Gilbert Burnham and Les Roberts have an op-ed in the Baltimore Sun:
Not wanting to think about civilian deaths in Iraq has become almost universal. But ignorance of the Iraqi death toll is no longer an option.
An Associated Press poll in February found that the average American believed about 9,900 Iraqis had been killed since the end of major combat operations in 2003. Recent evidence suggests that things in Iraq may be 100 times worse than Americans realize.
News report tallies suggest that about 75,000 Iraqis have died since the U.S.-led invasion. But a study of 13 war-affected countries presented at a recent Harvard conference found that more than 80 percent of violent deaths in conflicts go unreported by the press and governments.
City officials in Najaf were recently quoted on Middle East Online stating that 40,000 unidentified bodies have been buried in that Iraqi city since the start of the conflict. In a speech Sept. 5, Samir Sumaidaie, the Iraqi ambassador to the United States, stated that there were 500,000 new widows in Iraq. The Iraq Study Group similarly found that the Pentagon undercounted violent incidents by a factor of 10. Finally, last month, the respected British polling firm ORB released the results of a poll estimating that 22 percent of households had lost a member to violence during the occupation of Iraq, equating to 1.2 million deaths. This finding roughly verifies a less precisely worded BBC poll last February that reported 17 percent of Iraqis had a household member who was a victim of violence.
So multiple polls and scientific surveys all suggest the official figures and media-based estimates in Iraq have missed 70 percent to 95 percent of all deaths. The evidence suggests that the extent of underreporting by the media is only increasing with time.
Being forthright about the human cost of the war is in our long-term interests. How can military and civilian leadership comment intelligently about security trends in Iraq, or about whether any security policies are working, if they are not detecting most of the estimated 5,000-plus violent deaths that occur each week? Can American plans for the future of Iraq be respected within Iraq if they do not openly address the toll that they imply? Avoiding the issue of Iraqi deaths will likely come back to haunt us as young people in the Middle East grow up with ingrained hostility toward America.
(Via Stephen Soldz.)
"Can American plans for the future of Iraq be respected within Iraq if they do not openly address the toll that they imply? Avoiding the issue of Iraqi deaths will likely come back to haunt us as young people in the Middle East grow up with ingrained hostility toward America."
Yea. Great PR , Les. Let's immediately change the estimates on Iraqi dead numbers to 1.2 million because your dubious poll and a coupla of others disagree with our estimates as a way of winning the hearts and minds of Iraqi civilians in the middle of a war. What a sparkling idea that would be? Shall we try to figure out the consequences of this admission in the middle of this sorry mess?
Amazing. Obviously Les needed some Oped space face time.
O.K. Jc, let's just talk about the numbers, Les Roberts vs. you.
Les says 1.2M. You say, what? Be sure to include all your references.
Pop trot
I don't need to go with any numbers seing I don't know..... like Roberts doesn't either by the sounds of things.
Roberts is saying there have been possibly 1.2 million extra death in Iraq since the war. That means there has been an extraordinary 820 extra deaths a day in Iraq since the war began.
Have you even heard of one single day in this sorry mess of a war when there have been more than 250 Killed. I once heard that number over a two day period about 2 years ago.
I have no issue with the methodology used. He's just fudging the figures.
Polling firms run sample audits to verify if their pollers are actually doing the hard yards. What audit prcoess was carried out? You know?
It was nice to see Roberts got his methoid peer reviewed and all that. However I never read anywhere that there was any audit done. Nice.
Since JC is being such a prat, I suggest a trip to Coventry.
JC, every death in Iraq does not make it into the papers. But this has been explained to you many times before ...
Here is a major problem with the 9,900 Iraqi fatality number. If you assume the number is roughly 9,900, you are effectively saying that roughly 2.6 Iraqis have died per US soldier death since 2003 (the official number from icasualties.org as of today is 3816). To be pedantically clear, this number represents a ratio of US soldier deaths against TOTAL Iraqi deaths. If you think that roughly 4 Iraqis die per US soldier, the number would jump to 15,200 Iraqi deaths. If you say that roughly 4 Iraqis have died per coalition soldier death (total of 4121), then you have upped the comparative estimate to roughly 16,500. (Why do I choose 4? It is, admittedly, arbitrary. I'm just running the numbers.) If you say that 2 Iraqis have died per dead (3816) or wounded (27753) US soldier, you increase the estimate to ~63,000 Iraqi deaths. As a closer on this point, if you believe that the 9,900 fatality estimate is accurate, then you are saying that 0.3 Iraqis have died per US dead or wounded (to which I am more than slightly incredulous).
You might argue that the number YOU care about is the number of Iraqis killed by American soldiers, and not due to Iraqi-on-Iraqi (aka "sectarian") violence. However, a civilian killed during a time of social upheaval is a civilian killed, no matter the killer, and it doesn't take a good demagogue to lay each death at the feet of an occupying army (especially if the existing national government is seen as an emasculated puppet government). Put yourself (if you can think empathetically) in the shoes of a person who lives in an occupied country. Would YOU put the blame on civilian deaths upon a government that is seen to be a puppet of the occupier or at the feet of the occupying power? I would lay bets that most people (you included) would blame (for whatever reason) the occupying power. Under this simple scenario, each civilian death is a black mark against the occupier. This is the point to which the Op-Ed is directed (at least to my reading-between-the-lines).
Um... in reading my comments above, I would like to point out that I was originally writing to people who have Jc's apparent viewpoint in this issue. Not to Mr. Lambert or those who agree with the Op-Ed above. (Sorry for potential misunderstandings.)
jc:
As horrid as this suggestion is to myself, given your economics background, you should at least be able to bracket how many Iraqi deaths would be acceptable for the purported current policy goal? Consider it a full cost/ benefit analysis.
Then ask yourself, is this value exceeded by the second most standard deviation from the mean of the numbers cited by Les Roberts or ORB. If so, I would consider the polling accurate.
Mike
Re #8. I muddled up that second paragraph. What I meant was the values of the estimate exceed the second standard deviation, then one might understand the outrage.
Mike
Jc, you ask "Have you even heard of one single day in this sorry mess of a war when there have been more than 250 Killed." This implies that you believe these excess deaths are solely due to the direct effects of war. The excess deaths are due to a whole host of effects, some primary (bombings, shootings), some secondary (increased mortality from infectious disease, etc.), althought, truth be told, the 2006 Burnham et al. study did estimate about 90% of these were violent deaths. Just wanted to clear that little point up.
In regards to media reporting of violence, you're a bright fellow, can you give the discussion anything aside from the argument via personal incredulity? That argument gets really boring, really quick.
Why not just declare every person in Iraq effectively dead, lay the blame at W's feet and move on. I'm sure some methodology could be quickly cobbled together that would suit the low threshold of scrutiny displayed by anyone convinced that 1.2 million have perished to date. What's one additional order of magnitude to those anxious to bludgeon their political adversaries with missapplied statistical slight of hand?
1.2 million is roughly the same number dispatched by the Khmer Rouge during their four year reign of terror. They were systematically executing, starving and working to death thousands of people every week. Dose anyone really believe that our presence in Iraq is causing slaughter on this same scale?
As an aside, the population of Cambodia numbered ~7 million in 1975, while Iraq's population numbered somewhere around 27 million in 2003 (I think). Standardization of mortality rates by population size can be a useful tool.
Josephus,
I made no per capita comparison. The rate of deaths per week is independent of the size of the population. Does it make sense to you that the American presence in Iraq, devoted almost entirely to stopping violence, has resulted in the same rate of slaughter as the purposely murderous efforts of the Khmer Rouge?
Jc, I was specifically referencing your quote "Let's immediately change the estimates on Iraqi dead numbers to 1.2 million". Change the estimate from what? You obviously have another number in mind. What is it, and where id it come from?
Lance,
Prima facie, you are correct that the rates of death per week is not dependent on the population size ;) My point is that it is typically more acceptable to construct mortality statistics that are normalized to the entire population (In the case of Iraq, one might be tempted to exclude much of Kurdistan from this statistic). Doing this results in 4+ years of the occupation generating a normalized mortality rate approximately one quarter that of the 4 year reign of terror in Cambodia (if we have any confidence in our numbers).
You ask if it makes sense to me that such violence could occur as a result of the myriad factors of war and subsequent occupation. How am I supposed to know? All I can do is look at the studies. The Burnham et al. paper and ORB poll show an astounding level of violence. The methods that they use, especially those of Burnham and Roberts, have come under quite a bit of criticism, and as far as I can tell have weathered it quite well. Thus, I am inclined to rely on those numbers more as I try and understand what is happening halfway around the world. What use is it to play Gedankenexperiment if you have no sense of what is going on other than what spills out the mouths of a bunch of groups that have a vested interest how you perceive all things "Operation Iraqi Freedom"?
"Avoiding the issue of Iraqi deaths will likely come back to haunt us as young people in the Middle East grow up with ingrained hostility toward America."
Like they need any help. They've had decades of grievances with us of one sort or another. They've got the MSM helping them feel this way with their Palywood productions. And, well, heck, we're just the Great Satan. How can we win, except to roll over and show them our bellies?
Poor us. After all we've done for the Middle East, to think that people there might not like us. It must be that dratted media.
There is some spectacular avoidance going on, if the average American thinks 9,900 Iraqis have died violently from the occupation. Any booster of the American forces would imagine that the death ratio should be at least 10 to 1, and even Bush, years go, offered the IBC figures of the time (20 or 30,000).
A bad conscience leads to bad statistics. Innumeracy and avoidance are a fascinating combination. And note how no-one is seeking more accurate figures - where's the George Mason U. crowd?
"The evidence suggests that the extent of underreporting by the media is only increasing with time."
Roberts' point is backed by anecdotal evidence of reporters retreating to the green zone or some other fortress to protect themselves from random violence or targeted attacks.
It's foolish to put your confidence in media reports when the reporters themselves are isolated out of fear for their own lives.
Well, actually, the two studies published in the Lancet by Roberts have been subjected to *intense* scrutiny in an effort to shoot them down.
Tim
You first published Roberts report a few years ago. A quick calculation at the time implied the daily death rate under Roberts was around 250 per day. He's now ratcheted that up to 820 per day. At the rate roberts is going and allowing for compounding we will run out of Iraqi citizens and most of the Mid East by july 22 2011.
I can see the report in 2009:
Les roberts thinks the current death toll estimate of excess deaths is 39 million from a population of 27 million. Les thinks the birth rate will catch up ot the death rate in a few years to equalize the toll count estimate.
In other words at what point in time will the Roberts survey imply the entire Iraqi population are excess deaths.
I think it's important to know.
Whether that number is correct or not, there is a lack of will to find out from those prosecuting the Iraq war and I agree it should be in their long term interest to know. There would then be some slim hope that clear minded review might reveal important lessons.
It's pointless to argue with morons like jc, as they will never ever admit that they were wrong. All it does is tying up comment threads in pointless discussion.
And your comment was "pointful", Martin? The only thing I see as moronic is your abusive comment that adds nothing to the discssuion in terms of offering evidence to the contrary.
Roberts estimate is an outliar (sorry, wrong spelling). You need a little more than abuse to present a convincing case that the tail end of a bell curve is accurate, you dope.
It's also interesing it seems that it was the Baitimore Sun that carried the op-ed. Last time I looked the BS made the Guardian like a right wing rag.
I wonder how many papers turned him down in carrying that piece on their Op-ed. Anyone know if the NYTimes turned it down and why?
[It's also interesing it seems that it was the Baitimore(sic) Sun that carried the op-ed]
does anyone know which US town is home to Johns Hopkins University? could it be ... yes it could!
[Roberts estimate is an outliar (sorry, shit joke)]
no it isn't. It's right there or thereabouts with *lots* of other estimates.
The comments policy on Deltoid is in general against personal insults, but I think it is fair to say that for someone who seemingly has such strong opinions on the Lancet survey, for JC to be making so many false statements about it tends to support only the very most unfavourable assessment of his character, intelligence or both.
Of course , Dsquared, Roberts wouldn't have even thought of hawking the Op-ed piece to the Times or the Wash. Post just down the I 95 because he was being loyal to his home town paper. You know that for a fact?
While you're there you may want answer to these "facts".
When Roberts first published the massive death toll, the run rate was about 230 excess deaths per day. It is now running at 820 per day. This implies the death rate has taken off vertically over about 2.5 years. Do you or Roberts have any explantion for this incredible compound rate.
Let's assume he was right on his first count. I'm not saying he was but let's do so for the sake of argument.
That means, according to Roberts there have been 1,075,000 extra deaths since his very first "survey".
1,200,000-125,000= 1,075,000
That was about 2 1/2 years ago.
Therefore according to Les, the run rate in those 2 1/2 years has been close to 1,200 people a day.
Please discuss Dsquared. Let's not use Soviet Statisical analysis methods if we can avoid them please.
It is telling that detractors of Roberts' surveys have taken to ignoring even the anecdotal evidence coming out of Iraq to maintain their incredulity.
Kind of like the 911Truthers have to ignore the experts to maintain their belief that an airplane crash and fire could never destroy a skyscraper.
"This implies the death rate has taken off"
You got it.
Jc,
(1) The ORB poll (which estimated deaths due to direct violence) was finished in September of 2007, and the end of the first Roberts/Burnham study (which measured all excess deaths post invasion) was in September 2004, so the time interval is closer to 3 years.
(2) Their central estimate of the excess mortality without the contribution of Fallujah was 98,000, not 125,000, with about 90% of these deaths directly attributable to violence. With the inclusion of the Fallujah cluster, the central estimate jumps up to something on the order of 300,000 (I'm pretty sure the proportion due to direct violence gets much closer to unity).
The question of whether to include the Fallujah cluster actually makes a bit more uncertain the issue of a "jump" in mortality rate. Anyway, the point, for me at least, is that until I see some empirical evidence of these being VAST overestimates, my country and others have some real hard explaining to do to the people of Iraq, plain and simple. This explaining includes accurate accounting of the blood costs of modern urban war.
Jc at his best. he is showing off his total inability to make a point again!
even though he wrote about a quarter of the comments so far (7 of 29), and even though he has been called to state his oppinion on a valid deathtoll number or range of numbers, he has denied to do it.
understandable, as his highish selfperception will always take some massive hit, when he tries to state a fact. so he sticks to blurring nonsense.
of course every single argument he made in his 7 comments is plainly FALSE!
Let's immediately change the estimates on Iraqi dead numbers to 1.2 million because your dubious poll and a coupla of others disagree with our estimates as a way of winning the hearts and minds of Iraqi civilians in the middle of a war.
no one suggested to use the 1.2 mio number. Roberts is a serious guy, who uses a RANGE of number, to describe the deathtoll in Iraq. for an iraqi, it will be much WORSE to hear, the americans think that only 10000 of them have died, it confirms their feeling that an iraqi life is not worth a lot.
the underestimate of the deathtoll is influencing the way the war is fought and so for is the CAUSE of more death!
I don't need to go with any numbers seing I don't know..... like Roberts doesn't either by the sounds of things.
Roberts did a survey. you are talking out of you ass. slight difference.
Have you even heard of one single day in this sorry mess of a war when there have been more than 250 Killed. I once heard that number over a two day period about 2 years ago.
funny, but yes, i hear of such events. star with the Yazidi bombing.
http://tinyurl.com/39at3r
or th baghdad bridge stampede:
http://edition.cnn.com/2005/WORLD/meast/08/31/iraq.main/
ignorance is bliss! about how many dead by car accidents do you read in the NYtimes everyday?
Polling firms run sample audits to verify if their pollers are actually doing the hard yards. What audit prcoess was carried out? You know?
dear Jc, if you think hard, you will notice the difference between a commercial shopping mall poll on the flavor of cherry coke and a scientific study on killed family members in a warzone.
in science studies are affirmed or contradicted by REPLICATION. the lack of any replication attempt by deniers is very telling.
When Roberts first published the massive death toll, the run rate was about 230 excess deaths per day. It is now running at 820 per day. This implies the death rate has taken off vertically over about 2.5 years. Do you or Roberts have any explantion for this incredible compound rate.
the pentagon has confirmed a tripling of numbers of civilian casualties between jan 06 and the first half of 2007.
http://www.democracyarsenal.org/2007/09/fuzzy-numbers-a.html
(btw, the pentagon has "revised" this number, because it was contradicting the petraeus report..)
Les roberts thinks the current death toll estimate of excess deaths is 39 million from a population of 27 million.
is this an attempt to show of your ignorance? hint: please do some research on "limited growth"..
You need a little more than abuse to present a convincing case that the tail end of a bell curve is accurate, you dope.
good argument!
now all you have to do is to list all the peer reviewed studies on iraqi casualties rate, that make the rest of your "bell curve"
Martin, it's worth reviewing the discussion how to set up a 'killfile' script, so those with strong opinions and no data don't take over the threads. See post 50 and following of the 'Wonder if they'd bet on it' thread. It does wonders for your mental health. raises the proportion of data to invective, lowers your blood pressure, and washes dishes.
Iraq will haunt the US for years to come, longer if they cannot see the consequences. Some will argue that a million external refugees must mean the death toll is lower than claimed, but others may point out that it does take a lot to make a million people leave their homes over 2 to 3 years, and a high toll of violence in a civil war with the breakdown of society is one of those precipitants.
I don't know if the estimates are correct, I only know that they are estimates of form widely accepted in many other settings, and the argument from disbelief only tells what the writer is prepared not to believe. (I'm forced to belief my parents had sex 3 times, as I have 2 siblings, but certainly not more than that). As soon as a better estimate, using accepted methodology comes along, I'm willing to learn more about it. If it agrees with the Johns Hopkins estimates, that's important. If it disagrees, that's also important, and knowing why is crucial.
As for the multiplier from coalition deaths, remember that a) US forces are present but spend much time in isolated bases, not in the communities, b) death tolls don't include the 'contractors', who had 1001 deaths to the start of June 2007.
Jc asked:
Hmmm. [I wonder if any other rate has increased since August 2004](http://anonymous.coward.free.fr/misc/deathrates.png)?
Josephus:
At Robert's run rate we will run out of excess Iraqi dead in about 3 years time - that's all 27 million of them. There wouldn't be any people in the Mid in 10 years. This cumulative compounding is impressive.
Sod:
I'm sorry I find all your comments too boring to read and usually have nothing to add.
Robert
An increased US body count is mostly the result of roadside explosives.
"1.2 million is roughly the same number dispatched by the Khmer Rouge during their four year reign of terror. They were systematically executing, starving and working to death thousands of people every week. Dose anyone really believe that our presence in Iraq is causing slaughter on this same scale?"
They were also operating in a country with around 8 million people versus 25-28 million in Iraq.
@JC
And of course, that increase says nothing about the overall level of aggression, right?
Sod: I'm sorry I find all your comments too boring to read and usually have nothing to add.
plain truth. usually you have nothing to add. couldn t have said it better.
but let me sum up my post for you:
the few things you said were all WRONG.
Robert, An increased US body count is mostly the result of roadside explosives.
yes Robert, how could you dare to point out that US military casualties increased by a similar factor.
actually you would be hard pressed to find a SINGLE set of casualty numbers, that do NOT show an increase by 2 or 3 times.
"When Roberts first published the massive death toll, the run rate was about 230 excess deaths per day. It is now running at 820 per day. This implies the death rate has taken off vertically over about 2.5 years. Do you or Roberts have any explantion for this incredible compound rate."
Even the US military admits that there3 was a several-fold increase in violent incidents after the bombing os the al-Askaria mosque.
"There is some spectacular avoidance going on, if the average American thinks 9,900 Iraqis have died violently from the occupation."
The "average" American? Was this a scientific poll? In urban America, 90% of the persons polled "on the street" can't identify the Prime Minister of our closest cultural and trading partner, Canada. They can't tell you who the second president of the USA was, and they can't tell you what the Lancet says about numbers of deaths in Iraq.
Neither can the "average" person of any country that I'm aware of.
Ian
Do you think the deaths form that Mosque are what was missing from other tallies to reconcile with Robert's claims? Wow so about 1,1050,000 people died? That's a pretty big Mosque.
Ben,
Seriously Ben, no offense if you're Canadian but why would any American wanna know the name of the PM of the most boring nation on earth.
Do you recall the movie Canadian Bacon? Alan Alda was playing the US Prez having some poll troubles and he couldn't even get face time on CNN to declare war on Canada. No one cared, not even Canadians.
As for the second Prez. He's dead and can't cut taxes. Who cares?
JC, as is his wont, offers nothing more than an argument from personal incredulity. 'But... but... but... [spluttering] the numbers are really big, and the rate is increasing, so this must be wrong.'
Well, yes the numbers are really big, and the rate is increasing, and THAT IS THE FUCKING PROBLEM, JC!
Three separate broad sample-based estimates of mortality have now been conducted, and the three offer results that are consistent among them, and those results point to excess mortality in the vicinity of upwards of a million people to date. Every estimate deriving a lower number is from a method known to under-estimate by some unknown amount.
so JC, making an argument at odds with all the known applicable data (and no, counting daily reported deaths in the newspaper halfway round the world from the war, is not a good way to derive death estimates - duh!) - JC is reduced to filling up the thread with his unsupported blather hoping to overwhelm us by sheer weights of words.
Sorry, JC, lots of words wont bring the dead back to life, or reduce my country's culpability in allowing this fucking tragedy to happen.
Lee:
We don't know enough to make any estimate of the numbers dead. There have also been numerous estimates that have come suggesting the figure is around 50,000 (approx) others at 100,000.
Going from memory the UN estimate about year ago was about 75,000, if I recall correctly.
I don't know the number, you don't aand neither does Roberts. Roberts, however is playing agiprop for political purposes.
Josephus,
Your point is well taken. Still it assumes uniform violence around the entire country which as you point out for the Kurdish regions, among others, seems highly unlikely.
Ian,
As I replied earlier to Josephus, the rate of deaths per week is independent of population size. You still have to account for nearly five thousand deaths per week over the baseline mortality rate. There were vast "killing fields" in Cambodia full of the victims of the Khmer Rouge. Sadam is believed to have killed far less Shiites yet there are mass graves scattered around Iraq. Where are the 1.2 million EXTRA Iraqi bodies claimed in the Lancet study?
Check the cemeteries, Lance.
The "average" American? Was this a scientific poll?
AP-Ipsos:
The Associated Press-Ipsos poll on public attitudes about the Iraq war and U.S. policy in Iraq was conducted Feb. 12-15 and is based on telephone interviews with 1,002 adults from all states except Alaska and Hawaii.
results in this article:
http://www.msnbc.msn.com/id/17310383/
Jc, i know you are bored by my constant corrections to the nonsense you are writing. and it is even obvious, that you find FACTS to be boring, in contrast to your own "creative" vision of the world. but i can t spare you this:
Ian, do you think the deaths form that Mosque are what was missing from other tallies to reconcile with Robert's claims? Wow so about 1,1050,000 people died? That's a pretty big Mosque.
Ian used the term "AFTER" in his post. please learn to READ!
We don't know enough to make any estimate of the numbers dead. There have also been numerous estimates that have come suggesting the figure is around 50,000 (approx) others at 100,000.
LINKS, please! you are seriously mixing up BODYCOUNTS in the press, the famous "president Bush estimate" and a scientific estimate of the real deathtoll.
Going from memory the UN estimate about year ago was about 75,000, if I recall correctly.
again: COUNTS will NOT give you a real estimate of the situation!
and: LINK, please!
I don't know the number, you don't aand neither does Roberts. Roberts, however is playing agiprop for political purposes.
again: there is one minor difference: Roberts has conducted a scientific study on the issue. YOU HAVE NOT.
As I replied earlier to Josephus, the rate of deaths per week is independent of population size.
no. the number is basically meaningless without the population number. there might be more "common" criminal murders commited in China, than might occur in civil war in a tiny country.
Where are the 1.2 million EXTRA Iraqi bodies claimed in the Lancet study?
listen to Mr. Roberts:
City officials in Najaf were recently quoted on Middle East Online stating that 40,000 unidentified bodies have been buried in that Iraqi city since the start of the conflict.
"the American presence in Iraq, devoted almost entirely to stopping violence"
HA. HA. HA.
"We don't know enough to make any estimate of the numbers dead."
Really? Perhaps you missed paragraphs 3 and 4 of the linked piece, which suggest that quite a few people do actually know enough to make estimates of the numbers dead. In fact, sometimes it seems like the only people who don't think they know enough to make an estimate are those who'd like to ignore the issue. Which, ironically enough, was the point of the op-ed in the first place.
I think it says something about Roberts ability to be impartial that he thinks the February BBC poll and the September ORB poll roughly verify each other. Since "physically harmed" includes nonfatalities, the BBC poll implies a death toll less than half that found in the September ORB poll.
And neither really tells you what the number of civilian deaths was.
Wll: "...the BBC poll implies a death toll less than half that found in the September ORB poll."
Methinks you are making some strong assumptions about the BBC poll. In any case what Burnham and Roberts actually say is that ORB roughly verifies the earlier, less precisely worded BBC poll. That's not at all the same as saying that they verify each other.
I think it says something about Roberts ability to be impartial that he thinks the February BBC poll and the September ORB poll roughly verify each other. Since "physically harmed" includes nonfatalities, the BBC poll implies a death toll less than half that found in the September ORB poll.
as kevin pointed ou above, you misunderstood what he said.
are you awawre of the concept of error range?
And neither really tells you what the number of civilian deaths was.
was your killed son a terrorist? is NOT a good question for a door to door interview.
both polls do NOT support a "50000" deathtoll in Iraq till now.
QHow many members of your household, if any, have died as a result of the conflict in Iraq since 2003 (ie as a result of violence rather than a natural death such as old age)? Please note that I mean those who were actually living under your roof.
None 78%
One 16%
Two 5%
Three 1%
Four or more 0.002%
Given that from the 2005 census there are a total of 4,050,597 households this data suggests a total of 1,220,580 deaths since the invasion in 2003. Calculating the affect from the margin of error we believe that the range is a minimum of 733,158 to a maximum of 1,446,063
http://www.opinion.co.uk/Newsroom_details.aspx?NewsId=78
Lance,
You are talking about the same Pol Pot who was supported by the US, UK and other western countries at the United Nations when Viet Nam invaded the country in 1979 to put an end to cross border incursions? Perhaps your memory hole is much too large:
http://www.zmag.org/zmag/articles/hermansept97.htm
While on the matter of genocides and mass murder, how many Indochinese died under US bombs during the Korean and Viet Nam wars? Hazard a guess? Or how many victims were there under one of the world's biggest torturers and mass murderers, Suharto, also fully supported by the west for more than 30 years?
And let us not forget who armed and supported Saddam during the time he committed his worst crimes. Heck, Reagan's administration was largely responsible for taking Iraq off of the list of countries that sponsor terrorism in 1982 because the US wanted to support Saddam's war against Iran. So many crocodile tears, so little memory.
"Where are the 1.2 million EXTRA Iraqi bodies claimed in the Lancet study?"
The last tiem this que3stion was asked here I went throguh various media reports and assembled numerous first hand accounts of new cemeteries beign opend up; of gravediggers reporting they were burying for more people than pre-war.
I als pointedo ut tht reported increase in fatlaities recieved at the Central Baghdad morgure is consistent with the increased mortality reported by the Lancet studies.
If somr supporter of the sudanese government tried to deny the genocide in Darfur by demanding "where are the bodies?" how much credence would you give him?
How about the two million estimated dead in the congolese civil war?
One of the msot frustrating aspects of this debate is that we keep having to go over the same ground.
Supporters of the war reading the Lancet artivcle and go "these anti-American Saddam-lovers are tryimg to claim that the US murdered a million Ieaqis!!!"
No what the reports are saying is that TOTAL excess mortality is one million or more since the invasion. That includes Iraqi military casualties in the initial military campaign; civilian victims of the various ethnic and religious militias; victims of disease and victims of the massive and ongoing increase in plain old crime since the virtual collapse of the Ieaqi police.
Peopel keep saying "well we would never kill ommocent peopel like that" - apparently they have equal faith in the high moral standards of Al qaida in Iraq; the Mahdi Army; the kurdish Peshmurga and dozens if not hundreds of other Iraqi factions.
"Ian
Do you think the deaths form that Mosque are what was missing from other tallies to reconcile with Robert's claims? Wow so about 1,1050,000 people died? That's a pretty big Mosque."
Joe TRY to read for comprehension.
You said, in effect, "for the additional deaths between Lancet 1 and Lancet 2 to be correct, the level of violence would have ot have increased several-fold".
That's exactly what the US army has been reporting over the past year. The "surge" has only succeeding in returnign violence ot the pre-Samarra levels.
Let's try a thought-experiment, people.
You have a wretchedly poor country where over half the country is dependant on a government food dole to survive.
First, the dictator of the country, releases virtually the entire prison population and the majority of people in mental institutions.
Then he distributes millions of fire-arms to the population.
Then a foreign power invades.
The foreign power;
a. sacks the entire military and police force but leaves them in possession of their weapons;
b. cuts off the food dole for most of the population and scales it back severely for the rest
c. leaves weapons caches with millions more fire-arms unguarded.
What do you think is goin to happen?
Ian G., aren't you describing the scenario in various libertarian fantasy movies like Red Dawn? Why, naturally the freedom-loving inhabitants will drive out the invaders (preferably in one large battle, but if necessary in ongoing skirmishes), then the free market will spontaneously arise, there will be minimal government regulation, a basic police force, everyone will have health savings accounts, and Iraq will have the world's largest GDP and best educated population lowest infant mortality, etc within 15 years. They will erect statuses of Bush and Hayek at every street corner.
Why do you even have to ask? Isn't that how it always goes? It's like, an organic rule of history, dude.
Roberts and Burnham want people to be "forthright." Unfortunately, they still refuse to share the underlying data from their papers with all their academic critics. (They claim that the data from 2004 is "no longer available" --- I have no idea what that means --- and still won't show the 2006 data to statisticians like Gesine Reinert and others. Forthright is as forthright does.
For those who want another estimate, I still like Jon Pedersen's estimate of 100,000 violent deaths through July 2006. Perhaps members of the Deltoid community can explain how they know more about Iraqi mortality than Pedersen.
And, for those interested, I will be organizing a session on the Lancet surveys at the JSM in Denver in August 2008. If you are an academic with a paper to present, please contact me.
David, where is your call for Pedersen to release the underlying data behind his 100,000 number?
Is there any evidence at all that Burnham has refused to give the 2006 data to Gesine Reinert, other than David Kane's say-so? Has she even submitted a request?
"Perhaps members of the Deltoid community can explain how they know more about Iraqi mortality than Pedersen."
Well for my part I know better than to revise an "estimate" upwards by 50 percent without offering any explanation whatever for the revision. As those who have been paying attention will know, "Pedersen's estimate" jumped from 100,000 to "perhaps" 150,000.
To be fair to Pedersen, he never suggested that either number was actually an estimate - that's just another misrepresentation by David Kane. As the "perhaps" indicated, it was just a guess, an expression of a gut feeling. But I think that it showed poor judgement to oppose his gut feeling to a study by highly regarded researchers in a different field. He would hardly be impressed if an epidemiologist dismissed his work on living conditions in such an offhand manner.
1) Gesine Reinert is part of Michael Spagat's group working on Main Street Bias. (Or maybe Spagat is a part of Reinert's group. I don't know who, if anyone, is in charge. You can read about their attempt to get the data here.
2) Tim Lambert jokes:
Very funny! First, Pedersen's estimate is based on his overall experience with the region and with surveys especially his work on ILCS. Second, Pedersen, like any real scholar, has shared his ILCS data with everyone who has asked him for it. I have found the data to be quite helpful.
3) Contrary to Kevin's claim, Pedersen has never raised his estimate by 50%. He estimated 100,000 violent deaths (much less than the 600,000 from L2) and was comfortable with 50,000 non-violent deaths as a side-effect of the war and occupation, similar to what L2 found.
4) Kevin claims that I am misrepresenting Pedersen. Having met with Pedersen, I know that I am not. Here is the exact quote.
What word instead of "estimate" would Kevin like me to use? He is making a direct claim that L2 is wrong, at least when it comes to violent deaths. You may think that he has no basis for doing so, but try not to pretend that he didn't say what he said.
Pedersen is guessing, David. Nir Rosen is a reporter who's continued to go into Iraq and writes rather long stories in the Boston Review about it and he thinks the Lancet numbers are plausible, so if we're pitting experience against experience, I'd favor his. But that's not an argument that should convince anyone, any more than yours should.
Pedersen's data is from early 2004 and he admitted to Stephen Soldz that they didn't send teams into Fallujah at the time because it was too dangerous, and he also guessed they probably didn't sent interview teams into other areas that were violent at the time, and his ILCS question on violence was rather vague, so I don't think he actually has very much other than his opinion to go on.
Not that I have a firm opinion on the death toll--I'm a little unsatisfied with how the various pieces of evidence fit together.
David Kane: Contrary to Kevin's claim, Pedersen has never raised his estimate by 50%.
Oh yes he did. When questioned by Joe Emersberger, Pedersen switched from 100,000 to 150,000 for no apparent reason. Here is Emersberger's account which includes this question and answer:
Joe Emersberger: The Washington Post quoted you as [saying] that the Lancet study's findings [were] "high, and probably way too high. I would accept something in the vicinity of 100,000 but 600,000 is too much.' [...] Were you quoted accurately by the Washington post?
[...]
Jon Pedersen: yes, more or less, would perhaps say 150,000
David, the one who is trying "to pretend that he didn't say what he said" is you.
As to the MSB group, it is common knowledge that Michael Spagat asked for the data and was refused. Your claim upthread was that Gesine Reinert was refused. The link you give does not support that claim.
"Pedersen is guessing"
All of the "experts" are guessing. The most blatant reason that Lancet is a guess is that a group of interviewers just handed Roberts data and said they followed random selection methodology. That claim was not checked. Just believed. They said they recorded death certificates for almost all of them. That claim was not checked. Just believed. Nothing about any of it was ever checked. Just believed.
No survey researcher worth much and who had even the slightest interest in accurate results (rather than just usefully high results) would send interviewers out to pick respondents and do interviews without supervision, let them destroy the records of where they supposedly surveyed so that their work could not be checked, therefore have it obvious to all of them that their work would not be checked, and then just say 'we'll believe anything you give us'. Falsification by interviewers is far too common even when discouraged. That approach encourages it. And such risks would only multiply with such a controversial survey topic and with interviewers who live in the war zone and may have very strong interests in the outcome.
Any results would only be as strong as the very dubious belief that this approach would produce accurately randomized and faithfully reported survey data. (and this is aside from all the methodology issues that have been raised).
Then ORB doesn't even have a methodology. Can anyone tell me their procedure for selecting respondents? selecting locations from which to pick respondents? They just assert that their sample is representative, and then come back and say that well, it wasn't really and this is just an "estimate". You might as well pull a number out of a hat for as reliable as that is.
On another topic, a few above ask about where are all these bodies, and one answer is this 40,000 in Najaf claim. There have been a lot of stories about this cemetery. It's a large national cemetery, very holy in the Muslim world, where bodies are sent from all over Iraq (mostly from Baghdad), and even from outside Iraq. Such a figure would not support Roberts' speculations. It would probably be a lot more consistent with Pedersen's. Good thing Roberts and Burnham left these details out of their piece, I guess. The trick doesn't work as well if people know them (but then maybe R&B didn't know them either).
I don't think 40,000 "unidentified" bodies in the Najaf cemetary supports a violent death toll of 100,000.
As for the rest of your complaint, yeah, the interviewers could have lied. I think that's in the back of people's minds. There's always the possibility of fraud in any scientific paper--it's news to me that this means any published result is no better than a guess. It might have been a little hard for the Lancet authors to check up on the results, you know, given that there's a war on and unarmed Westerners wandering around Iraq might be noticeable to the wrong people. I have an idea though. Why don't you do it?
I don't think that
Donald, 40,000 to the largest national cemetary, with bodies coming from Baghdad and all over Iraq (and outside it) seems quite a bit more consistent with 100,000 than with 1 million. But then that's a number Pedersen gave a year ago. I suspect he does not believe nobody else has been killed since. I would also not take one comment from one report so literally. "unidentified" may be a generalization because most that are shipped there are that way. 40,000 is a nice round number too.
"There's always the possibility of fraud in any scientific paper--it's news to me that this means any published result is no better than a guess."
What I said does not apply to any published result. There are means by which such possibilities are minimized. And there are means by which they are maximized. Lancet includes nothing of the former and pretty much everything of the latter.
And if conditions do not allow for quality checks on the data, then the conditions don't allow the collection of quality data. Whether that's understandable or not, at the end of the day you still don't have quality data on which to draw conclusions.
Spell out what you mean. How did the Lancet bias things towards large numbers? I don't think the normal methods of ensuring there is no fraud could be employed in Iraq, but that doesn't mean it wasn't worth doing. Your idea seems to be that we throw up our hands and accept the much smaller numbers put out by the press, though there are good reasons for not trusting those.
And yeah, I know Najaf is a large cemetary, and kinda suspect that when someone says "40,000", they're probably rounding it. But yes, if they rounded it up from, say 4000 (an unusual arithmetic procedure) or if they wildly exaggerated the number that were unidentified, then I'd find the 100,000 figure more plausible.
You say all the numbers are guesswork, but it appears you only mean numbers that are much larger than the numbers you prefer. If you were a more straightforward agnostic about ALL the numbers, I'd find it hard to disagree with you.
Then ORB doesn't even have a methodology. Can anyone tell me their procedure for selecting respondents? selecting locations from which to pick respondents? They just assert that their sample is representative, and then come back and say that well, it wasn't really and this is just an "estimate". You might as well pull a number out of a hat for as reliable as that is.
this is not true. polling in Iraq is EXTREMELY difficult. the ORB was obviously doing the best they could, taking regional samples, mentioning the parts they had to exclude and doing proper work with the data and giving away the raw data.
if you want more, please do the research BY YOURSELF!
It's a large national cemetery, very holy in the Muslim world, where bodies are sent from all over Iraq (mostly from Baghdad), and even from outside Iraq. Such a figure would not support Roberts' speculations.
UNIDENTIFIED BODIES!"
why would people send unidentified bodies to the "very holy" graveyard?
that makes ZERO sense!
Donald, I don't know what assumptions you're making about this Najaf number, or why. UNAMI reports in 2006 said things like in Baghdad "the number of unidentified bodies was 1,229 in November and 1,397 in December"
If Baghdad sends their unclaimed bodies to this cemetery, which seems to be the common practice, 4,000 bodies would have been exceeded in pretty short order. 40,000 over the whole war does not really seem inconsistent with what's been reported about Baghdad figures. And that is besides that other morgues in (and outside) Iraq send bodies there too.
sod writes:
"why would people send unidentified bodies to the "very holy" graveyard? that makes ZERO sense!"
Well, read the many reports about it and tell the morgue in Baghdad and the guys who run the cemetery in Najaf that they make no sense:
http://www.timesonline.co.uk/tol/news/world/article634194.ece
Well, read the many reports about it and tell the morgue in Baghdad and the guys who run the cemetery in Najaf that they make no sense: http://www.timesonline.co.uk/tol/news/world/article634194.ece
this is ONE report of some sheik, driving there twice per week with ONE lorry. do the math!
it is absolutely impossible to bring the low deathtoll numbers and the "40000 unidentified bodies on Najaf graveyard" stories together.
if the iraqi total deathtoll so far was 50000 people, you need to assume that 80% of them were killed without being identified and 100% of those were send to Najaf for burial.
if the iraqi total deathtoll so far was 100000 people, you need to assume that twothirds of those are unidentified bodies and two thirds of those are driven to Najaf!
The "some sheik" is the guy in charge of collecting the unclaimed bodies from the Baghdad morgue (and others) to bring them to Najaf for burial:
http://www.cnn.com/2007/WORLD/meast/09/10/iraq.deadbodies/index.html#cn…
The type of figures I cited above are not inconsistent with the 40,000 figure. And that figure was given in one report last month. I don't know who was saying that the total iraqi death toll by last month was 50,000.
And for the 100,000 I think you need to do your math again.
There's no apparent inconsistency with the "low" numbers. It only appears that way if you think this is just any old cemetery where people from Najaf are buried, rather than the main destination for unclaimed bodies from Baghdad morgue and other places in Iraq.
40,000 unidentified bodies from Baghdad--you'd have to assume this is most of the unidentified bodies, that the number of identified bodies is comparable, and that the rest of Iraq is comparatively peaceful, or alternatively, that this is most of the bodies in Baghdad and the rest of Iraq is comparable. Then you'd get 100,000 or so.
Also, to believe the IBC numbers you'd have to assume that in one of the most dangerous wars for journalists in history, somehow they've managed to count the majority of the deaths and that a medical system which is falling apart is keeping score as well. Unlikely, no matter what one thinks of the surveys.
The "some sheik" is the guy in charge of collecting the unclaimed bodies from the Baghdad morgue (and others) to bring them to Najaf for burial:
the sheik is NOT in charge. he is a volunteer. we are talking about literally TONS of dead bodies being moved. nowhere in the text you quoted does it say that he is transporting ALL unclaimed bodies from Baghdad. nowhere does it say that he is transporting bodies from elsewhere in Iraq.
I don't know who was saying that the total iraqi death toll by last month was 50,000.
Iraq bodycount is cited quite often by lancet "sceptics" and has its lower number by 75000.
http://www.iraqbodycount.org/
And for the 100,000 I think you need to do your math again. There's no apparent inconsistency with the "low" numbers.
no, YOU do some math. i am really tired of people who deny to give any numbers. i gave you the calculations above:
you need to assume that two thirds of the victims of the civil war in Iraq are NEVER identified AND that two thirds of those are transported to Najaf.
again, among the victims, only ONE in three is identified. and two out of three unidentified bodies, where ever in Iraq they are found, are ALL transported to that ONE graveyard in Najaf.
there is ZERO evidence, that EITHER one of these claims is right!
Remember that Pedersen's estimate of 100,000-150,000 dead is more than a year old. Extrapolating by the change in the IBC total since then, that would be 200,000-300,000 today.
That number fits the March BBC-D3 survey as well.
Remember that Pedersen's estimate of 100,000-150,000 dead is more than a year old. Extrapolating by the change in the IBC total since then, that would be 200,000-300,000 today.
i think it is a major problem with the lancet "sceptics", that they do NOT update their numbers, but keep sticking to the lowest estimate in their mind.
it would be very reasonable to take some average between a "sceptic" number (one being based on serious analysis. the Pedersen number is NOT!) and some extrapolated Lancet number as a "conservative" estimate. we would end somewhere between 400000 and 800000.
the difference to the UPDATED lancet report, being a serious scientific study is striking!
talking about your numbers Will, how do you come to the conclusion that the last year brought as many deaths as the rest of the war?
does IBC have a better data overview than this one?
http://www.iraqbodycount.org/database/
considering your idea, i think it is dangerous to extrapolate real data from the IBC count, as it is media based. the IBC data has a Baghdad bias, so the offensive in Baghdad has most likely influenced IBC numbers MORE, than numbers all over Iraq. this should be taken into account.
It only appears that way if you think this is just any old cemetery where people from Najaf are buried, rather than the main destination for unclaimed bodies from Baghdad morgue and other places in Iraq.
late response to this part:
i doubt that many people are considering the Najaf graveyard as a sign of Najaf deaths alone. still a graveyard with 40000 "unidentified" bodies is SIGNIFICANT, when people run around using extremely low bodycount numbers.
Najaf graveyard population is inflated with dead pilgrims as well and the town has seen massive violence in the past.
still your claim that Najaf is receiving significant amounts of unidentified bodies from other parts but Baghdad is COMPLETELY unsupported by any sources!
Will, I think your extrapolations are too large. I think he said these things around late October and early November. I can't find IBC numbers from exactly that time, but this is a little earlier, and says about 43,800-48,700:
http://web.archive.org/web/20061009002723/http://www.iraqbodycount.org/
Today it's about 75,200-81,900. So you'd multiply by about 1.7.
sod writes:
"still your claim that Najaf is receiving significant amounts of unidentified bodies from other parts but Baghdad is COMPLETELY unsupported by any sources!"
The Times article I posted (page 2) discusses Sudani and colleagues' records on bodies from Falluja, Samarra. It's clear they're not just dealing with those from Baghdad and Najaf. This may also be useful:
http://www.globalsecurity.org/military/world/iraq/an-najaf-wadi-al-sala…
The majority would appear to be coming from Baghdad. With the type of figures that have been reported for Baghdad over the years (such as those I cited above), the majority of 40,000 could be accounted for rather easily without assuming some massive number way above those reported. Then others come from Najaf, Falluja, Samarra, elsewhere.
And I think you still need to do your math again. (Hint: 40,000 is not two thirds of 100,000)
From izzy's link:
"Shiites from all over the world, not only Iraqis or Iranians, but Shiites from Pakistan, India, Bahrain, all over the world go to Najaf and they ask to be buried in Najaf close to that mosque."
izzy, do you notice anything about that statement which doesn't square too neatly with your theory?
Here's a hint: "Shiites aspire to bury their dead in its cemetery, which stretches for miles."
Here's another: "It is the holiest and most highly sought-after burial place among Shiites."
Still struggling? How about this: "Corpses are brought from across Iraq, Iran and elsewhere in the Shiite world to lie close to Imam Ali, the cousin of Muhammad and his successor, whose remains are enshrined in a gold-domed mosque."
"Muslims strive to bury the deceased as soon as possible after death, avoiding the need for embalming or otherwise disturbing the body of the deceased."
http://islam.about.com/cs/elderly/a/funerals.htm
"It is important for Muslims to be buried as soon as they are considered deceased, and in some cases there have been recorded cases of Muslims been buried alive."
http://www.urbandictionary.com/define.php?term=Muslim+Burial
"Pre-Burial Rules and Rituals
In the Islam religion, it is customary to begin processes for burying the dead within 24 hours of the death. In non-Muslim countries, this can sometimes be a problem if the death occurs on a weekend or a holiday, when a death certificate can not be attained, and therefore the process cannot start. It begins with a ritual washing of the body; men washing men's bodies and women washing women's bodies. Either a man or woman may wash a child's body, and a husband is allowed to wash his wife's body, and vise-versa, if the need arises. It is preferred that the body is washed by close friends or relatives, rather than hospital staff or undertakers. Imams (prayer leaders of mosques) may also wash men, as well as midwifes may also wash females. Washing was traditionally done at the graveside,"
http://studentwebs.coloradocollege.edu/~l_jenkins/islam.htm
"Q: How soon the deceased be burried.
A: After the Janazah Salah has been performed the deceased should be buried as soon as possible."
http://www.isna.net/services/casc/questions.html
" in islam the dead bodies should be buried as early as possible after death,however this has no relation to the fate or ultimate destiny of the person.in islam we have no choice but to submit to the commandments of God if not then we are sinners and of course this will affect the ultimate destiny of the muslims. "
http://en.allexperts.com/q/Islam-947/burial-2.htm
Notice any pattern here which would indicate that izzy is a lying twit?
Are we really expected to believe that 40K bodies have been imported to Najaf in the middle of a civil war within one day of death?
Are the flying monkeys about to disembark from izzy's colon?
I don't notice any problem with my "theory" (it's more like straightforward observation I think). Presumably you're trying to say that the 40,000 number would exclude Sunnis and Christians. But if all or most of these bodies are "unidentified" it would usually not be possible to make such exclusions even if that was desired. Besides the links I posted (like the CNN one) say that they don't discriminate. They take whatever the morgues have to give them, regardless of sect, and want them all to have a proper burial.
The globalsecurity link was just to show how major a cemetery this is and how many different places both inside and outside Iraq have had bodies going to it over the years. The Times also illustrated this with a couple more specifics like Falluja (a Sunni area) and Samarra.
izzy #77:
Pedersen was responding to the survey done by Burnham et al (AKA Lancet II) which looked at responses up to June 2006. I thought it was probable that this was the date his own estimate refered to. If his educated guess of 100-150 thousand Iraqi dead was meant to apply to a later date, then extrapolating to the present would, as you suggest, yield a lower number than 200-300 thousand.
Will McLean wrote:
and
Before you guys get too far carried away (you're already carried away, I'm just trying to stop you from going too far), Pedersen's "educated" guess is unbacked by any data -- in Kane-speak, it's non-replicable, which is kinda amusing since David is Mr. Replication. In addition, as a basis for estimating post-invasion mortality, the D3 survey is inferior to the August 2007 ORB survey (which itself isn't ideally suited for this sort of thing. (BTW, the February ORB survey suffered from a similar problem as the March D3 survey -- ambiguity about the population at risk)). Bottom line? You guys are taking a wild-assed guess, inflating it by an only-slightly-less-wild-assed factor, and going "W00t!" cuz it's within a ballpark of a lousy estimate.
Perhaps it's time for a roundup of the claims about Najaf:
sod: Iraqis would not send unidentified bodies to this holy graveyard. That makes no sense.
Yes, they would and do.
sod: Bodies do not go there from other places but Baghdad. There's no proof of this.
Yes, they do. And yes there is.
Donohue: But they would only be Shiites.
No, they wouldn't.
And now there's...
elpi: bodies are buried within one day, so izzy is a liar!
unidentified/unclaimed bodies are not just stuck in the ground right away. They are kept for a time in hopes that someone will identify and claim them. If not, then they get sent for burial.
"There are two cemeteries, the first is in Karbala and the second in Najaf, both assigned for unidentified corpses," he told AFP. "After being left for three months at Baghdad's morgue, the bodies are brought here."
http://www.thenews.com.pk/daily_detail.asp?id=61133
I've heard shorter periods like three weeks also, but they are not buried within a day.
I wonder what will be next?
Robert says: In addition, as a basis for estimating post-invasion mortality, the D3 survey is inferior to the August 2007 ORB survey (which itself isn't ideally suited for this sort of thing. (BTW, the February ORB survey suffered from a similar problem as the March D3 survey -- ambiguity about the population at risk)).
I think the August 2007 ORB suffers from this problem too (among others). Johnny Heald described the two ORB polls on NPR:
"In March earlier this year we calculated that 26% of families had suffered a murder of a family member or relative. But that was their families. Iraqi families are very extended. So in August this year we polled 1600 Iraqis and asked them whether anyone who was living under their roof had died, been murdered sorry, since 2003. Not died as an accident, but being murdered."
http://www.npr.org/templates/story/story.php?storyId=14501232
Since we don't know exactly how big the extended family is with the March/February poll it's hard to covert the 26% into any particular death estimate. Thus the ambiguity. However, one thing we can know is that this should cover very much more than one household of six or seven people. As Heald says, Iraqis have "very extended" families, and the question allows for anyone that's related to the respondent in any way.
So, for their two polls to be consistent, the August poll which is attempting to limit to one household should produce a number at least a few times lower than 26%. But it produces 22%, almost the same number. Unless this can be explained just by bad sampling in ORB (I don't know what their sampling method was, they don't describe it), it looks like a lot of people in August were giving the same answers people gave in February, still responding about the "very extended" family, whatever relative they know who was killed, and we have the same ambiguity problem again.
But this would mean the August ORB estimate is exaggerated by at least a few times, since it rests so heavily on assuming that all the answers really do apply to only one household of six or seven people.
But then I also don't know where ORB even got their "4,050,597" household number to apply this assumption, or how accurate that might be either. They say this is from a "2005 census", but I'm pretty sure there is no 2005 census of Iraq. Les Roberts wasn't aware of one in October 2006 when he said: "The last census in Iraq was a decade ago" http://news.bbc.co.uk/2/hi/talking_point/6099020.stm
The Times article I posted (page 2) discusses Sudani and colleagues' records on bodies from Falluja, Samarra.
you got that one wrong. the article says:
The files created for each month show how the Sunni beheaders held sway when their stronghold of Falluja was stormed by US forces in 2004; how Shi'ite squads armed with electric drills stepped up their activities after an attack on the al-Askari mosque in Samarra last February; and how little difference the killing in June of Abu Musab al-Zarqawi, the leader of Al-Qaeda in Iraq, has made to the level of violence.
http://www.timesonline.co.uk/tol/news/world/article634194.ece?token=nul…
it is NOT talking about victims from Fallujah or Samarra showing up there, but of MORE victims being brought in, BECAUSE of the events at those places.
The majority would appear to be coming from Baghdad. With the type of figures that have been reported for Baghdad over the years (such as those I cited above), the majority of 40,000 could be accounted for rather easily without assuming some massive number way above those reported. Then others come from Najaf, Falluja, Samarra, elsewhere.
sorry, the number would still be horrible. a "majority"? so only 20000 unidentified bodies in a town of 500000? heck, even 5000 "original Najaf" unidentified bodies would be 1% of the population of that town!
some math: the sheik is driving there twice per week. over 4 years, that makes 400 trips, with an average of 100 bodies. (i consider ALL bodies coming from Baghdad)
your claim is, that every month from the beginning of the war, they have been transporting 800 unidentified bodies to najaf. on every trip with their "lorry", they transport and bury with their hands close to 10 tons of human remains.
And I think you still need to do your math again. (Hint: 40,000 is not two thirds of 100,000)
hint: my math is fine!
two thirds times two thirds equals 4/9s, which is reasonably close to 40%.
i have asked you several times to provide numbers, you have DENIED to do that. (of course they would look stupid)
1. you need to provide a guess, for the percentage of dead in this war, that end up as UNIDENTIFIED bodies.
2. then you need to provide a percentage number, for how many of that unidentified bodies end up in Najaf.
the two numbers need to multiply to 40%, which basically means that both of them must be higher than 50%!
(unless you go for a rather unlikely 40% times 100% scenario..)
sod: Iraqis would not send unidentified bodies to this holy graveyard. That makes no sense. -- Yes, they would and do.
hm. i wouldn t claim that i was right when i said that, but you made a rather wild claim about bodies all over the world being sent to Najaf. in that context my claim that they are NOT sending unidentified bodies makes sense.
http://scienceblogs.com/deltoid/2007/10/us_must_face_huge_death_toll_o…
"There are two cemeteries, the first is in Karbala and the second in Najaf, both assigned for unidentified corpses," he told AFP. "After being left for three months at Baghdad's morgue, the bodies are brought here."
doesn t this sound, as if they might be sending a significant number of unidentified bodies to Karbala? do you notice, what that does to your numbers????
izzy sez:
Yup, that's what I meant by "isn't ideally suited for this sort of thing". However, by clarifying that they meant "under this roof" it's better than either the February ORB or March BBC-D3 surveys.
You're assuming that everyone who answered in February interpreted family to mean "extended" family. Heald's point wasn't that everyone interpreted it that way but rather that some unknown fraction may have interpreted it that way. Without knowing that fraction (or the extent of extendedness), it's going to be hard to say what your multiplier ought to be -- however, presuming that their (unknown) sample design was the same in February and August, the two results suggest that in February not everyone interpreted family as "extended."
"doesn't this sound, as if they might be sending a significant number of unidentified bodies to Karbala? do you notice, what that does to your numbers????"
Try reading the article.
Robert, I see no indication that that was Heald's point. The February question said "family member or relative" iirc. That's asking them to report anyone they're related to in any way. Of course some might extend their answers to more distant relatives than others, but clearly the net here is substantially bigger than one household.
"However, by clarifying that they meant "under this roof" it's better than either the February ORB or March BBC-D3 surveys."
It would only be better if respondents chose to follow that clarification, and you had some way to know if they did, and how strictly. If they did not, as the earlier ORB and BBC polls suggest, it's a recipe for exaggerating estimates.
Try reading the article.
sorry, i missed the link. i am still unsure, what it is supposed to tell me. apart from adding another 2500+ unidentified bodies, making it 43% that you have to achieve by your calculation.
are you trying to tell me, that they moved only 2500 bodies to the FIRST graveyard they are "supplying" and slightly less than 40000 to their SECOND one?
izzy, grasping at strawz, inzizzted:
Yeah, getting the point of Heald's comment does sort of require that you know something about how survey are done and why a survey instrument would be changed after one iteration. And whether the net effect is "substantially bigger" is an empirical question that is answered by the data. The data suggest that the "extended" interpretation wasn't zero, but it also wasn't universal.
Perhaps, but since this wasn't a panel study, why would you think that the August respondents would answer a February question they weren't being asked? Dude, that's wack.
And whether the net effect is "substantially bigger" is an empirical question that is answered by the data. The data suggest that the "extended" interpretation wasn't zero, but it also wasn't universal.
It might suggest this if you considered it safe to assume the August question was sufficient to limit responses to one household. But I don't think that is safe at all. And the BBC and earlier ORB results are one reason.
But if I am understanding your point correctly, you're saying that when asked if a "family member or relative" was killed, people (and it would have to be most people to bring these into line) would just choose to limit their responses to one household for no apparent reason? So if their brother who lived in the next neighborhood was killed, they'd just say, "nah, I'm only going to report my house here"?
I don't think I'm the one grasping at straws.
"Perhaps, but since this wasn't a panel study, why would you think that the August respondents would answer a February question they weren't being asked? Dude, that's wack."
Silly. What i think is that people want to report their loved ones who've been killed, and want to report that they have suffered a loss, that this is the case whether the loved one formally lived in the same house or not, and that a simple "clarification" added to one question would often be ignored, or treated very loosely, as in "eh, he slept under my roof a few times".
Of course this does not mean they are literally answering the February question, but it means that the August question still often effectively results in the February answers.
sod: are you trying to tell me, that they moved only 2500 bodies to the FIRST graveyard they are "supplying" and slightly less than 40000 to their SECOND one?
That's what the article is telling you. Read it again, but all the way to the end this time:
"Once, all bodies just went to the massive graveyard of Najaf, but authorities there could not cope with the constant stream of corpses and now overflow cemeteries have been established in Karbala."
The one in Karbala is more recent, and much smaller.
izzy, in denial, declared:
Straw-graspers rarely do.
My claim (go upthread and check if you've forgotten) was simply that the February ORB survey and the similarly-worded March D3 survey were inferior to the August ORB survey for the purposes of mortality estimation. You, OTOH, have a much harder row to hoe: you've been denying that the August ORB is any better than the earlier ORB or D3 surveys, i.e., if it's not better it's either: 1) just as lousy, or 2) worse. In order to do that, you've got to believe either: 1) the respondents gave exactly the same answers without regard to the way the question was framed, or 2) the greater specificity of the August question resulted in larger bias in the other direction. There's no evidence that either of those things has happened; in fact, since the proportion decreased from 26% in February to 22% in August, the evidence supports the argument that at least some respondents did answer the re-worded question in the more restricted sense. So: no evidence that supports you, some evidence that refutes you, yet you're still holding your ground. Dude, you're grasping at straws.
I think you're shifting the issue around here somewhat. If you want to say the Aug is better than the Feb for mortality estimation, fine. In some sense you're probably right. A rock is probably better than a horse at being a basketball, for example. But someone else might say they're just as bad at it.
My claim (go upthread and check if you've forgotten) was simply that the February ORB survey and the similarly-worded March D3 survey were inferior to the August ORB survey for the purposes of mortality estimation. You, OTOH, have a much harder row to hoe: you've been denying that the August ORB is any better than the earlier ORB or D3 surveys, i.e., if it's not better it's either: 1) just as lousy, or 2) worse. In order to do that, you've got to believe either: 1) the respondents gave exactly the same answers without regard to the way the question was framed,
I don't think so. What I would need to believe is that, like the previous two polls, it's not clear exactly how many people or households are actually being represented in these answers.
or 2) the greater specificity of the August question resulted in larger bias in the other direction.
I think the greater specificity resulted in an illusion that the question no longer had the kind of uncertainty that the others did, and was therefore now suitable for mortality estimation. In that sense, it could be called "worse", as it leads to more hasty conclusions. From another point of view, maybe it's better.
"in fact, since the proportion decreased from 26% in February to 22% in August, the evidence supports the argument that at least some respondents did answer the re-worded question in the more restricted sense."
It does look like at least some answered in a more restricted way. But that's about it.
But as I said before, there's so many other unknowns about the ORB thing that such disparities could have other causes aside from the questions.
izzy, gnawing off his leg, wrote:
Dude, you yourself even quoted exactly what I wrote in #82 above: "as a basis for estimating post-invasion mortality, the D3 survey is inferior to the August 2007 ORB survey (which itself isn't ideally suited for this sort of thing. (BTW, the February ORB survey suffered from a similar problem as the March D3 survey -- ambiguity about the population at risk))." The guy who's been shifting the issue around is you, and you shifted so much that now you're stuck defending something that's indefensible. It's getting ugly in here. My advice? Stop thrashing around. You're hurting yourself.
Robert says: "grasping at strawz", "wack", "in denial", "Straw-grasper", "thrashing around", "gnawing off his leg", "defending something that's indefensible"...etc.
Robert, if you don't mind me asking, how old are you?
I think I could find more mature and rational discussion at a 5th-grade recess, dude.
In extremis, izzy whined:
Wait a second. Let me make sure I got this right: a guy who not only calls himself by the lower-case playground name "izzy" but also appears to know way too much about what 5th-graders discuss at recess is in a huff about the way he's getting addressed? Is that what's troubling you, Bunkie? Dude, your argument in #84 that the August ORB poll "should produce a number at least a few times lower than 26%" ain't well thought-out. If you want to be treated seriously, start making serious arguments. And stop whining. It's unseemly.
"Dude, your argument in #84 that the August ORB poll "should produce a number at least a few times lower than 26%" ain't well thought-out."
it's pretty straightforward, "dude".
"family member or relative" (A) is a far bigger pool of people than "living under your roof" (B).
For me (A) would be about 40 people and about a dozen households (not including distant relatives I don't really know well or see on any kind of regular basis). My (B) would be 4 people and one house. Each person's results would differ, but i think you need to twist yourself into quite a pretzel to make yourself believe that A will not be a substantially larger pool than B for almost everyone (I invite everyone to calculate their own numbers, A and B).
I think it is your attempts to evade this obvious point that are not well thought out, "straw-grasping", etc. This also tends to explain why you've resorted to name-calling and attributing crazed or irrational behavior to others when they aren't exhibiting any, dude.
But I think I've made my points. You can have the last words. I'm sure most readers here will be very impressed with whatever you come up with.
izzy, you are nitpicking and carefully choosing what parts of my posts to reply to. people reading this DO notice that you are avoiding the real points!
The one in Karbala is more recent, and much smaller.
yes i missed that one. but do you really consider the fact that they have to MOVE the graveyard for unidentified bodies, because another one with 40000 dead is OVERFLOWING as a sign to use the LOWEST estimate for war casualties that you are able to find?
http://www.thenews.com.pk/daily_detail.asp?id=61133
izzy fizzled out with:
Ooooo. A tad passive-aggressive, aren't we?
Well, since you're "giving" me the last word: the true amount of overstatement depends on the unknown proportion of February respondents who responded as if the question was about extended family rather than immediate family. Your little argument remains predicated on the assumption that the unknown proportion was close to 1, while the actual data from the more specific August survey suggests that it wasn't. That's a killer problem for your claim, which is why you've been trying to change the subject to the tone of my posts rather than their content. Dude, the August ORB survey has enough problems on its own, as I said in my first post on this topic. That the best criticism you can come up with against the August ORB survey is based on an earlier, rejected, question design from the February survey is a sign of desperation. That it's also unsupported by the data yet you continue to cling to it turns it from desperate to sad.
For me (A) would be about 40 people and about a dozen households (not including distant relatives I don't really know well or see on any kind of regular basis). My (B) would be 4 people and one house. Each person's results would differ, but i think you need to twist yourself into quite a pretzel to make yourself believe that A will not be a substantially larger pool than B for almost everyone (I invite everyone to calculate their own numbers, A and B).
i have serious doubts, that a comparison with western households/families makes much sense in this case. and i would want someone with some knowledge in arabic to take a look at the questions, before i start making wild assumptions.
but as in the case of the "40000 of 100000 war dead are unidentified bodies on the Najaf graveyard", izzy is missing the BIG picture here: NEITHER one of those polls supports a deathrate as low as 100000. FACT.
From the standpoint of precision, the August ORB survey *should* be superior for estimating mortality to their February survey, all other things being equal. Household members living under the same roof is a more precise definition than family/relative.
But all things aren't equal. The August survey had a number of problems. The Baghdad religion data seems to have been wrong, and quietly corrected after initial publication. Three governorates were completely unsampled. A large number of respondents in Baghdad seem to have refused to answer the survey at all. Baghdad gender demographics seem incompatible with the reported death rates. The *cumulative* death rate for Baghdad seems to have lurched up suddenly relative to the rest of Iraq, going from 120% of the Iraq average in the ORB poll taken in February to over 200% in the survey taken in August. ORB seems less than perfectly confident of the August results. ORB wrote:
*We have received a lot of interest in the recent 'casualties' poll that we carried out in conjunction with our local partners in Iraq - IIACSS. Our survey was carried out throughout Iraq and looked at the incidence of civilian deaths since 2003 - together with Iraqi's views on the displacement of people.*
*As with many other activities in Iraq polling has its own restraints and it is simply too dangerous for interviewers to operate in some areas. Meanwhile local authorities prevent interviewers from working in certain towns and districts. This means that we cannot gather opinion from the more volatile areas but, at the same time we have, so far, also limited coverage in rural districts. Both of these factors mean that any estimate of deaths will remain just that - an estimate.*
*While, for obvious reasons, we cannot boost our representation of people living in Iraq's most violent areas we have decided - following feedback from readers of our poll - to conduct a more extensive survey of rural areas to see how this may impact on our estimate. We are in the process of conducting additional interviews in rural areas of Iraq. Once this data has been verified and merged with our current data set we will post it here on the ORB website. We aim to be in a position to release this data within ten days i.e. first week of October.*
(end quote from ORB)
The first week in October ended October 7th, but no new data has been released by ORB
Will McLean wrote:
What are the denominators for your death rates?
Robert #103:
February Survey:
26% overall violent death rate per *family* in last three years. 31% in Baghdad, 24% outside.
August:
22% overall violent death rate per *household* since invasion. 48% in Baghdad, 12% outside.
Will:
Yeah, that's what I suspected you were doing. In order to evaluate that comparison (or other comparisons like that) between the February and August surveys, you also need to make an assumption about the ratio of "family" to "household" sizes in Baghdad and outside-Baghdad. Above, izzy was making an assumption about the relationship between "family" and "household" -- you're going one step more refined in trying to control for this by comparing within families and households but for your comparison to make sense you still have to worry that the "family" pattern is around the same as the "household" pattern.
In your post #102, you said "but all things aren't equal," then you proceeded to make a comparison as if they were close enough. That's probably wasted effort. As far as mortality estimation goes, I don't think the February ORB survey is usable at all; I think the August ORB survey is better, but only relatively and only a little. I consider both ORB surveys and the March D3 survey to be examples of what happens when people who usually run opinion surveys venture into trying to measure demographic events.
And your point about "Baghdad gender demographics seem incompatible with the reported death rates" was wrong when you first proposed it, and it hasn't gotten more correct since. The "gender demographics" you're talking about is the sex of the respondent, not the composition of the respondent's household.
Robert, do you think the polls have any value at all, if only as an order of magnitude estimate? I mean, to an amateur it seems like the August poll ought to be worth something, if they really did get a random sample of Iraqi households. People could be lying, of course (either the responders or the interviewers), but if that many families really did lose that number of people doesn't it at least give us a rough idea of the scale of the killing?
Oh, I certainly think that in aggregate they do have value, and I'm guessing the August ORB survey is much better than literal order-of-magnitude. Bottom line, their overall value is not so much in any particular estimate but rather in the trend they reveal. In aggregate, all of these surveys (the ILCS, the Roberts and Burnham surveys, the two ORB and the D3 surveys), plus the IBC, plus the coalition casualty rates, all draw a consistent picture: the toll on the civilian population, whatever it's level and whatever its source, is not only increasing, it's accelerating. That's the main story and there's really nothing that argues against it.
In a detached academic sense, of course I'm interested in the actual level -- and I certainly have ideas about how one could improve the surveys that have been done in order to produce better estimates -- but at this point it's pretty hollow arguing over whether the number of excess dead is 1.2 million or only half that. Frankly, it's pretty amazing that David Kane is trying to organize a session at the August 2008 JSM to criticize an estimate of 98,000 excess deaths which, by then, will have been published four years previously.
> Frankly, it's pretty amazing that David Kane is trying to organize a session at the August 2008 JSM to criticize an estimate of 98,000 excess deaths which, by then, will have been published four years previously.
Clearly, Kane is not interested in the truth of the matter. Interestingly, unlike other deniers, Kane seems to be primarily motivated by personal ambition, not ideology or passion. He seems to be trying to become the Right's leading authority on statistics - the man who debunked the lies of the Left. A JSM session would serve his personal aims well.
Sortition wrote:
The scary thing is that, considering the general level of statistical knowledge in the world, he might be able to pull it off.
I have no knowledge of his personal aims and motivations but here's something I find kinda disquieting: a couple of times over here on Deltoid he has said that he thinks of himself as a statistician. I understand that at the JSM he began his talk by telling everyone that he's not a statistician. So, when he's talking to real statisticians, he says he's not one; but when he's talking to folks he thinks aren't statisticians, he tells them he is. Hmmmm. I can't imagine how many actual academic submissions he'll get but it wouldn't surprise me if he got enough oddball submissions to fill out a session.
> The scary thing is that, considering the general level of statistical knowledge in the world, he might be able to pull it off.
Sadly, it's not even a matter of people being ignorant about statistics. It's all about media image - once someone "establishes credibility" (i.e., accepted by the media) as an authority on something, the facts don't matter at all. If Giuliani can be an authority on handling terror attacks, and Alan Dershowitz is an authority on human rights, then why can't Kane be an authority on statistics?
Sortition wrote:
Oh dear. I wish I could say you're an alarmist but I fear that you may be right. This may tie into the "John Mashey" thread.
I estimate that the Najaf Cemetery reports suggest a total violent death toll in Iraq since the invasion of 200-300 thousand.
http://willscommonplacebook.blogspot.com/2007/10/how-bad-is-it-in-iraq-…
I estimate that the Najaf Cemetery reports suggest a total violent death toll in Iraq since the invasion of 200-300 thousand.
sorry Will, but i think you got this one wrong. your estimate is based on a number of pretty wild assumptions! some pats i wasn t able to follow. (pretty tired at the moment anyway..)
BUT:
a total violent death toll of 200000 together with 40000 unidentified dead buried in Najaf alone gives you the same problem that izzy had above:
you need to bring up two percentage numbers that multiply to 20% for (a) the number of unidentified dead among all violent dead and (b) the percentage of those transported to Najaf.
It's way OT, but I should point out that Dershowitz is not and has never been an advocate of torture; he's said repeatedly that he's against it and if it were up to him it would be never done, anywhere. (Most recently, at Yale Law School, Oct. 11). "All of that said [the 'ticking time bomb' argument, etc.], the argument for allowing torture as an approved technique, even in a narrowly specified range of cases, is very troubling." ("Is There a Torturous Road to Justice?", Alan Dershowitz, LATimes, Nov 8 2001)
But unfortunately, toture is indeed carried out by essentially every nation where the opportunity arises; on the whim of bored night shift guards at Abu Ghraib, by sadistic professional torturers in Kahakhistan subcontracted by America so we can pretend our hands our clean, by policemen who can't think of any other way to keep control of the streets. What Dershowitz has said is that this haphazard approach where essentially any low level authority can try essentially anything they want with a low risk of being punished for it in order that we can pretend it doesn't happen, that if society is going to tolerate such practices, they should at least limit them with a set of well thought out predetermined policies, and/or with judicial review via "torture warrants", in the same way that the US hands out judicial death warrants. He feels that requiring somebody to take public responsibility would reduce the incidence of egregious practices.
"I have no doubt that if an actual ticking bomb situation were to arise, our law enforcement authorities would torture. The real debate is whether such torture should take place outside of our legal system or within it. The
answer to this seems clear: If we are to have torture, it should be authorized by the law.
Judges should have to issue a 'torture warrant' in each case. Thus we would not be winking an eye of quiet approval at torture while publicly condemning it." (ibid)
I don't necessarily agree, but I do find illuminating the way that the media and the public, even the intelligent folks, responded to this; by continuing to mouth the fantasy that torture never happens here, and if it does it's just the result of a rare disordered mind, and there isn't any systematic use of torture going on which needs to be reined in by regulations. Maybe Dershowitz is right after all.
Sod #113:
If the above seems too complicated, let me give you a simplified version that sets a crude upper limit on violent deaths in Iraq.
Pre-invasion, based on reports from cemetery workers at Najaf, it looks like about 2/3 of Iraqi Shiite dead were buried at Najaf. Conditions are more chaotic now, but the relative political and economic position of Shiites has improved, so it seems plausible that a similar or higher proportion holds.
In 2006, up to 2000 unidentified dead from Baghdad were buried at Najaf each month. Many of these would have been Sunni Arabs, so counting all post-invasion burials at Najaf as Shiite would significantly overcount the Shiite death rate. However, let us simplify and ignore this factor.
At their highest, burials at Najaf were 300 a day, 150 higher than the pre-invasion rate. The likeliest explanation for the disparity is an increase in violent deaths post-invasion.
The highest survey-based estimate of Sunni Arab deaths per household is twice the Shiite rate, and most surveys give a lower ratio. It appears that Sunni Arabs are about 75% as numerous as Shiites in Iraq. The highest survey based estimate for Kurdish deaths per household is 1/3 the Shiite rate. Iraqi Kurds seem to be about 1/3 as numerous as Shiite Arabs.
We then have the following upper bound for monthly violent deaths at their peak:
4,500 Shiites buried at Najaf.
2,250 Shiites buried elsewhere.
10,125 Sunni Arabs
750 Sunni Kurds
17,625 total.
This is 5.8 times the highest IBC monthly total. Applying the same multiple to the average of maximum and minimum IBC totals to date gives about 460,000. Since this ignores the problem of double counting unidentified dead at Najaf, and uses the highest surveyed value for Sunni Arab deaths it is likely to be a considerable overestimate.
Will, as off-the-cuff estimates go, this is interesting, but you're making one assumption that I think is wildly wrong: You're assuming that the proportion of dead being sent to Najaf has remained constant. That strikes me as really, REALLY unlikely. Damage to the infrastructure alone would probably mean that the proportion is lower now. More importantly, the level of violence and lack of security would probably result in a *very* drastic decrease in the proportion of burials dead sent to Najaf from outside Baghdad. You made the remark that "the relative political and economic position of Shiites has improved," but I think this is only true if you mean their position relative to the Sunnis, at the present time. That, however, doesn't matter: what matters is their position now, relative to their position in the 1990s. It seems likely to me that they are economically worse off now than they were before.
z,
One has to be quite naive to buy this kind of doublespeak ("I am against it, but it is the best option") - especially in the context of the time of publication - immediately after 9-11 and immediately after the capture of Khalid Sheikh Mohammed.
Robert Chung writes:
You "understand" because you were there? Because you talked to someone who was there? Because you imagined it? I certainly never said that I was not a statistician.
"Imagine" a bit harder. I have already signed up Jon Pedersen of Fafo and Safaa Amer of NORC. You think that these people are oddballs? Do tell! I also hope to have someone from Spagat et al. Perhaps Robert Chung can explain why he knows more statistics than, say, Gesine Reinert.
David, I was hoping you started out your talk with the truth: that in March 2003 the United States and its junior partner, Great Britain, launched an illegal war of aggression against Iraq that violated the UN Charter and the Nuremberg Code (both of which the US are signatories to) and which has led to the death of at least hundreds of thousands of Iraqi civilians. You could also have mentioned that the 'official' reasons of both governments for invading were based on serial lies, amplified by a culpable corporate state media apparatus, and that the real reason for invading was control of a region of vital interest to the US economy because it is a 'Source of stupendous strategic power' and 'The greatest material prize in history' (quoting the State Department, 1946). I would have hoped that you would have quoted influential planner George Kennan, who said that any country controlling this region had 'Veto power over the global economy' or Zbigniew Brezinski, who argued that US control over the region would give it 'Critical leverage'.
I'd be interested in know what part of your talk mentioned these vital points.
David Kane protested:
Well, thanks for the clarification! I'll have to forward your correction to the person who told me this.
I would never do that. First, because I don't know Gesine Reinert; second, because I don't claim to be a statistician. I do claim to be a demographer. And third, because I'm not into the credentialism one-up-manship crap. If I had been, you would've known more than a year ago that I was a demography professor--that you didn't is evidence that I don't think it's a particularly meaningful argument. But since *you* seem to think that sort of thing carries weight perhaps you can explain why you know more about demography than, say, yours truly. Wait, wait! Here's something better: I had the students in my undergraduate demographic methods course replicate the pre- and post-invasion crude death rates using Roberts' cluster-level data. You know, the ones you said no one could replicate. And we're only half-way through the first semester. Perhaps you could explain why they know more about mortality estimation than you.
And have your students replicated the associated confidence intervals? Do tell! Again, if you have a better source for the academic meaning of replication than this, please provide it. If your students (or you) can replicate the 8,000 to 194,000 confidence interval (or the confidence intervals associated with the pre or post war CMRs), please share the results withe the rest of us.
David Kane tried:
Nope. They're undergrads and we're only half-way through the first semester.
I may ask the first-year grad students to do it around the end of the semester. Plus, I sort of like the fact that you don't know how to do it.
But I'm glad you brought up Gary King. He's got a book coming out on mortality rates. Why didn't you ask him if the CMRs were replicable from the data provided before you claimed that no one could do it? Matter of fact, have you tried asking him about replicating the CIs? Let's see now, how many claims have you made that you've had to retract because you haven't done your homework? There's the "there was no peer review of the Roberts article", the "expedited review is highly unusual", the "no where, ever, on any subject, has there been a survey with response rates like these", the "no one can, or has ever been able to, replicate the CMRs", the "Roberts refuses to release the code needed to replicate the CMRs", the "survey workers were paid by Roberts", and, of course, the "I bet you don't know how to calculate a CMR, either." I'm sure I'm missing one or two.
I'm sure I'm missing one or two.
actually yu did a rather nice sum up Robert.
one could add: writing a paper, that demands that Fallujah datapoint MUST be included in the iraq deathtoll analysis..
Imagine" a bit harder. I have already signed up Jon Pedersen of Fafo and Safaa Amer of NORC.
people will join up, because it is an interesting topic to discuss.
i am not sure, how many of them know, that you are going to sell their participation as support for you wild thesis on the Michelle Malkin website.
Robert,
What an amusing collection of truths, half-truths and falsehoods that you have attributed to me, and all without the benefit of links to any proof! Let's start with just the first one. You claim that I insist that ""there was no peer review of the Roberts article." First, of course there was! I have seen snippets of the the peer reviews myself. Second, I unsurprisingly, never said that. Why do you insist on making things up? Also, "Roberts refuses to release the code needed to replicate the CMRs" is true. Don't believe me? Ask Roberts yourself. Respond on these two points and I'll take the time to point out your other errors.
David Kane claimed:
Nope. What's puzzling is why you think anyone would. The code needed to replicate the CMRs appears in cells T37 and U37 of the the "all data" sheet, and T36 and U36 of the "without Falluja" sheet of the Excel file right [here](http://timlambert.org/data/single%20sheets%20of%20Iraq%20data.xls) that Roberts provided to you. Anyone can open it up and see that they're there. Malkin could find 'em if she wanted. Fumento could find 'em if he wanted. Pedersen, Amer, and Gary King could. But I can sorta understand why you keep trying to suppress that information -- it does, after all, reveal that you're either: 1) incompetent; or 2) someone who, in the face of all evidence, continues to make unfounded claims. The latter, in turn, means you're either: 2a) completely delusional; or 2b) a liar. I wouldn't have thought that was such a tough choice, but it appears you went for door number 2. So although I can understand why you would want to suppress it, what I can't understand is why you think you can get away with it. My guess is 2a.
Let me remind everyone that David Kane created an R package based on this [spreadsheet](http://timlambert.org/data/single%20sheets%20of%20Iraq%20data.xls) and kept promoting it rather than the spreadsheet itself. Those with R could have verified this: Kane's package faithfully copied every line in the spreadsheet except the cells that showed how to calculate the CMRs. Then he told Malkin that Roberts refused to share that information, that the results were not replicable, and that one should be suspicious of the Roberts paper for that reason alone. If any of you use google or technorati to search for references to Kane's JSM paper, you'll find many blogosphere posts that repeat David's charge and conclude that Roberts was hiding something. Check it out for yourselves. In particular, I invite the "Lancet skeptics" who hang out here to check it out. There are many honest, good-faith critics of the Roberts and Burnham papers. David isn't one of them.
1) Yes, you can replicate the estimates but you can't replicate the associated confidence intervals. I have no problem with the estimates themselves. (And I thank you by name in the latest draft of the paper for pointing out this aspect of the spreadsheet.) The problem, as I show in my paper, is with the confidence intervals. You can't replicate those. A $100 to the charity of your choice if you can.
2) Stuff like this makes me not want to bother arguing with you.
This is just a lie, as anyone can see by downloading the R package that I created. Every byte from the spreadsheet that Roberts distributed is there. Check again if you don't believe me. Now, I did create an R dataframe in addition to the spreadsheet which included just the data from each cluster (without the summary data) but this is what any reasonable person would have done. It makes no sense to have a dataframe in which rows 1-33 are the results for individual clusters and row 34 is summary data. I don't see a single dataframe on all of CRAN that is organized like that.
David Kane danced around with:
David, you manage a capital fund so perhaps you've been misled into thinking you understand what motivates an obscure professor in a small field who lives on the edge of constant penury. Do you really think money is what drives me to do what I do? Dude, I don't argue for money; I argue for blood. There are no fights more vicious than academic fights since the stakes are so low. I'm just holding back until you once again make the claim that no one can replicate the CIs from the information provided. The truth may hurt, but it's this: you can't replicate the CIs from the information provided because you don't know what you're doing, in exactly the same you that you couldn't replicate the CMRs or the estimates from the information provided because you didn't know what you were doing. You're not qualified to be making this argument. That you continue to repeat the claim is delusional. You insult your Harvard PhD. You insult IQSS. You insult academia.
Well, gee whiz. I told you before that you can thank me for my comments as long as you include that my comments are disapproving. Could you please add that?
Yow. You labeled it a "cleaned-up" version of the data. Your "cleaning-up" left off the crucial formulas and then you claimed that Roberts refused to share them, or the data needed to reproduce the estimates. No reasonable person would do that. Strike that. No reasonable person with two contiguous neurons would do that. In fact, you repeated your lie just above:
This is the enduring claim that came out of your mess of a paper. You can't put the genie back in the bottle but you should make a good-faith effort to clean up your mess (not unlike the mess we created by invading Iraq). As far as I can see, you've made no good-faith effort to correct this.
Okay, you want to know how to get the CIs? Make a good-faith effort to fix the mess you've left behind and I'll tell you for free. Free, I say! That ought to motivate you cuz you get to keep your ginormous bet of one hundred American dollars. That's 69 euros at today's exchange rate! W00t! What I want is that you track down all the blogs that repeated your claim that the estimates were not replicable by anyone and tell them you retract that claim because you "cleaned-up" what was in the spreadsheet in front of you.
Robert,
Although there are many people in the Deltoid community who are worth arguing with, you are not one of them. My advice is to tone down your style lest these comments come back to haunt you someday. At your request, I have removed your name from the paper.
To others interested in this topic, a new version of my paper will be available soon.
David:
As I suspected: when challenged, you fold. Let's summarize for anyone who is still wasting time following this thread: it appears that you never talked to anyone who actually knew how to calculate a mortality rate. It appears that you never talked to anyone who actually knew how to calculate a CI around a mortality rate. You "cleaned up" Roberts' Excel spreadsheet and afterwards claimed that the formulas contained in it weren't there. You said, "I bet you're bluffing", then when I showed that I was not you said you never made the bet. You charged fraud on the part of the Roberts team. You have made demonstrably false claims. You refuse to correct your demonstrably false claims.
You do know that in fifty years your grandkids are going to be googling up your name and finding out that you didn't know how to calculate a mortality rate, right?