Spork Boards


Robert Taylor – December 12, 2007 03:15PM Reply Quote
Well, at least B?K! got a head start.

Dr. Strangelove – January 07, 2008 08:56AM Reply Quote
Well egg, the sample size has proved too little to be extrapolated.

tliet, that's just false. The sample size affects the precision (and results in the large bounds, i.e. 400,000 to 1,000,000), but not the accuracy. You have to show a real bias in the sample to affect the accuracy, which hasn't been done. This is all well understood (by the people who understand it), and has been widely discussed on the web.

tliet – January 07, 2008 02:38PM Reply Quote
Thank you for challenging me egg. My earlier criticism that the numbers were probably exaggerated should be changed in; the article correctly points out that amount cited is subject to sceptism because of the three main reasons mentioned there.

It's too bad really that the authors haven't really released much background stuff to better back up their claims, although I came across this page that has some interesting points and comments.

Also, while reading the Iraqy Body Count page I came across this page. And there are still people around that find it strange that the Muslim world is now a little more suspect of the 'free world'.

Edited 1 time(s). Last edit at 01/07/2008 02:39PM by tliet.

Dr. Strangelove – January 07, 2008 03:48PM Reply Quote
I think that paper gets so much flak simply because the numbers are so high. But I learned long ago that when an experiment gives you unexpected data, you should listen to it.

Also, back when this second study was released (which agree with the first study, which used a different protocol BTW), I did some quick back-of-the-envelope calculations comparing the results of the Lancet study with the death rate in the US due to automobile accidents. It turns out (IIRC) that if you believe the Iraq Body Count/US government/Iraq government figures on deaths in Iraq due to the war, that per capita rate was something like six times higher than the per capita US death rate for auto accidents. If the Lancet study's numbers are correct, that ratio is something like sixty to one.

So if the government numbers were correct, then living in Iraq for one week during 2003-2006 was about as deadly as driving US streets for six weeks. Doesn't quite seem right to me, somehow, but maybe war isn't as bad as I imagine.

tliet, both your links are the same.

Dave Loudin – January 07, 2008 04:03PM Reply Quote
Dr. Strangelove
But I learned long ago that when an experiment gives you unexpected data, you should listen to it.

So true!

Edited 1 time(s). Last edit at 01/07/2008 04:03PM by Dave Loudin.

rino – January 07, 2008 04:30PM Reply Quote
In America, the only respectable form of socialism is socialism for the rich.
I thought that was the stuff you swept away with big arm movements...

stan adams – January 07, 2008 04:35PM Reply Quote
I would think that your meaning of "experiment" is vitally important. To my way of thinking if you were setting out to document the horror of US involvement in Iraq it might be a good idea to go to great lengths to appear unbiased / impartial / beyond reproach in ones record keeping.

Doing otherwise sort of seems to make the whole exercise less than pointless.

Consider what would happen if some other (formerly) controversial data were compiled and reported upon. Say, for instance, some public health researchers set out to show profile the health of smokers. Perhaps the data would show that that cigarette smoking lead people to having lower stress levels and more easily maintain a healthy weight. They were funded by Phillip Morris. They may have been an outspoken critic of restrictions on smoking. They would not reveal all the details of their sample selection (did they pick out areas where cigarette sales were already booming? areas that had highly active heathly people?). Certain gaps in the raw data were unexplained. The data might still be valid and even truthful, but the possible motives would forever cloud the findings.

Really not helpful...

Dr. Strangelove – January 07, 2008 05:00PM Reply Quote
So what? Scientists aren't perfect, but I haven't read anything that indicates that this group's record keeping was sloppy or that their study was biased (intentionally or otherwise) in any way. As Les Roberts has pointed out on multiple occasions, if the skeptics really wanted to know the answer, they could do their own study. That's the way science works. Anybody can nit-pick any experimental protocol to death if they want. That's what's not helpful.

tliet – January 07, 2008 05:11PM Reply Quote

tliet, both your links are the same.

I see now and I've lost the link that contained the interesting story, damn.

Dr. Strangelove – January 09, 2008 02:27PM Reply Quote
Tim Lambert at Deltoid has long post dissecting the inaccuracies/innumeracies/mis-statements in the recent National Journal piece on the Roberts et al study of Iraq deaths that appeared in the National Journal and was linked to above.

Edit: and here's the response of the researchers themselves to the article.

Edited 1 time(s). Last edit at 01/09/2008 02:29PM by Dr. Strangelove.

El Jeffe – January 10, 2008 12:33PM Reply Quote
What a journey.

Two things:

1. Boom
2. Revised Iraq death number.


A new study suggests 151,000 people have died of war-related violence in Iraq since the U.S.-led invasion of that country. The study, which will be published Thursday in the New England Journal of Medicine, is based on an extensive survey of Iraqi households. The work, a joint project of Iraq's Ministry of Health and the World Health Organization, is the latest attempt to try to quantify how many Iraqis have died because of war-related violence. An earlier study by researchers from Johns Hopkins Bloomberg School of Public Health pegged the number of deaths at more than 600,000.

What a journey.

Dr. Strangelove – January 10, 2008 02:56PM Reply Quote
The coverage of the new study has been really horrible. Its results are actually in pretty good agreement with the original study done by Roberts et al and published in the Lancet.

The Lancet study estimated 400,000 - 900,000 excess deaths (of any type, not just due to violence), from the war.

The new study estimates 100,000 - 225,000 excess violent deaths due to the war. But it also (as mentioned in the washington post story, but not the NY Times story) estimates a 60% increase in non-violent deaths due to the war. So you can't compare the 151,000 number from the new study to the 600,000 number from the original study, as many people are doing (and was done in the quote Bill posted above). You have to include this 60% increase in non-violent deaths.

Fortunately, this is easy to do. The pre-war death rate in Iraq was about 5 per 1000 per year. For a population of 25 million, this works out to 125,000 deaths per year. A 60% increase in deaths means an extra 75,000 per year. Multiply by three (to account for the roughly 2003 - 2006 period covered by these studies) gives 225,000 excess deaths. So the new study results are really 325,000 - 450,000 excess deaths (of all types) due to the war.

Of course, this doesn't include the uncertainty in the 60% increase figure, so the real range is bigger, let's say 250,000 - 550,000. This is the number that needs to be compared to the Roberts study, so we have

Roberts: 400,000 - 900,000 excess deaths
WHO: 250,000 - 550,000 excess deaths (give or take).

Thus, it seems like this new study confirms the original study, for the most part. The most logical conclusion is that the true number lies in the overlap of these two ranges, i.e. 400,000 - 550,000. Coincidentally, I think that this is what I guessed was the true number back when the Roberts survey appeared. (I think I wrote here that I would guess that the true number was between 400,000 and 500,000).

Of course, the "liberal" media isn't reporting things this way. Some of it is surely due to their innumeracy, but not all, I'm sure.

Deltoid has some comments by Les Roberts himself on the similarities and differences between the two studies and their results.

Edited 1 time(s). Last edit at 01/10/2008 02:58PM by Dr. Strangelove.

stan adams – January 10, 2008 04:37PM Reply Quote
Regardless of innumeracy or policy bias it is "a better story" to have CONTROVERSY -- the sites (and regular ol' papers too) seem to believe they'll get more eyeballs by peddlin' two widely divergent numbers. The same motivation was in effect when the first study was reported on -- they wanted to imply that well over a half a million bodies had been piled up like so much cordword due to coalition munitions. trigger happy "private secuirty contractors", crazed insurgents, rogue elements from the former and/or reconstitued Iraqi armed forces et cetera...

That FED THE CONTROVERSY that Iraq was a complete f-up on the part of the administration AND THAT SERVED THE PURPOSES OF THE STUDY'S funder(s)/authors...

SoupIsGood Food – January 11, 2008 01:02AM Reply Quote
There is no controversy. It is, was, and will continue to be a complete f-up... only the most delusional of the neocons think otherwise. The only argument between the left and right over the war is is whether we should keep smashing our head against the wall... on the one hand, we're almost through! On the other, we're bashing our head against a brick wall. Regardless of how that plays out, even with the super-successful Surge, we're as far from a win-condition as we were when we first invaded.

So, studies now are simply that, studies. They measure the reality, rather than try to inflict perecpetion on reality, as may have been (but probably wansn't) the case earlier on. Knowing how much farther through the wall you need to bash is something worth knowing...

~ Soop

tliet – January 11, 2008 04:21AM Reply Quote
Also (this is a bit older news), the 'administration' is also still desperately trying to ignite another war using tired old tricks.

And indeed, as was already suspected the US navy isn't so sure about the footage themselves.


rino – January 11, 2008 11:38AM Reply Quote
In America, the only respectable form of socialism is socialism for the rich.
Is that the Dutch equivalent of CollegeHumor.com?

El Jeffe – January 11, 2008 11:54AM Reply Quote
What a journey.

What a journey.

Dr. Strangelove – January 11, 2008 12:11PM Reply Quote
DD over at Crooked Timber has a post on the new study that contains this:

The NEJM does have a much bigger sample size than either Lancet study (9345 households in 1086 clusters versus 1849 households in 50 clusters). But the randomness of the selection was seriously compromised – 11% of the clusters were too dangerous to travel to, and they have their data filled in by extrapolation from the Iraq Body Count website.

Huh, so this new study skipped the most dangerous sites and filled in that data from a website that's know to be a vast underestimate of the casualty rate? Do you think that might lead to some bias in the final results? The authors of the paper apparently acknowledge this, too bad the "liberal" media doesn't.

tliet – January 11, 2008 06:59PM Reply Quote

Is that the Dutch equivalent of CollegeHumor.com?

retecool.com is what used to be called a shocklog, shocking people with all kinds of crap. Lately tho, they make these kinds of analyses for fun and pleasure.

El Jeffe – January 13, 2008 01:40PM Reply Quote
What a journey.
Soros funded the one report of 650,000 killed.



“The authors should have disclosed the [Soros] donation and for many people that would have been a disqualifying factor in terms of publishing the research,”

What a journey.

Edited 2 time(s). Last edit at 01/13/2008 01:41PM by El Jeffe.

Dr. Strangelove – January 14, 2008 10:18AM Reply Quote
Bill, you're just a fount of FUD on this, aren't you.

One quick note about the Soros bugaboo. I commissioned L2. It was commissioned in Oct 2005, with internal funds from the Center for International Studies at MIT, of which I am executive director. The funds for public education (not the survey itself) came from the Open Society Institute in the following spring, long after things had started. Burnham did not know this (Roberts was not much involved at this point.) MIT was providing funds, that's all he knew or needed to know. There were other small donors involved too.


Sorry, only registered users may post in this forum.

Click here to login