The day after Haiti's elected president Jean-Bertrand Aristide was forced into exile by US troops in 2004, the British press was unmoved. Aristide had it coming, the Times thought, as he had been "increasingly despotic and erratic", while the Independent called him "a bloody dictator" and the Telegraph noted his "use of thugs to intimidate political opponents".
In a moment of unusual dissent, a letter in the London Review of Books by Peter Hallward, now of Middlesex University, tried to put events in context. "People with - generally tenuous - connections to Aristide's Lavalas party were probably responsible for around 30 killings in all the years he was in office." Set this against the 5,000 Lavalas supporters killed while Aristide was in exile from 1991-94, and the 50,000 deaths attributed to the Duvalier dictatorships.
Now we are learning more about bloodshed in Haiti. A study in the Lancet reports that, during the 22-month rule of the US-backed interim government that followed Aristide's departure, no fewer than 8,000 people were murdered in greater Port-au-Prince alone. Of these, 22 per cent were killed by the police, 26 per cent by anti-Aristide groups or the demobilised army, and 48 per cent by criminals. In addition, 35,000 women and girls were raped or sexually assaulted.
Professor Royce Hutson of Wayne State University, who co-wrote the study, said: "We didn't detect any Lavalas atrocities with regards to murder or sexual assault. We did detect some physical assaults . . . and some threatening behaviour."
In the entire US and British press three papers reported this: the Miami Herald, the Independent and the Guardian, and for the Guardian the story became news only after a letter to the Lancet challenged the report's credibility.
There is an echo here of the treatment given the October 2004 Lancet report that found 100,000 excess Iraqi civilian deaths since the 2003 invasion. This was dismissed at the time and has generally been ignored ever since, even though earlier mortality studies in the Democratic Republic of Congo by the same lead author, Les Roberts, and using the same methodology, were cited by the likes of Tony Blair and Colin Powell.
Last February, Roberts returned to the subject of Iraq, this time estimating that the civilian death toll there could have reached 200,000-300,000. Having examined the media's preferred study, produced by the Iraq Body Count (IBC) website - which puts the number of civilian deaths, based on reports by the media, at 43,000-48,000 - Roberts and colleagues found that "it cannot be more than 20 per cent complete".
This prompted a debate, and in a June letter to IBC, Roberts described his team's reaction to the website's most recent defence of its work: "We decided that it was so devoid of credibility, and so laden with self-interest rather than the interest of the Iraqis, it did not merit a response."
That judgement, from some of the world's leading epidemiologists, could hardly be more damning. And hardly of less interest to the mainstream media, which ignored it.
David Edwards and David Cromwell edit www.medialens.org and are the authors of "Guardians of Power: the myth of the liberal media" (Pluto, £14.99)