All coups seek to topple the political leadership, but they differ in terms of their leaders. While soldiers spearhead a majority of coups, a small number is led by civilians. And whereas high-ranking officers are the largest group among coup leaders, mid- and low-ranking soldiers account for a substantial share of putsches. Several datasets have recently offered data on the identity and political aims of coup leaders, to study the origin and outcome of different types of coups. However, these datasets have important limitations in their scope and how they address differing organizational structures of militaries across countries and time. This article therefore introduces a novel dataset on the identity of the leaders of 474 coups from 1950 to 2020 that distinguishes between coups led by civilians and military officers, as well as between coups by junior, mid-ranking, and senior officers. We discuss how the dataset complements previous data, present patterns across time and space, and show that successful and failed coups by senior, mid-ranking, and junior officers entail different prospects for post-coup democratization. The article underlines the importance for refined empirical measures and theoretical arguments in coup research.