"It's Dangerous to go Alone!" Scientific Excellence of PhD Holders and Their Mentors - Network Analysis of Croatian Doctoral Students

Collaboration between researchers is seen as paramount in advancing science, be it because of its potential to offer novel ideas crossing scientific fields, because of its ability to boost productivity by having researchers who work well together, or by introducing and retaining new scientists to a wider network of peers, allowing them to access knowledge and information otherwise unavailable. This article sets to explore a specific part of scientific collaboration: mentor-protégés collaboration, especially during the formative years of a protégé's scientific career, during the writing of their dissertation and five years after. Gathering data on scientific productivity from the publication repository aggregator Croatian Scientific Bibliography, mentor-protégé collaboration was explored in order to test whether mentor's productivity could affect his or her protégé's productivity. Analysis of variance and linear regression analysis confirm that mentor productivity is positively correlated with protégé productivity, especially in the case where mentors are highly productive (stars) scientists. Additionally, network values such as betweenness and weighted degree centralities are explored in order to test whether mentors' values will affect protégé's position in the network. While mentor-protégé betweenness centrality values are found not to be correlated, weighted degree measures do seem to play an important role in introducing protégés to their mentor's network of peers and co-authors.