Student Network Centrality and Academic Performance: Evidence from United Nations University

In this paper we empirically studied the relationship between network centrality and academic performance among a group of 47 PhD students from UNU-MERIT institute. We conducted an independent email survey and relied on social networks theory as well as standard econometric procedures to analyse the data. We found a significant reversed U-shaped relation between network centrality and students’ academic performance. We controlled our results by several node’s characteristics such as age, academic background, and research area. Additional evidence shows that there is a negative impact of age on academic performance at PhD student level. Contributions of this paper can refer to the input into studies that aim to explore peereffect. Also it contributes to the methodological approach by combining elements of network analysis and econometric theories. This study demonstrates that when evaluating the impact of network centrality on performance, there is no significant difference between various network centrality measurements.

JEL : D85, I21, I23, L14

Keywords : Networks analysis, Network centrality, Peer-effect, Academic performance

UNU-MERIT Working Papers ISSN 1871-9872