Peace Science Made Accessible, Understandable, and Useful.

Identifying the Most Effective Form of Intervention to Mitigate Mass Atrocities

Identifying the Most Effective Form of Intervention to Mitigate Mass Atrocities

Photo credit: Hugo

This analysis summarizes and reflects on the following research: Broache, M., & Cronin-Furman, K. (2020). Does type of violence matter for interventions to mitigate mass atrocities? Journal of Global Security Studies, 1-9.

Talking Points

  • The type of violence characterizing mass atrocities—ethnic identity-based violence or violence against political opponents—influences the effectiveness of different forms of intervention to mitigate mass atrocity violence.
  • “Naming and shaming” campaigns—as measured by Amnesty International reports and press releases—appear to decrease the severity of mass atrocities in cases of politicide (violence against political opponents) but not in cases of identity-based violence.
  • Military intervention against perpetrators is the only form of military intervention that appears to decrease the severity of mass atrocities, and it seems to have this effect only in cases of identity-based violence, not in cases of politicide (violence against political opponents).
  • Neutral military interventions and military interventions on the side of perpetrators, as well as economic sanctions, do not seem to be effective at mitigating mass atrocities, either in cases of identity-based violence or in cases of politicide.
  • In cases where the international community is poised to intervene in the face of mass atrocities in a particular country, concerned actors should consider the nature of mass atrocity violence before deciding which type of action will have the best chances of decreasing violence against civilians.


Finding out how to effectively stop, or at least diminish the severity of, mass atrocities is one of the most daunting challenges faced by the international community. Does “humanitarian” military intervention protect civilians or put them at greater risk? Do the costs to perpetrators of economic sanctions or human rights organizations’ “naming and shaming” campaigns translate into lower levels of anti-civilian violence? Moving beyond the parameters of previous research, Michael Broache and Kate Cronin-Furman examine whether the type of atrocity being perpetrated against civilians might influence the efficacy of different forms of international intervention. 

Mass atrocities: “[L]arge-scale, systematic violence against civilian populations.” Normally includes the four crimes identified in the 2005 UN World Summit outcome document: genocide, war crimes, crimes against humanity, and ethnic cleansing.

Straus, S. (2016). Fundamentals of genocide and mass atrocity prevention. Washington, DC: United States Holocaust Memorial Museum.

By “atrocity type” the authors mean the nature of targeting—are the targets of mass atrocities members of a particular ethnic identity group (identity-based violence) or simply political opponents of the perpetrator (politicide)? Although they acknowledge the possibility that identity-based violence and politicide may not meaningfully differ, they point to distinctions highlighted in previous research, including the suggestion that ethnic identity may be seen as “less malleable than political affiliation,” thereby leading to more “intense” or “intractable” violence, and that ethnic identity may be more visible than political identity. Consequently, the authors develop two hypotheses about how different forms of intervention might have different effects depending on whether mass atrocities are characterized by identity-based violence or politicide:

  • Hypothesis 1: Military interventions will be more effective in cases of identity-based violence. According to the authors, effective military interventions protect or defend “potential victims” while also “degrad[ing]” perpetrators’ military capability, requiring the intervener to clearly distinguish between victim and perpetrator groups—something they reason is easier to do in cases of identity-based violence.  
  • Hypothesis 2: Nonmilitary interventions—in particular, economic sanctions and/or “naming and shaming” campaigns—will be more effective in cases of politicide. The authors think economic sanctions or “naming and shaming” will only work if perpetrators are sensitive to the costs these impose, and they think perpetrators will only be sensitive to these costs if they are at least somewhat willing to compromise with their opponents. They reason that political differences are more conducive to compromise, while ethnic identity differences—perceived as “intrinsic”—are less so. Therefore, perpetrators will be more sensitive to the costs of these nonmilitary interventions in cases of politicide than in cases of identity-based violence.

To test these hypotheses, the authors draw on a few datasets: the Political Instability Task Force State Failure dataset (for identifying the severity of mass atrocity cases year to year, 1975-2008, and, with some revisions, cases of identity-based violence versus politicide), the International Military Interventions dataset (for identifying instances of military intervention, divided into three categories: antiperpetrator, neutral, and properpetrator), an economic sanctions dataset (counting the number of “sanctions imposed by other countries and international organizations”), and a “naming and shaming” dataset (counting the number of Amnesty International reports and press releases coming out each year on a given country). The authors run different statistical models to see if there are any significant relationships between different intervention types and severity levels of mass atrocities (based on the number of deaths per year), distinguishing between cases of identity-based violence and cases of politicide. They find two significant relationships: Antiperpetrator military interventions—but not neutral or properpetrator military interventions—appear to decrease the severity of mass atrocities in cases of identity-based violence, and “naming and shaming” campaigns appear to decrease the severity of mass atrocities in cases of politicide. In more concrete terms, the authors find that the chances of a hypothetical case of identity-based violence declining in severity increases from 16.5% (without an antiperpetrator military intervention) to 22% (with an antiperpetrator military intervention). Likewise, the chances of a hypothetical case of politicide declining in severity increases from 63% (without any Amnesty International reports released) to 66.5% (with only one Amnesty International report released).

Although these findings are broadly consistent with the authors’ hypotheses, the authors do point out that they were unable to test their proposed mechanisms—the reasoning they suggested for explaining the relationships they found. Nonetheless, they argue that “interventions to halt mass atrocities should not be treated as a one-size-fits-all solution” since, according to their analysis, “antiperpetrator military interventions have ameliorative effects only in cases of identity-based violence, while naming and shaming helps only in politicides.”

Informing Practice  

Although one of the authors’ central takeaways is that there is a meaningful difference between ethnic identity-based violence and politicide with respect to the effectiveness of different forms of intervention, there is another, perhaps more radical insight lying just under the surface. To notice it, we need to carefully attend to the details of the study—in particular, how the authors measure their variables. To measure the nonmilitary intervention they call “naming and shaming,” the authors count the number of reports and press releases Amnesty International publishes each year about the country in question. As noted above in the hypothetical scenario they play out, while an antiperpetrator military intervention increases the chances of mass atrocity de-escalation (in cases of identity-based violence) by 5.5 percentage points, the release of just one Amnesty International report increases the chances of mass atrocity de-escalation (in cases of politicide) by 3.5 percentage points. Allow that to sink in for just a moment. Think of the massive human and economic resources needed to intervene militarily in a country, not to mention the enormous loss of human life (whether people intentionally targeted on the “perpetrator” side, people unintentionally harmed or killed in “victim” or other groups, or people in the intervening force) and destruction of physical and social infrastructure that military intervention entails, however humanitarian its goals. Then think of the resources needed and risks taken to research, write, and publish one human rights report—not negligeable of course, as human rights advocates are often targeted by repressive governments, but still nothing close to the enormous human and economic costs of military intervention. Even for those of us convinced of the power of nonmilitary efforts at violence mitigation and conflict transformation, the fact that just one report could have nearly the same effect as an entire military intervention is astounding. (This leaves aside for the moment the question of whether we can reasonably count as a success a military intervention that diminishes the severity of mass atrocities against some while contributing to other forms of human, social, and material destruction in a country, with unknown long-term consequences.) It leaves one to wonder what predicted substantive effect a multi-report “naming and shaming” campaign would have—not to mention a coordinated campaign orchestrated by a coalition of governmental and nongovernmental actors, rather than just one (albeit prominent) NGO.

Furthermore, “naming and shaming” (as measured by Amnesty International reports) is one of only two forms of nonmilitary intervention examined in the research, along with economic sanctions.  Left out are additional, on-the-ground, nonmilitary approaches for protecting civilians, like unarmed civilian protection (UCP) and nonviolent resistance. While UCP may occasionally include “naming and shaming” elements (depending on the organization), it differs from human rights report-writing and publicity in that UCP teams are on the ground, living in affected communities, strategically using their presence and engaging in activities to prevent further violations against civilians—and in the long-term even contributing to the restoration of the social fabric. And although the purpose of nonviolent resistance is not explicitly civilian protection, recent research shows that mass atrocities happen much less frequently in cases of nonviolent resistance than in cases of armed rebellion—highlighting the protective effect of nonviolent discipline in a resistance movement and of international efforts to support such nonviolent movements. The success of even limited (and remote) “naming and shaming” activities suggests that much more attention should be devoted to these other even more robust nonmilitary options for protecting civilians. [MW]

Continued Reading

Perkoski, E., & Chenoweth, E. (2018, May). Nonviolent resistance and prevention of mass killings during popular uprisings. ICNC Special Report Series, Volume 2. Retrieved on September 18, 2020, from

Ackerman, P., & Merriman, H. (2019, May). Preventing mass atrocities: From a Responsibility to Protect (RtoP) to a Right to Assist (RtoA) campaigns of civil resistance. ICNC Special Report Series, Volume 3. Retrieved on September 18, 2020, from

Johnson, C. (N.d.). Unarmed civilian protection (UCP): A concise overview. Retrieved on September 21, 2020, from

Atrocities Prevention. (2020, July 1). Warning statement on the potential for mass atrocities in the United States. Retrieved on September 21, 2020, from


Amnesty International:

Nonviolent Peaceforce:

International Center on Nonviolent Conflict:

Bridge Alliance:

Key Words: mass atrocities; civilian protection; violence prevention; identity-based violence; politicide; military intervention; naming and shaming campaigns; economic sanctions

Next article Special Issue: Local, National, and International Peacebuilding
Previous article Is Leadership Decapitation Effective at Sustainably Ending Civil War?