A couple of years ago, we had a debate on the Lancet study for the number of Iraqi civilian deaths during a certain timeframe. I suggest everyone interested in the subject read that debate, it was a good one. This study essentially argued that more than 10 times as many Iraqis had died since 2003 than any other source (pro-government or anti-war) had previously alleged. This study is often cited by certain forum members here.

The National Journal has now compiled a comprehensive list of the numerous methodology concerns. They are overwhelming. Some of the organizations that compiled these inaccuracies are opponents of the Iraq war.



I can’t help but note that this sound refutation took so long to come out. The Lancet study had a very negative impact, the figures cited throughout the media both here and abroad. I believe that both the figures and the timing of the study's disclosure of those figures were heavily motivated by the election cycle at the time. I can't help but wonder if this long-overdue refutation is influenced by the election cycle as well?

1) Will this refutation make headlines the way the Lancet study did?
2) Does this confirm or change your views of the Lancet study, or do you find the refutation suspect? Please explain your reasoning.
3) Why is this being released now?

My first question is Mrs P, before I comment on your questions- where is the opposing publlished piece that actually refutes the science, in other words, who has published the rebuttal in an actual journal of science- the national journal is NOT a scientific journal, it is a media piece. good or bad, I don't see an actual rebuttal based on repeating the experiment at this point. This, unless you can show me the actual science journal that refutes the study- is still nothing more than an op-ed piece.

In thier "about us"-

WHO WE ARE & WHAT WE DO
National Journal Group Inc. is a leading publisher of magazines, newsletters, books and directories for people who have a professional interest in politics, policy and government. Based in Washington, D.C., National Journal Group Inc. is committed to providing publications and services that are nonpartisan, reliable and of the highest quality.

There is nothing there like a science journal, it is simply editorial comment.

Editorial comment is less than useless if an opposing study has been done- it is simply calling into question the characters who have done the study as dishonest. The John Hopkins study has passed muster as a peer reviewed case- unless I am missing the link that a peer reviewed article rebutted the Lancet?

All I see here is more character assasination and calling into question the motives or outright acccuses the science journal of doctoring evidence.

If there is a peer reviewed rebbuttal, we need to see it- otherwise, we are just looking at another smear campaign with no credibility, or should be given no credibility, as this is not science- it is media op-ed.

If there is a peer reviewed rebuttal tha shows the flaws in methodology and changes the numbers- I will hold the Lancet as the only science article actually published.

If there is no peer reviewed study Mrs P- I would have to say, really, you know better. This is a political magazine, NOT a journal of science, and should be treated as such.

I will retract that if there is a peer reviewed article that I am missing, through misreading or stupidity on my part, at which time, you will have my apologies for messing up your thread.

Not yet published, but according to one of its authors, "accepted subject to minor revisions at the Journal of Peace Research," is this paper on Bias in epidemiological studies of conflict mortality. The Journal of Peace Research is peer-reviewed, and published bimonthly by the International Peace Research Institute in Oslo.

The authors make, in my non-expert opinion, a good case for a "main street bias" in the Lancet paper's household sampling method.

A very weak rebuttal. The only real rebuttal is to replicate the study- which is always the standard by which a peer reviewed article needs to be judged. Nit-picking at parts of the methodology, without actually replicating the study, and fixing the suspected methodology, is not acceptable.

If they have a problem with methodology, they have all the time in the world to "fix" this methodology and replicate the study.

I genuinely feel that folks that find fault with this study are afraid to do just that- because it will prove the Lancet study correct, if they attempt to replicate the study in an ethical and scientific manner.
http://www.washingtonpost.com/wp-dyn/conte...0602279_pf.html


For instance- the Cloning controversy and hoax. Some poeple did have some questions about the South Korean methodology, and when repeated- it was found to be a hoax- that is shy it is so very rare that published articles are thoroughly debunked- because in a peer reviewed journal, only science is allowed- if you try to inject opinion or bias, it will be found out.

Until the research is repeated, and fails under scrutiny of being falsifiable, then it stands.

It may very will be that the Lancet is wrong, it would be hypocritical of me to say otherwise. I don't think it will revise it down in any meaningful way though. And, it may very well revise it up. I bet those casting stones are a bit worried about that as well.

I'm not sure how the study can be "replicated" as the researchers themselves will not share their raw data*, making it impossible to check the same houses. The closest thing that can be done to "replicate" this study was done by the United Nations Development Program and their findings were a quarter of the number the Lancet study cites.

Researcher: Hello Iraqis terribly sorry to bother you but we're going to have to reinvade and kill you all over again. Seems our methodology was a bit dodgy and well we're not really sure how many of you we killed. So, if it's not a bother - I believe your religion has some sort of reanimation rites where you create zombies and... oh... oh that's not you that's other brown people. Darker? Yes. Of course. Well then... hmmm. We're in a bit of spot; yes? Then could you dig up your relatives and we'll have another go at counting them?

See, CR, I don't think that's going to work. Lancet needs to prove their numbers and "open the books" not the other way around. The numbers they released are complete bull pucky and are off by magnitudes. As much as some would like to have been killing 500 Iraqis a day they weren't quite hitting that number. It was a political stunt, and a failed one at that.

2) Does this confirm or change your views of the Lancet study, or do you find the refutation suspect? Please explain your reasoning.

This confirms my view. As I noted when we originally discussed this, their methodology was highly, highly flawed, and done in such a way as to offer opportunities to skew the data and confirm their biases. I'm not going into all the detail, but the Lancet obviously uses this methodology to inflate death counts in war (as they did in Congo). The only logical conclusion is that they are anti-war and want to overstate war's consequences. Which may be a noble goal, but does not excuse manipulation of statistics.

Remember, they not only overstated the death count by a factor of 10 or 20, they said that the vast majority (92%) of "excess deaths" were due to bullets bombs from invading forces. That's just not believable, given that this is the conflict with the greatest use of "smart weapons" in the history of mankind. Just like their first study, when they estimated 1,000 Iraqis were dying every day, and somehow no one anywhere saw these deaths occuring.

I'm not sure why a "peer review" is needed to debunk a study that was biased and funded by biased-actors like George Soros in the first place. The guy who commissioned the study has a book out entitled The 100 Ways America is Screwing up the World. I trust that it was not, um, "peer reviewed."

edit just to add an example - The accidental deaths included 15,000 Iraqis killed by U.S. vehicles in road incidents — extrapolated from five death reports.

The Lancet study extrapolated 15,000 Iraqi deaths from vehicle accidents with US soldiers, based on a grand total of 5 reports. It obviously never occured to them that, with 15,000 Iraqis being killed by US vehicles, a whole bunch of US soldiers would be dying from these accidents as well - you know, when a jeep or humvee hits an Iraqi car, soldiers get injured too. Unless they believe that the psycho-killer US soldiers are running down civilian pedestrians? Again, just ridiculous.

This is even a crappy new low for you- there is no need to invade, they just need to go over with some researchers that are just as willing to put thier life in danger to gather thier own data using the same methods, and only changing the methodology areas that they felt was skewed. Don't have the guts to repeat the study- don't do the criticism. There are plenty of right wing think type orgs out there with a bunch of corporate money, they are perfectly capable of funding thier own study. Cluster survey science is sound- if they have a problem with those that have gathered the data, send in thier own poeple to gather it- s

You first statement shows amazing ignorance- the poeple that did the survey were not invading or killing anyone- nor do they have to count bodies. The same info gathering info can be used again- the numbers won't change significantly now, except to go up, if the Lancet study was correct, or not up or down significantly when considering the difference in time,

IF the Lancet study says 650k dead, those poeple are going to remain dead, and the survey should be repeatable, minus the methods you have problems with.

you don't have to dig anyone up to repeat the study, you don't hve to invade anything, you don't have to count bodies- you just need to perform another cluster survey. You statement shows amazing ignorance of the methods anyway- so why bother comment if you are that amazingly ignorant of cluster surveys?

Until you have some evidence that "it is a failed political statement"- you should keep your teeth closed and perhaps learn what the hell you are talking about.

IN fact, if it bothers you, you can certainly volunteer to go over and repeat the study yourself. The poeple that were involved in the evidence gathering were brave enough to go house to house, or are you too much of a coward to do what a college student is willing to do?

Places have tons of money and resources that they could easily replicate this study/cluster survey- instead, they want to smear the study instead of re-doing the survey for themselves, and eliminate the info gathering issues that they say are wrong, If the poeple in the Lancet study were able to do it- then you can repeat it too BA, but apparently, the conservative smear types don't have the cojones to do it. Wonder what they are afraid of?



Of course it confirms your view- you didn't want to accept the story in the first place, not because of the science, but because you don't like Goerge Soros. So why doesn't the big money groups like Haliburton, or any of the massive corporate pirate types fund thier own studies CW? They have more money than Soros, and a vested interest in debunking it.

Soros funded it because the right wingers are afraid the truth might not be too bearable to them. Smear the researchers because of where the funding came from is not science either.

You need "peer review" because it WAS published in a peer reviewed, very prestigious and well known science journal, NOT because of who funds it. GW himself can ask for and fund a competing study, and I would accept it if it were to be published, because it has to meet the criteria for publishing, that is the way science works. That is why there are so very, very few hoaxes in science today, compared to how many articles are published- because some bright graduate student will know he can get a great new job, funding for research etc, if you are able to debunk or refute former research- that is the great check and balance of science-

That is why, when there is a real hoax or "doctoring" of data, or real bias injected- it is so quickly found out- because it is so easy to repeat it in the "peer review" publishing proccess. The very act of publishing has a "bias filter" simply because it someone can come right behind them and discover this bias, and then repeat the survey, and show under no uncertain terms that the bias was there.

If they just want to cast stones, and don't have the cojones to just do it- they need to shut the hell up, because they are no better than Rush or Bill, they are just character assasination 'infotainment".

if there are political ramifications and fallout from the "constroversial" findings- then it is easily repeatable- are you telling me that conservative think tanks in the nation can't hire some folks to repeat this methodology and survey?

Ya, I am sure Soros has far more resources than all the chickenhawks in the country?

Haliburton has the kind of money that could buy Soros in a day- why haven't they done anything to refute this- they already have employees in country that could replicate the same survey, with very little re-alocation of resources.

Just sour grapes and character assasination simply because you can't stand the fact that we may have harmed that country in a massive and pervasive manner. That country is still far worse off under our occupation than it was under Saddam, and the truth hurts.

It is hard to find a country that America has done more good than harm CW- and if that FACT bothers you, perhaps you need to ask your elected officials to stop doing the things that harm other poeple?



Lancet has already proved thier numbers by publishing. It doesn't work that way BA, and your ignorance of science is showing. They have met thier burden of proof, and if someone doesn't like it- too bad, it is encumbant upon those that don't like the results to hire poeple in Iraq, just like JH did, and repeat the survey. If you don't like that, the deal with the numbers that are there.

IT is no different than if I came up with some breakthrough in physics, and I believe some of the results is proprietary to my study. After I published, you will have to have yor own lab, and use your own resources to refute or confirm my breakthrough. This is EXACTLY what happened in the cold fusion and cloning hoaxes- someone said "that don't seem right"- so a competing entity, I believe, in this case , a competing University- and they easily proved thier articles and findings to be a hoax, even though the original entities that submitted thier false findings to be published kept a large part of thier research secret- because that is the money making part of the research if true- the debunkers of the hoax simply followed the published materials methodology and debunked the findings. The south Korean researcher that lied about the cloning NEVER TURNED OVER HIS RAW DATA, those that did not agree had to create thier own data- and that is consistant with all research in the world, for the most part, especially when the raw data is proprietary.

All that has to happen is someone to follow the article and repeat the survey- and it simply hasn't happened, despite groups in this country that have more than enough resources to repeat it.

Until someone repeats the survey- you are dealing with character assasination and smear technices, and there is no debunking science whatsoever. IT is all op-ed, and that is the end of it.

That only makes sense if Halliburton has a unit of statisticians, analysts and surveyors. Barring that, how in the world could they replicate the Lancet survey?

Apparently you need : D to recognize sarcasm...

****

The science and methodology IS being debunked. Are you even reading the original article. Why isn't Lancet giving over their raw data? What are they afraid of?

But as I mentioned before, the UN Development Program did the same type of cluster survey, and the results were drastically different. Otherwise, how can this particular survey be repeated? You are basically demanding the impossible and then saying that until the impossible is completed, no other evidence matters. If you are suggesting that they conduct another similar survey, well, mission accomplished. This is not a physics or chemistry laboratory experiment, it's a statistical analysis using survey data. Unless they go to the same homes (impossible) it cannot be repeated exactly.

Guha-Sapir, the director of the World Health Organization's Collaborating Center for Research on the Epidemiology of Disasters, is in the process of completing a paper outlining numerous mathematical and procedural errors in the Lancet II article, as well. There was the aforementioned HICN study. The United Nations Development Program study. All conspiring against this team of researchers and their work?

Here's a pertinent question.

Even if Lancet is refuted successfully, (which seems probable), will the fact that several thousand Iraqi citizens died instead of several million detract any horror from the fact that those several thousand are still, in fact, dead?

Even if the numbers are off, there's always a stigma that clings close to civilian deaths in wartime. Lancet was, in all probability, constructed precisely for that purpose: to enflame antiwar sentiment with horrific "statistically accurate" numbers. But no matter the actual sum of those numbers, the import that they carry is still the same.

One million, one thousand--these are only relevant from our perspective. On the face of it, one million is obviously "worse" than one thousand deaths. But tell that to the families of the deceased.

No, I linked to my post which details my concerns with the statistics (or "science" as you call it). Cluster sampling, small samples, questionable techniques, unbelievable response rate of 98%, failure to 'reality check' things like 15,000 Iraqis killed in vehicle accidents with troops, failure to collect demographic data (to help fraud-proof the study), including deaths from a car bombing in July, when their study supposedly ended June 30 ... Those are substantive questions of the findings, and have nothing to do with Soros.


Dude, they didn't include the clusters in the study as published, so any "peer review" has to accept the research at face value. It would have only taken a few pages, so why didn't they include it?

No, actually it's not.

The Wall Street Journal sheds some light on this deeply flawed "study". Turns out the Minister of Health under Saddam was the main researcher and he has never released his raw data. As for the peer review that CR keeps bleating about - it wasn't a standard peer review it was an expedited peer review... Expedited for the US elections. Enough!

As for whether it is 10,000 or 1,000,000 it absolutely matters. Stop making up facts to enflame.

Link to wsj: (mods please fix I am on a Blackberry)
http://mobile2.wsj.com/device/html_article...ew_and_outlooks

Well at least we know why the Iraqi Ministry of Health numbers pre-war were off: the same guy conducted these studies. Pre-war he had a vested interest in underreporting deaths and post-war he had an interest in inflating the number of deaths. Mr. Lafta laughably says that the research he did before was unrealiable, but we should believe him this time--without seeing the raw data, of course.

no other entity has gone out and attempted to replicate or do thier own study. They only they they even try to do is attempt to tear down the study- how far off was the lancet study? Was there 350k dead or 650k dead? Oh, you don't know? You are guessing, right? Why? Because no one has done as an exhaustive of a survey?

IT is all meaningless opinion until someone does thier own study using what they believe to be either the correct methodology, or comes up with another methodology that refutes the study, but does as an exhaustive survey/

Right now, as I stated before, all we have is someone doesn't "feel" the study was right. There is no counter-argument except opinion, at best, at this time.

Once again, when another survey is published, and the results are wildly different than the Lancet survey (I would say on the magnitude of +or- a factor of x2- either over 1 million deaths, or under 375k deaths), then, the Lancet survey stands as the only scientific survey to model the Iraqi dead.

Poeple emotionally don't want to believe it, so they try to nit-pick it, but until someone goes out and takes a nice, large sampling using similar methodology, minus the repairs needed as mentioned here, then it is all just guessing and hoping the Lancet study is wrong.

Mrs P- you said the UN is actually doing a survey right now? Did I miss that on the links?

The US goverment has refused to even try to count the dead- considering how GW likes to hide bad news as best he can under the cover of "classified information"- I would GUESS that the GW admin is afraid to commision a real scientific survey- thinking that the Lancet is right maybe? Or close enough that it still looks like a major wipe out of a large chunk of the Iraqi population.

I will tell you, my own doubts with this story didn't revolve around the methodology as much as my thinking that a lot of the "dead" counted were actually refugees to Syria and Iran and what not. Then a bunch o of the runners got caught up as being counted as being dead.

Sounds plausible enough, but once again, my guess is worth 0 here as well, considering no one has the guts to go out and complete a competeing survey. At this time, the Lancet survey is the only one that has been completed, and we have to work with those numbers until someone comes out with a peer reviewed competing survey.

This study was so flawed from the beginning it was a joke – yet much of the left biased press grabbed onto it and would not let go.

Naturally when the date is refuted it would never make the front page of the NYT or any media that had latched onto it.

I think we have to be honest here, there is really no way at this point in time, to conduct a survey even using the same methodology as Lancet used. With the large numbers of displaced Iraqi's, the sectarian turf battles between the Ministries and the insourcing [if you will] of health care and other civil services by the various militias............I don't believe anybody could get more than a ballpark figure. The only play left in the game is to attampt to determine, based on ground realities and survey methods used by Lancet, if the survey was at all accurate. From my perspective [and not being a stats analyst], that's all the Lancet study was........a ballpark guess.

Looks like the World Health Organization has come out with a study: Iraqi Death Toll at 151,000. Clearly, this isn't at the same level of 600,000+ Iraqi dead in the Lancet study but it's still 6 digits. I do think there were serious methodology problems with the Lancet study. Still, the numbers in this latest study are sobering.

I will accept the UN study as close enough.

Somewhat of a phyric victory though eh?

Lancet study was 5 times off, perhaps- however, we are still talking about 150k dead. Not something to be proud of.

CR:

The problem with the Lancet study, well, you served, yes? Does war follow a bell curve? So why we would expect any one or twenty or forty places to be fairly representative of anywhere else? That was and is the fundamental flaw. I otherwise wonder why no one stepped back and asked the question, are you asking me to believe that with 600,000 total excess death that we have more deaths than combined war deaths during the US Civil War? And also more excess death than the combined death toll of Hiroshima, Nagasaki, Tokyo, Dresden, Hamburg and Manila?

And, yes, 100,000 is sobering, but if these are excess deaths and not war deaths, then we need to ask some questions, first and foremost being, while not being proud of the human tragedy, how can we fairly be blamed not only for the whole Shia-Sunni affair, but also the both Shia and Sunni v. Marsh Arabs affair? There comes a time when the people of the world, and Iraq, can no longer blame us for all of the excess death, since in certain instances it's them killing each other and without any reference to us whatsoever. The other way of putting the matter is that one can blame the warden all one wants for releasing the inmates, but in the end, it is the inmates who are the ones primarily responsible for their future criminality [or not].

2) Does this confirm or change your views of the Lancet study, or do you find the refutation suspect? Please explain your reasoning.

I'm not getting into the politics of the situation but when the figures first came out they seemed to me to be awfully bloated, particularly because people who were keeping records on the ground were coming up with figures that were much smaller. I mean how many dead folks are there going to be that aren't going to be recorded at a public burial site or at a hospital?

From the original link.


It appears that this one guy who ran the survey wouldn't let independent outsiders look at his original notes. It's hard to see how this study can be taken seriously if there is no meaningful peer review allowed.

I'm more inclined to take seriously the folks who counted the bodies on the ground than those got their information through interviews, without there being positive independent confirmation of the dead persons.

However one can see part of the discrepancy. The Lancet Report included soldiers and insurgents while the Iraq Body Count tried to restrict it to civilians
http://nationaljournal.com/img/njgraphics/080104_munro1.gif

Somebody should have done a gut check on this one.


It seems to me good intentions were involved but not a lot of peer review or common sense or respect for more conventional approaches were included. It's an example of where so called science leads folks astray. You can get lost in the numbers and miss the corrective of good old seat of the pants common sense.

Edit.
Just caught the WHO report from a preceding post.

Now that makes more sense. A more extensive interview process wedded to reports from morgues and hospitals.

The Lancet study stated a 95% probability that 650,000 people have been killed by this war. However, the numbers could have gone either way with a decreased probability those numbers are accurate.

Nobody believed that study, but a new one comes out with greatly reduced numbers and people gobble this up without question. Why doesn't this new study face the same scrutiny? Assuming standard statistical modeling, there's a 95% chance 150,000 have been killed. But as with the Lancet study, the numbers could have gone either way as well.

As usual, people are believing what they need to believe. In my opinion, the first study was not flawed because nobody argues with those methods and results whenever they are used somewhere else.

You're completely discounting the facts that the Lancet Study was not properly peer reviewed but, admittedly, expedited to coincide with the US Election cycle and that the lead researched has never released the raw data. I won't even get into the funding or the inconsistencies of the lead researcher's previous work.

That's why Lancet was discounted by people who took a second to look past the ridiculous numbers posited.

I don't know if this helps but if you read the WHO link it gives roughly a range of 50,000 on the downside and 70,000 on the upside. This study also employed a considerably larger group of interviewees and the figures roughly check out with morgue and hospital figures etc. Frankly I haven't got a clue where you come up with the 95% probability, but then I've never had a class in statistical analysis.

As far as the methods and results apparently the first study hasn't been subject to full peer review so yes there is an argument with the methods and results.

You have a bell curve in a cluster survey- with a "% of confidence" throughout the curve. For the Lancet study- there was a 95% confidence that the number was around 650k dead. the farther you move from that number- the less the % of confidence. There is the number of 150k in the lancet study as well- it just has a very low % of confidence. Cluster surveys are pretty dead on accurate, but they are not infallible either- but they are used for about anything when you want to track populations- like, oh, do you want to open a new Wal-mart in "X" location- you hire a firm that DOES know stat analysis, and you pay them good money to get it right. Lot's of money is gambled on these surveys, world whide- because they are so very accurate.

The Lancet study has been attacked because they don't like the poeple doing the survey- NOT because the science is wrong, or thier methodology is wrong.

There is a strong possibility still that the Lancet survey is right, but on the lower side of the curve- say, if 250k were at 68%, and the WHO study 250k is at 68%, it very much mor likely that the real number is closer to 250k.

Considering how many non-fatal US casualties there are in Iraq- I believe the number is over 50k now, and 10 times that number for Iraqi dead is not even a bad guess either. Considering that a large majority of those injuries are not survivable without the immediate attention a US soldier recieves when wounded or injured in some way.

so, I accept the UN numbers as a good scientific rebuttal- but I also accept the Lancet as good science- and reject the usual character smear that comes whenever a study comes out that casts negative light on bad policy.

I am sure the actual number lies somewhere north of the UN study, somewhere south of the Lancet study.

While appreciate your 180 on this the preceding post basically reads:

Fake but accurate.

You are discounting the fact that you have no idea what you are talking about. The Lancet study is the only peer reviewed study out of all the Iraq casualty studies.

Back in 2004, the Lancet released a study showing upwards of 100,000. Nobody argued or complained. Only when the 650,000 value with a 95% probability became public, did people start screaming. And as I said before, nobody is questioning the newest report of 150,000. Let's pretend that's true: How do you square the first Lancet study being accurate at 100,000, but see only a 50,000 value change 4 years later while violence continually increased?

When I read the report "discounting" the Lancet study, I was prepared to admit I was wrong. I though they had actually found something. But they didn't. They are calling foul while using "the liberal George Soros" type invective to make a point.

From wikipedia:


It has been peer reviewed and the only people calling foul are people who have no statistical knowledge. I develop SPC software. You can see whatever you want with statistics, but it doesn't hide the probabilities. The Lancet group did nothing wrong or out of the ordinary. That entire "main street" premise has been debunked which makes the latest counter-argument highly suspect along with their political descriptions.

was not properly peer reviewed but, admittedly, expedited to coincide with the US Election cycle

"Properly" peer reviewed? Is that the new standard? I didn't know there were peer reviews and proper peer reviews.

I suggest everyone wanting to actually know the truth should read this very spirited debate among statisticians regarding the Lancet study. Only one person says the study is flawed. Everyone else is trying to teach him how the study was done correctly in very scientific detail.

Edited to add this passage:

I have a general beef with journalism regarding a few topics. Apparently you go to someplace called "J-school," and you are supposedly qualified to report on things like the "Lancet study." No journalist has apparently ever had a class in statistics, and they never report on things like this correctly. They buy hook, line and sinker "studies" by biased actors like the CSPI or the "Lancet" study, then they hardly ever retract when things are proven wrong. Partly due to bias, partly due to being fed a sexy story, partly due to a lack of knowledge of the subject, whether statistics, economics, etc.

This leads to stories like "despite lower crime rate, more people are in jail" when a 3rd-grade reasoning would tell you "more people are in jail, ergo crime is down." Hardly any story on trends takes into account population shifts, or indexes cost increases to inflation ... the list goes on and on.

I have been interviewed for attribution a few dozen times, and it was almost always painful to help the reporter understand the specific discipline that he/she was querying me about. I think that we are better off reading blogs and trade rags on a given subject (like this one), because at least you hear arguments from both sides, played out by people who know the subject. This is of course complicated in political arguments when certain fields are dominated by one side or the other, e.g., business = supply-siders, sociologists = liberals, etc.

So, CNN and everyone said "600,000 dead" in Iraq, and they have moved on. It matters not what we say here, nor what the statisticians come out with. The story is out there, and everyone has either forgotten it, or doesn't care, or accepted it as fact, and we can't change that.

Two co-authors, Gilbert Burnham and Les Roberts of Johns Hopkins University, told the reporters that they opposed the war from the outset and sent their report to the Lancet on the condition that it be published before the election. Mr. Roberts, who opposed removing Saddam from power, sought the Democratic nomination for New York's 24th Congressional District in 2006. Asked why he ran, Mr. Roberts replied, "It was a combination of Iraq and [Hurricane] Katrina."

Then there is Lancet Editor Richard Horton, "who agreed to rush the study into print, with an expedited peer review process and without seeing the surveyors' original data," report Mr. Munro and Mr. Cannon. He has also made no secret of his politics. "At a September 2006 rally in Manchester, England, Horton declared, 'This axis of Anglo-American imperialism extends its influence through war and conflict, gathering power and wealth as it goes, so millions of people are left to die in poverty and disease,'" they write. See YouTube for more.

We also learn that the key person involved in collecting the Lancet data was Iraqi researcher Riyadh Lafta, who has failed to follow the customary scientific practice of making his data available for inspection by other researchers. Mr. Lafta had been an official in Saddam's ministry of health when the dictator was attempting to end international sanctions against Iraq. He wrote articles asserting that many Iraqis were dying from cancer and other diseases caused by spent U.S. uranium shells from the Gulf War. According to National Journal, the Lancet studies "of Iraqi war deaths rest on the data provided by Lafta, who operated with little American supervision and has rarely appeared in public or been interviewed about his role."

My God, you're right! I was doing some research and found out that on top of all that, Mr. Roberts regularly - and with blatant disregard of the law - tears the tags off his mattresses. I feel so ashamed.

Really, there's a lot of invective without any science. I posted a link where a group of statisticians - guys who work at universities and write books on the subject - who almost all (except for one and he won't provide his credentials) have studied the report and found the results to be reasonable. They have some minor critiques, but nothing that would have swayed the result.

My point is, SPC (statistical process control) is not for the faint of heart. I sell SPC tools on my engineering website for other software developers to use. So, I know a little about it. But compared to the level Roberts and his people use, I'm an idiot when it comes to his type of analytical process.

So, when you can't argue on the merits of the science, you use a swift boat tactic to extract the same results. I'm willing to be persuaded that the study is flawed because the numbers seem high to me, but that's mostly what I want to believe. I read the study just like tons of other statisticians and peer review groups and saw nothing that jumped out as "fatally flawed'. And I still haven't. This latest report discredited the people - not the science.

DR, would you do something about your post above. It's stretching things out and making it difficult to follow.



Maybe there is a discrepancy between how the survey was approached in theory and in practice. Obviously something is wrong if the WHO using a much greater group of Iraqi interviewees and backed up by the public on the ground records ended up with a radically lower number of Iraqi violent deaths.

"Nobody argued or complained" with the first Lancet study? Of whom are you speaking? I know it couldn't be the folks at AD, because we had a debate on this one too. In fact, YOU started that thread. The first Lancet study didn't require proof of death. For that matter, there was no proof of the reason of death other than the relative's good word and who is going to admit that their dead family member (supposing he/she actually existed and died in the first place) was an insurgent in the middle of a warzone?

Your own link above offers a list of criticisms for the first study...it differed markedly from all other sources, from humanitarian agencies to the UN.

Ok, Mrs. P - fair point. What I should have said, was nobody in the scientific community argued and complained. When I was doing the research for this study, I saw lots of references to the first study (referenced in your link), but haven't seen anything that discounted that study. But in re-reading the thread you linked to, there were some familiar responses (from the same people using them here today):








And my response:



So, I would be very interested in a scientific response to the first study as I think you have a valid point. I assumed nobody proved the first study wrong from a scientific standpoint and that may be a false assumption. Again, I'm willing to be persuaded that one or both of these studies is flawed. Not from a propagandist, but from a scientist. And the first study is critical because if correct, the 150,000 figure being used now is far too low.

From your wiki reference - Fred Kaplan called foul - and he's a journalist. But your reference contains this:

PERSONAL ATTACK REMOVED

I said I am willing to accept a WHO study- not that I am prepared to throw out the ONLY peer reviewed study, with no scientific peers TO DATE willing to actually throw out the study, or question it really- as DR has pointed out- again.

The entire problem with this debate is that poeple that are COMPLETELY ignorant of statistical analysis are commenting on it- PERSONAL ATTACK REMOVED

DR has pretty much laid it out here- I have only had, at this point, 2 college level stats classes, and they are difficult and enlightening. But I do have enough passing knowledge of statistical analysis to understand you have no freakin' idea what you are talking about- as usual.

PERSONAL ATTACK REMOVED

I said I am willing to accept the UN numbers, but I am still not prepared to throw out the Lancet study as the only peer reviewed article ( and what the hell does Not Properly Peer reviewed- it is or it isn't, it is like being a little bit pregnant ) that presents the most scientific study to date. It is still better than the UN results, as far as an actual survey based on peer reviewable science- the UN study is NOT peer reviewed- that is why I had such a hard time finding it in peer reviewed journals when Mrs P was talking about it early on.

Without it being peer reviewed- it can be argued that the UN report is worthless, and not scientific, after all, the UN has a vested interest in seeing thier #1 cash cow not look TOO bad- oh, and since they all work for the UN, I guess we should just toss out the whole survey anyway- because, you know, it's the UN

BA- until you come up with something other than some media critiques- you know, you might try cracking open the Lancet and going through the original study- I am sure your local university has a copy, so you don't have to subscribe, mine does, I have read it. I am going to bet money you haven't.

AS far as the Lancet study goes- it is NOT "refuted"- at this point- that is a lie- I have seen no journal refutation to date, and I used the Pub-med journal search, probably the best in the world- so I will let DRs comment stand on this one:

Ok, Mrs. P - fair point. What I should have said, was nobody in the scientific community argued and complained. When I was doing the research for this study, I saw lots of references to the first study (referenced in your link), but haven't seen anything that discounted that study.

What is scientific about anything you have posted? I am sure the actual number lies somewhere north of the UN study, somewhere south of the Lancet study <--is that your scientific contribution? Is that what Statistics 101 and 102 has given you the ability to do?

When I (and publications like the WSJ, et al) say not properly peer reviewed it means that the standard peer review process was not followed. A little pregnant and a little peer reviewed are not scientifically comparable. Being pregnant is a an either or proposition - peer reviewing isn't. Perhaps a Statistics 201 would have let you in on that.

No amount of Bell Curves or clusters can change the fact that the report was intentionally rushed to publication to coincide with the US Election cycle. If Pat Roberts were to have a peer reviewed report published proving that evolution doesn't exist would you blindly accept it - I mean it is peer reviewed after all.

The problems with Lancet are in methodology and the lack of raw data. I would expect a man of your scientific reverence would also have issues with this. Do you have any scientific idea what a half a million bodies looks like? How much area they would take up? Earlier in the thread someone posted that they thought the numbers were cooked in order to enrage the public but that was OK because it was a good cause.

However we both agree the Lancet number is wrong. You feel it's somewhere south and I feel it's nowhere near. So when I tell your statement reeks of fake but accurate - I stand by that.

The WHO report was published in the New England Journal of Medicine. It is one of the most popular and widely-read peer-reviewed general medical journals in the world.

Richard Horton, editor of the Lancet has said this about peer review:


LINK

Peer review often does not review data, it reviews methodology. The only data available to reviewers is the data available in the study, which is often summarized (as is the case here) and not raw data where suspect data can be identified.

Peer review certainly does not make a study unassailable.

Interestingly, Horton said a few months before this study was published in 2006 that Bush and Blair were "killing children" in Iraq.

Thanks Mrs P- I must have not typed in the right criteria for the search- "violence and fatality" don't come up with anything on Iraq- just on the US and western countries, mostly dealing with domestic violence

That being said- I DON'T discard the Lancet study OR the WHO study. No scientist would- because it would be dishonest to do so at this point- because, for the same reasons poster's here didn't like the Lancet study- are the same issues that the WHO study MIGHT have- for instance-

(from the link provided by Mrs P, thanks again)

Assessment of Mortality Data

Of the 1086 originally selected clusters, 115 (10.6%) were not visited because of problems with security. These clusters were located in Anbar (61.7% of the unvisited clusters), Baghdad (26.9%), Nineveh (10.4%), and Wasit (0.8%). Since past mortality is likely to be higher in these clusters than in those that were visited during the IFHS, we imputed mortality figures for the missing clusters in Anbar and Baghdad with the use of information from the Iraq Body Count on the distribution of deaths among provinces to estimate the ratio of rates of death in these areas to those in other provinces with high death rates. Data from the Iraq Body Count were used to compute ratios for death rates in Anbar and Baghdad, as compared with the three provinces that contributed more than 4% each to the total number of deaths reported for the period from March 2003 through June 2006.

For instance, we compared the ratio of the rate of death in Baghdad relative to the rate in three high-mortality provinces reported by the Iraq Body Count (3.08) with the rate ratio reported by the IFHS for the same provinces (1.56). To obtain the same ratio, overall mortality in Baghdad would need to have been 1.97 times as high as that in the three other provinces on the basis of the visited clusters only. This corresponds to a rate of death in the missing clusters that is 4.0 times as high as that in the visited clusters; the corresponding numbers for Anbar were 1.43 and 1.70, respectively (Table 1 of the Supplementary Appendix). This adjustment involves some uncertainty, since it assumes that completeness of reporting for the Iraq Body Count is similar for Baghdad and other high-mortality provinces. Since the Iraq Body Count did not collect information on age and sex for all deaths, the adjustment factors were not stratified according to these factors.

and again:

In general, the underreporting of deaths is likely to be common in household surveys. The most serious concern is household dissolution after the death of a household member. Several demographic assessments have suggested that there has been an underreporting of deaths in the IFHS


Here is the deal with the WHO study- which, like the lancet, I am not REFUTING- I am just showing that it may be more or less flawed than the Lancet study-

They rely on Iraq Body count for some of thier numbers- which, also as a political body, may have a vested interest one way or the other as well. - But I would be as guilty as BA if I just discounted the study in the NEJM as he did with the lancet- with no real science, just media, as my rebuttal

The nice thing about peer review though Amlord- it DOES do a remarkable job of "outing" those that use false data to skew results. NO ONE has scentifically argued this = only bloggers and editorialists have done this.

Again, fromt the site Mrs P posted-

Uncertainty in the missing cluster-adjustment factors was difficult to quantify, since we assumed that the excess risk of mortality in missing clusters in Baghdad and Anbar was normally distributed, with standard deviations of 0.2 and 0.1, respectively. All analyses were performed with the use of Stata statistical software, version 9. All P values are two-sided.

It does have it's flaws too- it assumes an awful lot in the most violent sectors, and I believe the Lancet study has the same critique in this area- though, obviously, the Lancet study info gatherers were alot more brave than the UN gatherers.

Peer review is absolutely "unassailable" Amlord- it is infinately the opposite- you can "assail" it anytime, by simply repeating the experiament/ procedure for themselves- to date, and, in fact, it looks like, as a second look has the Lancet study doing more to address the weakness in the WHO study- they went more into "dangerous" area and did not rely so much on IBC.


I will accept the WHO study as I will the Lancet study- as science properly done, with controversial results leading to Op-Ed pieces in the paper- and, for the most part, the criticisms of the Lancet study DON'T come from science- they come from media- a HUGE difference.

NO science is "unassailable"- it is ALL subject to review, debate and rebuttal- at this time, there are not really any good "rebuttals" to either study, and the only thing that either study really points to is the need for more studies.

Also- it is important to note that not all survey's are "science subject to peer review" - but they can be usefull anyway. For instance- the "box store boom" up here is based on a study commissioned by a retailers association detailing the projected population growth and buying trends. Never was published in a journal- but has proven pretty darn accurate, and, it is accepted as scientific enough for large corporations to risk billions of dollars on the surveys-

and the same methods are used as in the Lancet survey for Iraqi deaths as used by Walmart to determine if the population growth will fuel the need for another box store.

99% of the criticisms of the Lancet study are based on op-ed pieces with no basis in science, they just try to make it look that way.



and that would be both accurate and precise, it just evokes emotional response. The citizens of Iraq, including the children, are far worse off under the US occupation than under Saddam Hussien. It evokes emotional response- but it does not change the fact that it is true. Bush and Blair's policy killed far more children under the occupation than under Saddam. Our actions certainly have led to a large number of children to die.

There is nothing dishonest about that statement, so what is the problem here- you don't like the messenger?

Well, the question here is "which children". Obviously, the Sunnis, who had a fairly "nice" if you want to call it that, time under Saddam. But he was killing Shia and Kurdish children just as efficiently, if not more, than our dear blundering president and the slightly-more-competent former PM. Instead of focused genocide on a single ethnic group, the war has unleashed a free-for all, with children of all ethnicities, Sunni, Shia, and Kurds alike, getting killed equally. Now isn't that just heart-warmingly egalitarian?

Bush commits his acts out of political clumsiness rather than actual menace. Blair...is inscrutable, but not sadistic. Not in public, anyways. His private life is his own business.

Until you can show me some real data stating that George HW Bush and Tony Blair enacted policies with the specific intent of killing children, and then managed to rack up a body count higher than the (ex) President of Iraq himself, I'll take your statement with a pinch of salt. A bulldozer of salt. A really-really-big bulldozer.

You are the one that devined intent- not the original quote- he didn't say "GW couldn't wait to get to Iraq so he could kill some children"- he said "Blair and Bush are killing children"- I would devine the intent as "GW and Blair's policies in Iraq are killing children"- which I find to be true.

However- GW and Blair, no matter which study we go by, have enacted a policy that caused the total chaos and civil war we have today. I mean- you didn't exactly have millions of refugees headed to Syria under Saddam. Sure, the guy was brutal and evil- that being said- the entire country is worse off under the US occupation than under Saddam- and that does NOT make me happy.

It is as I have said many times- we, as a county, and being a patriot myself, keep making the same buttinski mistakes, over and over. It is hard to find a country we have entered post Korean war that we have been "the good guys"- Grenada, perhaps Serbia, well, that is about the end of that list.

As a country, we ABSOLUTELY HATE taking responsibiity for our countries behavior outside our borders. We have one study that says there are 650k more poeple dead under the US occupation than it would have been under Saddam. I think it approaches the magic "million