Although psychology has recently witnessed a burgeoning interest in the predictors of social and political action generally, little research has considered the psychological mechanisms by which people come to choose extreme or radical forms of action. How and why do groups come to favor radical or extreme solutions (radicalization) over conventional political pathways (politicization)? Theory in both political science and psychology suggests that social interaction plays an important role, but this has never been demonstrated experimentally. Results (N=114) show that social interaction can lead to both politicized and radicalized solutions but that radicalization rests on the perception that extreme action is legitimate. The findings provide the first experimental analog of the group-based dynamics that underpin political engagement and political extremism.