Dynamics of Conflict-Related Sexual Violence

This project has addressed why sexual violence occurs during conflict and repressive spells, and what might be done to reduce or eliminate it. The project has produced several academic articles and two original datasets, and also updated an existing leading dataset on sexual violence in war. One of the articles from the project is a review in Annual Review of Political Science which we expect will help set priorities and shape the research agenda on sexual violence in the years to come. Results and insights from the research have also been communicated to a wider audience through op eds, blogs, and the media; as well as in briefing for policy makers and government officials in Norway and internationally. We have focused on three main challenges. First, a fundamental challenge for research on conflict-related sexual violence is methodological -- how to establish reliable systematic evidence and to overcome problems of potential data biases which can occur i.a. due to sexual violence being fraught with taboos, the conflict contexts in the violence occurs, and the phenomenon being potentially overlooked and discounted. Existing data was therefore limited. In response to these challenges we have collected new data and also conducted a study where we propose to use so-called Latent Variable Modeling (LVM) as one technique to handle uncertainty in estimates of sexual violence in conflict. Uncertain estimates is an inherent problem in research on sexual violence because of taboos and other factors that make observation hard. An article about this problem and possible solutions was published in Journal of Peace Research. Another main challenge is substantive -- the lack of understanding of the relationships between different forms of sexual violence and other parts of the repertoire of behaviors in contentious settings. Although sexual violence is often studied as an aggregate category, different forms of sexual violence might have both different causes and consequences. This will be masked when we use aggregated measures. To provide a solution to this challenge, we collected new disaggregated prevalence data on different forms of sexual violence. A third main challenge has been the lack of evidence-based policy to solve the problem of sexual violence in conflict. The project has focused in particular on the claim that ending impunity is the solution to sexual violence. This claim was previously not systematically tested. We published an article in Journal of Conflict Resolution where we find that in situation with amnesties, sexual violence by rebel groups tends to continue or increase. However, we do not find evidence of a deterrent effect of trials. This is an important finding to be aware of, given the centrality of ending impunity in the public debate and among policy makers on how to deal with the aftermath of wartime sexual violence. In the project period we have updated the sexual violence in war (SVAC) dataset to also cover the years 2010-2019 and written a Policy Brief and a blog post to present results from this update. We find that sexual violence has been, and continues to be, a problem in conflicts on all continents, that states are frequent perpetrators of such violence, but that that an increasing number of rebel groups also engage in such atrocities in recent years. We also have a data collection on repertoires of sexual violence (RSVAC) to disaggregate different forms of violations. A paper presenting the new data was published in Journal of Peace Research. We have also collected new data on sexual violence by state actors within and outside of war contexts. The new dataset on State-perpetrated Sexual Violence (SPSV), and has global coverage from 1980-2016, and is the basis for an article and a book which will be published after the project period.

The project has produces two major data collections, six academic articles in high ranking journals, and a number of other written outputs, media appearances, blog posts and a podcast for a wider audience. Further publications are anticipated after the project period, including two academic books. The project should significantly advance the study of sexual violence in conflict settings, both during outright war and not least outside of active conflict situations. It should also direct the research community to a greater focus on states as perpetrators of sexual violence, and shape the direction of the academic literature on sexual violence in conflict in years to come. The datasets that the project has provided are going to be of benefit for research communities interested in political violence, and the advancement in thinking through what type of data is needed to assess particular questions and how we deal with observability challenges is going to improve the quality of future research by project members and the wider scholarly community. In particular do we anticipate that the addition of data on state perpetrators inside and outside of war is going to open up a new research frontier that could bring a similar growth in academic research that we saw after the publications of the Sexual Violence and Armed Conflict dataset (SVAC) in 2014 (Cohen & Nordås 2014), which has to date been cited 303 times according to Google Scholar, and given rise to a large number of studies. The project can also show how trials have limited deterrent effect on subsequent use of sexual violence by armed organizations, which should inform policy. Results and insights from the project is already being communicated to key policy makers, such as the UN Special Representative on Sexual Violence in War, Pamela Patton, to the Norwegian Ministry of Foreign Affairs, and to the US State Department.

The purpose of this project is to address why sexual violence occurs during conflict and what might be done to reduce or eliminate it. We will focus on tackling the main challenges facing the literature on conflict-related sexual violence: The most basic challenge is methodological -- how to establish reliable systematic evidence and to overcome problems of potential data biases on sexual violence in contentious settings. Reporting biases can occur i.a. due to sexual violence being fraught with taboos, the conflict contexts in the violence occurs, and the phenomenon being potentially overlooked and discounted. We will therefore map (RQ1) which data generation biases exist, how do they potentially skew our results, and how can they be overcome? The second challenge is substantive -- the lack of understanding of the relationships between different forms of sexual violence and other parts of the repertoire of behaviors in contentious settings. Sexual violence is generally studied in isolation or as an aggregate category, although there can be important dynamics between different contentious repertoires and the different forms of sexual violence (e.g. rape, sexual slavery, sexual torture) might have both different causes and consequences. We therefore ask: (RQ2) what is the strategic relevance and function of different forms of sexual violence in contentious politics? Based on this and the organizational, ideational, and other characteristics of perpetrator groups, we are able to ask the critical question of: (RQ3) which interventions and/or policies will significantly reduce or eliminate the use of sexual violence atrocities? The answers to these three related questions will be sought through intensive data triangulation, cross-case comparative analyses using advanced statistical techniques, as well as in-depth micro-level studies. It will be critical to advancing the understanding and possible prevention of this heinous war crime.