Disciplinary collaboration rates in the social sciences and humanities: what is the influence of classification type?

Using different methods to assign disciplines to publications can influence bibliometric analyses. In this study, we test the influence of applying two different types of classification on the disciplinary collaboration rates of researchers from the Social Sciences and Humanities. Two different classification types are contrasted: organisational classification, which assigns discipline(s) based on the discipline of the unit(s) of the authors, and cognitive classification, which considers the discipline(s) assigned to the channel of the publication. The data set is based on a comprehensive local database of SSH research in Flanders, Belgium. Applied to collaboration, the two classification types both show an overall increase in co-authorship in SSH during the studied period. For certain periods, however, they reveal clearly dissimilar trends, especially for publications written by Humanities scholars: while the Humanities according to the cognitive classification have reached a plateau in co-authorship, collaboration rates in the Humanities according to the organisational classification continue to increase. We show that these variations are due to an increase in the proportion of publications of Humanities researchers outside Humanities channels. As such, the comparison of classification types can provide a deeper understanding of disciplinary differences in the evolution of co-authorship.