BBC News publishes my response to Les Roberts

After a prompt from an advocacy group, BBC’s Paul Reynolds added comments from epidemiologist Les Roberts to a BBC News piece on the WikiLeaks Iraq war logs. I complained to Paul that Roberts was using the space to reheat his old attacks on Iraq Body Count (IBC), and that his comments (with one exception) had little relevance to the subject matter of the article.

I also pointed out (with examples) that most of Roberts’s claims were either misinformed, unsubstantiated or simply false. Paul, to his credit, immediately published my comments on the BBC News page. But what he published was an edited version of comments which I’d already whittled down (to get the basic points across to someone who wasn’t already familiar with the studies/background). Here is the original version of my comments:

[Les Roberts] “A) It is likely that the IBC and Wikileaks reports tend to have the same lens (many reports coming from the Government of Iraq, oversampling of Baghdad, oversampling of the largest events and missing single killings).”

Robert Shone: On the previously unknown 15,000 deaths, IBC point out that: “The majority of these new deaths come from small incidents of one to three deaths”. If the WikiLeaks report had the same “lens” as IBC on size of incidents, as claimed by Roberts, this wouldn’t be the case.

[Les Roberts] “B) We have shown that most violent deaths in the press over the first 4 years of the war are not in IBC because they (cautiously) required multiple press reports and (unavoidably) used a few search terms that did not capture all events. (see attached)”

RS: Roberts hasn’t “shown” this anywhere (unless the “attached” which Roberts mentions contains something new and previously unpublished – what was this “attached”?). I suspect he’s making similar unfounded claims to the ones he makes in C), which I’ve already shown is false.

[Les Roberts] “C) In Baghdad where we believe the press coverage was by far the best, we showed that most violent deaths reported in phone and Skype interviews were not in IBC.”

RS: That’s false. Roberts’s study found that “38%” of deaths “were absolutely not in the [IBC] database”. The majority (62%) were included by IBC or can’t be ruled out as not included by Roberts’s Mickey Mouse study (which used a truly massive, comprehensive sample of only “18 primary interviewees”! – see my blog post for more details). The only way Roberts can make the above (false) claim is by ignoring the records in IBC’s database which came from morgues, etc (and which therefore Roberts couldn’t match, due to lack of detail). It’s very dishonest. Les Roberts’s study: http://pdm.medicine.wisc.edu/Volume_23/issue_4/siegler.pdf (page 3)

[Les Roberts] “D) There are just so many things that are not consistent with 120,000 deaths! The ORB 11/07 and BBC (see: http://news.bbc.co.uk/1/shared/bsp/hi/pdfs/19_03_07_iraqpollnew.pdf ) polls that are completely at odds with the IBC implication that 1 in 20 or 1 in 25 Iraqi households have lost someone to violence. The ORB implication that 1 in 4 households have lost someone matches all the ground reports I hear. You cannot have the Iraqi Ambassador reporting half a million new war widows or UNICEF speculating that there are a million orphans if there are 120,000 war deaths.”

RS: The comment about widows is completely bogus, as around 490,000 women would be widowed over a seven year period regardless of war (making a pro-rata population comparison to rate of widowhood in the US, for example – an over-simplistic comparison, to be sure, but it does underline the basic point that other things than war create widows). This issue has been debated a lot, and the conclusion among the informed seems to be that there’s currently no way of knowing how many of the widows are due to the recent conflict, to previous conflicts (80s & 90s), other factors, etc. As for the rest of Roberts’s comment here, this is where some real “balance” would come in useful. For example, the ORB study has recently been convincingly demolished in a peer-reviewed study published in Survey Research Methods [2010] Vol.4, No.1. Even Les Roberts’s colleague, Francesco Checchi, thinks there are serious problems with it (he was quoted as saying so in a recent BBC report*). So it might be better to use a more credible study, such as IFHS, as a comparison. So much for “balance”.

*Checchi said the ORB figure was “implausible”, that it had a “major weakness” (echoing the Survey Research Methods study findings). He added that the Iraq death count was “likely to be between 200,000 and 500,000”. (BBC World Service, 27 Aug 2010)

………………………………………………………

Published 8.43am October 28 2010