Peace Science Made Accessible, Understandable, and Useful.

Why Do Some Protests Escalate to Violence?

Why Do Some Protests Escalate to Violence?

Photo credit: Jonathan van Smit

This analysis summarizes and reflects on the following research: Ives, B., & Lewis, J. S. (2020). From rallies to riots: Why some protests become violent. Journal of Conflict Resolution, 64(5), 958-986.

Talking Points

  • A protest is more likely to escalate to violence a) the more recently it has faced state repression and b) when it is spontaneous rather than well-organized.
  • Recent repression decreases the costs of violent protest relative to the costs of nonviolent protest, thereby “attract[ing] violence-oriented participants” to an otherwise nonviolent protest and heightening the risk of violent escalation.
  • A spontaneous, less-organized protest lowers the barriers of entry for violence-oriented participants, while also reducing the ability of fellow nonviolent protesters to influence the behavior of these violence-oriented participants.
  • Violent escalation of protests may be less the product of deliberate strategy than previously realized.


Recent recognition of the greater relative effectiveness of nonviolent resistance has heightened interest in the factors that lead to violent escalation in nonviolent movements. We know that maintaining nonviolent discipline can attract the broad-based support and participation necessary for a resistance movement’s success. We also know that the use of violence on the part of a resistance movement can strengthen the opponent’s/state’s position by reinforcing its narrative, generate greater repression, and diminish movement support and participation. Why, then, do some protests still “turn” violent? While most previous research has focused on the strategic level of nonviolent campaigns, Brandon Ives and Jacob S. Lewis instead focus on the level of individual protests and the role spontaneous, non-strategic decisions might play in violent escalation.

The authors propose two main hypotheses to explain when an individual protest is more likely to become “violent” (by which they mean protesters engaging in the intentional infliction of physical injury or property damage):

Hypothesis 1: “Protests are more likely to escalate into violence when repression has been recently used.” The logic here is that violence-prone individuals are more likely to participate in protests if the relative costs of violent versus nonviolent resistance are low. Generally speaking, the costs to protesters (especially “the risk of arrest, detention, or death”) of violent resistance are higher than the costs of nonviolent resistance (meaning the relative costs of violence versus nonviolence are high). The costlier nonviolent resistance becomes—through government repression—the smaller this difference in cost (between violent and nonviolent resistance) becomes, potentially making violent resistance look more attractive in comparison to nonviolent resistance. Therefore, recent repression of nonviolent protests may prompt those more oriented towards violence to join in protests and engage in violent tactics.

Hypothesis 2: “Protests are more likely to escalate into violence when they are unorganized rather than organized.” The authors posit that highly organized protests—with “identifiable hierarchical structures,” “coherent end goals,” and “dense networks of activists who know one another”—are better positioned to keep violence-prone actors out of their ranks, influence participant behavior, and provide participants with greater certainty about “what kind of event this will be.” These factors contribute to the maintenance of nonviolent discipline. Accordingly, unorganized protests may be more susceptible to the presence of violence-prone individuals and have weaker influence on participant behavior. Furthermore, weaker (or non-existent) relationships between protesters and greater uncertainty about the protest’s goals and behavioral expectations may make individuals more likely to respond with violence to tense confrontations with security forces.

The authors test these hypotheses with data from the Social Conflict Analysis Database (SCAD) on protests across Africa 1990-2015. The database includes 6,036 nonviolent protests in 48 countries that either did or did not escalate to violence. To assess the relative cost of violence to nonviolence (for Hypothesis 1), they measure the amount of time since the last use of repression against a protest in the same administrative district. With reference to protest organization (Hypothesis 2), they designate each protest as either “organized” or “spontaneous,” depending on whether “clear leadership or organization can be identified” (according to the SCAD).

Statistical analysis supports their two main hypotheses. First, they find that “protests are more likely to escalate into violence if recent protests were repressed by the government.” The chance that a protest will escalate to violence goes from 11% six months after the last incident of government repression to 17%, 20%, and 25% a month, week, and day, respectively, after the last incident of government repression. Next, they find that “spontaneous protest events are more likely to escalate to violence [with a 20.8% chance] than organized protests [with a 12.6% chance].”

To flesh out their proposed explanations, the authors examine the #FeesMustFall protests in South Africa during October 2015. The protests against increased university fees grew quickly but lacked central organization, despite the presence of key student leaders. Initially, nonviolent protests did not result in a change to the proposed fees, but they did meet with repression in the form of arrests, as well as “stun grenades and rubber bullets.” Small groups of students then began to engage in “spontaneous burning and physical attacks.” The authors conclude that “[t]his violence did not occur as a result of strategy” but rather as a result of insufficient organization and protesters’ perception that “remaining nonviolent already entailed high costs” and was also “somewhat ineffective.”

The authors conclude that a protest is more likely to escalate to violence a) the more recently it has faced state repression and b) when it is spontaneous rather than well-organized. Recent repression decreases the costs of violent protest relative to nonviolent protest, thereby “attract[ing] violence-oriented participants.” And a spontaneous protest lowers the barriers of entry for violence-oriented participants, while also reducing the ability of fellow nonviolent protesters to influence the behavior of these participants.

Informing Practice

Given the strategic utility and even protection value of nonviolent discipline, the question remains: Why do individuals still resort to violence in an otherwise nonviolent protest when doing so is not effective or protective? Despite a strong focus in social science research on the strategic decision-making of “unitary, rational actors,” this research indicates that the answer lies instead at the non-strategic level. Individuals employing violence at protests are not necessarily thinking strategically, motivated as they may be by self-preservation or justice or righteous anger. Still, while some of these uses of violence may be purely expressive, others, even if not strategic, may nonetheless rely on unexamined assumptions about the efficacy of violence. This dimension emerges only briefly in this research study—with its emphasis instead on the relative costs of violence versus nonviolence—when the authors identify an additional factor motivating South African activists to use violence: the belief that nonviolence had become not only costlier but also ineffective. The frequent rush to this conclusion must be addressed head-on. Nonviolence is often not given the sort of lee-way or time that violence is to “work.” If a nonviolent movement has not achieved its objectives in a few short days or weeks or months of protest (a tactic that does not nearly exhaust the list of over 200 documented methods of nonviolent resistance), often some activists will begin to question the efficacy of nonviolence—and then simply assume that violence will be more effective. These tactics may then be given years to “work,” without the same impatience to switch back to nonviolent resistance when they haven’t.

To counter such moves to violence and thereby strengthen a movement’s ability to transform the status quo, activists must therefore continuously work to challenge the deeply engrained assumptions and narratives about the efficacy and protection value of violence that many fellow activists may hold, especially in response to repression. Likewise, activists must work to embed nonviolent discipline in the movement’s identity and strategy through trainings, pledges, symbolism, and other efforts to create movement cohesion. When and if not everyone is convinced, activist leaders must insist on clear boundaries between different activist groups in their actions, so that the public does not easily mistake one group for the other, even if they may share common goals. Finally, activists can take various measures to diminish the chances that agents provocateurs will succeed at spreading violence within their movement. These steps will increase the chances that, even in the face of state (or non-state) repression, a movement can maintain nonviolent discipline and thereby have a better chance of creating the change it seeks. [MW]

Continued Reading

Peace Science Digest. (2019, November 1). Limiting state violence and activist violence in nonviolent resistance struggles. Retrieved on April 20, 2020, from

Peace Science Digest. (2020, May 1). What accounts for the shift from nonviolent to violent resistance in the Syrian uprising? Retrieved on November 13, 2020, from

Peace Science Digest. (2020, May 1). How do “violent flanks” affect the outcomes of nonviolent campaigns? Retrieved on November 13, 2020, from

Peace Science Digest. (2019, March 28). Making civil resistance work against rightwing populism. Retrieved on April 20, 2020, from

Perkoski, E., & Chenoweth, E. (2018, April). Nonviolent resistance and prevention of mass killings during popular uprisings. ICNC Special Report Series, Volume No. 2.

Pinckney, J. (2016). Making or breaking nonviolent discipline in civil resistance movements. ICNC Monograph Series. Washington, DC: ICNC.

Chenoweth, E., & Stephan, Maria J. (2011). Why civil resistance works: The strategic logic of nonviolent conflict. New York: Columbia University Press.

Gupta, S. (2020, June 5). What the 1960s civil rights protests can teach us about fighting racism today. (An interview with Prof. Omar Wasow.) ScienceNews. Retrieved on November 13, 2020, from

Hastings, T., & Vogel, A. (2020, forthcoming). Agents provocateurs, violent flanks, and nonviolent movements: A historical and strategic perspective. Washington, DC: International Center on Nonviolent Conflict. Retrieved on November 13, 2020, from

Organizations: International Center on Nonviolent Conflict:

Keywords: nonviolent/civil resistance, protests, violence, escalation, repression

The above analysis is from Volume 1, Issue 1, of the Peace Science Digest.

Next article Why Do Individuals Support Political Violence?
Previous article Power and Co-optation in Protester-Police Interactions in Nonviolent Protest