The report “Selection of Research Data” contains general guidelines on how to determine what research data should be preserved for the long term and what data should not. These guidelines can be used by individual researchers or research groups, researchers who co-operate within a collaboratory, research institutes, university faculties and national or international organisations focusing on a specific scientific/scholarly discipline and bodies funding research. The checklist is also suitable for use by managers of archives, research repositories, and heritage institutions. This study shows the situation in the area of selecting research data, based on a survey of the literature, interviews with key players and the experience gained by DANS and the 3TU Data Centre. The main issues are considered. The most important reasons for preserving research data for the long term have been formulated. The checklist based on these can be used as a guide for creating guidelines for specific scientific/scholarly disciplines. The report also describes the various roles of people involved in selection. Attention is paid to the best moment in the digital life cycle of research data for selecting them.