What is co-authorship?

Science and technology policy academics and evaluators use co-authorship as a proxy for research collaboration despite knowing better. Anecdotally we understand that an individual might be listed as an author on a particular publication for numerous reasons other than research collaboration. Yet because of the accessibility and other advantages of bibliometric data, co-authorship is continuously used as a proxy for research collaboration. In this study, a national (US) sample of academic researchers was asked about their relationships with their closest research collaborators—some with whom respondents reported having co-authored and some with whom respondents reported not co-authoring. The results suggest there are numerous dimensions of co-authorship, the most influential of which is informal and relational and with little (directly) to do with intellectual and/or other resource contributions. Implications for theory and practice are discussed. Generally we advise academics and evaluators interested in tracking co-authorship as a proxy for collaboration to collect additional data beyond those available from popular bibliometric resources because such information means better-informed modeling and better-informed policy and management decision making.