Ex-combatants who fought with the Salvadoran Army during El Salvador’s 1980–92 civil war often recall being ‘captured’, rather than recruited, suffering beatings and humiliation in the course of training, and fighting without a sense of purpose or direction. Those who served with rebel forces, by contrast, recall fatigue and frustration with new routines, but seldom hazing or abuse; most also recalled deep, ongoing instruction about the purpose and goals of the war. This comparison highlights the broad variation in armed groups’ formal institutions for socialization, a topic that political scientists have only recently begun to examine in depth. The Salvadoran case also emphasizes some shortcomings of the existing literature, which may elide the differing effects of different formal institutions, treat individual institutions as operating independently on combatant behavior, and/or fail to map complex causal processes intervening between institutions and behavior. This article takes as its starting point the observation that many armed group institutions – including recruitment, military training, political training, and disciplinary regimes – are components of the process known more generally as ‘combatant socialization’. Examining specific institutional processes associated with combatant socialization allows for the generation of more refined and specific theories of combatant socialization as both a causal variable and an outcome. At the same time, treating armed group institutions as related elements of a broader process, rather than as fully separate institutions, may also advance understandings of the effects of these institutions. I demonstrate that the implementation and content of formal institutions for socialization varied significantly both across and within groups in El Salvador; building on this analysis, I lay out several potential directions for comparative research.