Lancet/IFHS (Reductio Ad Absurdum)
Posted by ALP on January 21 2008, 16:45 » Uploaded 21/01/08 16:56

The attempts by Lancet supporters to see "agreement" between the Lancet 2006 study and the new WHO/IFHS study are slightly comical (the two studies differ in their estimates by 450,000 violent deaths). Aly, a poster at Medialens, mistakenly imputes a figure of 400,000 excess deaths to the IFHS study (mistaken for the reasons given in Will McLean's piece, posted earlier).

Aly goes further and writes that the IFHS study "is in reasonable agreement with L1 and L2 on the issue of total excess mortality"

By that, he means the excess deaths figure of 400,000 somehow "agrees" with the Lancet 2004 and Lancet 2006 studies.

This is curious, since L1 and L2 don't agree with each other at all, as you can see from Tim Lambert's recent extrapolated-to-present figures:

Lancet 1 excess deaths: 420,000
Lancet 2 excess deaths: 1,200,000

Aly thinks that his 400,000 figure "agrees" with L2 because it falls within the range estimated by L2: 392,979 - 942,636 excess deaths).

But by that kind of "logic" one could argue that a study estimating zero violent deaths (and less than 10,000 excess deaths) "agreed" with L1, since L1 estimated between 8,000 and 194,000 excess deaths).

The problem with these "agreements" between studies is that they depend on a "dartboard" interpretation of the massive ranges provided. And this is precisely the type of interpretation that elicited such moral outrage from Lancet defenders when Fred Kaplan first applied it. One shouldn't, they argued, see the bottom figure (eg L1's 8,000 or L2's 392,979) as representative of a likely estimate.

Bottom line:
Lancet 2: 601,000 violent deaths
IFHS: 151,000 violent deaths

The studies disagree in a big way. But make no mistake, 151,000 is a mass slaughter, a calamity, an unimaginable scale of suffering. It's not pro-war propaganda, as some who are addicted to quoting bigger numbers (eg millions) seem to think.

 

Comment 01 – ALP January 21 2008, 17:01

I'd add that the same people who wish to see (or rather hallucinate) "agreement" between WHO/IFHS and Lancet also wish to see disagreement between WHO/IFHS and Iraq Body Count (IBC).

These folk - eg at Medialens and a few blogs - have things exactly backwards, presumably as a result of the massive levels of cognitive dissonance which the IFHS release triggered.

The WHO/IFHS figures largely agree with IBC's position and confirm IBC's data. The IFHS estimate is about three times that of IBC's count (for the period in question), but IBC's is for civilians only, remember (unlike IFHS). And IBC have always stated that they necessarily undercount (since "many, if not most" deaths go unreported). Taking into account these factors, IBC's count (and their estimation of the level of their undercount) is vindicated by the new study.

Medialens's insistent claim that IBC capture less than 10% of violent deaths has been destroyed (together with most of their accompanying "arguments" against IBC) by the new study.

Comment 02 – Roger W January 21 2008, 18:07

I've noticed that a number of Lancet fan blogs are going all out to find fault with this latest survey. It's interesting. I advise my students to not let their scientific brains fall to the emotion of politics. Not so long ago, the Lancet fans were condemning anyone who dared to criticise +peer-reviewed+ science (i.e. the Lancet study). The moral outrage was palpable. And now here they are, going to all lengths to pick holes in a peer-reviewed paper that happens not to conform to their prejudices.

Comment 03 – Russ Bridger January 21 2008, 18:45

It doesn't surprise me in the least that the Medialens believers have gone into cognitive dissonance meltdown over this. A credible, scientifically authoritative estimate which doesn't fit their dogma, and yet which doesn't fit a pro-war dogma either. How can that be? Time for some really creative damage limitation.

Comment 04 – Julie January 21 2008, 21:13

I don't understand these comparisons between Lancet '04 and Lancet '06. The last time I looked they were claiming that those two studies reinforced each other when you allowed for the deaths occurring after the '04 study. But now they contradict each other by 800,000 deaths?!?

I think some people are playing games with these big numbers for political reasons, or simply for face-saving. Let's remember we're talking about the deaths of human beings.

Comment 05 – Peter January 21 2008, 13:41

It's sad to see Medialens subscribers become so desperate that they claim 400,000 excess deaths is in "agreement" with 655,000 deaths (that is what they are effectively saying). A discrepancy of a quarter of a million deaths of human beings - if that's an "agreement", I'd hate to see a disagreement.

I agree with Julie - there's something very unpleasant about the way in which they juggle these huge numbers around in such a misleading way, presumably in order to save face.

Tim Lambert (another Medialens reader) claims that IFHS vindicates Lancet 2004, even if not Lancet 2006:

"If you think that the IFHS study is reasonable then you must conclude that Lancet 1 has been confirmed and the critics of Lancet 1 were wrong."

But this claim by Lambert requires that Falluja is excluded from Lancet 2004. This is in complete contradiction to earlier attempts to show Lancet 2004 in "agreement" with Lancet 2006 by including Falluja in 2004.

It seems that Falluja is included or excluded depending on what you want to prove.

The straight, direct comparison between Lancet 2006 and IFHS is that IFHS estimates 450,000 less violent deaths than Lancet 2006. That's simply an undeniable fact, unlike the big number juggling that's going on.

Comment 06 – Woofles January 23 2008, 11:48

ALP, in an earlier thread you wrote:

"... what is Medialens's response to the Neil Munro article (which has now come to their attention)? Do they attempt to refute Munro's arguments or question his data?"

I guess you got your answer. You also wrote:

"If one is interested in 'truth', one can't have the luxury of ignoring inconvenient facts (of the type that Munro presents)".

You were very keen to draw attention to Munro's article here and elsewhere. Have you changed your view of the "'truth'" of Munro's "arguments", "facts" and "data"?

W

Comment 07 – ALP January 23 2008, 14:02

In answer to Woofles - Medialens have relayed claims by John Tirman, on Soros, etc, but unless I'm mistaken, they haven't dealt with any of those "inconvenient facts" that Munro presented concerning the science and the scientific criticisms. Their instinct is rather to continue to point out the political leanings of Munro and the NJ.

As for the last paragraph of your post, you sound like a certain persistent little troll who has been trying to lay some sort of feeble guilt-by-association smear on me. Let me be clear: It shouldn't be difficult for those with more than two inches of forehead to distinguish facts which are worth addressing from dubious political insinuations. Munro's article contains both.

And if you've been paying attention, you'll see that both Tim Lambert and Les Roberts have now admitted that they were wrong, and that Munro was right, concerning one of those "inconvenient facts" which certainly warranted drawing attention to.

Note also that you (and the troll) were highly selective in quoting me. In the same post that you and the troll quoted, I also wrote the following:

...time would be better spent in evaluating facts, data, etc, than in casting doubt on people based on their political views.[...]

to be taken seriously, research into the true scale of the calamity (particularly the application of epidemiological techniques to war zones - a new, tentative, imperfect area of research) requires that the kind of issues raised in Munro's article (the problems with methodology, unverifiable data, etc - not the political views of the studys' authors) are taken on board and answered, not dismissed with the wave of a hand...

Clear enough for you?

Comment 08 – Raoul Djukanovic January 23 2007, 15:16

Any word on the source, structure and content of the "review" you trailed here earlier, Woofles (in a copy/paste from JHU to MLMB, as per its eventual writeup alertwise)?

http://www.mediahell.org/community/08011201.htm

Comment 09 – Woofles January 23 2008, 16:28

ALP

"Their instinct is rather to continue to point out the political leanings of Munro and the NJ."

I don't think that was the main focus of the alert - it pointed out that, according to Tirman, Munro and Cannon were responsible for deliberate distortion, in fact "a disgraceful lie", in smearing the Lancet team re: Soros. That's nothing to do with "political leanings".

"... unless I'm mistaken, they haven't dealt with any of those 'inconvenient facts' that Munro presented concerning the science and the scientific criticisms".

True - if you ignore the links provided to the Lancet authors and Tirman analysis, which deal with them comprehensively.

You're willing to recognise, "I suspect Munro was going for a hatchet job regarding the Soros aspect)" now that it's undeniable. Not quite your stance before the ML exposure when, with not a caveat or caution in sight, you were encouraging people to read Munro's dreadful piece.

Raoul, sorry mate, you've lost me.

W

Comment 10 – Roger W January 23 2008, 17:48

Woofles,
Your cowardly attempts at smearing ALP don't fool anyone. ALP's post (which you obviously read) predates all the detailed responses to Munro's piece. In that post, of 6th January, ALP makes it absolutely clear that he disapproves of the "casting doubt on people based on their political views", and that what should be taken on board from Munro's article is "the problems with methodology, unverifiable data, etc - not the political views of the studys' authors".

That's what ALP wrote about Munro's article on January 6th. What part of that don't you understand? Why are you insinuating in a cowardly way that ALP shares some of Munro's alleged guilt over the Soros funding claims?

I'm fairly new to this debate, but I'm already sick of this kind of thing. Really, you're fooling nobody, Woofles.

Rog out.

Comment 11 – Russ Bridger January 23 2008, 18:22

Woofles has outed himself as a pathetic accusatory jerk who doesn't have the nads to accuse people directly.

Say what you really want to say, Woofles, don't dress it up in that snakey, reptilian, insinuating manner which seems so common among you Medialens losers.

Comment 12 – Julie January 23 2008, 19:02

Sorry, Woofles, but there is indeed further focus on the politics of Neil Munro and the Nat Journal in the ML article, contrary to what you say. The longish second part has this as a main theme, presenting a quote describing Munro as "a militant right-winger" (militant?) and presenting his views on Iraq in general, Osama bin Laden, etc. Yes, it's terrible to be a rightwing militant (whatever that means), but it doesn't say much about the specific points that Munro raised about the science eg like the fairly important point on which Les Roberts now admits to being wrong (on recording demographic info).

Comment 13 – Raoul Djukanovic January 23 2008, 21:32

Woofles,

You write:

I don't think that [pointing out political leanings] was the main focus of the alert - it pointed out that, according to Tirman, Munro and Cannon were responsible for deliberate distortion, in fact "a disgraceful lie", in smearing the Lancet team re: Soros. That's nothing to do with "political leanings".

Perish the thought that ML would ever stoop so low as to smear people so remarkablifiedly, especially not when they oppose the same wars (and write cogent and insightful commentary on preventing future ones): http://members5.boardhost.com[...]

As for the "review", you cited the one cited by JHU's statement (which you cross-posted, and which later rocked up in the alert, in copy/paste form). I'm asking if you know any more about its source, structure and content, given that you thought it important enough to trail here.