Email Me When A New Comment Is Made
Show Only Poster Name and Title Newest to Oldest

Pages: PREV 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 NEXT

shek 10/15/2006 9:14:58 AM

Eu4ea,

Two questions for you (in addition to the original benchmarking question I posed):

1. Roberts et al spent nearly an entire paragraph on page 7 of the study benchmarking their "excess" deaths figure to the Congo, Vietnam, Timor, and the Darfur (although they don't really discuss at all motivating why the gross figures from these conflicts provide a proper benchmark). From this paragraph, it is clear that they understand that benchmarking is a standard in statistical analysis. Given this, why do you think they make no attempt to benchmark their pre-invasion figure given that there are published figures out there to benchmark against?

2. Why does Roberts et al only use approximately 14 1/2 months of pre-invasion data points given that they argue that their findings, to include their 2004 findings, don't suffer from recall bias:

The striking similarity between the 2004 and 2006 estimates of pre-war mortality diminishes concerns about
people’s ability to recall deaths accurately over a 4-year period. Likewise, the similar patterns of mortality over time documented in our survey and by other sources corroborate our findings about the trends in mortality over time.

Do feel that this short time period (after all, they use over 3 years of data for their post-war death rate) presents a bias? Why or why not?

Thanks.

Shek

Quote Reply

eu4ea 10/16/2006 4:39:14 AM

Cool, it looks like we're all back and Sysops is on top of the missing thread thing.

I do recall the missing 'is Iraq in a Civil War Iraq?' thread, S-2 - you are right about that, it just vanished one day. It was a good thread, too; hope Sysops can bring it back, or we can start a new one on the same subject. That might be the better option; there's quite a bit of new data.

Re: the Lancet study, Shek's recent points were why they would spend time in their report comparing their findings to what we know about mortality in other wars (Congo, East Timor, Vietnam, etc.) and why they only used 14 months of pre-invasion data as their benchmark for the expected death rate.

As I see it, the answer to the first one is quite straightforward; it was an external 'sanity check' to see if their findings were reasonable. The question there was "are the results we are obtained consistent with what we've seen in previous conflicts?", and the aswer to that was a resounding "yes" - 2.5% of the population dead over 3+ years of armed conflict seems to be highly consistent with what we have seen elsewhere. If anything, it's unusualy low.

The implication (which they did not state) is that numbers previously presented (such as Bush's claim last December that around 30,000 Iraqis had died) are extremely unusual. 0.125% of the population dead over 3 years of fighting is on the face of it, absurd. It would be the first time in history that we see anything like that.

Regarding the question of why they only used 14 months worth of pre-war data as their benchmark, I dont have a good answer for that. I imagine they could have used more, dont know what their reasons were. The related question, however, is how much of a difference this would have made.

Someone, I think it was EW3 went through and calculated it using a 5 year average of CIA figures - I believe the figure he came back with was 6.1 deaths/thousand pop. (it's in the missing threads). The benchmark the Lancet used was 5.5 deaths/thousand. While significant, the 0.6 difference is not central; the average post-war mortality measured by the Lancet was 13.2 dead/thousand. Hence even if we apply the more stringent benchmark, the difference would account for less than 1/20th of the measured total.

Which comes back to the central point, which is that even allowing for the issues we've located so far (the clustering, the empty-house problem, the two excluded datasets, and the lenght-of-benchmark) the Lancet study is still by a *long* shot the best study of post-war mortality I have seen so far.

Most of the rest have glaring base errors - for one, most are simple secondary source aggregates, which is simply not a relevant method, by any stretch. Not only that, their results cluster around 0.1 to 0.4% of the population dead in 3+ years of armed conflict, while the CIAs numbers actually claim that mortality has steadily gone down over the course of the war. To put it mildly, those results are radically out of synch with what we've seen in other conflicts. Further, the documentation about their study design, field work and statistical analysis is either non-existent or a dramatically lower quality than the Lancet's.

Taking all that, at this point it looks like on one side we have a well-designed, well documented study published in a top tier scientific journal that produced results that are very much in line with what we have seen previously.
On the other we have a collection of methodologically unsound, poorly documented studies published across a motley collection of websites and newspapers whose main conclusion is that what we are seeing in Iraq is a level of mortality that's radically lower than any other armed conflict we've ever seen, by a factor of 10-20.

Given that, if we had an office pool going, I'm very clear about where I would put my money.

Heart,

eu4ea

Quote Reply

Ashley-the-man 10/16/2006 11:53:17 AM

“October 16, 2006: The British medical journal, The Lancet, has again turned over its pages to political propaganda pretending to be science. The latest report claims that a very flawed survey of the Iraqi population proves that military and terrorist operations have killed over 600,000 Iraqis in the past three years. Several things should be noted. First, the normal death rate of the Iraqi population would leave about 550,000 dead since early 2003. Second, the terrorist, and counter-terrorist, violence in Iraq is largely restricted to four of the 18 provinces. About a third of the population is involved, mainly because Baghdad is a principal battleground. But the Lancet study implies that a third of the population has suffered these losses, which means over seven percent of the people living in that area would have died since 2003. That's a lot of bodies. Where are they? Where are the standards required for statistics and data in a study like this? No matter, the Lancet did a similar study in 2004, just before the U.S. presidential elections. That study was eventually discredited, just as the recent one will be.”

In my post in the Lancet threat - that is apparently being held captive in Area 51 – I posited that half the population was suffering these “655,000” killed. That was simply a ballpark wiff but Strategy page has provided better totals with the same insight. A region with 7.5% killed – violently – would make for some significant migration figures. In my city of 200,000, 7.5% would mean 15,000 headless bodies turning in trailer parks, hundreds of potholes with adjacent shrapnel damaged buildings and a thriving u-haul franchise. Not that Baghdad is not suffering many deaths, but when does a level of killing impact society to a major degree.

Dunnigan has written that military units will break after a predictable number of casualties. When does a society break? I am most interested in an analysis of the effect that the casualties and anticipated duration will have on Iraq. Maybe four million more exit visas for Sunnis will quell the violence.

Quote Reply

swhitebull 10/16/2006 12:44:30 PM

David Frum - writer of the "axis of Evil " speech, specululates on the actual interviewing techniques by the locals that the survey team hired to collect the data:

link

swhitebull

Quote Reply

Plutarch 10/16/2006 1:26:58 PM

Do we have reliable data for Anbar province? Does the Sadrist Health Ministry keep stats on the Sunni death toll? Name for me the government's liaison to Anbar, or the morgue directors for Samara, and Ramadi? I doubt we have any hard data out of the most violent province in Iraq. I would also dispute the notion that only 4 provinces are violent; Basra has as many civilian deaths as Diyala (one of the four violent provinces), and only slightly less violent than Anbar, according to Brookings. Does that seem accurate to you? Wouldn't Anbar be twice or three times or more violent than "stable" Basra?

As for the bodies argument, we still haven't dug up all of Saddam's bodies, nor is there a full accounting for Rwanda. Speaking of Rwanda, the genocide there killed 937,000 people according to a census using population samples (not an actual counting of the bodies), and resulted in ten percent of the total population being killed in just four month. Most were Tutsi, which had according to some demographics 77 percent of their population wiped out. Yet the Tutsis survived and so did Rwanda. By comparison 7.5 percent of the Sunni population is not that significant.

http://www.hrw.org/reports/199...

Third World countries have high birth rates which lead to positive population growth, Rwanda quickly recovered form its genocide and has a 2.5 percent growth rate. The DRC has a growth rate of over 3% even with a violent war that has killed over four million people, its death rate is 13.27 per 1,000 or 822,740 per year (according to CIA). That includes 100,000 AIDS deaths per year. But CIA's death rate is underreported; for Iraq it lists 5.37 deaths per 1,000 population, which translates to 1

Quote Reply

shek 10/16/2006 1:57:24 PM

Ashley,

Check your math. A CRM of 5.37 per 1000 equals approximately 140K per year when I calculate it.

Shek

Quote Reply

Ashley-the-man 10/16/2006 2:34:43 PM

Ashley,

Check your math. A CRM of 5.37 per 1000 equals approximately 140K per year when I calculate it.

Shek"

Sorry, you lost me with your analysis.

Quote Reply

Plutarch 10/16/2006 2:49:21 PM

Ashley,

Check your math. A CRM of 5.37 per 1000 equals approximately 140K per year when I calculate it.

Shek

Oh ooopps forgot the extra 0. 139,620 dead per year is correct. However I still suspect the death rate for 2006 is underreported. In July 2000 Iraq's death rate was 6.4 per 1000, 6.21 per 1000 in 2001, etc.under Saddam, yes, but internally stable as well.

http://www.photius.com/wfb2000...

I have also seen UN estimates that Iraq during the 1980s (full blown war and genocide) had a mortality rate of 8.1 per 1000. The only questions on the study are is the baseline underreported at 5.5 per 1000 (we could say it is 7.5/1000 to keep it in line with previous numbers, and that would reduce the number of excess deaths by several hundred thousand) , and is the sample biased. It is doubtful that Iraqis who had their relatives killed by Saddam would under report those deaths in a confidential study.

There are currently 100 deaths per day in Baghdad, also there are reports that there are at least one violent death every hour in Basra, and at least that many (24) in the rest of Iraq. So at least 55,000 Iraqis will die in political violence per year or close to 200,000 Iraqis dead for the first four years of the war, and those are the numbers we know of. Not every Iraqi shot in the back and dumped near his family's house is going to be reported in the media or in the morgue. The Iraqis certainly cannot trust the police to help them "solve" the murders.

Quote Reply

TAC II 10/16/2006 3:44:12 PM

Iraq Body Count Press Release 16 October 2006

Reality checks: some responses to the latest Lancet estimates

Hamit Dardagan, John Sloboda, and Josh Dougherty

Summary

A new study has been released by the Lancet medical journal estimating over 650,000 excess deaths in Iraq. The Iraqi mortality estimates published in the Lancet in October 2006 imply, among other things, that:

On average, a thousand Iraqis have been violently killed every single day in the first half of 2006, with less than a tenth of them being noticed by any public surveillance mechanisms;
Some 800,000 or more Iraqis suffered blast wounds and other serious conflict-related injuries in the past two years, but less than a tenth of them received any kind of hospital treatment;
Over 7% of the entire adult male population of Iraq has already been killed in violence, with no less than 10% in the worst affected areas covering most of central Iraq;
Half a million death certificates were received by families which were never officially recorded as having been issued;
The Coalition has killed far more Iraqis in the last year than in earlier years containing the initial massive "Shock and Awe" invasion and the major assaults on Falluja.

If these assertions are true, they further imply:

incompetence and/or fraud on a truly massive scale by Iraqi officials in hospitals and ministries, on a local, regional and national level, perfectly coordinated from the moment the occupation began;
bizarre and self-destructive behaviour on the part of all but a small minority of 800,000 injured, mostly non-combatant, Iraqis;
the utter failure of local or external agencies to notice and respond to a decimation of the adult male population in key urban areas;
an abject failure of the media, Iraqi as well as international, to observe that Coalition-caused events of the scale they reported during the three-week invasion in 2003 have been occurring every month for over a year.

In the light of such extreme and improbable implications, a rational alternative conclusion to be considered is that the authors have drawn conclusions from unrepresentative data. In addition, totals of the magnitude generated by this study are unnecessary to brand the invasion and occupation of Iraq a human and strategic tragedy.

Download Press Release (pdf) ...

Quote Reply

swhitebull DAMN - TAC. Using the LEFT to Clobber the Study 10/16/2006 3:53:15 PM

It is just SO unfair to be using a left-wing website to question the veracity of the Lancet study.

Here's another link to them as well: link

swhitebull - wonder if IraqBodyCount will be attacked as a shill for the Administration now.

Quote Reply

PREV 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 NEXT

Reality checks: some responses to the latest Lancet estimates

Summary

Contents