A cluster analysis of scholar and journal bibliometric indicators

✍ Scribed by Massimo Franceschet

• Publisher: John Wiley and Sons
• Year: 2009
• Tongue: English
• Weight: 140 KB
• Volume: 60
• Category: Article
• ISSN: 1532-2882
• DOI: 10.1002/asi.21152

✦ Synopsis

Abstract

We investigate different approaches based on correlation analysis to reduce the complexity of a space of quantitative indicators for the assessment of research performance. The proposed methods group bibliometric indicators into clusters of highly intercorrelated indicators. Each cluster is then associated with a representative indicator. The set of all representatives corresponds to a base of orthogonal metrics capturing independent aspects of research performance and can be exploited to design a composite performance indicator. We apply the devised methodology to isolate orthogonal performance metrics for scholars and journals in the field of computer science and to design a global performance indicator. The methodology is general and can be exploited to design composite indicators that are based on a set of possibly overlapping criteria.