| WIKI | FAQ | Tech FAQ | |
 http://csua.com/feed/ | |
| ||||||
| 5/26 |
| 2005/6/17 [Reference/Military, Politics/Foreign/MiddleEast/Iraq] UID:38179 Activity:high 63%like:38170 |
6/17 [Hey, I was challenged to read this. I did and here are my comments.
Please don't delete it for at least a day or so.]
I've read the Lancet study (Oct 2004) and it does seem to have some
problems. In particular:
- the "100,000 dead" number comes from a 95% conficence interval
spanning 8,000 to 194,000. Pretty big range there.
- "Interviewers were initially reluctant to ask to see death
certificates because this might have implied they did not believe the
respondents, perhaps triggering violence. Thus, a compromise was
reached for which interviewers would attempt to confirm at least two
deaths per cluster."
- "We think it is unlikely that deaths were falsely recorded.
Interviewers also believed that in the Iraqi culture it was unlikely
for respondents to fabricate deaths.
"It is possible that deaths were not reported, because families might
wish to conceal the death or because neonatal deaths might go without
mention."
- "When violent deaths were attributed to a faction in the conflict or
to criminal forces, no further investigation into the death was made
to respect the privacy of the family and for the safety of the
interviewers.
- "Of these, two were attributed to anti-coalition forces, two were of
unknown origin, seven were criminal murders, and one was from the
previous regime during the invasion."
This suggests that violent crime was mixed in with war-related deaths.
- The slate article (http://slate.msn.com/id/2108887 that criticizes
the study points right to the biggest problem: the cluster
reassignments:
"During September, 2004, many roads were not under the control of the
Government of Iraq or coalition forces. Local police checkpoints
were perceived by team members as target identification screens for
rebel groups. To lessen risks to investigators, we sought to minimise
travel distances and the number of Governorates to visit, while still
sampling from all regions of the country. We did this by clumping
pairs of Governorates."
So their random sample is reduced dramatically.
Like all statistial analyses, the results can be hugely varied
depending on methodology. I see the Lancet study as seriously flawed
and the claim of 100,000 extra dead invalid. See the slate article
for a link to the .pdf of the study and read it yourself. -emarkp
\- for a long comment like this, why dont you put a link to
~ping/lancet.blurb. people can put short followups here.
I've read the Lancet study (Oct 2004) and it does seem to have some
problems. In particular:
\- Why dont you move long comments like to an external file?
- the "100,000 dead" number comes from a 95% conficence interval
spanning 8,000 to 194,000. Pretty big range there.
- "Interviewers were initially reluctant to ask to see death
certificates because this might have implied they did not believe the
respondents, perhaps triggering violence. Thus, a compromise was
reached for which interviewers would attempt to confirm at least two
deaths per cluster."
- "We think it is unlikely that deaths were falsely recorded.
Interviewers also believed that in the Iraqi culture it was unlikely
for respondents to fabricate deaths.
"It is possible that deaths were not reported, because families might
wish to conceal the death or because neonatal deaths might go without
mention."
- "When violent deaths were attributed to a faction in the conflict or
to criminal forces, no further investigation into the death was made
to respect the privacy of the family and for the safety of the
interviewers.
- "Of these, two were attributed to anti-coalition forces, two were of
unknown origin, seven were criminal murders, and one was from the
previous regime during the invasion."
This suggests that violent crime was mixed in with war-related deaths.
- The slate article (http://slate.msn.com/id/2108887 that criticizes
the study points right to the biggest problem: the cluster
reassignments:
"During September, 2004, many roads were not under the control of the
Government of Iraq or coalition forces. Local police checkpoints
were perceived by team members as target identification screens for
rebel groups. To lessen risks to investigators, we sought to minimise
travel distances and the number of Governorates to visit, while still
sampling from all regions of the country. We did this by clumping
pairs of Governorates."
So their random sample is reduced dramatically.
Like all statistial analyses, the results can be hugely varied
depending on methodology. I see the Lancet study as seriously flawed
and the claim of 100,000 extra dead invalid. See the slate article
for a link to the .pdf of the study and read it yourself. -emarkp
\_ emarkp, i think you need to be a little less verbose.
\_ So does the Lancet when they're making up numbers. etc. -emarkp
\_ the death amount is 40k to 150k at a 85% confidence interval
and 60k to 120k at a 75% confidence interval. The report did
not claim that all ~100k were killed by American action, on the
contrary, many were known to have died due to the unstable
security situation. If the study is so flawed, why doesn't anyone
else do one to debunk it. We know the Pentagon has its own
numbers, why don't they release them?
\_ I think these numbers: http://www.iraqbodycount.net
are more reliable. -emarkp
\_"Our maximum therefore refers to reported deaths - which
can only be a sample of true deaths unless one assumes that
every civilian death has been reported. It is likely that
many if not most civilian casualties will go unreported by
the media. That is the sad nature of war." -FAQ from that
site. So they are a very reliable count of what they are
counting, which is known to be an underestimate of the
total number of people killed.
\_ Unless deaths are over reported. -emarkp |
| 5/26 |
|
| slate.msn.com/id/2108887 Lancet, reveal s that this number is so loose as to be meaningless. The report's authors derive this figure by estimating how many Iraqis die d in a 14-month period before the US invasion, conducting surveys on h ow many died in a similar period after the invasion began (more on those surveys later), and subtracting the difference. That differencethe num ber of "extra" deaths in the post-invasion periodsignifies the war's to ll. But read the passage that cites the calculati on more fully: We estimate there were 98,000 extra deaths (95% CI 8000-194 000) during the post-war period. Readers who are accustomed to perusing statistical documents know what th e set of numbers in the parentheses means. It means that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000. Imagine reading a poll reporting that George W Bush will win somewhere b etween 4 percent and 96 percent of the votes in this Tuesday's election. You would say that this is a useless poll and that something must have gone terribly wrong with the sampling. The same is true of the Lancet ar ticle: It's a useless study; The problem is, ultimately, not with the scholars who conducted the study ; It's hard to conduct reliable, random surveysand to e xtrapolate meaningful data from the results of those surveysin the chao tic, restrictive environment of war. However, these scholars are responsible for the hype surrounding the stud y Gilbert Burnham, one of the co-authors, told the International Herald Tribune (for a story reprinted in today's New York Times), "We're quite sure that the estimate of 100,000 is a conservative estimate." Yet the text of the study reveals this is simply untrue. Burnham should have sai d, "We're not quite sure what our estimate means. Assuming our model is accurate, the actual death toll might be 100,000, or it might be somewhe re between 92,000 lower and 94,000 higher than that number." Not a meaty headline, but truer to the findings of his own study. Here's how the Johns Hopkins teamwhich, for the record, was led by Dr. L es Roberts of the university's Bloomberg School of Public Healthwent ab out its work. They randomly selected 33 neighborhoods across Iraqequal- sized population "clusters"and, this past September, set out to intervi ew 30 households in each. They asked how many people in each household d ied, of what causes, during the 14 months before the US invasionand h ow many died, of what, in the 17 months since the war began. They then t ook the results of their random sample and extrapolated them to the enti re country, assuming that their 33 clusters were perfectly representativ e of all Iraq. This is a time-honored technique for many epidemiological studies, but th ose conducting them have to take great care that the way they select the neighborhoods is truly random (which, as most poll-watchers of any sort know, is difficult under the easiest of circumstances). There's a furth er complication when studying the results of war, especially a war fough t mainly by precision bombs dropped from the air: The damage is not rand omly distributed; One of the 33 cluste rs they selected happened to be in Fallujah, one of the most heavily bom bed and shelled cities in all Iraq. Was it legitimate to extrapolate fro m a sample that included such an extreme case? More awkward yet, it turn ed out, two-thirds of all the violent deaths that the team recorded took place in the Fallujah cluster. They settled the dilemma by issuing two sets of figuresone with Fallujah, the other without. The estimate of 98 ,000 deaths is the extrapolation from the set that does not include Fall ujah. What's the extrapolation for the set that does include Fallujah? it's impossible to fi gure out how to extrapolate from it. A question does arise, though: Is t his difficulty a result of some peculiarity about the fighting in Falluj ah? Or is it a result of some peculiarity in the survey's methodology? The survey team simply could not visit some of the randomly chosen clusters; the roads were blocked off, in some cases by coalition checkpoints. So the team picked other, more accessible are as that had received similar amounts of damage. In any case, the detour destroyed the survey's randomness; In other cases, the team didn't find enough people in a cluster to interview, so they expanded t he survey to an adjoining cluster. Again, at that point, the survey was no longer random, and so the results are suspect. Beth Osborne Daponte, senior research scholar at Yale University's Instit ution for Social and Policy Studies, put the point diplomatically after reading the Lancet article this morning and discussing it with me in a p hone conversation: "It attests to the difficulty of doing this sort of s urvey work during a war. No one can come up with any credible estimate s yet, at least not through the sorts of methods used here." The study, though, does have a fundamental flaw that has nothing to do wi th the limits imposed by wartimeand this flaw suggests that, within the study's wide range of possible casualty estimates, the real number tend s more toward the lower end of the scale. In order to gauge the risk of death brought on by the war, the researchers first had to measure the ri sk of death in Iraq before the war. First, Daponte (who has studied Iraqi population figures for many years) questions the finding that prewar mortality was 5 deaths per 1,000. According to quite compreh ensive data collected by the United Nations, Iraq's mortality rate from 1980-85 was 81 per 1,000. From 1985-90, the years leading up to the 199 1 Gulf War, the rate declined to 68 per 1,000. After '91, the numbers a re murkier, but clearly they went up. Whatever they were in 2002, they w ere almost certainly higher than 5 per 1,000. In other words, the wartim e mortality rateif it is 79 per 1,000probably does not exceed the pea cetime rate by as much as the Johns Hopkins team assumes. The second problem with the calculation goes back to the problem cited at the top of this articlethe margin of error. Those mysterious numbers in the parentheses mean the a uthors are 95 percent confident that the risk of death now is between 1 1 and 23 times higher than it was before the invasionin other words, a s little as 10 percent higher or as much as 130 percent higher. report estimating that, of the 800 Iraqis killed in last April's siege of Fallujah, 572 to 616 of them were civilians, at least 3 08 of them women and children). The IBC estimates that between 14,181 and 16,312 Iraqi civilians have die d as a result of the warabout half of them since the battlefield phase of the war ended last May The group also notes that these figures are p robably on the low side, since some deaths must have taken place outside the media's purview. So, let's call it 15,000 orallowing for deaths that the press didn't rep ort20,000 or 25,000, maybe 30,000 Iraqi civilians killed in a pre-empti ve war waged (according to the latest rationale) on their behalf. That's a number more solidly rooted in reality than the Hopkins figureand, gi ven that fact, no less shocking. War Stories: Our Subtle New Strategy for Winning Iraqi Hearts and Min ds Kaboom! The most dismaying thing I've read in a while is a Page One story in the May 17 Philadelphi... |
| www.iraqbodycount.net Surely there must be ma ny, many more civilian deaths than you've published. We are not a news organization ourselves and like everyone else can only base our information on what has been reported so far. What we are attem pting to provide is a credible compilation of civilian deaths that have been reported by recognized sources. Our maximum therefore refers to rep orted deaths - which can only be a sample of true deaths unless one assu mes that every civilian death has been reported. It is likely that many if not most civilian casualties will go unreported by the media. |