[Schrodinger's Cat]

Originally Posted by russotto

OK, now I've looked at the study. That range is the double-ended 95% confidence interval (excluding Fallujah). That's not so great. Then you get into the problems with methodology, which UT has outlined quite well.

Furthermore, I see no claim in the study that these were civilian deaths. That appears to be someone else's addition.

The study is junk.

From the introduction on page one of the study:

One project has kept a running
estimate of press accounts of the number of Iraqi citizens
killed by coalition forces: at present, the estimated range
is 13000–15 000 (http://www.iraqbodycount.net). Aside
from the likelihood that press accounts are incomplete,
this source does not record deaths that are the indirect
result of the armed conflict. Other sources place the
death toll much higher.14 In a recent BBC article decrying
the lack of a reliable civilian death count from the war in
Iraq, Ken Roth of Human Rights Watch purports that it
will not be possible “to come up with anything better
than a good guess at the final civilian cost”.

In the present setting of insecurity and limited availability
of health information, we undertook a nationwide
survey to estimate mortality during the 14·6 months
before the invasion (Jan 1, 2002, to March 18, 2003) and
to compare it with the period from March 19, 2003, to
the date of the interview, between Sept 8 and 20, 2004.

One hopes that your comprehension of statistics is better than your reading comprehension, especially if your profession requires the use of statistical methods. Frankly, I have no problem with the 95% cl. Maybe you wouldn't either if you actually read the paper.