Measuring Interdisciplinarity: A Graph-Based Analysis of Brazilian Academic Committees

This study presents a scientometric analysis of interdisciplinary scientific collaboration in the composition of academic committees in Brazil. The research employed an extensive dataset, obtained from CAPES and Sucupira platforms, which store information on Brazilian academic output: master's and doctoral dissertations and theses. The collected data enabled the examination of relationships between 66 Knowledge Areas (KAs). Applying graphs to model the relationships among committee members, we distinguished each member according to their assigned role: candidate, evaluator, advisor. Three "academic graphs" were designed to represent these relationships: evaluations, invitations, and co-participations. As a method, we adapted a Knowledge Discovery in Database (KDD) process, systematizing the necessary steps to enable the desired measurement. Applying statistical techniques, the research results revealed significant patterns of interdisciplinarity in the composition of the committees. It was possible to identify KAs that collaborate more frequently (Ecology, Education, Linguistics, Zoology), as well as those that demonstrate less interaction, in addition to the relative role of each as a hub for other areas (Biochemistry, Education, Chemistry, and Sociology). This study contributes to a deeper understanding of the role of defense committees in the advancement of Brazilian science, emphasizing the relevance of interdisciplinary collaboration in the evaluation of academic output.