How have the responses been at the conferences where you have presented results of your survey (supervised by Royce Hutson)? I understand you presented results that were not part of the paper published in the Lancet.
We did present additional information at other conferences as well. For instance, Dr. Hutson spoke at the American Public Health Association conference about the identity of the victims of human rights violations. You'll remember that our piece in the Lancet focused on answering the question "who were the perpetrators?" At APHA, Dr. Hutson spoke about "who were the victims?" Not surprisingly, people who were politically affiliated -- with any party -- were significantly more likely to be murdered. Also, individuals from households affiliated with the Lavalas and Lespwa political parties were more likely to be arrested than people from other political parties and those who had no political affiliation at all.
People have been very responsive to our study, especially concerning our use of GPS. Dr. Hutson just got back from presenting our study at a conference in Geneva where the participants were primarily people in this field, that is, people who study war-related human rights violations, the effect of small arms on the population, mortality, etc. He can probably tell you more about that.
Traditionally mortality during armed conflict has been estimated by using passive accounts: relying on reports from key informants (such as health workers, community leaders, the Red Cross, etc) or through documents such as analyzing morgue records or media reports. The Iraq Body Count website is a good example of document analysis. They count the numbers of those killed in the Iraq war by using news media reports.
But the problems with this type of passive counting are numerous. For instance, if we rely only on key informants to supply us with the numbers of those killed or raped, then we may miss people whose human rights violations was never reported to that key informant (for instance, women who were too embarrassed to report being raped, or families who never took their family member’s body to the morgue because they didn't have the money to pay the mortician). Key informants may not have access to accurate numbers or they may inflate or deflate their estimates because of their own personal or political bias.
Counting human rights violations using media reports has drawbacks as well. One of the criticisms of the Iraq Body Count website is that the deaths of many Iraqis are never reported in the national news. Someone may die from injuries sustained during a gun battle in which they were an innocent bystander, and they may die weeks or months later. Does every death get reported in the media? Of course not, usually only those that are most newsworthy make it into the press.
People are starting to recognize that the human cost of armed conflict can’t be calculated by simply counting the number of combatants who fall on the battlefield. Armed conflict has devastating consequences to families and communities. Haiti is a good example. In 2004 we saw whole communities which were displaced because of the fighting. This was not a bloodless coup. People died. Children were raped. People’s homes were looted and destroyed. People were arrested but never charged with a crime and held in jail for long periods of time. Those who care about politics can argue until the cows come home about whether the coup was justified or whether more or less human rights violations happened after the ouster of President Aristide. But the fact remains that serious human rights violations including rape, murder, and illegal detentions were rampant during the 22-months after Aristide left Haiti.
Let me make another point here. Some critics have argued that the human rights violations against some people were justified or that any murders after the coup were in self-defense against the "tyranny of Lavalas." Let me quote an email that I received recently from an American who travels frequently to Haiti. He wrote: “I can understand their [referring to the perpetrators of human rights violations] desire to rape and pillage. Aristide and his goons persecuted the Haitian people for years. He [Aristide] ordered his chimere gangs to murder anyone who wouldn’t submit to his authority and then he gave them the guns to do it! If the people felt like taking things into their hands and getting some revenge, then so be it.”
Rape is never justified. I don’t care who you are or what your politics are, it is never okay for you to force anyone to have sex with you. In the same way, killing your political opponents, burning down their house, destroying their business, stealing their car, kidnapping their child, beating and torturing them – all of these things are wrong. There is never any justification for violating someone’s human rights. If you have a political disagreement with someone, work it out in the realm of politics, don’t rape their wife or murder their child.
The argument that Haitian people who committed human rights violations were motivated by revenge is also very racist and xenophobic. It implies that the Haitian people are incapable of controlling themselves while other (read: civilized) peoples work their problems without extracting revenge. This just isn’t true. The individuals and groups who engaged in massive human rights violations in Haiti did so knowingly and purposefully. They didn’t accidentally or in the heat of the moment kill people; no, those who tortured, raped and killed did it to terrorize a population.
People who commit massive and systematic human rights violations do so because they think they can get away with it and because they find it more expedient to cultivate political power through the violent suppression of dissent than by fostering a democratic society that promotes the free and open exchange of ideas.
Has anyone proposed testing the results of your survey by redoing it? I am thinking in particular of critics of the study. Can your survey be checked at this point or would recall bias be too big a problem?
No, no one has proposed redoing the survey to me. In any survey about past events there are going to be issues with recall bias. Recall bias happens with a survey respondent’s answer to a question is influenced by their own memory of what happened. For instance, if I ask you what you had for breakfast this morning, you’d probably remember. But, unless you eat the same breakfast every morning, if I asked you what you had for breakfast 30 days ago, you probably wouldn’t be able to remember.
We assume that when very significant events happen in a person’s life, they are able to remember the date or month and year in which they occurred. But as time passes, a person’s memory of the exact details of the event may fade. I remember getting my wallet lifted at Boston’s South Station 16 years ago, but I don’t remember how much money was stolen, the month it happened, or how I managed to take the train home with no money to pay for my fare.
Recognizing this, it's important to design a survey so that detailed questions are asked about recent events and only more general questions are asked about events that took place a long time ago. We used another technique to improve the reliability of subject recall as it related to event dates: each interviewer had a calendar that had been filled in with the dates of significant events in Haiti over the pervious two years. The calendar included events such as major storms, holidays, important things that were reported in the news media, etc. When a person reported a human rights violation, we tried to verify the date by asking what events occurred in the same time period and whether the violation happened before or after events for which we knew the exact date. Of course, no system is perfect, but researchers have developed techniques like this to improve the reliability of information gathered from survey respondents.
Could our survey in Haiti be replicated? Of course it could. We gave a very detailed explanation of our methods in the Lancet article so that people could replicate our type of study in Haiti or elsewhere if they wanted to use the Random GPS Coordinate Sampling methodology.
Of course, in doing our study we didn’t ask about the years before the departure of Aristide. Our study was not motivated by the desire to prove a particular political point; rather, we wanted to know what was going on in Haiti at the time and what the extent of the human rights problem was so that services could be developed to address the needs of survivors. So for us, it wasn’t so important to ask questions about two or three years prior. But if someone wanted to, they could. The information may not be as reliable as more recently recalled events, but I would assume that someone who has been sexually assaulted, shot, tortured or had their family member murdered would be able to recall the general details such as the date and the perpetrator five or ten years after the event. Less significant events may be harder to recall and the information might not be as reliable. Anyone who wanted to replicate our study could do it easily from our description in the Lancet article and I’d encourage them to really delve into the literature and see what the research says about recall bias related to the events we studied.
How do you respond to people who note that (in two cases) research published in the Lancet over the years have been exposed as fraudulent. The implication is that therefore what the Lancet publishes cannot be trusted. Is there a medical journal with a much better track record than the Lancet?
The Lancet is up there with the New England Journal of Medicine and the Journal of the American Medical Association. It’s a distinguished, reputable journal that has taken a very proactive approach to publishing public health studies such as ours, Les Roberts and his colleagues’ study on mortality in Iraq, studies about conditions in the Palestinian Territories, and others that deal with significant human rights issues.
People uphold the New England Journal of Medicine and say that it is a more trustworthy source than the Lancet but the New England Journal of Medicine has also withdrawn papers because of fraud. The news media has this problem as well; look at the New York Times, the New Republic and others who have had to withdraw reporting because it was fabricated. In research articles, people get caught often times because, using statistics, we can tell when data is fabricated. In my experience, fabricated data is never as real as real life.
Was the Haiti Human Rights Survey analyzed in this way?
Using statistics? Of course. We turned over the dataset to someone else, a professor who is an expert in this area, who not only reran the analysis we did for the article but also ran a series of statistical tests to see if the data was manipulated in any way or if there was bias associated with any one or a group of interviewers. And of course no irregularities or problems were found, which is why this accusation is a dead issue with all except a handful of individuals outside the academic world who have strongly-held opinions about particular political events in Haiti.
In looking at this issue of reliability, I think an important thing to remember is that you can’t judge a research article by looking at the beliefs or presumed beliefs of the authors. After all, just like most people get into cancer research because they think cancer is a bad thing, most people who conduct human rights research do it because they think rape and murder are bad things.
To know if you can trust a research article, you must first read the study (not the media reports about the study or someone’s convoluted list-serv post about the study) and know that you understand it. Then look at whether it has been peer-reviewed. Our study was reviewed by something like a dozen people, all of whom are experts in their field and all of whom made comments for suggested additions and revisions that strengthened the final paper. Look at the methodology. Is it sound? Does it use accepted practices? If it uses something new and different, is it well explained? Are there any problems in the methodology? Does the article use published peer-reviewed literature to back up the methodology and findings of the study?
What most surprised you about the way the corporate press responded to your survey?
I wouldn’t say that I was particularly surprised one way or the other. I mean, we all know that human rights violations happened in Haiti during and after the coup. Journalists know this as well, whether or not they report it, they know what happened. And the mainstream North American press responded to our study the way they have responded to most news about Haiti: with silence.
I was a little surprised that the Miami Herald ran a letter to the editor from Gerard Latortue decrying our study and using my name in an accusation about children being trained in terrorism. If the former prime minister had an objection to our study, then his administration should not have authorize us to conduct the study. We did get written permission from the upper level of Latortue’s administration prior to conducting the Haiti Human Rights Study and they were fully informed (in writing) of who was conducting the research and of the fact that I previously volunteered at the Lafanmi Selavi Center for Street Children. But then once Prime Minister Latortue heard that the results of our study included massive human rights abuses by government officials under his watch, then he objected to the study.
But overall I was most surprised and encouraged by the response our study received within Canada. Several newspapers, including the Montreal Gazette, reported on the study and the Canadian military responded by ordering an investigation into the actions of their soldiers who were part of the Multinational Forces in Haiti after the departure of President Aristide. I would hope that other countries and their military forces would be as open to investigation and willing to consider how their soldiers have violated the human rights of ordinary citizens while on peacekeeping missions.