Why do ordinary people become supportive of violent, extremist ideologies? Over the past several years, tens of thousands of individuals across the world have become attracted to propaganda disseminated by the Islamic State (ISIS), and many have left their home countries to join the organization. This dissertation closely examines possible explanations for pro-ISIS radicalization in Europe and the United States. I argue that anti-Muslim hostility is an important driver of pro-ISIS radicalization, leading individuals who feel isolated to become attracted to the organization’s propaganda. I also contend that groups like ISIS are aware of this pattern, and thus seek to purposefully provoke hostility against potential supporters by carrying out terrorist attacks. I maintain that efforts to stop radicalization should focus on ways to reduce hostility and increase the inclusion of minorities in the West. The various dissertation papers empirically examine different aspects of these arguments. In the first paper, I examine whether anti-Muslim hostility might be driving pro-ISIS radicalization in Europe, by analyzing the online activity of thousands of ISIS sympathizers in France, Germany, Belgium, and the United Kingdom. Matching online radicalization indicators with offline data on vote share for far-right, anti-Muslim parties, I show that the intensity of anti-Muslim hostility at the local (neighborhood/municipality) level strongly correlates with support for ISIS on Twitter. In addition, I show that events that stir anti-Muslim sentiments, such as terrorist attacks and anti-Muslim protests, lead ISIS sympathizers to significantly increase pro-ISIS rhetoric, especially in areas with high far-right support. In the second paper, I argue that armed groups strategically use terrorism to manipulate levels of anti-Muslim hostility in Western countries. I test whether terrorism leads to greater expressions of anti-Muslim hostility using data on thirty-six terrorist attacks perpetrated by radical jihadists in the West from 2010 to 2016, examining how they shaped anti-Muslim attitudes among individuals in targeted countries. I find that individuals systematically and significantly increase the posting of anti-Muslim content on social media after exposure to terrorism. The effect spikes immediately after attacks, decays over time, but remains significantly higher than pre-attack levels up to a month after the events. The results also reveal that the impact of terrorist attacks on anti-Muslim rhetoric is similar for individuals who already expressed hostility to Muslims before the attacks and those who did not. Finally, I observe that the impact of terrorist attacks on anti-Muslim hostility increases with attacks resulting in greater numbers of casualties. In the third paper, I examine what might be done to stop online radicalization and support for ISIS in the West. I collected data on community engagement events performed in the United States by the Obama Administration, which aimed to increase trust and relationships between the Muslim population and the American government, and combined them with high-frequency, geo-located panel data on tens of thousands of individuals in America who follow Islamic State accounts on Twitter. By analyzing over 100 community engagement events in a Difference-in-Differences design, I find that community engagement activities are systematically and significantly associated with a reduction in pro-ISIS rhetoric on Twitter among individuals located in event areas. In addition, the observed negative relationship between community engagement activities and pro-ISIS rhetoric is stronger in areas that held a large number of these events.