Two sides of the same coin: can campaigns generate support for both human rights and retributive violence?

ABSTRACT

Do framing strategies that are effective at encouraging pro-social behavior, such as participation in human rights campaigns, also mobilize support for violence within the same subjects? We use an experimental research design to examine individuals’ reactions to personal, humanizing narratives about past victimization. Participants are randomly assigned to one of eight treatment groups, which variously highlighted the humanity of the subject, the intensity of the past violence, and/or an evocative photograph of the subject that underscores her loss and vulnerability. We expect narratives that emphasize the subject’s humanity will encourage the audience to see the subject as innocent and as a victim, but also to feel angry about her experience. As a result, individuals will be more likely to defend the subject’s human rights, and to condone her use of retributive violence. We find that humanizing narratives lead respondents to simultaneously support a human rights appeal on the subject’s behalf and her use of retributive violence. Perceiving the subject in the narrative as innocent or as a victim mediates these effects, but anger often does not.

KEYWORDS:

• Human rights
• violence
• survey experiment
• framing
• non-state actors

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Notes

¹ This research was performed under College of Wooster Human Subjects Research Committee approval IRB #2016013.

² Scholars have written critically of the ethics of crowdsourced research, and MTurk, in particular (Marinova 2016). In addition to the low wage and poor quality work, Marinova (2016) criticizes the cover of legitimacy that academics provide to corporations like Amazon’s MTurk by using its services to conduct research. We have reflected on these critiques carefully; agreeing with many of the authors’ conclusions, we are actively considering alternative online sites for future data collection initiatives.

³ For a thorough discussion of the strengths and limitations of MTurk for data collection in the social sciences, see Paolacci, Chandler, and Ipeirotis (2010); Berinsky, Huber, and Lenz (2012).

⁴ We reversed the order of the response options from what was presented in the original survey, so that a larger value indicates more support for Leila’s involvement in the violent resistance.

⁵ Examining the full range of emotional reactions confirms that, in general, seeing any of the treatments generates negative emotions. For example, we observe increases in feelings of fear, sadness, frustration, shock and disgust, but decreases in positive or neutral feelings such as happiness, calmness and indifference.

⁶ Factors such as sex, age, race, political party affiliation, religiosity, education level, previous exposure to violence (Richards, Morrill, and Anderson 2012), and one’s sense of agency (McEntire, Leiby, and Krain 2015) affect one’s attitudes regarding human rights violations and willingness to act. However, because respondents were randomly assigned, any observed difference in treatment effects cannot be attributed to socio-demographic factors. Models with these controls are in the online appendix.

⁷ Robustness tests with all other mediators included in each model are provided in the online appendix. Results are substantively similar.