous. Therefore, they are not interchangeable. That is why the words proof, prove or variations thereof are generally not applied to the results of statistical studies.
And you are still dead wrong either way. That there were at least 8,000 excess deaths is not even remotely the only statistically significant finding of the study in question.
Could there be 30,000 dead from air strikes with only half of these recorded by the IBC? I don't see why not. Why do you think the press will be able to record all of them?
TallDave blithered, No, I had a point -- freedom matters, and S Korea is the proof. Sorry if that was too hard for you to decpiher.
Of course freedom is better for GDP. But Iraq isn't free right now either. Furthermore, saying "Iraq is better off free, ceteris parabis" is a nice warm fuzzy statement, but it says very little about prospective costs and benefits of the Iraq invasion, especially when adjusted for risk.
Tim:
Further to my original post in this thread, is there any Falluja-inclusive interpretation of the study data which would account for Garfield's explanation that his 30,000 death extrapolation from coalition bombing was the result of his mistaken inclusion of Falluja?
I don't see any.
liberal blathered:
Of course freedom is better for GDP
All right! We've finally gotten the Left to admit that! High-fives all around!
But Iraq isn't free right now either.
Hmmmm, elections, right to assemble, free press, free speech, free economic policies... things aren't perfect, but they're the freest Arab nation around by a long shot.
"Iraq is better off free, ceteris parabis" is a nice warm fuzzy statement but it says very little about prospective costs and benefits of the Iraq invasion, especially when adjusted for risk.
You are indeed a true master of the obvious. That was my point in bringing up the S Korea / N Korea dichotomy: to provide an example of what the benefits might be.
And if postwar Iraq ends up anything like S Korea, historians will be giggling at how much time and energy was spent debating whether invading Iraq was a good idea.
Shirin,
If one wants to communicate clearly and effectively, it is important to choose the appropriate words and terms. Proof and statistical significance are two very different concepts, and the two terms are not synonymous. Therefore, they are not interchangeable.
That's just stupid. They have very similar meanings and are often used interchangeably. If you find the slightest hint of ambiguity or informal speech incomprehensible that's your problem; I'm not going to waste my time debating the meaning of "is." But as I said, even accepting your point doesn't change anything.
That there were at least 8,000 excess deaths is not even remotely the only statistically significant finding of the study in question.
As usual, you tell me I'm wrong but can provide no evidence in support. Good Lord, I'm debating a fourth-grader.
@ Tim
You write:
" Could there be 30,000 dead from air strikes with only half of these recorded by the IBC? I don't see why not.
Tim how can you write something like this and expect us to take you seriously? The IBC has NOT recorded 15,000 dead from airstrikes. In fact, an enormous number of dead recorded by the IBC database are those innocents who have been murdered in car bombings, suicide bombings, intimidation shootings by terrorists, etc. You constantly criticize others for playing fast and loose with facts and then you try to pull this on us. In fact, (if you actually look at the IBC database), only a fraction of the IBC count is attributable to airstrikes. That fact only widens the gap between the IBC and the Lancet figures.
---Ray D.
For the record: The IBC database.
Pascal:
the implications of the very wide confidence interval.
Thank you, I'm glad to see there are some sensible people here.
Ray D., does the term "lower bound" mean anything to you?
If so, then you know why your comments about IBC versus Lancet do not merit a response. If not, you need to learn.
So Kevin,
Let's cut to the chase:
Do you believe that there have been 100,000 "additional" civilian deaths in Iraq because of the war as the Lancet study estimates? If so, how has the media missed so many violent deaths caused by airstrikes? And what does "lower bound" have to do with any of this?
Secondly, Tim is the one claiming that the media and IBC could have missed half of "30,000" airstrike related deaths.
---Ray D.
Kevin, does the term "very wide confidence interval" mean anything to you? How about "useful result?" If not, you need to learn.
Ray,
I believe Kevin is referring to the fact the number still falls above the lower bound of the confidence interval, which up until now they've been loathe to acknowledge the existence of.
LOL You gotta love how the Lancet study supporters want to claim the 100,000 number is significant despite the huge confidence interval, but when it's challenged they want to say "but we have a huge confidence interval, so you're within the lower bound! Nyah nyah nyah!"
Tim is the one claiming that the media and IBC could have missed half of "30,000" airstrike related deaths.
Airstrikes account for a rather small fraction of the IBC figures (about 1,000.)
TallDave,
No, the terms proof and statistically significant do NOT have very similar meanings. They represent two very different concepts. There may be some applications of statistics in which the word proof can correctly be used (I hope the scientists here will enlighten me if that is the case), but I cannot recall ever seeing the word proof used in a report of a statistical study in relation to its results. I also cannot recall anyone with even a basic understanding of stats trying to use the word proof interchangeably with the term statistically significant.
"If you find the slightest hint of ambiguity or informal speech incomprehensible that's your problem;"
TallDave, this is not a case of informal speech or of ambiguity in language. Those are both easy to deal with among people of good will. This is a case of someone trying to sound as if he knows more than he does about a subject, and repeatedly displaying his lack of knowledge. Your insistence that a 95% CI constitutes proof of something, and your subsequent insistence that statistical significance is synonymous with proof are just two of the examples of a lack of understanding of statistics that you have placed in abundance on these pages. Several people have pointed out your errors to you, but you continue to repeat the same ones and add new ones to the list daily.
"I'm not going to waste my time debating the meaning of "is." "
TallDave, this is not about the meaning of "is". This is about the meaning of two different, non-interchangeable words that describe two very different concepts, one of which is applicable to the results of a statistical study, the other of which is not.
"you tell me I'm wrong but can provide no evidence in support."
Others here have recommended that you take a course in elementary statistics. That will provide you with the knowledge you need to see the statistically significant findings of the study in question. Additionally, there is plenty of information here from a variety of very highly qualified individual. Anyone who is able to absorb and understand this information can see what are the statistically significant findings of the study. I do not see any point in taking more of my time, and filling more of the space Tim so generously provides to repeat information that is already here in abundance.
@ Tim,
You yourself have stated repeatedly that violence caused most of the "additional" deaths estimated. Again, that would mean a minimum of 50,000 violent deaths, of which well over 30,000 have gone completely unrecorded by the world media or Iraqi authorities.
As far as I can conclude from looking at the numbers, the study implies that something like 60 to 70% of "additional" deaths were due to violence, which would mean 60,000-70,000 estimated "additional" deaths were due to violence. Again, if we are to believe you, the vast majority of those violent deaths have been missed entirely by the world media and Iraqi authorities. And all of this is based on 46 actual pre war deaths and 89 (or 146) post "invasion" deaths. The whole thing is based on a sample group that is simply too small and too subject to anomalies.
---Ray D.
By way of contrast the IBC lists over 1600 dead from car or suicide bomb.
Sorry, it is 142 if we count Falluja, not 146.
Pascal, do you have a link to that breakdown of the IBC #s?
you need to see the statistically significant findings of the study in question.
The statistically significant findings of the study are defined by the confidence interval. The confidence interval in this case is bounded by 8,000 and 194,000 (ex Falluja.) Inclusive of Falluja the confidence interval is many times wider, and its lower bound is in all likelihood below zero. The 'statistical likelihood' (again ex falluja) of a number between 88-108k deaths is only 16%.
Ray D, I have a handy dandy spreadsheet I can send over if you shoot me your email.
Ray D.,
I believe that the Lancet study estimate was the correct figure for them to report given the data, which was gathered in the best manner that could reasonably have been expected. If you know of a better methodology tell us about it (in less than 1,000 words, please).
The media misses many violent deaths caused by aerial weaponry (and by other means) because few reporters are foolhardy enough to wander around a district which is under attack counting bodies.
The figures provided by IBC are, at best, a lower bound for the true number of deaths, which is of course unknown.
My previous comment about reporters is woefully incomplete - there are many, many reasons why the media probably fail to pick up more than a fraction of the violence in Iraq. But that should be obvious to any thoughtful person and the other sort is not worth bothering with.
Shirin,
Sigh... they are used interchangeably all the time. Yes, a 95% confidence interval generally constitutes proof. This is very, very, very basic statistics. Here's a half-dozen examples.
The results of radiocarbon measurements at Arizona, Oxford and Zurich yield a calibrated calendar age range with at least 95 confidence for the linen of the Shroud of Turin of AD 1260 - 1390 (rounded down/up to nearest 10 yr). These results therefore provide conclusive evidence that the linen of the Shroud of Turin is mediaeval.
http://shroud.typepad.com/topics/
Empirical Proof for Presenting Screen Captures in Software Documentation ...There was no statistically significant difference for correctly locating
http://www.techcomm-online.org/issues/v51n2/full/0649.html
Proof in statistics is about induction: reasoning from effects back to causes. In logic this is the source of many fallacies, but is essential in real life. The best one can say in a statistical proof is that what we see is unlikely to have happened by chance. Although you can never be entirely certain of anything, you can at least know how likely you are to be wrong. A 5% significance means that you are wrong one time in twenty - good enough?
http://www.meandeviation.com/tutorials/stats/notes/outline.html
Still, many parapsychologists assert that the current body of experimental evidence provides statistically significant proof of psi. .
http://www.google.com/search?q=%22statistically+significant%22+proof&hl…
http://www.tabexperts.com/ToxicExposuresII.htm
Litigation in matters of toxic exposures usually hinges on proof of a causal link... The investigation determines whether a statistically significant ...
https://faculty.washington.edu/petej/Hamrick%20-%20Ergo%20Savings.pdf
... Proof that Ergonomics Works:. Combined Results of Over 100 Independent ...assessment scores were statistically significant. ...
I'm not wasting my time with someone who provides no arguemnts besides ridiculous semantic quibbles and unsupported "refutations." If you want to debate on theat level, I refer you to the nearest elementary school.
If you know of a better methodology tell us about it
Easy. More samples, and include all the provinces.
My previous comment about reporters is woefully incomplete - there are many, many reasons why the media probably fail to pick up more than a fraction of the violence in Iraq. But that should be obvious to any thoughtful person and the other sort is not worth bothering with.
Any thoughtful person would realize there are local media in Iraq that would pick these things up. Even in Iraq, you can't blow things up without people noticing.
TallDave wrote, Hmmmm, elections, right to assemble, free press, free speech, free economic policies... things aren't perfect, but they're the freest Arab nation around by a long shot.
Absolute nonsense. The freest Arab country, even with the Syrian occupation, is clearly Lebanon.
The notion that any of those freedoms you listed apply to Iraq, in the middle of an occupation and rather bloody insurgancy, is laughable.
" 'Iraq is better off free, ceteris parabis' is a nice warm fuzzy statement but it says very little about prospective costs and benefits of the Iraq invasion, especially when adjusted for risk." You are indeed a true master of the obvious.
Hardly. It doesn't appear obvious to you, because you don't seem to be aware of the costs and risks (e.g., full-blown civil war).
That was my point in bringing up the S Korea / N Korea dichotomy: to provide an example of what the benefits might be. But there were many paths by which Iraqi freedom could have become more likely, and the future prospects for Iraqi freedom are hardly on solid ground at this point.
@ Kevin:
The main thing I would improve about the study is the size of the sample group, which I would make much, much larger. The small size of the group is the factor that renders the rest of the study, its methodology and its results virtually meaningless because even tiny changes in data could have radically altered the results.
Is a larger study possible? It may be difficult but I think it is. I think a respected, independent body should conduct it. And why not have several large studies? We would certainly get a better idea of what has really been going on in Iraq.
"The media misses many violent deaths caused by aerial weaponry (and by other means) because few reporters are foolhardy enough to wander around a district which is under attack counting bodies.
Actually that is false. A lot of reporters are foolhardy enough which is why quite a few have died to date. Even if no journalists are present at the site of a bombing, what is keeping Iraqis and others who witness the bombing from telling someone in the international media? I'm not saying the media has been able to report all violent deaths, maybe they've even missed a significant number, but tens of thousands? That argument is just ridiculous Kevin.
---Ray D.
TallDave, you are a vertitable compendium of standard-issue talking points, myths, misconceptions and notions, every one of them conceived in the absence of even the minimum knowledge of Iraq.
liberal:
You're joking, right? It's illegal even to assemble in Lebanon and the gov't is currently appointed by Syria.
Yes, those freedoms do exist in Iraq even in the midst of the insurgency, and there is no insurgency in large parts of the country. If you choose to simply ignore reality I guess I don't know how I can argue with that.
But there were many paths by which Iraqi freedom could have become more likely
It's easy to speculate that there might be ways the Hussein regime could have voluntarily granted Iraqis freedom and democracy, but harder to come up with any reasonable scenarios that don't involve military force.
Shirin,
That's a fairly typical post by you.
When you understand Socratic debate and want to have a real conversation, let me know.
I agree Ray, as I've been saying let's have a real study, or studies, with a reasonabe consensus beforehand that the parameters are fair.
TallDave wrote, It's illegal even to assemble in Lebanon and the gov't is currently appointed by Syria.
You mean to tell me Lebanon doesn't have elections?
Perhaps the elections in Lebanon aren't 100% free and fair. Then again, neither are elections in the US (where the winner of the popular vote for President wasn't seated in 2000). Nor are they in Iraq, where (a) the Sunnis weren't represented in the vote, and (b) the vote itself was very possibly tampered with to deflate the majority of the winning Shiite slate.
Yes, those freedoms do exist in Iraq even in the midst of the insurgency...
There's no freedom to assemble when you might be shot by the occupying authority if you don't make it home past curfew.
Continue hallucinating if you will. I prefer to remain in the reality-based community.
TallDave wrote, When you understand Socratic debate and want to have a real conversation, let me know.
LOL!
Mirror time.
liberal,
OK, now you've gone off the deep end. The law says the electoral college chooses the US President, not the popular vote. Are US elections perfect? No, for instace there is a growing consensus that JFK was not legitimately elected due to massive fraud in IL and TX. But they're very good in all but the closest elections, and comparing them to Lebanon's just makes you look silly.
Sorry, with that comment you've made the "not worth talking to" list. Enjoy your "reality."
TallDave continues to impeach his own knowledge of stats and highlight his ignorance of peer reviewed research everytime he says something like:
The Lancet study found at least 8,000 excess deaths (at 95% confidence)
1) 8000 represents the 97.5% lower bound of the distribution; the 95% lower bound is somewhat higher.
2) These types of figures are always reported as mean +/- deviation. That is the convention in peer review as well as the main stream media.
TallDave's criticism would seem less politically motivated if he were on a systemwide crusade to change this convention, rather than simply attacking the Lancet paper. I look forward to reading his letter to the NYT arguing that GWB's approval rating of 48% +/-4% should be reported as "at least 44%".
Could there be 30,000 dead from air strikes with only half of these recorded by the IBC? I don't see why not.
Tim,
The IBC breaks down civilian casualties by "weapon" employed. I count somewhere around 1200 deaths by "aerial assault," "air strike" and variations thereupon. I am generously assuming that all such assaults are as described rather than misattributed mortar attacks or misdirected anti-aircraft rounds.
I suggest you re-read the IBC database before making ridiculous statements like the one above. Nowhere near 15,000 deaths by air strike have been recorded by the IBC.
motivated if he were on a systemwide crusade to change this convention,
This is completely idiotic. TallDave's comment concerns the absurd width of the confidence interval, which is itself a function of the small sample size. The inclusion of the Falluja outlier would magnify this CI many times. You should ask yourselves why the authors included a CI for the "risk of death" net of Falluja but no similar figure for the raw excess death CI.
The media misses many violent deaths caused by aerial weaponry
"The media" is hardly the only source for IBC variables. As noted, they also admit thousands of undifferentiated morgue and hospital reports. Media sources for the IBC include both embedded and unembedded arabophone sources, including Al Jazeera.
"The inclusion of the Falluja outlier would magnify this CI many times."
Fascinatingly, those who continue to point this out never bother to include the fact that it would also increase the midpoint value by a factor of about three.
Fascinatingly...
More fascinating still is the idea that almost 1/3 of the Fallujah cluster should be described as "excess dead" according to your own 200,000 figure. BUt critics are happy to point out this number, because it can't possibly be true. Even the study's authors have disowned this figure.
The reason the Fallujah outlier was discarded for the purposes of the central estimate was simply to preserve the credibility of the study. It had nothing to do with science, or they wouldn't have bothered to calculate much less report the IFC risk CI in the editorial/summary.
"almost 1/3 of the Fallujah cluster should be described as "excess dead" according to your own 200,000 figure."
Sorry, but I am having problems making sense out of this statement, so I can't tell whether to concede the point or to contest it. Would you please clarify what you mean here?
"The reason the Fallujah outlier was discarded for the purposes of the central estimate was simply to preserve the credibility of the study."
How exactly do you know this as a fact? The thing is, it is conventional practice to discard outliers such as the Falluja case, so without convincing information to the contrary, the most likely explanation is that they were simply following conventional practice.
PS Pascal, it is also not unusual at all to include outliers in the discussion of a study, even though they are not normally included in the results.
How exactly do you know this as a fact?
I know this as a fact because Les Roberts has admitted as much in interviews on the topic [and also in an email to me: he wrote that "not giving a number or range that could have been discredited and undermined the validity of the study seemed then and now like better service to the Iraqis." He also disowns the 200k figure entirely. Perhaps Roberts intended for it never to be published, which invites the question of why the falluja-inclusive risk WAS.]
The 200k figure would make up nearly 1/3 of the 739,000 people represented within the Falluja cluster. Do you really think 1 in 3 civilians in that province alone were killed?
What is "normally done" in highly unusual stats-normative studies of war casualties is unknowable. What should be done is the point of this discussion.
Again, Shirin, a confidence interval for "risk of death" was expressed in the editorial summary NET of Fallujah, why was no similar figure offered for the "excess death" variable? Good luck finding one.
Dave
I find the statments in that wikipedia article that living standards in north Korewa were higher than in south Korea in the 1950's and 1960's highly surprising.
Take a look at virtually any account of pre-WWII Korean history. virtually all the major political, cultural and economic centres of Korea are located in the South. When the Japanese colonised Korea they put most of their efforts into developing the southern areas which were closer to Japan.
Northern Korea, since at least the fall of the Koguryo dynasty, has been an economically underdeveloped backwater which has a much harsher climate than the south and was exposed to military threats from China and other hostile powers.
The Korean War was fought primarily on the territory of South Korea, this may have caused enough damage to temporarily push southern living standards below those in the north but if so it was a historical fluke.
It's nice to think that democracy and free markets automatically lead to higher economic growth and, in general it may be true. But there are plenty of counter-examples such as Pinochet's Chile or Lee Kwan Yu's Singapore on the one hand and pre-reform India on the other.
Pascal, it is conventional practice to exclude outliers exactly because they skew the results and therefore undermine the validity of the study, so it sounds as if they were indeed following conventional practice, and for the conventional reasons.
"What is "normally done" in highly unusual stats-normative studies of war casualties is unknowable."
What is normally done in studies that estimate population values based on sample data is hardly unknowable. What is done in epidemiological studies is also hardly unknowable. It is a real stretch to argue that this study is somehow so different from statistical studies in general and epidemiological studies in particular in order to make your point.
Pascal, the IBC does not report "weapon" for all entries and some entries are mixtures of various weapons. You can identify at least a 1,000 dead from air strikes, but that is clealy a lower bound.
Shirin, the heterogeneity of the distribution implied by the study results directly informs the reliability of its confidence measures and therefore its epistemic value. You seem unwilling to acknowledge this point altogether. Other informed commentators (dsquared, eg) have accepted that the norm-based confidence measures hold little value here for the same reason.
Epidemiological surveys of war-related casualties are quite rare. "Conventional practice" means nothing. All but one of the studies cited as precedent in the Lancet addressed regions where disease was the major cause of death. The only other study (re:Kosovo) was an order of magnitude more precise than this one.
But the point you continue to ignore is that the outlier wasn't really discarded at all. Constant reference is made to it throughout the summary. Numerous confidence intervals inclusive of falluja are cited throughout. Study-wide conclusions are represented based on its results (including the deceptive claim re: aerial attacks.) The Lancet presentation of the raw data is the only one likely to have been read by anyone.
You can identify at least a 1,000 dead from air strikes, but that is clealy a lower bound.
Tim, I can identify at least 4 or 5,000 where airstrikes are clearly NOT the cause of death. 3,000 or so are unequivocally from insurgent sources (IED, suicide bomb, truck bomb etc) The other bulk entries contain tags such as "deaths from gunfire" "60% from gunshot" etc that obviously rule out air assault. The overwhelming number ARE tagged as resulting from gunfire or some variation thereupon. I am happy to send you an itemized spreadsheet if you doubt me, or you can do what i did and collate and itemize the page results yourself in excel.
Correction: upon reviewing the IBC I seem to have slightly undercounted the total for US airstrikes. The correct number I believe is 1500 or so. My unofficial tally of US-mandatory deaths is 2,263 as of early February, versus 3100 or so insurgency-only. The balance are undifferentiated by source, although the vast majority of those are designated SPECIFICALLY as having resulted from some type of gunfire. Nowhere near 15,000 have died from air assault and no fair reading supports this claim.
Tim:
My apologies for asking this question a second time, but do you have any response as to whether Garfield's explanation for his mistake in the EPIC interview concerning ex-Falluja bombing deaths is supported by the study data?
It seems obvious that the data does not support his explanation, and in fact his error actually originates in the study's Interpretation statement on page 1.
Mike, including Falluja puts the bombing deaths way over 30,000, but if you included just some of it, or were conservative in your calculations i suppose you could get 30,000.
Pascal, I did a tally of the IBC data, classifying them into one of three categories: air strikes, not air strikes, and undetermined (these were ones with no weapon given or a mixture of air strikes with other deaths). I found about 2000 from air strikes, 5000 not air strikes and 11000-12000 undetermined. So the range for air strke deaths in the IBC data is 2000 to 14000. The upper end of the range is about half of 30,000. It's more likely to be in the lower or middle of the range, so I modify my earlier comment to:
"Could there be 30,000 dead from air strikes with only a third or a quarter of these recorded by the IBC? I don't see why not."
It stikes me that air strike deaths would be concentrated in areas where the writ of the US forces, the Iraqi government and international organizations would be small.
11000-12000 undetermined
Some of the largest bulk entries from Baghdad morgues contain qualifying information alluding to main cause of death (generally "gunfire, explosions, etc.") You are assuming (with your lower "quarter" guess) that no fewer than 5,000 out of 10,000 or so deaths undifferentiated by the IBC result from US airstrikes. This is baseless speculation and actually contradicted by line item examination of the data. The only quantitative breakdown of such entries lists "55% by gunfire." I note the database seems to have been considerably revised since 2 months ago -- one line item claiming 60 percent of the largest bulk entry was due to gunfire seems to have disappeared. This is unhelpful to say the least. Now I suppose we are free to assume they are at least half due to airstrikes, ignoring all other sources altogether (artillery, mortars, RPGs, conventional explosives etc etc etc.)
the range for air strke deaths in the IBC data is 2000 to 14000 is somewhat less ridiculous than your initial 15,000 claim, but still baseless guesswork.
Please note as well that Fallujah is included in all IBC figures.
I found about 2000 from air strikes, 5000 not air strikes
Applying the same ratio of 2/7 to the undifferentiated "upper bound" yields around 3500, or a total of 5500 civilian dead due to air strikes including Fallujah. The Lancet estimate ex-Fallujah is over five times as large.
Tim:
Thanks for the reply.
I don't see what calculations are available from the study data, conservative or otherwise, which would lead to a 30,000 fatality bombing estimate. If you included all the non-coalition bombing deaths, but eliminated say, the male bombing deaths from Falluja and kept only the Falluja women and children deaths, you still wind up with a bombing extrapolation of around 100,000.
Not that there would be any basis or justification for Garfield to consider including a partial Falluja count. I don't see any other explanation for Garfield's error. It had to have been arrived at from the Interpretation statement, not from any influence of the Falluja cluster.
In any event, the 30,000 bombing estimate has been deleted from the EPIC interview, although the text of the interview still states that a majority of the 57,600 violent deaths resulted from coalition air strikes.