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 :)

 

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/public... - 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.

 

”http://crookedtimber.org/2005/...”

 

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 governm

 

The countdown is on......

 

 

"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.

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

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

catch me at geronimo_skipper@hotmail.com