How international is internationally collaborated research? Heritage composition of Russia's international collaboration network

International research performance indicators attain increased attention in science policy. More studies point to ongoing global bias in production, composition and assessment of research performance metrics (Rafols et al., 2012; van Leeuwen et al., 2001). This research examines international collaboration. It is an indicator routinely operationalised as co-authorship of articles between organisations by taking organisational address as a proxy of the collaborating country (Katz and Martin, 1997). We use geographical approximation of author heritage rooted in the morphology of the surname and find that in a significant minority of internationally collaborated papers, co-authors are likely to have the same origin. In other words, we observe an overestimation in the international collaboration indicator. The findings indicate that if a significant share of international collaborations of a national research system occurs with researchers previously affiliated with this system, internationally coauthored publications are therefore only 'inter-national' on a formal inter-organisational level. This contributes to the evidence that stresses more complex nature of scientific collaboration (Bozeman and Corley, 2004) and may have fundamental implications on the use of international collaboration indicator.