The Strategypage is a comprehensive summary of military news and affairs.
 News As History - May 17, 2008

Advertisement


Advertisement


Advertisement



New Strategy - Wargames at Discount Prices
1.Horent Leader
2.Harpoon 4: Modern Tactical Naval Warfare
3.Empires In Arms

4.Gallic Wars
5.Fast Action Battle: The Bulge
6.Campaigns of King David
7.Queen of the Celts
8.Danube Front '85
9.Axis and Allies: Guadalcanal
10.Guns of August

100+ Computer and Board games all with free shipping.
 
 
 

Utah SEO Firm

Xango

Smiley Gifts for Babies

Military History | How To Make War | Wars Around the World Rules of Use
Iraq Discussion Board
Sign In   Return to Topic Page
Subject: Where for art tho Lancet threads?
EW3    10/14/2006 6:52:31 PM
Is SP getting hacked.
We've had 1 thread lock up so it can't be posted to.
And the two replacement threads hacked.
Seems like right after sheck makes cogent points about the fallacy of the report, the thread he does it on disappears.
 
Quote    Reply
 Latest
 News
 
 Most
 Read
 
 Most
 Commented
 Hot
 Topics

Email Me When A New Comment Is Made
Show Only Poster Name and Title     Sort in Reverse Order Posted

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14   NEXT
Pseudonym       10/14/2006 7:30:20 PM
It's funny how threads like this always disappear right after the bogus claims are debunked....

For example right after i posted the zombietime link to show the clear and incontrovertible evidence provided in numerous photo's and a video to debunk the article posted on the Aussie thread, it mysteriously disappeared.  It was there the whole time they kept calling me crazy and saying the Israeli's were deliberately targetting Red Crescent Ambulances and innocent Lebanese civilians, and then lo and behold Zombie has assembled all the evidence whose surface I was just scratching and...

POOF, it's gone.

This isn't the first time a thread went missing when hard fact hit the floor, like the hi-res photo of the ambulance that the ICRC took off its site to "keep the moral high ground" lol, I haven't said anything as of yet, though this is the second or third thread I've seen started to comment on missing threads who mysteriously always have a common theme in them, but I have a feeling one of the admins perhaps has an agenda.
 
Quote    Reply

eu4ea       10/14/2006 8:25:17 PM
Yeah, three Lancet threads are just disapeared over the past few days.  Odd &annoying.

Re: whether this is some kind of censorship - I have to admit the thought did cross my mind, bt ultimatelly I think it's pretty random. 

Heck, the nature of the beast is that we each tend to think that we're arguing pretty effectivelly, so the conclusion that 'Someone who disagrees with me knocked off my thread - censorship!' is pretty natural...

Heck, I thought I was doing a good job of defending the Lancet's methodology and results - thou I'm pretty sure that EW3 and to a lesser extent Shek and others would disagree with that... :)

Heart,

eu4ea

 
Quote    Reply

shek       10/14/2006 9:00:23 PM
Pretty random?  Three threads on the exact same subject in 3 or 4 days when I've never seen threads disappear before?  Seems to be a pretty clear pattern emerging.  Anyways, given that I don't like retyping, I saved my post and am now ready to react to hacking on a moment's notice :)
 
Quote    Reply

shek       10/14/2006 9:00:47 PM
 

Eu4ea,

 

You can easily find the CIA figures by Googling “CIA Factbook death rate 200x”. It is interesting how you demand links and yet when you quoted a CIA figure of 5.0 (one which I couldn’t find or verify) there was nary a link. In any event, I’ll provide the 2006 link – https://www.cia.gov/cia/publications/factbook/geos/iz.html - you’ll find the other years I quoted in one of the first five hits from Google when you use the search string I provide above (substituting the specific year for “x”). However, haggling over links really isn’t important to my argument. 

 

What is important is the fact that you endorsed a CIA figure and now you want to discredit a CIA figure. You cannot have it both ways, unless you can demonstrate that the pre-war methodology was satisfactory and that there was some change that caused the methodology to be incorrect. Given that you are asking for the methodology, it sounds like you had no basis to endorse the pre-war figure.


I do not know the methodology that the CIA used, and given that we didn’t have agents in Iraq pre-war, I really can’t imagine how a black box figure coupled with a lack of agent infrastructure in Iraq is that defendable. So, I find the CIA figures pre-war to be dubious (and while we certainly have the agent infrastructure now in Iraq, the post-invasion figures are still black box, and so I don’t put weight on them as well).

 

So, this leaves us to question the fundamental underpinnings of the Roberts et al numbers. They are based on comparing the mortality rate of pre-war Iraq with post-invasion Iraq to derive the “excess” death figure. However, when you benchmark their pre-war numbers, they underestimate these numbers. Due to this, the published results overestimate the “excess” deaths figure. For example, if we use the UNPD numbers from below, you could estimate that they overestimate the “excess” death figure by nearly 281K alone by underestimating the pre-war death rate. What is even more important about this is that the 95% CI now moves the lower limit much closer to 0 (it would be approximately 112K if the variance were to remain the same). Thus, the finding of significance at the 5% level becomes much more sensitive to rejection if consistent biases exist in the study (i.e. the dropping of 3 clusters where it is reasonable to assume to that the figures would be biased towards overestimating the “excess” death figure based on the geographic location). 

 

I will quote from some comments made on the econ blog Crooked Timber that were from the original survey.

 

link

 

Comment 99. “But there are other numbers for the before the war scenario. In the 1980s, the last time international agencies could check any real numbers, Iraq’s mortality rate was always in the region of 7 to 8 (per 1,000 people). According to the World Bank’s “World Development Indicators” link report, the 2002 figure for Iraq was 8 as well. And if you look at some other reports from various reports, you will find that most reports put the figure somewhere around 8.

 

Comment 106. Some more data on the mortality or death rate per 1,000 in Iraq before the war: Source: UN Population Division & Unicef Downloadable for example at: pdf.wri.org/wr98_hh2.pdf Mortality rate in Iraq: 1975-1980: 8.8 1995-2000: 8.5 1980-1991 doesn’t produce any good data because of the Gulf wars; “smart sanctions” were introduced in 1996, so sanctions don’t have a great influence on the data (beyond what was the government’s responsibility). Average figures are especially valuable as they eliminate statistical outliers. I think it can be safely said that Iraq’s “natural” mortality rate is somewhere around 8 per 1,000. Of course, Les Roberts et al. think that the mortality rate in 2002 was 5 per 1,000. Here’s the challenge: Can anyone find anywhere in the academic literature and scientific databases a source which would correspond to Roberts et al finding of 5 (+/- 10%) for a year before the Iraq invasion? All the statistics for various years before the invasion lie somewhere around 8, which is the best scientific data we’ve got so far. Up to now, Roberts et al are the only one who claim it was 5. Now the question is, which claim looks more likely, is based on more solid methodological grounds, is validated by reiterated tests, and provides greater overall consistency with other comparative historical and cross-country datasets? I think it’s pretty obvious that Roberts’ sample is the outlier here, not all the other studies before.

 

To overcome this weakness in the study, you’d have to find some academic literature that finds a crude mortality rate similar to the Roberts et al finding of 5.5 per 1000. Good luck, as I have yet seen anybody provide these. 

 
Quote    Reply

EW3    shek/EU   10/14/2006 9:08:32 PM
 
The countdown is on......
 
 
 
Quote    Reply

S-2    eu4ea Reply   10/14/2006 9:40:57 PM
"Random"?  Perhaps.  This has happened not two months ago with YOUR topic about civil war.  In fact, I asked you about it, as it occurred while you were on vacation, if you'll recall.
 
Don't take this as a suggestion that I think you're up to something nefarious.  Not by a long shot.  Still, something's up that's inexplicable.  Sure hope SYSOPS comes clean.
 
Quote    Reply

eu4ea       10/14/2006 9:45:59 PM
EW - tick... tock... tick... :)

Shek - I guess maybe I'm just naive or something, but it does seem to me that the likeliest explanation is probably a bug...  Occam's razor and all that. Thou, if you are right, that *is* totally unacceptable. 

Heck, those were 3 good threads, where lots of people had thought about it, posted, done their resesearch etc...

Heart,

eu4ea

 
Quote    Reply

shek       10/14/2006 9:50:18 PM
Eu4ea,
 
A bug that specifically targets Lancet threads while leaving others on the same board intact?  Not so random in my book.  Maybe the CIA analyst that runs the crude mortality numbers didn't take too kindly to my words ;)
 
Shek
 
Quote    Reply

shek       10/14/2006 10:17:43 PM
EW3,
 
In one of the MIA threads, you had asked about cluster sampling.  I couldn't find any good internet sites that explain it in detail well.  If you'd like, I could email my notes which include some of the basic derivations that prove that under cluster sampling you'll get much larger variances than you do with simple random sample.  If so, I'll have to scan them in at work next week.

Shek
 
Quote    Reply

EW3    shek   10/14/2006 10:23:45 PM

catch me at geronimo_skipper@hotmail.com  
 
Quote    Reply

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

 

“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.  




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. 
 
 
 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 14,000 deaths per year (from all causes), we know that at least double that many Iraqis have already died this year through violence alone/  By contrast we can look at the death rates for two other Muslim Third World countries.  Egypt, which is relatively stable, (little or no violence) has a death rate of 5.23 per 1000 pop.  Afghanistan by contrast is listed with a death rate of 20.34 per 1000.  It is difficult for me to imagine that Iraq would be considered as stable as Egypt is , and that the violence in Iraq does not lead to a spike in the death rate that would be more than the average Third World Muslim country.  Also it is hard to believe that Afghanistan would be a more violent place to live than Iraq. Of course there are many more diseases in Afghanistan than Iraq, but there is a lot more violence in Iraq than Afghanistan.
 
 
 
Quote    Reply
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14   NEXT

StrategyWorld.com© 1998 - 2008StrategyWorld.com. All rights Reserved. StrategyWorld.com, StrategyPage.com, FYEO, For Your Eyes Only and Al Nofi's CIC are all trademarks of StrategyWorld.com Privacy Policy