The murder of George Floyd ignited one of the largest mass mobilizations in US history, including both non-violent and violent BlackLivesMatter protests in the summer of 2020. Many have since asked: did the violence within the largely non-violent movement help or hurt its goals? To answer this question, we used real-world data (ACLED, 2020) about the location of all BlackLivesMatter protests during the summer of 2020 to identify US counties that featured no protests, only nonviolent protests, or both nonviolent and violent protests. We then combined this data with survey data (N = 494, Study 1), data from the Congressional Cooperative Election Study (N = 43,924, Study 2A), and data from Project Implicit (N = 180,480, Study 2B), in order to examine how exposure (i.e. living in a county with) different types of protest affected both support for the key policy goals of the movement and prejudice towards Black Americans. We found that the 2020 BLM protests had no impact on prejudice among either liberals or conservatives. However, they were, even when violent, able to increase support for BlackLivesMatter’s key policy goals among conservatives living in relatively liberal areas. As such, this research suggests that violent, disruptive actions within a broader non-violent movement may affect those likely to be resistant to the movement. We connect these findings to the notion of disruptive action, which explains why these effects do not materialize in reducing prejudice, but in generating support for important policy goals of the movement.