Many countries have introduced national evaluation exercises. The methods commonly applied have been developed for the natural and life sciences and are criticized for being ill-adapted to the social sciences. After giving an overview of different classifications of countries' evaluation procedures, we present an empirically derived typology of national research evaluation systems that takes into account whether adaptations to disciplinary specificities, such as social sciences knowledge production and dissemination practices, are considered. The analysis shows that national evaluation systems are diverse and complex combinations of research evaluation practices at the national, regional and institutional level, including the influence of the international level on research funding of projects and careers. We conclude that diversity rather than standardisation in evaluation practices is key to reach the aim of evaluation: to provide researchers and university administrators with a tool to improve their organisation and research practice.