PUB - Publikationen an der Universität Bielefeld

A powerful strategy within the study of collaboration in science is to posit that co-authorship patterns represent social networks. It is prerequisite to an application of Social Network Analysis (SNA) to define the network entities. A network analysis of the inter-institutional collaboration in COLLNET on the basis of co-authorships was conducted. The study reveals that it is crucial whether the co-authorship itself is seen as an author's relational property or as a social event that brings the authors together. The former possibility is represented by a onemode network in which each author can be related to each other author. Quite distinct from that are two-mode networks, the latter approach. They consist of two single data sets in which relations are only possible between different sets. Different modes of representations require different network approaches. One is that co-authorship networks are seen as one-mode networks, which has the advantage of the application of a variety of measures. In contrast, twomode networks, the other option, cannot be analysed by standard techniques but its distinctive features demand a new conceptualisation of measures. In conclusion, the two-mode perspective is more promising because it allows a dual perspective on collaboration in science which includes researchers as well as their scientific output.