Viral Violence:
The Effects of Police Violence Framing, Group Identity, and Militarization on Public Outrage and Perception of Police 

By Neil Fasching

Thesis Supervisor: dr. Bert N. Bakker







Watch Neil’s 5 minutes pitch presentation


As exemplified by the recent protests and riots following the murder of George Floyd, the framing of police conduct can have a profound impact on the behavior and attitudes of the general public. However, the extent of this impact has yet to be investigated. Using an experiment fielded in the United States, I investigate the degree to which the framing of police violence affects the public’s outrage and perception of the police. Further, I extend this literature by building on intergroup conflict theory (Mackie, Devos, & Smith, 2000; Tajfel, 1981) and by replicating recent work into the effect of police militarization (Mummolo, 2018). And finally, I test how police violence, group identity, and militarization interact in influencing the public’s anger and outrage as well as attitudes towards the police.



Originally, the design of this thesis was digitally innovative. Respondents were to be hooked up to physiological measurement equipment that would track their arousal while watching videos of police conduct, some of which would depict police violence. However, due to constraints caused by COVID-19, this design had to be abandoned, and a more socially-responsible design was adopted that did not require in-person contact. As such, 2x2x2 between-subjects survey experiment was chosen that manipulates militarization (no militarization vs. militarization), group identity (in-group vs. out-group) and violence (no violence vs. violence). There were two different types of stimuli for this experiment: photos and vignettes. The photos were of police officers with two different conditions that differed the level of militarization of the police.

The second manipulation was the vignette briefing describing the events of a nondescript protest. In this vignette, there were two different manipulations. First, the issue that was being protested about was manipulated. The group was either an anti-abortion group or climate change group. The issues of climate change and abortion were chosen as they are highly salient in the US and are top issues of importance for voters (Gallup, 2021; Tyson, 2020). Using respondents’ support for these issues that was asked at the beginning of the survey, it was possible to establish whether this group represents a respondent’s in-group or out-group. Second, the actions undertaken by the police during the protest were also manipulated. In the control condition, the police were non-violent, while in the treatment condition, the police were violence.



In terms of main effects, the results indicate that individuals in the police violence condition were significantly more outraged than individuals in the police non-violence condition (β = .24, p < .001). As moral outrage was rescaled to be between 0 and 1, the effect of police violence on outrage is quite large and represents a 24 percentage point increase in moral outrage compared to the condition when there was no violence. Further, individuals in the police violence condition held a slightly more negative perception of the police than individuals in the police non-violence condition (β = -.04, p < .001). Contrary to the findings of Mummolo (2018), I find no statistical difference in the perception of the police between individuals in the police violence condition and the police non-violence condition (β = -.01, p = .35). While I hypothesized that people would be more willing to share content on social media when the content depicted police violence, the results indicate the opposite is true: people were 4 percentage points less likely to share content on social media when the content depicted police violence (β = -.04, p < .001).

In terms of two-way interactions, outrage is greater when the violence is against one’s in-group (vs. out-group) (β = .05, p < .001). Said another way, police violence against one’s out-group increases public outrage by roughly 20 percentage points relative to non-violence, but police violence against one’s in-group increases public outrage by roughly 26 percentage points relative to non-violence. At the same time, I found no significant interaction between police violence and group identity on the perception of the police (β = .0002, p = .98).




Gallup. (2021, Jan). The abortion issue in presidential elections. Author. Retrieved


Mackie, D. M., Devos, T., & Smith, E. R. (2000). Intergroup emotions: Explaining offensive action tendencies in an intergroup context. Journal of Personality and Social Psychology, 79(4), 602–616. doi: 10.1037/0022-3514.79.4.602 

Mummolo, J. (2018, August). Militarization fails to enhance police safety or reduce

crime but may harm police reputation. Proceedings of the National Academy of

Sciences, 115(37), 9181–9186. doi: 10.1073/pnas.1805161115

Tyson, A. (2020, Oct). How important is climate change to voters in the 2020 election?

Pew Research Center. Retrieved from how-important-is-climate-change-to-voters-in-the-2020-election/