Do voters regard male and female candidates equally? Does apparent religiosity of candidates help or hurt their electoral chances? Where biases exist, what explains them? We present a novel explanation of political bias, drawing from role congruity theory. It posits that political contexts shape citizens' perceptions of qualities that make a “capable leader,” which subsequently drives their willingness to vote for candidates. Evidence from a survey experiment embedded in the 2012 Tunisian Post-Election Survey demonstrates that this theory explains biases based on gender and religiosity better than dominant modernization and social identity theories. Moreover, these mechanisms are also likely to drive political biases related to other features and in other countries. This has important implications for policymakers aiming to reduce political biases in Tunisia, the MENA, or globally. It should encourage them to pay careful attention to stereotyped traits of underrepresented groups and successful leaders, and to use institutional solutions (e.g., electoral quotas) to shape expectations about underrepresented groups and leadership.