Scott Ritter Says Don't Even Bother

[Guest phil gollin]

Dave Dash :

 

Re. the 100,000 casualties claim. Whilst I can understand that it is contentious, it certainly hasn't been shown to be invalid. By the same token, it is only an estimate done by survey and CANNOT be regarded as rigourously accurate.

 

Two points though :

 

1: It deliberately did not take into account the higher casualties in places like Faluja - so it is NOT biased by such areas.

 

2: You mention how many people would know if so many people were killed (leading to a larger insurgency) - THAT is exactly how the survey was put together.

 

I am not saying it st particularly accurate or not, BUT as the only vaguely neutral comprehensive assessment then it does have to be taken into account.

 

The one set of criticisms that I would like to see (I haven't seen them but they may exist) is from a set of respectable public opinion companies (like Gallop, etc....) who could genuinely assess whether the sampling was statistically relevant and the questions asked unbiased and the results "valid".

[Jeff]

It's simply impossible and frankly consciously intellectually dishonest to try and establish casualty figures by asking people what they heard in a rumour mill. "Well my cousin Abdul heard form his barber that the guy's neighbor knew someone who heard that 25 people died when the Americans bombed a house." That "study" isn't worth the paper it's printed on.

[swerve]

It's simply impossible and frankly consciously intellectually dishonest to try and establish casualty figures by asking people what they heard in a rumour mill. "Well my cousin Abdul heard form his barber that the guy's neighbor knew someone who heard that 25 people died when the Americans bombed a house." That "study" isn't worth the paper it's printed on.

 

I think it was supposed to be limited to personal knowledge, not hearsay. Otherwise there'd be a serious problem with multiple counting, & the authors know that & acknowledge it.

 

The serious arguments about validity (i.e. by statisticians who've looked at the methods used) are about how representative the sample is & the reliability of the methods used for calculating a multiplier, not its honesty.

 

My opinion is that the high end of their range seems far more than I expected, & I'm unwilling to accept it unless it's confirmed independently, which I think unlikely. The low end? I don't know. Perhaps we'll never know.

[JOE BRENNAN]

The serious arguments about validity (i.e. by statisticians who've looked at the methods used) are about how representative the sample is & the reliability of the methods used for calculating a multiplier, not its honesty.

146417[/snapback]

Jeff was speaking of intellectual dishonesty, not dishonesty in the sense of cooking up numbers. But if you [use] methods of knowably dubious effectiveness in the circumstances, and then moreover if recipients or your results don't use the proper skepticism, that amounts to intellectual dishonesty by the end of the chain, and that's what he have with that study.

 

I think it was a US newspaper consortium did an estimate of civilian deaths (sticking to the word "deaths" and avoiding the the ambiguous "casualties" is also a good idea in these cases) in defeat of the Taliban in Afghanistan that came out in the 100's I think. So it dropped right off the radar screen as a story naturally. About the same size population, much shorter campaign, but that's a huge ratio.

 

Another example of problems with this sort of thing is the 1991 estimate of Iraqi deaths, in that case mainly military so not as politically charged. 100k became an accepted figure for that for awhile in the absence of anything better, but basic common sense or rough knowedge of military history suggested at the time, and some later info suggested more strongly, that was a greatly inflated estimate.

 

I agree with Jeff, not worth paper it's written on from what I know. And just because one person or group, without provable literal dishonesty, take a stab, it doesn't mean everyone has to use that figure in the absence of any other. The fact they took the stab and others didn't can show that they are not as careful with facts as others, who might have looked at it and said "this cannot be nailed down to any acceptable degree of accuracy at this point in this situation, we'll simply muddy the waters trying". Some don't care if they muddy the waters as long as they don't outright lie, and some don't care to look closely at the likely quality of such a number. Those are examples of intellectual dishonesty.

 

Btw I don't see how limiting to personal knowledge would avoid massive duplication. I saw X people (I think) die in an explosion (which could have set off by the insurgents, although I guess in this sort of thing that's the Coalition's fault too), so did a lot of other people, how do you parse that all out? Falling back to one of my favorite topics, supposed methods of winnowing out duplication of eye witness accounts were usually highly ineffective v. (knowable objective in that case) truth of for example how many planes AA gunners or bomber gunners downed, the winnowed results were usually still highly inflated. And in that case there was more of a presumption of some attempt at professional detached observation by the witnesses.

 

Joe

Edited February 17, 2005 by JOE BRENNAN

[DaveDash]

Not to mention the nature of the insurgency and Muslim fighting in general makes the differences of a 'civilian' and 'combatant' very thin sometimes.

 

I've read numerous reports from news sources and soldiers about children, women, and so forth both in Somalia and Iraq taking up arms and weapons or 'getting in the line of fire' in a not so accidental manner.

[swerve]

Btw I don't see how limiting to personal knowledge would avoid massive duplication. I saw X people (I think) die in an explosion (which could have set off by the insurgents,

 

Joe

 

Joe,

 

The study recorded all deaths (not "casualties") reported by members of surveyed households as having occurred in those households during a stated period, from all causes. Names, dates, ages & causes were recorded. In each locality, a few households were asked if they had death certificates. 80% of those asked for them had death certificates which confirmed their information.

 

I don't find the high end of the range credible, & I think there must be something seriously wrong with the study, but that's not a reason to attack it for imagined faults, as both you and Jeff have done by misrepresenting the methods ("I saw X people (I think) die in an explosion, so did a lot of other people") without bothering to make any checks at all, when a few minutes with a search engine would have told you the truth. In the circumstances, talk of "intellectual dishonesty" doesn't sound very convincing. I realise that it's carelessness on your part, not dishonesty, but it does show the perils of hasty judgements made in ignorance.

 

The range of excess deaths they came up with was from 8000 to 194000. When you have a range that wide, I'd wonder whether it's worth publishing, but when you've been funded to do the study maybe you're expected to.

[JOE BRENNAN]

both you and Jeff have done by misrepresenting the methods

carelessness on your part,

 

146494[/snapback]

No just elementary filtering in a world with limited time, lots of interesting things to learn, but a sea of bullsh*t one can waste a lot of time on. Thanks for the info, but you haven't said anything to make me think that study should be part of my analysis of the Iraq issue, which is where I was before; besides the fact that it's out there and affects other people eager to believe it, which again I already knew. You don't seem to be willing to support it yourself so bottom line, once all your curtsying to the science and fairness is finished, what's the difference?

 

Anyway specifically you said it didn't use hearsay, I simply pointed out with a hypothetical that that by itself logically wouldn't prevent a massive overstatement. I will freely stipulate I don't think it meets the hurdle of looking into closely, I'm not generally averse to looking into things.

 

And I don't think it's gained much more credibility than stories that US nuclear testing started the tsunami. Not comparing motives of compilers there, just the similar reality that US-haters believe it, while others tend to seriously doubt it for the top line reason that the numbers are so at odds with the reported progress and nature of the war openly covered by a pretty US-skeptical media, as compared w/ common sense to better known results in similar situations.

 

And intellectual dishonesty is exactly publishing ranges of 0-200k with knowledge of the human tendency to pick the middle of ranges combined with people at the next stage splitting the difference and using it as fact. I'm saying nothing much, reliable and passing basic smell tests is known based on that, doesn't seem a hasty or careless "judgement".

 

Joe

Edited February 17, 2005 by JOE BRENNAN

[Legion 46]

Scott Ritter is just trying to remain "relevant."