Back to Messageboard | Home | Site Map  
 

Professor Marc Herold is unimpressed with Medialens  
Posted by ALP on October 25 2007, 16:27 » Uploaded 25/10/07 20:04  

As mentioned in a previous post, the website, Medialens, has continued its long smear campaign against Iraq Body Count with, among other things, an attempt to use a quote from Professor Marc Herold to discredit IBC:

IBC's methodology was devised by Marc Herold [...] It was Herold's Afghan Victim Memorial Project that inspired John Sloboda to set up IBC. Herold's “most conservative estimate” of Afghan civilian deaths resulting from American/NATO operations is between 5,700 and 6,500. But, he cautions, this is “probably a vast underestimate”. [...] There is no reason to believe that the application of the same methodology in Iraq is generating very different results. But IBC has never, to our knowledge, accepted that their own count is "probably a vast underestimate"... [Medialens Alert, 3/10/07]

This is wrong in several respects (a few already mentioned here). Professor Herold has written to ZNet (who republished the Medialens piece) to correct the errors made by Medialens. As a result, the ZNet editor has had to insert a prefacing remark to the offending paragraph in the Medialens piece, pointing to the correcting comments from Herold:

[ZNet Editor’s note] Marc Herold, referenced above, has sent the following clarification:

"I helped form Iraq Body Count in early 2003. The inspiration for IBC was not my Afghan Victim Memorial Project but rather the data base at http://pubpages.unh.edu/~mwherold, specifically Daily Casualty Count of Afghan Civilians Killed by U.S. Bombing (Copyright © 2004 Marc W. Herold). The Afghan Victim Memorial Project was only begun in September 2004 (at http://pubpages.unh.edu/~mwherold/memorial.htm ).

I did indeed employ the phrase “probably a vast underestimate,” but let me explain. First, the population density of Afghanistan in areas where most of the fighting has been taking place since the fall of Kandahar around December 10, 2001, is extremely low. Hence, the numbers of civilians killed in US/NATO operations is nothing in the order of Iraq (which is far more urban). I do employ media AND OTHER NON-MEDIA REPORTS (including from persons on-the-ground at times when available). Secondly, by that phrase I am thinking of a maximum order of twice the number of deaths I capture and report. As to the inference implied, my efforts (begun in October 2001) to “count the dead” in Afghanistan should not and cannot be used to critique, invalidate, dispute that which Iraq Body Count has been undertaking." http://www.zmag.org/content/showarticle.cfm?ItemID=14007

What I find "remarkable" about this is that Medialens would be so transparently lazy/sloppy in their research. After running a smear campaign against Iraq Body Count since early 2006, they must realise their claims on these important matters are going to be closely scrutinised by well-informed people. Yet still they can't get the facts right.

COMMENTS Post comment

 

Comment 01 – Russ Bridger October 26 2007, 09:17

The boys at Medialens have self-promoted their little smearfest all over the internet, and Herold's corrections will probably only appear at ZNet. And so... the web of distortions peddled by Medialens continues...

Rave reviews for Medialens on CloneNet:

"A major contribution..." - Norwich Buddhist Society

"Narrow [...] intolerant" - George Monbiot

"Lovely Stuff" - Shakin' Stevens

Comment 02 – Peter October 26 2007, 11:21

One of their traits is that they don't correct their own errors, even when the case against them is overwhelming. It seems they'd rather have their own subscribers thinking something which isn't true than have them thinking that Medialens have got it wrong.

The errors raised here will be an interesting test for them. Will they issue a clarification to their subscribers pointing out Herold's response to their misuse of his quote (the only really honourable route - and, thankfully, the route taken by ZNet) or will they hope that nobody notices?

Somehow, I suspect the latter will be the case. I also doubt that the few Medialens fans who post here (dav, Stephen, Woofles?) will encourage them to correct their errors. One mustn't upset one's leaders, after all.

Comment 03 – Ken Farrell October 26 2007, 12:15

From my (admittedly fairly limited) experience, the Medialens MO in these cases is to make vaguely suggestive, insulting remarks about the people who raise these errors. Always politely worded, of course, but the result is, in effect: "don't listen to this person, as he/she is an agent of darkness".

Comment 04 – momo October 28 2007, 14:22

interesting exchange in the comments section here on this point :

http://www.democratsdiary.co.uk[...]

Comment 05 – Peter October 28 2007, 16:08

I've copied the thread from democratsdiary.co.uk to a new thread at Media Hell, as David Wearing (who runs the democratsdiary blog) didn't want to continue with it.

Comment 06 – SimonC October 29 2007, 09:36

To paraphrase "ALP" - what I find remarkable about this is that MediaHell would be so transparently sloppy in their arguments.

So what if ML got the name of the project wrong - they were quoting the Toronto Star. Your vitriol would be better directed there. It has no real effect on the central argument.

The central thrust of the ZNet preface is Herold trying to validate his methodology in Afghanistan. He reaffirms (on ZNet) that his methodology probably resulted in vast underestimate. MLs point is that, if his methodology was used in Iraq by IBC, they believe IBC to be a vast underestimate too. I believe it too. Note that, at first sight, Herold's Afghan methodology appears MORE likely to sweep up civilian deaths than the more stringent/restrictive IBC methodology. He, at least employed "other non-media reports", thus giving him a wider pool of data to draw from.

Your argument that somehow ML have made a factual error is (except for the name of the project, which is not really important) unfounded. Nothing in Herold's preface detracts from the ML argument. Oh, he may make claims that he under-reported by a factor of half and no more, but THAT IS JUST HIS OPINION! He is no epidemiologist. He is a professor of economics. When he has scientific, peer-reviewed evidence that his figures account for half the actual deaths, he should share it. So should IBC.

And you should put up real arguments, not unsubstantiated vitriol.

Comment 07 – ALP October 29 2007, 10:20

SimonC wrote:
"MLs point is that, if his methodology was used in Iraq by IBC, they believe IBC to be a vast underestimate too."

You should get your facts straight, Simon. IBC used the same general approach as Herold, but they didn't use the same methodology.

SimonC wrote:
" Your argument that somehow ML have made a factual error is (except for the name of the project, which is not really important) unfounded."

Actually, they made several factual errors. And not just in quoting the wrong project and in stating that IBC used the same methodology as Herold.

For example, Medialens wrote: "There is no reason to believe that the application of the same methodology in Iraq is generating very different results."

In fact there are a number of reasons to "believe" that "very different" results would be generated in Iraq. And not just because of the difference in methodologies. As Herold notes, for example, the conditions in Afghanistan are very different from those in Iraq. Even your expert epidemiologists would acknowledge this. Perhaps Medialens should have consulted one before making the above lazy/sloppy statement.

I've documented some of Medialens's other factual errors (not Herold related) here.

SimonC wrote:
"...what I find remarkable about this is that MediaHell would be so transparently sloppy in their arguments."

Again, you should get your facts straight, Simon. "MediaHell" made no arguments on this subject.

Simon is, as readers will now be aware, another Medialens disciple - reflexively defending Medialens against all criticisms, without doing his homework first.

Comment 08 – Stephen October 29 2007, 12:22

ALP writes:

I've documented some of Medialens's other factual errors (not Herold related) here.

Actually, dav pretty much demolished that list of "factual errors".

http://www.mediahell.org/[...]

[Last line of above post removed – we don't allow speculation about the identities of other posters - Board admin.]

Comment 09 – SimonC October 29 2007, 14:49

ALP, I am aware that ML have not got many friends here, or at IBC. I am reminded of the old adage that if you tell a man something he knows, he will thank you for it. Tell him something he doesn't know and he will resent you for it.

Hence the current attitude to MLs stance over IBC. It's fundamental claim is that IBC is fatally flawed (it's numbers are too low). Consequently, IBC is used by the great and the good as a propaganda tool to cover up their aggression in Iraq. ML believe (as do I) that IBC do not have a vested interest in exposing the true number of Iraqis killed - rather, as the current establishment darlings they are quite enjoying their 15 mins of fame.

Now, with that clear there is very little point in flogging the tired nag that is methodology again. We all know IBCs methodology well by now - reports of a death in 2 reputable (generally western) MSM outlets. Herold went further to find his bodies - he SPOKE to people. Apart from that, please explain how the methodologies differ in any meaningful way.

And please give concrete, scientific reasons why conditions in Iraq and Afghanistan would generate different levels of "vast discrepancy". You say that there IS reason to believe results in Iraq would be different but that is just opinion. Give us facts. Talk of less built-up areas in Afghanistan is fine - but is it more or less likely to result in deaths, and in those being reported? I doubt you will be able to give concrete reasons and for good reason - neither Herold's nor IBCs methodologies are accepted ways of assessing excess mortality in a war zone. The Lancet's are. ORB comes close. I'll say it again - IBCs are NOT.

You can wrap yourself up in semantic arguments as much as you want - either you agree with the scientific, epidemiological view that IBCs figures are too low or you don't.

It is a shame that I am to be labeled as a ML disciple - I think my arguments pretty much stand on their own two feet.

Comment 10 – sonny October 29 2007, 19:24

"Herold went further to find his bodies - he SPOKE to people. Apart from that, please explain how the methodologies differ in any meaningful way."

Like so many other things being imputed to Herold, I think you're wrong that he's saying he "went further" than IBC. What he's saying is that he does not only use media reports, as implied, again falsely, by ML. IBC also says this. I don't think what he's saying is that he "spoke to people" either.

You and Media Lens should have done a little homework about this before making false claims about it. You wouldn't have even had to do that much. To find some of the methodological differences would just take, for example, reading the Toronto Star article where ML lifted the quote. Herold lists three reasons there for his comment. One is: "Herold's figures also do not include the victims of the Taliban."

One aspect of Herold's count is that it is limited to victims of US and allied bombings or ground attacks. IBC includes killings by any perpetrator.

As for differences with Afghanistan and Iraq that can create different results, Herold's second reason is that a lot of violence in Afghanistan has taken place in very remote mountainous areas, a condition which does not quite apply to Iraq.

The type of reasons Herold is giving do not really apply directly to IBC, but the ML brain shut down after finding their propaganda tool in "vast underestimate". No need to read the rest of the article or understand anything Herold was actually saying, or what his methods are.

I think you're labelled as an ML disciple because you show all the traits. You make silly excuses for ML's ignorant errors. You don't know what Herold's methods are. You don't know what IBC's are. You don't know that there is no one "accepted way" to measure deaths in a war zone. And you probably never even heard the term "epidemiology" until October 2004. You do seem to know what you want to believe, but that's about it.

Comment 11 – Raoul Djukanovic October 30 2007, 07:13

It is a shame that I am to be labeled as a ML disciple - I think my arguments pretty much stand on their own two feet.

Is that why you recycle their "mainly Western" sources fallacy then?

Comment 12 – SimonC October 30 2007, 11:57

sonny, that is a opinion piece. However, it doesn't answer my questions. Apart from it being your opinion (and that of ALP and Herold) that Iraq and Afghanistan are different, you fail to show how they are different. You fail because there is no accepted methodology to assess excess mortality by using reports from mainly western, english-speaking press reports.

What difference does it make if Taliban deaths are not included in Herolds figures? It hardly seems important when the whole point is that the entire methodology is flawed.

"And you probably never even heard the term "epidemiology" until October 2004. You do seem to know what you want to believe, but that's about it."

And that has relevance in what way?

Raoul, the sources are not in dispute. IBC list them in detail here:

http://reports.iraqbodycount.org/[...]

If the sources are not mainly western media outlets, I cannot see from this evidence where I got that wrong. I am not closed-minded on this, if you could point out the error in my conclusion I would be grateful.

IBC do a comprehensive job of covering reports of deaths in Iraq as they are reported. They generally report the lowest levels of mortality. Their methodology is acknowledged as being a "vast underestimate"of deaths. They are used by the establishment, who instigated an illegal war of aggression, as a true measure of mortality. Hopefully you see the problem here.

Comment 13 – Raoul Djukanovic October 30 2007, 17:18

If the sources are not mainly western media outlets, I cannot see from this evidence where I got that wrong. I am not closed-minded on this, if you could point out the error in my conclusion I would be grateful.

The implication, explicit in the early alerts, was that the Westernness of the sources meant they were recording fewer deaths than some other, unnamed, foreign media, which captured far more.

This is a figment of the Davids' imagination, as their last alert conceded (though they projected it back into a "challenge" from IBC). There are no such other sources and the "Western" news agencies are the most comprehensive ones I know of.

This isn't to say they capture all the deaths. But then IBC used to say they expected them to miss "many if not most". If ML is reduced to playing semantic games about the difference between this phrase and vast, I think we can judge for ourselves the meaningless of their protracted smear campaign.

None of which implies that the IBC tally is anything other than an undercount, as they've always said they would be (although I think they could make that more obvious on their website and have repeatedly suggested they do jsut that).

It's hardly the moral offence of the century, however, and even if they didn't exist, the world wouldn't suddenly "know" a million Iraqis had died (or do anything differently as a result).

Comment 14 – marc herold October 30 2007, 20:08

Hello compagneros:
If your wage is $20,000 a year and it is increased to $40,000 a year, I believe that represents a vast raise. Ask anyone who experiences such a change. I simply used the word "vast" in that sense, but it was highjacked by Media Lens and used as a propaganda weapon to critique IBC.
Counted civilian impact deaths in Afghanistan number some ~6,000-7,000; further indirect civilian deaths at some 20,000.Until someone can show me, I repeat show me, that 3-10 times more civilians were killed directly by US/NATO actions since October 7, 2001, well until then, I don't care which so-called experts are cited, etc. Naturally, counts are different from estimates. My approach relies on much more than pure media reports, everyone capito? Indeed, I include photos supplied by NGOs working in high-conflict zones.
Does anyone READ what I wrote in the much-maligned correction: "First, the population density of Afghanistan in areas where most of the fighting has been taking place since the fall of Kandahar around December 10, 2001, is extremely low. Hence, the numbers of civilians killed in US/NATO operations is nothing in the order of Iraq (which is far more urban)." For numbers approaching 70-100,000 in Afghanistan to hold, entire sections of Kabul, Kandahar, etc. would have had to be flattened (as they were in the intra-mujahideen fighting of the early 9990s). But none of that happened since October 2001.. For this comment to be even considered, does it need to be peer-reviewed? Get real. For 2-3 decades, I have fought and will continue to fight such infatuation and awe of peer-reviewed journals. By that token, terrific journals like Monthly Review, New Left Review, NACLA's Report on the Americas, etc. would count for nothing.
As far as the much ballyhooed Guatemala study, I'd be delighted to shred it when asked. And for one Simon C. above, since when are economists (and engineers - I have an electronics engineering degree...so much for this stupid pedigree nonesense...anyone on the so-called Left should be ashahmed of themselves for playing that game) an inferior breed (including in their mastery of mathematics) to epidemiologists? This type of gratuitous inuendo is pathetic.
Professor Marc W. Herold, University of New Hampshire

Comment 15 – SimonC November 1 2007, 10:21

Thanks for the reply Marc. I am sorry you infer that your pedigree is being attacked - not my intention at all. You are clearly not an inferior breed. I truly apologise for any offence caused. I only intended to point out that, if you want an idea of excess mortality, an estimate carried out in Iraq by epidemiologists using accepted international methods, complying with accepted international standards and peer-reviewed is as near to best practice as possible. Leaving ego aside for one moment your methodology isn't best practice, is it? There are just too many unknowns there. I would be grateful, though, if you could shed light on some of them:

Are you now stating definitively that the true total of excess deaths in Afghanistan/Iraq is actually twice what you/IBC are reporting? (I infer this from your salary analogy).

Are you stating that your methodology is more likely to catch reports than the IBC methodology? (After all, as you state, your sources were more varied/less rigorous).

You asked for proof that your figure in Afghanistan could be 3-10 times higher than your methodology identifies. As far as I know that isn't available. It is, however, available for Iraq. Is this acceptable evidence? If not, why not?

Raoul, I am not sure you answered my point. You claimed the "mainly western sources" was a fallacy, which I recycled. When I asked for clarification you changed direction - are you stating here that there are only western sources:

"There are no such other sources and the "Western" news agencies are the most comprehensive ones I know of."?

I don't think anyone at ML would argue with this. Their concern is the inherent bias in the western sources.

Good point on the million casualties, though.

Comment 16 – ALP November 1 2007, 11:58

SimonC wrote:
"Leaving ego aside for one moment your methodology isn't best practice, is it?"

The "best practice" for counting actual deaths is obviously different from the "best practice" for sampling, conducting interviews or statistical extrapolation. "Leaving ego aside", perhaps you could recommend to Professor Herold (based on your vast experience in the field) improvements he might make to his practices for counting deaths in Afghanistan.

Simon wrote:
"an estimate carried out in Iraq by epidemiologists using accepted international methods, complying with accepted international standards and peer-reviewed is as near to best practice as possible".

Simon, again, you need to check your facts. We don't actually know whether Lancet 2006 complied with "accepted International" methods/standards, since nobody (outside those who conducted the study) knows, for example, what sampling procedures they used for most of Iraq.

Given the importance of randomness of sampling, this issue couldn't be more important. It certainly isn't "best practice" to present an extrapolation from 300 recorded deaths to an estimated 600,000+ without demonstrating how random sampling was achieved.

Also, in an email to me, Gilbert Burnham, the lead author of the Lancet 2006 study, wrote that because there are so few surveys done in conflict situations, and even fewer done in urban, middle-development countries, there is no "standard" or norm for these studies.

The idea that you can import epidemiological methods (which weren't designed for use in conflicts, and which are not well-validated in conflicts) to estimate deaths caused by violence in non-"standard" conflict situations, whilst not publishing the details of the sampling methodology used, and then declare it "best practice" - well, it's not "credible".

Comment 17 – Raoul Djukanovic November 1 2007, 17:12

I don't think anyone at ML would argue with this. Their concern is the inherent bias in the western sources.

Sorry, but this is meaningless. The bias inherent in news agency reports can only really be discerned in framing - it's extremely rare that they don't report facts somewhere, if those facts are known about. Unless you can demonstrate that they're suppressing information (as opposed to burying it, together with its meaning, as Chomsky and Herman charged) then there's nothing to be concerned about, assuming your concern stems from wanting to help IBC be as comprehensive as possible (as opposed to telling everyone to be a propaganda modellist).

That a news agency bureau chief concedes he doesn't report every death he hears about is hardly surprising - since when was it the job of journalists to act as mortuary registrars? In any case, IBC has always said large numbers of deaths are bound to go unreported, so the semantic disputes over the meaning of "vast" (and the implied moral deficiencies of everyone who's not given to "good" smears and misrepresentations in the name of "helping Iraqis") seem a bit of a sick joke to me.

Comment 18 – sonny November 1 2007, 20:41

"You asked for proof that your figure in Afghanistan could be 3-10 times higher than your methodology identifies. As far as I know that isn't available. It is, however, available for Iraq. Is this acceptable evidence? If not, why not?"

I would say it is definitely not available for Iraq. To the extent that survey estimates could provide such proof, those that have been put forth so far for Iraq differ quite sharply (including the two Lancet ones, contrary to deceptive claims by its expert authors about that). So surveys could not provide such proof here, unless one just discards those which run contrary to the "proof" (which is what you appear to do).

But how much proof is there in what's left, even after the inconvenient ones and contradictory evidence are sent down the memory hole?

With Lancet, a group of epidemiologists from the United States employed a group of anonymous interviewers from Iraq to do a survey.

Then the latter tells the former that they did interviews in 47 neighborhoods in Iraq. Is there proof of that? No.

They say that respondents for these interviews were selected at random. Is there proof of that? No.

They say they confirmed reports of deaths in these alleged interviews using death certificates. Is there proof of that? No.

The authors claim this alleged survey produced a random sample which is suitable for extrapolation. Is there proof of that? No.

What is most sorely lacking with all this is, well, proof. There isn't a scratch of actual proof for anything about it.

As for "accepted international methods" which are "near to best practice", as ALP points out there is little that is standardized for measuring violent deaths in a war zone. Bombings and shootings is not what epidemiology was designed to deal with in the first place. Like a number of methods and approaches, it could provide insight into such a topic, but that does not make it "accepted" above all others (and by whom anyway?). For that matter, some "leading experts" dosn't seem to find the Lancet methods all that "acceptable" or trustworthy, and seem to also have little problem "accepting" those of IBC, such as this one: http://www.taipeitimes.com/News/[...]

Also, "accepted" or "best" practice for survey research includes such measures as discussed here: http://www.amstat.org[...].pdf
Take a look at it. Not only is the Lancet approach not "near to" this best practice, it turns these "accepted" measures for ensuring reliable survey data on their head, almost completely on every single point. For this and other reasons, one could not use the Lancet study as "proof" for things like you claim, because it is entirely devoid precisely of proof.

Comment 19 – finn mccool November 2 2007, 16:22

As far as I can understand, it would make no sense whatever to interpret these comments about how many deaths their counts catch as referring to total excess mortality, but as it doesn't appear to be as clear to everyone, there is perhaps something I'm missing here.

 

Post comment

   
Name:
Email:  
Display email
address?
  Don't display    Display
Lifespan of comment   Delete after 3 weeks    Keep permanent if post is permanent  

Comment:

 
Optional link URL:
eg "http://mediahell.org"
 
Optional link text:
eg "Media Hell"
 
 

 

  Messageboard Back to top