Normalized International Collaboration Score: A Novel Indicator for Measuring International Co-Authorship

International collaboration on research publications is increasingly evaluated as part of a raft of performance measures. Levels of international co-authorship have increased substantially over the last few decades and vary substantially by research field and publication type; however, these variations are not typically accounted for by international collaboration indicators. In this research-in-progress paper, we introduce a novel metric, the Normalised International Collaboration Score, which adjusts the number of countries appearing on publication records using baselines relevant to the subject, age and type of the publication. A pilot analysis shows that these baselines vary substantially and that the application of this metric yields very different results to a more common measure of international collaboration. The limitations of the metric are discussed, along planned extensions for the full version of the study, as well as the relationship between normalised collaboration and citation. Conference Topic Indicators Background and Purpose Measuring international co-authorship The availability of author address metadata on publication indices such as Web of Science and Scopus allows the analysis of patterns in co-authorship, including the collaboration by authors from different countries on research outputs. This approach has been used in many studies for decades (such as Glänzel & De Lange, 1997; Narin, Stevens, & Whitlow, 1991; Nederhof & Moed, 1993) and metrics describing international collaboration now appear regularly in bibliometric handbooks (Colledge, 2014; Rehn, Kronman, & Wadskog, 2007) and in reporting tools such as Thomson Reuters’ InCites, Elsevier’s SciVal and SCImago’s Journal & Country Ranking. Such publications tend to receive higher levels of citation, an effect that is not due to the increased propensity for self citation arising from additional authors (Van Raan, 1998), but likely rather shared experience, knowledge and equipment. Analysis of international co-authorship metadata has highlighted other important aspects of collaboration. Firstly, levels of international collaboration have increased substantially over the last quarter century (Leydesdorff & Wagner, 2008); and secondly, levels of international collaboration vary by field of research (Frame & Carpenter, 1979). A report on Thomson Reuters’ InCites (retrieved 7 January 2015) indicates that 2013 articles, reviews and proceedings papers in Tropical Medicine involved international collaboration 46.7% of the time, while for History, this was only 4.3% of the time. Even within Medicine, Emergency Medicine saw only 9.9% foreign collaboration, far lower than Tropical Medicine. Variation is significant over time, with Astronomy & Astrophysics international collaboration rising from 19.4% in 1993 to 45.0% in 2013. To these two aspects, we must add publication type; 2013