Network and information-theoretic studies on the effects of research environment in scientific careers: geography, prominence and gender

Identificador:  TDX:4226

Handle: https://hdl.handle.net/20.500.11797/TDX4226

Autores:  Danús Amengual, Lluís

Resumen:
Science is the result of the coordinated effort of many individuals, and as such it is a complex social system. Consequently, its features depend on the interactions between many actors and its evolution cannot be explained by the behavior of the different actors composing it and thus, its study is extremely challenging. In this thesis we address three different questions about the development of science with a network science and an information theoretical approach. The first question is how different institutional and research cultures, such as those in North America and Europe, affect the way scientists collaborate with each other. We observe how despite being equal at first glance, prominent researchers based in North America display a different collaborative behavior than those based in Europe, with the latter establishing significantly denser collaboration networks than the former. The second question concerns the characteristics of prominent researchers in their early career. We found that those who later became part of the most cited authors in their fields, start with an initial advantage in terms of publication venues, citations and institutional prestige that is sustained over time, to the end that one can estimate correctly their later performance. The last question we analyze is the effect that the research environment and gender have in the selection of a research topics in the early career of young faculty. Analyzing the topics of research before joining a department, we observe how male early-career faculty approach the research topics of the department at higher rates that their female counterparts despite having the same number of collaborations overall and that those collaborations are career age-shaped in the case of female researchers. In conclusion, we show how more nuanced analyses reveal effects that do not appear with more traditional approaches.