Measuring epistemic success of a biodiversity citizen science program: A citation study

Materials and methods The ‘‘Vigie-Nature” program Vigie-Nature started in 1989, with the aim of improving our knowledge of biodiversity by engaging a large diversity of people (amateur naturalists, managers of green areas, pupils, farmers and other citizens) in the collection of field data in the whole French territory (for more information, one can refer to http://www.vigienature.fr/fr/presentation-2831). Vigie- Nature mobilizes non-professional citizens to quantify the abundance of species, in order to characterize the evolution of ecosystems in terms of species assemblage. It gathers different participative observatories, focusing notably on pollinator insects (Spipoll project), birds (Bird- Lab), or urban vegetation (Sauvage de ma rue). The program is managed by the MNHN, with the collaboration of naturalist associations. Data quality is controlled by scientific experts, which verify species’ identifications. The researchers use these collected data to advance research in ecology and conservation science by publishing results in scientific journals such as Biodiversity and Conservation, Journal of Ornithology, Biological Conservation, Landscape Ecology, Global Change Biology, and Frontiers in Ecology and the Environment. Between 2007 and 2019, 171 articles were published within the frame of this program. The corresponding references are available in a dedicated data base (https://www.zotero.org/groups/85788/vigienature/items/collectionKey/4M3I4ZUR). Within this body of publications, 123 articles present scientific results on the basis of collected data,

while the remaining articles propose meta-reflections about citizen science. The Vigie-Nature program is thus well adapted for the purpose of scientometric studies since it allows to pursue statistical analysis on the basis of a large sample of publications. Whereas studies such as Oden- wald’s (2018) rely on aggregated data from distinct citizen science programs, the bulk of Vigie- Nature’s publications is large enough to allow for a study of the scientific impact of one specific project. Data collection To conduct our study, we considered the 123 scientific articles in the field of ecology and envi- ronmental sciences from the Vigie-Nature bibliography published between 2007 and 2019, with a total of 5813 citations in peer-reviewed journals (hereafter, VN group of articles). Fig 1 gives the distribution of this sample with respect to publication year. To collect the number of citations on each year post-publication for each of these publica- tions we used the Web Of Science (WoS) database—a reliable and user friendly platform, widely considered the ‘‘gold standard” for citation studies [31]. We first computed the number of self-citations (we considered an article A is self-cited by B when at least one author of A is also an author of B) in order to allow for a more comprehensive quantification of the scientific impact of Vigie-Nature’s result. We found a self-citation rate of about 11% in average, which appears to be close to the average value of self-citations in contemporary science (about 12%

following [32]), but lower than the self-citation rate in the ecological sciences (about 16% fol- lowing [33]). S2 Appendix compares the citation profile obtained with and without self-cita- tion. In the following, we do not exclude self-citation from the citation rates. We then calculated the number of citations for each article from their year of publication (shifted to year 0) up to 2019. This produces data on a range from 0 to 12 years post-publica- tion, with 123 articles for the first point (time = 0 year) and 6 articles (corresponding to those published in 2007) for the last one (time = 12 years). This method of calculation might intro- duce a bias because a paper published, for instance, on 2019-01-01 is treated as identical to a paper published on 2020-12-31. This bias may be expected to be influential for the first point PLOS ONE A citation study of a biodiversity citizen science program PLOS ONE | https://doi.org/10.1371/journal.pone.0258350 October 11, 2021 3 / 13