NWB2023_Evaluating the diversity of scientific discourse on twenty-one multilingual Wikipedias using citation analysis
Scientific research is often cited in Wikipedia pages, however little is known about these citations across different Wikipedia languages and the range of research being used only in non-English Wikipedias. This study aimed to investigate the scale of research citation across twenty non-English Wikipedias and the degree to which they cite a different set of research.
We analyzed Wikipedia citations of over 4 million publications in 21 different Wikipedia languages, including over 1 million journal papers and 2 million books. A minimum of 80,000 citations was required for a language to be included in the analysis. The study analyzed the percentage of publications cited in English Wikipedia compared to non-English Wikipedia, as well as the percentage of publications cited in a single non-English Wikipedia compared to multiple non-English Wikipedias.
The study found that approximately 40% of publications were only cited in non-English Wikipedia, either in a single Wikipedia language or in multiple non-English Wikipedias. There were significant differences between the number/percentage of publications shared in English vs non-English Wikipedia, but no significant differences between non-English Wikipedia in a single language compared to those cited in multiple non-English Wikipedias. The study also found that the percentage of books cited in non-English Wikipedia was higher than that of journal publications.
The study concluded that non-English Wikipedia provides a unique voice for research literature and that there is a significant amount of unique publications being solely cited in non-English Wikipedia. These findings suggest that each Wikipedia makes a unique contribution to the representation of knowledge within its stakeholder groups, and underscores the importance of multilingualism in knowledge sharing.
