SpatialDan,
Comment Posted By Bill Arnold On 11.10.2006 @ 17:49
The concern here is that both the overlapping deaths error and the spatial correlation error are likely to produce overestimated death rates.
Why wouldn't spatial correlation error also be likely to produce underestimated death rates?
DrSteve,
Well, yes and no. Let’s not let “a sample of 12,000†become the meme here.
Agreed. I was commenting more on the practical considerations. Sample size 12,000 and tight geographical clustering were chosen primarily to "balance the need for robust data with the level of risk acceptable to field teams" - the implied consideration is that the more the field teams roam, the more likely they are to be killed.
(2) whether there’s overcounting of shared members from families spilling across households (selection of adjacent structures beginning with a randomly selected house);
This is the one that immediately bothered me as well. As you say, it should be easy enough to verify.Re larger sample size;
Comment Posted By Bill Arnold On 11.10.2006 @ 15:18
A sample size
of 12 000 was calculated to be adequate to identify a
doubling of an estimated pre-invasion crude mortality
rate of 5·0 per 1000 people per year with 95% confidence
and a power of 80%, and was chosen to balance the need
for robust data with the level of risk acceptable to field
teams. Sampling followed the same approach used in
2004,8 except that selection of survey sites was by random
numbers applied to streets or blocks rather than with
global positioning units (GPS), since surveyors felt that
being seen with a GPS unit could put their lives at risk.
The use of GPS units might be seen as targeting an area
for air strikes, or that the unit was in reality a remote
detonation control.
Margin of error? Did they mean 95% Confidence Interval? Surveys and polls have margins of error; statistical studies do not.
Comment Posted By Bill Arnold On 11.10.2006 @ 14:28
Inexcusably sloppy language on my part, sorry. Yes, 95% confidence interval.
It's an 8 page paper, and quite readable. The margin of error is considerably lower than the margin of error for the 2004 estimate. The last page of the paper outlines the reasonable statistical reasons to doubt the conclusions.
Here
Comment Posted By Bill Arnold On 11.10.2006 @ 13:21
is the link again.
FAR EAST NUCLEAR DOMINOES READY TO FALL
Also bookmark www.armscontrolwonk.com - they tend to have pretty good technical analyses, and more important they often provide Google Earth kmz files of interesting sites. (Even civilian spy sat resolution is useful.)
Comment Posted By Bill Arnold On 9.10.2006 @ 14:35
POLL LEAK TO FOX MAY SIGNAL LEADERSHIP STRUGGLE
Tradesports is my oracle when intuition is murky.
The market there atm has Senate staying in Republican control 75/25, and House staying in Republican control about 40-60.(A little depressing to this Democrat...)
Comment Posted By Bill Arnold On 5.10.2006 @ 21:33
DECLASSIFIED NIE LEAVES UNANSWERED QUESTIONS
Bush charged at the news conference that political opponents leaked select parts of the National Intelligence Estimate to media organizations last weekend “to create confusion in the minds of the American people†in the weeks before the Nov. 7 mid-term elections.
The POTUS (Bush) questions the timing. Hmm. :-)In any event, it's good to see the declassified Key Findings section. (I do wonder how close it is to the original.)
Comment Posted By Bill Arnold On 26.09.2006 @ 22:24
HAS THE POPE THROWN DOWN THE GAUNTLET TO ISLAM?
I have zero trouble believing that this was a simple blunder. The Pope is the head of a religion that is a contemporary force for peace, and it's hard to believe he would have deliberately risked even a low probability of igniting mass violence against Christians. There is also the "principle of least malice" that should be applied here. (This is a particularly useful filter to apply when thinking about one's ideological opponents.)
Comment Posted By Bill Arnold On 18.09.2006 @ 21:14
"THE PATH TO 9/11" PRE POST MORTEM
Two points:
Comment Posted By Bill Arnold On 8.09.2006 @ 18:27
(1) Remove the made-up facts (counterfactual scenes), and the docudrama is hard to object to. Some of us left-of-center people will object anyway, but our objections will have less substance.
(2) The stories, which seem true, of how the pre-production copies were almost entirely circulated to R-wingers suggest either (a) strong right-wing bias, or (b) a marketing campaign leveraging right-wing bias, or (c) both, and (d) counting on free publicity due to complaints from the other end of the political spectrum. This is strongly suggestive of a political project, along the lines of a Michael Moore film.
(Regular reader, not a member of your target audience.)
Comment Posted By Bill Arnold On 2.09.2006 @ 19:18
Thank you for this. It's scary and sad that unconditional support for the current, disgraceful US policy on detainees has become one of the tests for right-wingness.