- In Alerts 2004
- Post 04 November 2004
- Last Updated on 04 November 2004
- Hits: 11785
By way of a splendid coincidence, Media Lens received this delayed response from the BBC on November 3:
"I am writing in response to your email to Caroline Hawley dated 13/10. Apologies for the delay in replying. The estimate for the number of people killed by Saddam Hussein is based on the figure from Human Rights Watch. They estimate that as many as 290,000 Iraqis were 'disappeared' by the Iraqi government over the past two decades, and that many of these 'disappeared' are those whose remains are now being unearthed in mass graves in Iraq.
"The investigators are working for the Iraqi Special Tribunal which was is under the jurisdiction of the Iraqi interim government.
Notice the lack of scepticism about the methodology used in arriving at the figures, and about the credibility of the Iraqi Special Tribune - there are no reservations, doubts or caveats.
On October 13, we had written to BBC correspondent Caroline Hawley asking the basis for her claim that Saddam Hussein had killed 300,000 Iraqis. In her BBC News report, Hawley had also expressed no reservations about the figure.
By contrast, a BBC website review of last week's Lancet report of 100,000 excess civilian deaths in Iraq since the invasion, included several balancing comments:
On November 2, we received this response from Channel 4 News science reporter, Tom Clarke:
"The weakness I refered to in my piece is a direct reflection of the limitiations highlighted by the researchers themselves in the discussion of their results. These limitations were also referred to by the lancet's editor Richard Horton in his commentary that accompanied the article. I direct you [to] the entire of page six of their paper.
"Remarkably, neither the MoD nor the FCO keep a tally of civilian casuaties in Iraq, and Downing Street dismissed -- pretty much out of hand, the only attempt to do so. By pointing these facts out in my report, I hoped the viewer might draw their own conclusions. My piece reported on the best attempt so far to quantify civilian casulties in Iraq, yet would have been seriously biased, had I not made plain its significant limitations (a study with a 95% confidence interval for the final death toll rangining from 8,000 to 194,000)."
One of the report's authors, Richard Garfield, has responded to the same challenge:
"Research is more than summarizing data, it is also interpretation. If we had just visited the 32 neighborhoods without Falluja and did not look at the data or think about them, we would have reported 98000 deaths, and said the measure was so imprecise that there was a 2.5% chance that there had been less than 8,000 deaths, a 10% chance that there had been less than about 45,000 deaths,....all of those assumptions that go with normal distributions.
"But we had two other pieces of information. First, violence accounted for only 2% of deaths before the war and was the main cause of death after the invasion. That is something new, consistent with the dramatic rise in mortality and reduces the likelihood that the true number was at the lower end of the confidence range. Secondly, there is the Falluja data, which imply that there are pockets of Anbar, or other communities like Falluja, experiencing intense conflict, that have far more deaths than the rest of the country. We set aside these data in statistical analysis because the result in this cluster was such an outlier, but it tells us that the true death toll is far more likely to be on the high-side of our point estimate than on the low side." (Professor Richard Garfield, email sent to Media Lens reader, October 31, 2004)
It is important to recognise that providing 95% confidence intervals is standard in scientific analysis of data. While Clarke was correct to mention these confidence intervals, he misrepresented the limitations of the report in stating that "the study's main weakness... is that it multiplies a small sample across the whole of Iraq".
This was not, in fact, a serious weakness. In response, we repeat here the comments made by report author Dr. Gilbert Burnham, quoted in part one of this alert:
"Can one estimate national figures on the basis of a sample?
"The answer is certainly yes (the basis of all census methods), provided that the sample is national, households are randomly selected, and great precautions are taken to eliminate biases. These are all what we did. Now the precision of the results is mostly dependent on sample size. The bigger the sample, the more precise the result. We calculated this carefully, and we had the statistical power to say what we did." (Dr. Gilbert Burnham, email to David Edwards, October 30, 2004)
Lancet editor, Richard Horton, emphasised these points in his editorial:
"This remarkable piece of work represents the efforts of a courageous team of scientists. To have included more clusters [of households interviewed for the survey] would have improved the precision of their findings, but at an enormous and unacceptable risk to the team of interviewers who gathered the primary data. Despite these unusual challenges, the central observation - namely, that civilian mortality since the war has risen due to the effects of aerial weaponry - is convincing". (http://image.thelancet.com/extras/04cmt384web.pdf)
In addition, to reiterate, the twin facts that:
(a) violence accounted for only 2% of deaths before the war but was the main cause of death after the invasion, and
(b) the high death-rate Fallujah data was discounted in the estimated tally of 98,000 excess deaths, mean that any movement away from this figure is likely to be on the high side of the estimated range, thus +greater+ than 98,000 deaths.
The Lancet report authors candidly and correctly weighed up +all+ the possible limitations of their study. Having done so, the estimate of 98,000 deaths is, based on all the evidence and its careful interpretation, a +conservative+ estimate. The researchers also, correctly, call for verification of these results by an independent body such as the World Health Organisation. For Channel 4 news to dismiss so easily the care, courage and scientific rigour of this report is remarkable.
In a BBC Newsnight debate - the only debate we have seen on the Lancet report - Michael Clarke of the International Policy Institute challenged the Lancet's editor, Richard Horton:
"This 100,000 figure, remember, has a huge margin of error. It ranges from 8,000 - which is half what most of the rest of us think it is - to 194,000. And what they've done is split the difference and said, 'Well, we think about 98,000 with some measure of confidence,' because there is a sort of a confidence statistical factor built into that. But that really isn't, to my mind, very credible." (BBC2, Newsnight, November 2, 2004)
"Well that's not true. What you just heard isn't a correct summary of the research."
Horton began explaining that the figures were higher than previous estimates because this was the first empirical research of Iraqis themselves carried out in Iraq. Newsnight anchor, Gavin Esler, then interrupted, starkly revealing his failure to understand the figures:
"But you haven't got 100,000 death certificates, you haven't got 100,000 bodies. You've got somewhere between 8,000 and 194,000 is where you've put it, and you've gone in the middle."
Horton continued his reply:
"But that again is a misunderstanding of the figures. The most likely estimate of excess deaths is 98,000. It's +not+ right to say that it's equally likely it could be between 8,000 and 194,000. The most likely figure is 98,000, and as soon as you go away from that figure, either lower or higher, it's much less likely it will be much lower or higher."
Remarkably, Clarke then instantly renounced his claim that the authors had simply "split the difference", before moving on to a second claim:
"Statistically, that's a reasonable assumption, but the numbers are too small. I mean remember, seven people - very brave people, who did a series of interviews in 33 different places - but seven people, spent a month interviewing the Iraqi families that they went to visit. And on the basis of interviews, and what Iraqi families told them, they've made these extrapolations. But the numbers are really very small. The number of confirmed deaths that they look at is 61 - not 61,000, or 610, but 61."
Of course the idea that the "numbers are too small" is +exactly+ the kind of issue that the serious scientists, peer reviewers and Lancet editors involved in producing the report would have focused on first in considering the merits of the research. According to the report's authors, it is simply not true that "the numbers are too small". Based on the chosen sample size, the 100,000 figure can be advanced "with great certainty". This is simply a matter of scientific logic.
Richard Horton concluded his comments by noting the shocking truth that "there has been no debate" on the Lancet report.
The Charnel House - A Simple Question From A Couple Of Amateurs
Ahead of the war, almost all journalists accepted that Iraq under Saddam Hussein was one vast torture chamber for the civilian population. Journalists and politicians talked often of a "charnel house", a "giant Gulag" and so on. In October 2003, Labour MP Ann Clwyd said:
"In June this year I stood on the edge of Saddam's killing fields. I saw the skeletons of men, women and children being dug up in this enormous mass grave. I do not believe, and neither do you, that we should turn a blind eye to such atrocities." (Paul Eastham, 'Blair safe as tears wash away Labour rebellion over Iraq,' Daily Mail, October 2, 2003)
In June 2003, the Telegraph's Con Coughlin wrote:
"Another day and another mass grave is unearthed in Iraq". He added: "So many of these harrowing sites have been uncovered in the two months since Saddam's overthrow that even the experts are starting to lose count of just how many atrocities were committed by the Iraqi dictator's henchmen... If this were Kosovo, the Government would be under fire for not having acted sooner to prevent the genocide."
"Having just returned from three weeks in post-liberation Iraq, I find it almost perverse that anyone should question the wisdom of removing Saddam from power." (Coughlin, 'So what if Saddam's deadly arsenal is never found? The war was just,' The Sunday Telegraph, June 1, 2003)
Journalists and politicians appeared to truly believe that genocide was ongoing in Iraq immediately prior to the invasion. And yet it appears not to have occurred to even one of them to check these claims against the findings of the Lancet report, which studied Iraqi deaths in periods before and after the invasion, interviewing 8,000 people around the country.
The point is that the report has produced a thorough scientific analysis of deaths for a period of time when Saddam Hussein was said to have been on a murderous rampage. It took a matter of moments for us - a couple of amateurs with virtually zero resources - to write to the report authors and ask them what they had found:
"Did your research uncover evidence of mass murder by Saddam Hussein's regime in the year prior to the invasion? There have been government and media claims of tens, even hundreds, of thousands murdered, whereas Amnesty International told me they estimated the figure to be in the hundreds. Did your research cast any light on that?" (David Edwards to Dr. Les Roberts, October 30, 2004)
The next day, we received this reply from the report's lead author, Dr. Les Roberts:
"There was one reported killing by Saddam's folks in the first days of the war (considered post-invasion) and there were a couple people who disappeared during the invasion (all adult males). We did not count disappearances as deaths. Thus, no, we have no evidence of that. That does not prove it did not happen. If it was only hundreds of deaths, our small sample probably would not have detected it." (Email to David Edwards, October 31, 2004)
Not one journalist has commented on the significance of the absence of evidence for claims that Saddam Hussein was involved in mass killing in the year prior to the invasion. Nor, in the case of reports of 300,000 alleged victims of Saddam Hussein, have there been consistent expressions of scepticism about, indeed dismissal of, 'dubious' methodology, statistical analysis and interpretation of results.
Coda - What Lester Luborsky Saw
Consider that, right up to the very eve of war last year, Tony Blair, Jack Straw and others insisted that "even now" a peaceful solution was possible. It was simple - all Saddam Hussein had to do was to comply fully with UN resolutions, cooperate with UN weapons inspectors, and give up his weapons of mass destruction. If that happened, the troops amassed in the desert would all go home - everyone would be happy. Journalists consistently reported Blair's claim that he was "going the extra mile for peace" as credible and sincere.
And yet, to our knowledge, not one journalist ever discussed what would happen in the event of a peaceful outcome. What would actually have happened if Iraq had somehow been able to comply fully by US-UK standards so that weapons inspectors were allowed to continue their work to the point where Iraq was given a clean bill of health? Would sanctions then have been lifted, leaving Saddam Hussein in power, Iraqi oil out of Western hands, and "coalition" forces wandering happily home? Would the US have been required to pick up the tab for sending a quarter of a million men around the world only to bring them straight back?
The reason no journalist ever discussed this possibility, we believe, was because they knew the US-UK governments would never permit such an eventuality. In other words, at some level, journalists +knew+ war was inevitable.
But, paradoxically, many of them also, simultaneously, +did+ believe that Blair and Straw were sincere in talking of a peaceful resolution. The intense propaganda brainwashing that persuades us of the fundamental benevolence of our leaders was in collision with the obvious realities. Media reporting reflected the propaganda but steered clear of areas of thought where the propaganda was quickly rendered absurd. And journalists were able to do all of this without knowing that they were doing it. But how on earth does this work?
In the 1960s, psychologist Lester Luborsky used a special camera to track the eye movements of subjects asked to look at a set of pictures. Subjects were asked to rate which pictures they liked and disliked. Three of the pictures were sexual in content, with one, for example, showing the outline of a woman's breast, beyond which a man could be seen reading a newspaper. The response of some subjects was remarkable. They were able to avoid letting their gaze stray even once to the sexually suggestive parts of the pictures. When later asked to describe the content of the pictures, these subjects remembered little or nothing sexually suggestive about them. Some could not even recall seeing the pictures. Psychologist Daniel Goleman explains:
"In order to avoid looking, some element of the mind must have known first what the picture contained, so that it knew what to avoid. The mind somehow grasps what is going on and rushes a protective filter into place, thus steering awareness away from what threatens." (Goleman, Vital Lies, Simple Truths - The Psychology of Self-Deception, Bloomsbury, 1997, p.107)
Like all of us, journalists need to believe they are honest, reasonable human beings. There is a problem, however - career success depends on their not finding +too+ much to criticise in the operation of elite state-corporate power. The art of successful mainstream journalism is the art of reconciling these two irreconcilables without admitting the lie to conscious awareness.
As with the individuals in Luborsky's experiments, journalists somehow understand what is going on and rush a protective filter into place, steering awareness away from what threatens. It is an awesome and intriguing phenomenon of human psychology.
The consequences are awesome, too - incinerated children, dismembered bodies, and families smashed and broken in depthless grief.
The goal of Media Lens is to promote rationality, compassion and respect for others. In writing letters to journalists, we strongly urge readers to maintain a polite, non-aggressive and non-abusive tone.