Impact of supervisors' research style on young biomedical scientists' capacity development as measured by REDi, a novel index of crossdisciplinarity.

The challenge for medical schools in Japan is to develop research activities for innovation. This study aimed at analyzing the connection between the research output of "promising researchers" (next-generation leaders in terms of research activity) and their supervisors' past research activities to identify those factors that impact researchers' performance. Activity was analyzed from the viewpoints of productivity, coauthorship networks, and research impact using a novel index called the Research Diversity Index (REDi) that quantifies crossdisciplinarity. Research funding, which has not yet been fully utilized in correlation studies of the characteristics of authors, was also considered in this study. For the promising researchers extracted using betweenness centrality scores within coauthorship networks, there were diachronic correlations between the records of the promising researchers and those of their supervisors. Supervisor leadership as measured by the number of last-authored publications and extent of networking had a positive effect on the promising researchers productivity. Supervisors' research style of integrating knowledge from multiple fields, as measured by REDi, was negatively correlated with the publication impact of promising researchers, suggesting that REDi is useful as a novel indicator of research quality not being captured by existing indices. It was also noted that establishing an academic presence through extensive collaborations could be advantageous for obtaining research funding, especially from top-down government programs. The possible implications of this study for promoting research activities are the importance of incorporating new doctorates into research groups at an early stage and that of promoting interinstitutional, crossdisciplinary collaborations. Classification code: MSC: 62P10 Applications of statistics to biology and medical sciences; meta-analysis.JEL: Z1Z10 Cultural Economics • Economic Sociology • Economic Anthropology- General.