Rethinking Wartime Rape

Recent evidence gathered by Harvard’s Dara Kay Cohen suggests a new paradigm is necessary.

What drives wartime sexual violence? Conventional wisdom points to the sudden breakdown of government, primordial ethnic hatreds, or even patriarchal society writ-large. But a closer look at variation across cases tells a different story.

Three top experts in this field have recently published a United States Institute of Peace (USIP) Special Report1 debunking some of the most commonly held notions of wartime sexual violence. According to the report, sexual violence does not occur in all ethnic wars and does occur in many conflicts where there are no ethnic tensions. In Sri Lanka’s civil war for example, a classic case of secessionist ethnic conflict, rape by the LTTE was rare, whereas in Sierra Leone ethnicity was not a factor and rape was widespread.

Misconceptions, it turns out, abound. Armed groups in many societies with patriarchal norms frequently do not engage in rape. Wartime sexual violence is far more common among state militaries than rebel groups, perpetrators are not always men, and victims include both men and women.2

What explains this turnaround in thinking?

One recent set of contributions comes from recent research by Harvard’s Dara Kay Cohen. Cohen, assistant professor of public policy at Harvard Kennedy School and one of the authors of the USIP Special Report on wartime sexual violence, argues that instead of focusing on macro-level explanations, such as societal levels of gender inequality, we should be investigating group dynamics, where the real motive behind wartime rape can be understood.

In a study3 published in the August 2013 American Political Science Review, Cohen analyzes an original cross-national dataset covering the eighty-six major civil wars spanning three decades and turns conventional thinking on its head by suggesting that perpetrators of rape are commonly driven by the need for social cohesion rather than ethnic tension, lust, or hatred for women. Using data collected from US State Department Human Rights Country reports, she finds a correlation between the abduction of fighters—which tends to result in low social cohesion—and the prevalence of wartime rape by both state armies and rebel groups.

Drawing from fields like criminology and sociology, Cohen reasons that gang rape—a particular form of public rape by multiple perpetrators that occurs more often in war than in peacetime—serves the purposes of establishing individual status in a group, as well as the social cohesiveness of the group as a whole. “De-emphasizing the attacks' sexual nature,” she says, “psychological and sociological studies of gang rape find that perpetrators experience increased mutual esteem, and that rape serves as an act of camaraderie.”4

Her “combatant socialization” model holds up statistically across cases as well as within them. By combining survey data with in-depth interviews conducted during fieldwork, Cohen shows how her research can shed new light on now infamous cases of wartime sexual violence, such as in Sierra Leone.

Interviewees in Sierra Leone reported feeling frightened and isolated when they were first abducted. Hence, armed groups face a challenge: how to create a coherent force out of a group of frightened strangers who feel no loyalty towards the group of which they are now a member. Gang rape is one such method.…By participating in group rape—and perhaps by bragging about the individual rapes they have committed—combatants signal to their new peers that they are part of the unit, and are willing to take risks to remain in the group. Thus, rape is part of the process of hazing new recruits and of maintaining social order among existing members.5

Cohen’s findings suggest we need to move beyond a moral diagnosis of perpetrators as social deviants. Instead, we should try to understand how group dynamics and context can sometimes make otherwise rational people commit terrible atrocities.

This behavioral account of wartime rape allows for new ways in thinking about both the victims and perpetrators of rape: one of the observable implications of the combatant socialization argument is that women who are abducted into armed groups to serve as fighters face similar pressures to participate in gang rape that men face. In another recent study6 published in World Politics, Cohen investigates the role of female combatants in the perpetration of wartime rape in Sierra Leone, where population-based survey data has shown that about one in four of all incidents of reported gang rape were perpetrated by groups that included women. She documents chilling first-hand accounts of the brutality of female combatants, like the RUF leader Adama “Cut-Hand,” and, through interviews with former fighters, identifies the variety of ways in which women directly participated in rape—by locating, restraining, and/or physically violating their victims.7

How we think about the causes of wartime rape matters. By focusing early warning mechanisms to look for indicators of rape, we might stand a better chance of preventing these horrible crimes in the first place—or at least intervening more quickly.

In Syria for example, where the United Nations Office of the High Commissioner of Human Rights and the IRC have reported the prevalence of rape against men, women, and children as a significant feature of the ongoing conflict, much is at stake.

The “Women Under Siege Crowdmap”—a crowd-sourced participatory map that collects, geo-locates, and displays real-time data about the instances of rape in the on-going civil war in Syria—shows how harnessing the power of the crowd, social media, and new technologies can increase situational awareness. Cohen‘s research suggests that if such crowdmaps focus their data collection on reports of instances of abduction by armed groups during armed conflict, perhaps policy-makers and practitioners would be better informed about the likelihood, location, and severity of future instances of rape.

Accuracy and generalization, however, continue to be a considerable challenge for real-time data collection—as Cohen herself acknowledges—and collecting information about sexual violence even after the fact proves to be difficult. For starters, there is not a standard definition of “sexual violence,” so scholars have been hindered in their attempts to hypothesize and test relationships between predictive factors and levels of sexual violence. There are also concerns about reporting biases due to the social stigma for victims of rape, as well as the incentives that affect the reporting of rape by news media and human rights organizations. These concerns have been explored in the context of Liberia by Cohen and colleague Amelia Hoover Green in a 2012 study8 published in the Journal of Peace Research.

Wartime rape is difficult to study because strong assumptions often prevent analysts from asking the right questions. “One of the reasons we don’t have much information about male victims or about female perpetrators,” says Cohen in an interview, “is because assumptions about gender are so deeply embedded, even in the researchers’ own minds, about who is a perpetrator of rape and who is a victim of rape that it was not until quite recently that researchers started designing surveys about sexual violence that didn’t assume the perpetrator was male and the victim was female.”9 Assumptions are powerful and affect the way we collect data.

Finally, when data are properly collected, what we do with it matters. Information helps identify trends and construct belief systems about causes and consequences of wartime sexual violence, which in turn affects policies regarding post-conflict reconciliation, transitional justice, and reconstruction and reintegration efforts. If Cohen is right, we’ve got to reconceptualize our understanding of the role women play, not only as victims, but also as perpetrators of rape. This suggests more work is needed on making Disarmament, Demobilization, and Reconciliation (DDR) programs more responsive and reflective of the numerous roles of women in war.

Perhaps most importantly, assuming that rape is ubiquitous in war suggests that it is inevitable, which greatly limits our perceptions about what we can do about it. But wartime rape is neither ubiquitous nor inevitable, according to Cohen and her colleagues. Systematically testing conventional wisdoms about root causes allows for more evidence-based policies and priorities—such as faster warning and response, and more inclusive post-conflict programs—to take shape. Recent scholarship suggests that we take a more cautious approach in our attempts to uncover the causes of complex behavior across time and space. Cohen’s new research is a welcome step in the right direction.

Joseph GuayCommunications Specialist, Weatherhead Center for International Affairs.


  1. Cohen, Dara Kay, Amelia Hoover Green, and Elisabeth Jean Wood. 2013. Wartime Sexual Violence: Misconceptions, Implications, and Ways Forward. United States Institute of Peace. Washington, DC: United States Institute of Peace. Copy at
  2. Caulfield, Kristin and Megan Countey. “In Conversation with…Dara Kay Cohen. Centerpiece. Spring 2013, Vol. 27 No. 2.
  3. Cohen, Dara Kay. 2013. Explaining Rape during Civil War: Cross-National Evidence (1980–2009). American Political Science Review 107, no. 3: 461–477. Copy at
  4. Ibid., 464.
  5. Ibid., 465.
  6. Cohen, Dara Kay. 2013. Female Combatants and the Perpetration of Violence: Wartime Rape in the Sierra Leone Civil War. World Politics 65, no. 3: 383–415. Copy at
  7. Ibid., 384.
  8. Cohen, Dara Kay, and Amelia Hoover Green. 2012. Dueling Incentives: Sexual Violence in Liberia and the Politics of Human Rights Advocacy. Journal of Peace Research 49, no. 3: 445-458. Copy at
  9. Caulfield, Kristin and Megan Countey. “In Conversation with…Dara Kay Cohen. Centerpiece. Spring 2013, Vol. 27 No. 2.

Image of Dara Kay CohenDara Kay Cohen is a Faculty Associate of the Weatherhead Center for International Affairs and an assistant professor of public policy at Harvard Kennedy School. Research Interests: International relations; international security; civil war; political violence; violence during conflict; gender and conflict; West Africa; Timor-Leste; and El Salvador.