Photo Credit: Trokilinochchi. Attribution 2.0 Generic (CC BY 2.0)
The following analysis appears in Volume 4, Issue 5 of the Peace Science Digest.
Citation: Traunmüller, R., Kijewski, S., & Freitag, M. (2019). The silent victims of sexual violence during war: Evidence from a list experiment in Sri Lanka. Journal of Conflict Resolution, 63(9), 2015-2042.
Keywords: Sri Lanka, sexual violence, civil war, gender, research methodology, list experiment, LTTE
Although sexual violence is known to be a common weapon of war, it remains difficult to establish its occurrence in specific contexts due to underreporting. This silence around sexual violence can stem from victims’ feelings of shame or guilt or from a well-founded fear of stigmatization or repression, especially when the perpetrators remain in power. In particular, Sri Lanka is a context where there has been uncertainty about the extent of sexual violence in that country’s civil war (1983-2009) between the Sinhalese-majority-controlled government and the Liberation Tigers of Tamil Eelam (LTTE, the primary armed group fighting for a Tamil homeland), though recent reports indicate that sexual violence may have been systematic and asymmetric. Accordingly, the authors examine the “scope, distribution, and determinants of sexual violence” during the Sri Lankan civil war, using a research method intended to overcome some of the anticipated underreporting: a list experiment.
Rather than ask respondents directly whether they experienced sexual violence during the war, this list experiment presented individuals with a list of items and asked how many—not which—of these they experienced during the war. One group of respondents received a list that included three items, none of which had to do with sexual violence, resulting in a numerical response between 0 and 3. The other group of respondents received the same list but with an additional item about whether they personally experienced sexual assault, resulting in a numerical response between 0 and 4. By comparing the higher average number of the second group with the lower average number of the first group, the authors were able to determine the percentage of respondents who experienced sexual violence during the war—the only difference between the two lists that could have accounted for the different numerical averages. The list experiment was embedded in a broader face-to-face survey of Sri Lankans across the country in 2016 (1,800 respondents total) that also asked direct questions about whether individuals experienced or witnessed sexual assault. Additionally, the survey gathered respondents’ demographic information, along with information on whether respondents had been displaced or had assisted and/or participated in the military or other armed group during the war.
A few key findings stand out. First, the list experiment reveals that sexual violence was in fact much more prevalent than direct questioning would suggest, with about 13.4% of the population estimated to have experienced sexual violence during the war, compared to 1.4% of the population when direct questioning is used. Second, comparing the list experiment results with those from the direct questions reveals which groups are most vulnerable to sexual violence, as well as which groups are most hesitant to report these experiences. For instance, non-combatants who assisted the military or other armed groups were the most vulnerable to sexual violence, at about 42%, though only 3% reported personal experience with sexual violence in the direct question. Additionally, although it is widely assumed that women are more vulnerable to sexual violence than men, 20% of men are estimated to have experienced sexual violence compared to 9% of women, though only 0.8% of men and 1.8% of women admit to experiencing wartime sexual violence in response to the direct question.
Third, digging deeper into the relationships between variables like ethnicity, armed group assistance, gender, and displacement, the authors provide support for more specific determinants of wartime sexual violence in the Sri Lanka—though their victim-centered data can provide only indirect support for claims about the perpetrators or their intentions. In particular, 52% of Tamils who assisted armed groups (presumably the LTTE or associated Tamil armed groups) experienced sexual violence, compared to 20% of non-Tamil supporters of the military/armed groups and 11% of Tamils who did not support armed groups. Additionally, although men and women among the non-displaced population experienced sexual violence at roughly equal levels (12% and 14%, respectively), among those who were displaced, men had a much higher rate of sexual violence at 31% as opposed to 10% for women. Taken together, these findings support the argument that government forces “perpetrated sexual violence asymmetrically and strategically against collaborators of the LTTE” as they were fleeing the war zone to “coerce confessions, degrade suspects, and discourage broader Tamil involvement with the LTTE.” Although the finding on the higher rate of sexual violence against displaced men runs counter to expectations, it makes sense in light of the fact that men would be more likely to be seen by government forces as potential LTTE collaborators.
In sum, this research provides compelling evidence for the value of list experiments as a means for uncovering sensitive information such as wartime sexual violence—with troubling findings on the use of sexual violence as a weapon of war in the Sri Lankan case.
In the context of the Sri Lankan civil war,
- A list experiment is an effective research method for uncovering sensitive information, as its use suggests that sexual violence was much more prevalent during the Sri Lankan civil war (affecting about 13.4% of the population) than direct questioning would indicate (at 1.4% of the population).
- 52% of Tamils who assisted armed groups experienced sexual violence, compared to 20% of non-Tamil supporters of the military/armed groups and 11% of Tamils who did not support armed groups, suggesting that government forces “perpetrated sexual violence asymmetrically and strategically against collaborators of the LTTE.”
- Despite a roughly equal rate of sexual violence among men and women in the general population (12% and 14%, respectively), among displaced Sri Lankans, men had a much higher rate of sexual violence at 31% as opposed to 10% for women, suggesting that sexual violence was employed mostly against men with suspected LTTE ties as they were fleeing the war zone.
A few insights from this research stand out for war and violence prevention. First, this research highlights the political and ethical implications of methodological choices. With a sensitive subject such as sexual violence, it matters how researchers seek out information. Different research methodologies will get very different results: whereas a direct survey question about sexual violence may lead researchers to conclude that sexual violence was not very prevalent in a particular context, a list experiment—which does not require respondents to directly reveal their experience with sexual violence—is likely to reveal a more accurate estimation of its prevalence. And this knowledge can then lead to policy changes and calls for greater accountability on the part of perpetrators. This serves as a useful reminder not only to other academic researchers but also to organizations in the field carrying out their own research to inform programming that careful deliberations must go into research design—and that a list experiment may serve as a helpful methodological tool in some cases.
Second, the finding that displaced men were sexually assaulted at a much higher rate than displaced women has important policy implications and should lead organizations and policy-makers to re-examine their own assumptions about women being the prime targets of sexual assault. As Charli Carpenter highlights in her work on gendered assumptions about vulnerability in the context of the Bosnian War, not critically examining these assumptions can have grave implications, as it did in Srebrenica where the emphasis on protecting women and children left men at greater risk for the massacre that eventually took place. Of course, the distribution of vulnerability among genders will vary from one context to another, but the present research serves as a useful reminder that organizations should employ inclusive language and practices in their protection work and in their work with sexual assault survivors in particular, so that all genders find their services accessible. In short, organizations and policy-makers should remember that gender-informed programming entails, among other things, attention to the different experiences of various genders, not just attention to the experiences of women, though of course that remains vitally important.
Third, it is worth noting that any “power” sexual assault may have over its victims—especially its male victims—it gains from patriarchy. The “feminizing” or “emasculating” effect of sexual assault is particularly potent as a form of degradation—beyond the obvious physical pain and harm entailed—only in a world where gender hierarchies are firmly entrenched. Therefore, key to the effort to limit the use of sexual violence in wartime is a dismantling of this hierarchy where being “feminized” is experienced as an insult. The research findings on the high prevalence of sexual violence against men but the extremely low rate at which men report it (in the direct survey question) provides a striking illustration of these patriarchal forces at work—and the powerful silencing effect these can have especially on men who face a great deal of pressure to appear virulently masculine. As a start in countering these forces, perhaps the revelation of the widespread prevalence of wartime sexual violence among both men and women—at least in the Sri Lankan context—can create space for this experience to be destigmatized and discussed.
|Patriarchy||an unequal social system where men and masculinity are highly valued and in a dominant position and women and femininity are devalued and in a subordinate position.|
|Gender hierarchies||the privileging of masculinity and its associated traits over femininity and its associated traits, whether or not these traits are attached to actual women and men, such that actors or actions marked as “masculine” have greater access to power and/or are more favored and those marked as “feminine” are marginalized, denigrated, or excluded.|
Human Rights Watch. (2013, February 26). “We will teach you a lesson”: Sexual violence against Tamils by Sri Lankan security forces. Retrieved on October 30, 2019, from https://www.hrw.org/report/2013/02/26/we-will-teach-you-lesson/sexual-violence-against-tamils-sri-lankan-security-forces
Cronin-Furman, K. (2017, November 16). Are Sri Lankan officers ordering soldiers to sexually assault Tamil detainees? The Washington Post. Retrieved on October 30, 2019, from https://www.washingtonpost.com/news/monkey-cage/wp/2017/11/16/are-sri-lankan-officers-ordering-soldiers-to-sexually-assault-tamil-detainees/
Carpenter, R. C. (2003). “Women and children first”: Gender, norms, and humanitarian evacuation in the Balkans 1991-95. International Organization, 57, 661-694.
Sexual Violence Research Initiative (Sri Lanka): https://www.svri.org/documents/sri-lanka
 The way the list experiment and direct questions were worded, this sexual violence could include domestic violence perpetrated during the war as well as more directly war-related sexual violence at the hands of armed actors.