, Volume 101, Issue 2, pp 1345-1360
Date: 07 May 2014

Comparative study on structure and correlation among author co-occurrence networks in bibliometrics

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This paper introduces author-level bibliometric co-occurrence network by discussing its history and contribution to the analysis of scholarly communication and intellectual structure. The difference among various author co-occurrence networks, which type of network shall be adapted in different situations, as well as the relationship among these networks, however, remain not explored. Five types of author co-occurrence networks were proposed: (1) co-authorship (CA); (2) author co-citation (ACC); (3) author bibliographic coupling (ABC); (4) words-based author coupling (WAC); (5) journals-based author coupling (JAC). Networks of 98 high impact authors from 30 journals indexed by 2011 version of Journal Citation Report-SSCI under the Information Science & Library Science category are constructed for study. Social network analysis and hierarchical cluster analysis are applied to identify sub-networks with results visualized by VOSviewer software. QAP test is used to find potential correlation among networks. Cluster analysis results show that all the five types of networks have the power for revealing intellectual structure of sciences but the revealed structures are different from each other. ABC identified more sub-structures than other types of network, followed by CA and ACC. WAC result is easily affected and JAC result is ambiguous. QAP test result shows that ABC network has the highest proximity with other types of networks while CA network has relatively lower proximity with other networks. This paper will provide a better comprehension of author interaction and contribute to cognitive application of author co-occurrence network analysis.