Interview by Kristin Caulfield & Megan Countey
Dara Kay Cohen has always been interested in studying violence, but she explored several different directions before she found her current path. Initially, Cohen considered becoming a lawyer working within the US context on sexual assault and violence issues. While attending Brown University for her undergraduate degree, Cohen volunteered as a rape crisis counselor and interned at the Rhode Island Attorney General’s Office in the Domestic Violence/Sexual Assault Unit. After graduation, she went to work as a paralegal in Washington, DC, for two years in the Department of Justice’s Counterterrorism Section.
Cohen had only been part of the small unit for two months when the United States was attacked on September 11. She began to think more about the issues that the Counterterrorism Section was exploring: questions about terrorism, political violence, and victimization. Cohen realized that she wanted to focus on approaching these issues from the perspective of a scholar rather than as a lawyer.
With this focus in mind, Cohen was admitted to the political science PhD program at Stanford University to study American politics and international relations. But before she started at Stanford, she went on a trip through Europe and spent a week in Sarajevo. Learning about the war there, and its horrific conflict-related sexual violence, reawakened her original interests. Her experience in Sarajevo stayed in the back of her mind during her early years at Stanford and while she studied with one of her advisors, Harvard graduate Jeremy Weinstein, who had analyzed insurgent groups as part of his dissertation. Hearing him discuss his research on civilian abuses and his fieldwork with former fighters inspired Cohen to focus her own dissertation on wartime sexual violence.
Tell us about your current projects.
I’m currently working on two projects that examine sexual violence during wartime. The first is a book project where I study rape during civil wars. There are two main parts to the book project. The first is a set of comparative case studies based on my fieldwork interviewing ex-combatants and non-combatants in three post-conflict countries: Sierra Leone, El Salvador, and East Timor. The second part of the book project is a statistical analysis where I use original cross-national data to examine all major civil wars from 1980 onward and test arguments about correlates of mass rape during wartime.
My second project is serving as a principal investigator on a project funded by the National Science Foundation with partners based at the Peace Research Institute of Oslo (PRIO). We are collecting very detailed data about sexual violence at the level of the armed group. Using a broad definition of sexual violence, and a different definition of civil war—that includes much smaller scale conflicts—we are gathering available information from a number of publicly available sources. These are the most detailed cross-national data ever collected on wartime sexual violence, and our hope is that these data will be of use to both the academic and policy communities.
You have written about the challenges of gathering accurate data on sexual violence. What are some of the concerns you’ve raised?
In terms of the quality of the data, sexual violence is an incredibly challenging topic for a number of reasons. One reason is that there isn’t a consensus definition of what sexual violence is—and as a result, it’s been difficult to accumulate knowledge in this realm of research. Some scholars define sexual violence as rape, while others include a wider range of violations, such as forced abortion, forced sterilization, or forced marriage. Still others might include sexual insults and forced undressing—things that don’t involve direct force, but are clearly sexualized.
Beyond definitional issues, there are biases in reporting. A great deal of the existing data on lethal violations comes from news reports, so scholars will scour newspapers and code how many people died in a particular incident, and then use that data to estimate the number of deaths of a conflict-year. But it is not possible to do that with sexual violence because rape is not reported in the same way. One way of dealing with this challenge—and I’ve done this in my own research—is rather than measuring the number of victims or incidents as scholars have done with lethal violations, is to instead code qualitative descriptions of the relative severity of the rape in a particular conflict-year. For example, was rape described as being very infrequent, common, or on a massive scale? These sorts of qualitative descriptions are not precise, but they help to avoid the concerns of defining exactly how many victims were attacked.
Although it is often correctly assumed that rape is generally underreported, that is not always the case. There are complicated politics around rape statistics. I recently co-authored an article that focused on a controversy about the extent of sexual violence in Liberia. My colleague and I were struck by a statistical estimate of wartime rape in Liberia that has started becoming conventional wisdom: seventy-five percent of women in Liberia were raped during the war. This statistic has been repeated again and again—and after examining the sources for the estimate, we concluded it was impossible. To put the seventy-five percent number into context, in other cases of mass rape, very well-done population-based surveys have uncovered something like eight to ten percent of the population of women who might report conflict-related rape. We investigated the seventy-five percent claim a bit and it seems that it was based on a misreading of a particular survey. The interesting question we explore in the article is not so much why this survey was misread, but how did that seventy-five percent statistic start gaining such traction? Why didn’t people question it? And most importantly, what are the effects of statistics like that on the victims of sexual violence in Liberia and in future conflicts?
And finally, there are also biases in the ways researchers approach questions of sexual violence. One of the reasons we don’t have a lot of information about male victims or about female perpetrators of sexual violations is that assumptions about gender are so deeply embedded, even among researchers, about who is a perpetrator and who is a victim. In cases where researchers have asked more gender-neutral and open-ended survey questions—such as the number of perpetrators and the sex of perpetrators—they’ve had surprising findings. Studies in Liberia and in the DRC have found both male and female victims who report both male and female perpetrators.
The other related issue is not so much the way that researchers collect data, but the way that victims report it. For example, victims who have experienced a great deal of violence might be more likely to report lethal violations rather than sexual violence. There are some cultures where women may be more likely to report violations against their male relatives than violations that either they or their female relatives experienced. Finally, if a woman is raped in a private room by a soldier it might be unlikely that she would report it. However, if someone is raped by multiple perpetrators in front of their village or family—there are witnesses. They may be more likely to report that violation. There is no easy way to correct for those biases but it is important to be aware of them.
You recently wrote a blog post arguing against the claim that wartime sexual violence is decreasing and the difficulties in accurately analyzing temporal trends.
There is currently a great deal of discussion and debate in the field about trends over time. The authors of the November 2012 Human Security Report argued that because there are fewer wars now and that these wars are generally less lethal, that it is very unlikely that wartime rape is increasing, and actually, they claimed that it is probably decreasing. My blog response, coauthored with two colleagues, was that we cannot assume that rape and killing are perfectly correlated. If we look at specific conflicts, for example, there are clearly cases where killing and sexual violence are not correlated and actually follow very different temporal patterns. More broadly, my coauthors and I argued that on a global scale, it doesn’t really matter if incidents of rape are declining or increasing. What matters is that we know that there are a number of current conflicts with reports of mass rape. For the individual victims in, for example, the DRC or Syria, the focus on the global trend is meaningless. It may actually be harmful, if policymakers believe the problem is getting better, and funds are directed away from mitigating the consequences of sexual violence.
Your work has a lot of policy implications. Are you aiming to impact policy?
That’s one of the great joys of working in a policy school. I have a unique opportunity—and am actively encouraged—to translate some of my work to readers outside of peer-reviewed academic journals. I very much welcome engagement with the policy community about the implications of my findings.
In terms of policy implications, one of the important things I’ve learned from my book project is that much of conventional wisdom about the causes of rape during wartime is based on the few cases that are widely studied: Bosnia, Rwanda, and increasingly, the DRC. These common arguments include country-level features: countries that have more pronounced gender inequality may be more likely to experience rape during wartime. They also include features of the conflict itself. Some argue that ethnic wars, or wars experiencing genocide, may be more likely to have rape during wartime. What I do in my book project is look at the entire universe of cases of civil wars to determine if those arguments can explain, in general, the phenomenon of wartime rape. I find that a lot of the conventional wisdom, even if it explains individual cases, is not strongly associated with rape in wartime in general. For example, while ethnic hatred clearly played a role in rape in Rwanda and Bosnia, I do not find that ethnic wars are systematically associated with rape in wartime.
The key puzzle is really why, even within the context of the same war, do some armed groups rape and others do not? My argument is that the most important source of variation is on the level of the armed group. For example, in the case of Sierra Leone, there were two armed groups with almost identical types of fighters—same age group, same religion, same backgrounds, same professions before they became fighters—but one group committed a massive amount of rape and the other mostly refrained from committing rape. I argue that the key to understanding variation in wartime rape really lies in understanding the differences between armed groups. I ultimately argue that armed groups that recruit their fighters through kidnapping use rape as a socialization tool to create cohesion; I show that groups that use voluntary means for recruiting fighters are much less likely to be reported as perpetrators of rape.
And that’s unfortunately a hard sell to the policy world because it’s not immediately obvious what sorts of interventions should then follow. Although I’ve written policy briefs to brainstorm some potentialy useful policy interventions, I’m hesitant to make very specific recommendations, in part, because I don’t really see that as my role. What I hope instead is that my findings inform policymakers about the patterns in—and correlates of—wartime rape. If I can provide a better sense of the overall variation—where rape is reported, the various forms of sexual violence that were committed, who the perpetrators and victims are—then this could ultimately lead to more evidence-based policy.
You’re researching some of the worst aspects of humanity. How do you cope with the emotional impact of this subject? What advice would you give to undergraduates who are considering this difficult course of study?
Overall, I have found studying this topic inspiring rather than depressing. I did all of my fieldwork after wars had ended, and it is easier to interview someone about the violence they experienced years later—when you’re seeing them in their house and they are surrounded by their children and have, in a way, moved on. Although initially I was anxious about interviewing perpetrators of atrocities, one of the things that I learned, and what I end up arguing in my book, is that many times people have to be coerced to commit terrible violence. I strongly disagree with scholars who argue that violence, including rape, is part of human nature. In fact, most men—and the arguments are mostly made about men—even when given ample opportunity, even when they’re armed, even when their victims are not, don’t take advantage of that opportunity. There was something very enlightening about hearing during my interviews that many men were themselves victims of violence and many of them described—especially in the contexts that I’ve been studying—having been abducted, kidnapped, and strongly pressured into committing rape. Interviews with perpetrators are especially important. Much of what we know about atrocities during wartime, including rape, is extrapolated from victims’ testimonies. Both the research community and the advocacy community have been focused on victims—which is a good and important thing to do. But the next frontier for research is not to just assume the motivations of the armed groups based on the statements from victims, but to talk directly to the members of the armed groups and try to understand motivations for violence from their perspective. Most importantly, was rape something that was ordered, or was it simply tolerated? To really get to the root causes we need to talk directly to the perpetrators.
On the other hand, doing this work can get to be very difficult. Probably one of the reasons that I felt prepared is because I had previously worked as a domestic-violence advocate. I was used to hearing people describe terrible violence they had experienced. But it can be grueling to do those types of interviews. There are cases of journalists who have interviewed victims of wartime violence and who develop post-traumatic stress disorder (PTSD). I tend to be hesitant in encouraging young students to take that on,s and I do worry about students being prepared for the sorts of things they will be facing.
However, for students who prefer not to research violence through interviews, there is a wealth of existing data about human rights violations, including court transcripts, Truth and Reconciliation Commission victim statements, and other types of archival data like police reports that can be analyzed by undergraduate students interested in these issues.
Tell us something about yourself that most people won’t know in your academic world.
I love to bake. If I hadn’t become a professor, I might have become a pastry chef—I have a pretty serious dessert-making hobby. I like to experiment, and I look for recipes that are especially challenging or unusual. What’s really satisfying is that you have a finished product in a few hours. That never happens with academic research.
Dara Kay Cohen is a Faculty Associate of the Weatherhead Center for International Affairs and an assistant professor of public policy at Harvard Kennedy School. Photo credit: Megan Countey.