When the numbers don’t add up: Researchers and media struggle with stats on sexualized violence

By — February 7, 2012

In January 2011, The Economist published the number of women raped in six conflicts, including an estimate of 500,000 women raped in the Rwandan genocide of 1994. Many readers may have taken these statistics at face value. In fact, however, estimates of rape in Rwanda range from 250,000 to 500,000 and are based on the number of reported pregnancies from rape, which underestimates prevalence.

Unfortunately, quantifying sexualized violence just isn’t straightforward. Data collection is complex and full of obstacles. All research comes with caveats. But more so than with other topics, data on sexualized violence is limited by a reluctance to report on the part of survivors and ethical and logistical difficulties that restrict its collection on the part of researchers. Despite all this, there is increasing demand by stakeholders such as NGOs, policymakers, and the media to put a number to the sexualized violence occurring in recent conflicts.

Young woman in hospital room

Six men kidnapped Chibalonza, 26, seen here at a hospital in South Kivu, eastern Congo, in 2008. They held her for nine months in an underground hole in the forest, where she was raped repeatedly. (Lynsey Addario/VII)

Survivors tend to underreport their experiences of sexualized violence due to stigma, shame, fear of disclosure to family members and the community, the threat of violent retaliation by perpetrators or abuse by intimate partners, and lack of confidence in the justice system. This means that any estimates obtained from police reports or other entities such as United Nations missions, or studies that summarize the number of survivors that seek care at health care facilities for sexualized violence-related injuries (or NGOs and other social services for psychosocial support), will be low estimates of the true prevalence of sexualized violence in any conflict.

In fact, according to studies by Michelle Hynes of the Centers for Disease Control, Jeanne Ward of the Reproductive Health Response in Conflict Consortium, and others, only about 6 to 7 percent of women in the East Timorese and Rwandan conflicts who suffered sexualized violence reported it to authorities or sought medical attention.

Researchers can provide more accurate estimates using population-based studies, which are costly and time-intensive because they typically entail randomly selecting households, going door-to-door, and spending hours interviewing individuals in each household. These surveys are detailed and representative of the general population and thus necessary for understanding the root causes, dynamics, and scope of the issue. However, due to the time it takes to implement them, more rapid assessments such as interviewing key informants, a method used during the Bosnian conflict of the early 1990s, are often necessary in emerging crises to mobilize resources.

Other factors limit researchers’ ability to provide population-based estimates of sexualized violence in conflict. Researchers are bound by ethical obligations to ensure survivors’ privacy and minimize any potential risks. This means interviewing respondents without others overhearing and keeping their identities and any information disclosed confidential. Ensuring privacy is particularly problematic in settings where we may expect the highest rates of sexualized violence, such as in internally displaced person (IDP) camps—these camps are crowded and tent walls are thin.

Minimizing risks to respondents means that researchers must provide contact information for referral services for health care or psychosocial support, which may be difficult to access or nonexistent. Furthermore, research must not be undertaken if there are no potential benefits of doing so to the survivors or their communities. Such benefits might be indirect, such as future policies on gender-based violence at the national level or community interventions focused on prevention.

For all these reasons, the estimates that we do have on sexualized violence in conflict are often flawed or incomplete. (For a more detailed description of flawed estimates in recent conflicts, see this article I published in the Bulletin of the World Health Organization with Peterman of the International Food Policy Research Institute in December.)

The media can also play a part in misrepresenting numbers. Reporters have rampantly cited a statistic they did not fully understand, claiming that 75 percent of Liberian women had been raped during the civil wars, according to an upcoming article in the Journal of Peace Research. In fact, more accurate estimates based on population-based studies suggest that the estimate is somewhere between 9.2 and 15 percent, according to research published in the Journal of the American Medical Association in 2008 and 1998.

As more information on sexualized violence is released, what we already knew is confirmed—that it is all too common. And while the Democratic Republic of Congo is often referred to as the “rape capital of the world”—with nearly four women raped every five minutes, according to my research—it is not the only country with high rates of rape. Researchers have estimated that Swaziland and Botswana have annual rates of rape of 4.7 and 4.6 per 100 women aged 18 to 49, respectively, according to research published in the American Journal of Public Health in June. These rates are higher than national rates for DRC, but still lower than rates in North and South Kivu, the provinces in DRC most heavily affected by conflict, which have a combined rate of 5.7 rapes annually per 100 women aged 15 to 49, my research has found.

And lest we try to convince ourselves that widespread sexualized violence is only a problem of poor, developing countries, the Centers for Disease Control estimates that 18 percent of women in the U.S. have been a victim of rape or attempted rape in her lifetime. That’s about one in five.

Sexualized violence is a problem that deserves attention so we can assist survivors and craft interventions aimed at prevention. Effective policies and programs are possible only when a more complete description of the problem is informed by rigorous research and conveyed accurately by the media.

Tia Palermo is assistant professor in the Graduate Program in Public Health at Stony Brook University (SUNY). She holds a Ph.D. in public policy from the University of North Carolina at Chapel Hill. Her research on gender-based violence has appeared in peer-reviewed journals and has been covered by the media, including The New York Times, CNN, NPR, and BBC News.