Comparing the SDG-classification systems of Dimensions, InCites and SciVal for the University of Helsinki

Abstract

Introduction. We determined how the publications of the University of Helsinki were attributed to the United Nations sustainable development goals (SDGs) in three different databases. The databases were Dimensions, Scopus and Web of Science. We also looked at the overlaps.

Method. DOIs for the publications were collected from our current research information system. These DOIs were then used to create publication sets in the databases. SDG-attributions of publications in these publication sets were collected for analysis.

Analysis. Total publication counts and overlaps were calculated with simple Excel functions. This was done for all publications and separately for only those publications that were found in all three databases. Venn graphs were plotted for each sustainable development goal.

Results. The analysis revealed significant differences between the attributions. Overlaps tended to be low usually covering only a few percent up to, at best, half of the publications. Total number of publications also varied with manyfold differences. The attributions were also found to be rather unstable due to updates in attribution systems.

Conclusions. Sustainable development goal attributions are very database dependent and remain unstable. At present, there is no responsible way to track them for evaluative purposes. It is unlikely that the situation will change in the future. This is due to their general scope and a lack of stabilising mechanisms.