Why is mass commitment to democracy ? defined as an abstract preference for democracy over non-democratic regimes and the rejection of non-democratic alternatives in the face of crisis ? higher in some new democracies on average and lower in others? Similarly, why are some citizens within countries more committed to democracy than others? To answer these questions, this dissertation proposes a new lifetime political learning model that synthesizes the insights of the existing theoretical perspectives with leading models of political learning. Specifically, this lifetime learning model stipulates that democratic commitment is the product of a weighted average of information imparted by the political contexts experienced by individuals and the political subgroups to which they belong. The extent to which political contexts provoke pro-democratic political learning is determined by factors including socioeconomic modernization, governmental performance, international diffusion of democratic norms, and the occurrence of events such as mass protests that provide information about the desirability of different regimes. Crucially, the age at which individuals encounter such political contexts, their level of political sophistication, and the subgroups to which they belong condition the contexts' impact on their regime preferences. Since the model integrates macro-level political contexts with the micro-level learning process, it can explain not only why countries differ in levels of democratic commitment in the aggregate, but also why individuals within countries express divergent views about democracy. Empirically, the study first leverages over three decades of comparable survey data from Spain to conduct an intensive test of the theory. This analysis shows that attitudes toward democracy are formed and evolve throughout the life cycle, political sophistication and identification with salient subgroups moderate the learning process, and modernization, civil violence, international diffusion, and government performance explain the evolution of democratic support over time. Second, analysis of short-term panel and repeated cross-sectional data from 19 Latin American countries broadly confirm the generalizability of the Spanish results, and also underline the importance of mass mobilization and modernization in determining the degree to which different social classes and ideological groups become more or less committed to democracy in different countries.