This presentation showcases the value of ETDs ‘to improve graduate education and scholarship’ in a cross-country, policy context. ETDs provide the basis to reliably identify key dimensions of the dissertation (e.g. year, subject matter, relation to other publications, length and composition, degree of inter/cross-disciplinarity, etc.). They thus provide a unique vantage point to better understand the changing role that the dissertation plays in the context of the PhD candidate’s career, how it differs in different country and disciplinary contexts, how it is changing, etc. Using data-base architecture being developed in RISIS2, the presentation presents a current cross-country effort to design and set up a curated dataset of digital dissertations to study changing research careers across countries and time. The infrastructure consolidates ETDs, links information with other information (onward publication streams), provides table-graph functions and ontology tools to understand the field, and provides secure authentication and access protocols. An objective is to inform science policy and to chart how the landscape is changing and to study what that means for science.