Open Science and Authorship of Supplementary Material. Evidence from a Research Community

Overall, the most frequent variation events are removals, while shuffles immediately follow. This could suggest that dataset and software may require skills that not every author in  has or that, by any means, some of them have a scarce interest in appearing in the supplementary material. The nature of shuffles is even harder to grasp. While this could be accidental in some cases, we believe that in many occurrences, this is intentional; in fact, 93 shuffles out of 168 (55.35%) involving datasets involve non-adjacent authors, making it harder to happen by mere chance.

Additions are the least frequent variation events; nonetheless, this suggests the existence of a considerable "submerged workforce" taking part in research support activities while being underrepresented in literature and hence in traditional rewards and evaluation frameworks.

On a side note, the software has a higher percentage of variation events than datasets, potentially taking to the extreme the aforementioned considerations. Unfortunately, the few data points do not permit us to move further hypothesis.

Generally speaking, a considerable amount of pairs presents substantial variations between the authors participating in the publication and the authors participating in the dataset (or software) deposition, thus indicating the presence within the examined research community of an alternative accreditation scheme in parallel to the traditional one (i.e., same authors in literature and supplementary material). This confirms our initial hypothesis and poses the premises for a longitudinal, large-scale analysis of the phenomenon.