“…For example, the use of some specific sources such as recommended articles from F1000 (Bornmann, 2014;2015;Bornmann;Haunschild, 2015;2018) or articles from specific journals such as PLoS One (Zahedi et al, 2014c;De-Winter, 2015;Costas, 2018) and Journal of the Association for Information Science and Technology (Bar-Ilan et al, 2019) produce specific results that can hardly be generalised. Another important bias is introduced when local studies are analysed, mainly from emergent regions such as Latin America (Maricato;Filho, 2018;Alperin, 2015) or from non-English-speaking countries such as Spain (Torres-Salinas et al, 2016;Torres-Salinas et al, 2018b), Peru (Eléspuru-Briceño; Huaroto, 2016) or China (Wang et al, 2016). The selection of specific formats such as books (Hammarfelt, 2014;Torres-Salinas et al, 2017a;2017b;Torres-Salinas et al, 2018a) and datasets Peters et al, 2016;Kraker et al, 2015) also limits the extrapolation of the results.…”