Options
Connected for Success? : How Network Centrality on ResearchGate Relates to Bibliometrics, Altmetrics and Webometrics
Type
presentation
Date Issued
2015-06-27
Author(s)
Abstract
Academic social network sites (SNS) are booming. A recent large-scale survey published in Nature indicates that almost 90 percent of researchers in science and engineering and more than 70 percent in the social sciences, arts and humanities are aware of ResearchGate - next to Academia.edu the largest academic SNS with more than 6 million users. However, only limited research has been carried out on academic SNS. Although a vivid community creates and implements alternative measures of scientific impact with social media data, little use has been made of the potential of academic SNS as a data source. Consequently, few studies employ person-based metrics that cover users' social capital in the form of structural indicators and network statistics (centrality, density, homophily, clustering). This contribution draws on extensive data from ResearchGate to address this issue and add a relational component to altmetrics research. It includes a follower/following network of 302 nodes on ResearchGate: the complete faculty of a Swiss public university who are members on this academic SNS as of early 2014. We describe the overall network with classical metrics of social network analysis and compute the centrality of each individual node. Results indicate low density, high institutional homophily, a skewed degree distribution and many isolates. We then compare the structural properties of individual nodes with other metrics of influence. To do so, the network data is complemented with detailed attribute data, such as department affiliation, gender and position within the university hierarchy. Moreover, we collect researchers' activity on ResearchGate, bibliometric information, webometrics and altmetrics, i.e., the prominence of their publications on general and specific social media platforms. We evaluate whether the relational aspect of influence in the form network centrality correlates with activity, bibliometric, webometric and almetrics indicators as well as personal attributes. Significant and intermediate correlations between activity and centrality are found, while the correlations between centrality and bibliometric as well as altmetrics are weaker but still significant. No significant correlations between webometrics (coverage of publications on general social media platforms, like Twitter and Facebook) and network centrality occur. The analysis suggests that network centrality is distinct but correlated with (bibliometric) output metrics and therefore worthy of inclusion in future altmetrics studies.
Language
English
Keywords
Altmetrics
ResearchGate
Social Network Sites
Social Media
Bibliometrics
Webometrics
Social Network Analysis
Centrality
HSG Classification
contribution to scientific community
HSG Profile Area
SoM - Business Innovation
Refereed
No
Event Title
XXXV INSNA Sunbelt Conference
Event Location
Brighton
Subject(s)
Division(s)
Eprints ID
241796