Abstract

Scientific authorship has become a contested terrain in contemporary science. Based on a survey of authors across fields, we measure the likelihood of specialist authors : people who only made specialized contributions, such as data, materials, or funding; and “nonauthor collaborators” : those who did significant work on the project but do not appear as authors, across different research contexts, including field, size of the project team, commercial orientation, impact of publication, and organization of the collaboration. We find that guest and ghost authors are common, with about one-third of publications having at least one specialist author and over half having at least one nonauthor collaborator. We see significant cross-field variations in both overall rates and types of specialist authors and nonauthor collaborators. We find there are generally fewer specialist authors among highly cited papers and more graduate student nonauthor collaborators in single location projects. The results suggest authorship practices vary across fields, and by project characteristics, complicating the use of authorship lists as a basis for evaluation. We discuss implications of these findings for interpreting author lists in the context of science policy.