Argumentation From Evidence: A Bibliometric Analysis of the Knowledge Structure of an Interdisciplinary Research Field

Argumentation from evidence is treated as a core scientific practice in contemporary science education, yet the term argumentation is used with different meanings across several research communities. This creates difficulty in delimiting the literature when a review is undertaken. This study maps the knowledge structure of the literature on argumentation associated with data and evidence, using metadata from Scopus and four science mapping techniques implemented in VOSviewer. A high-sensitivity query on the TITLE-ABS-KEY field returned 14,938 records, with no restriction on year, document type, subject area or language. The keyword co-occurrence network retained 1,000 keywords in five clusters; the co-citation network retained 251 cited references in seven clusters; the co-authorship network retained 13 authors in three clusters; and the document citation network retained 21 documents in six clusters. The four analyses converge on a single observation: the field is a set of communities that share terminology rather than an integrated research area. The science education cluster in the co-citation network has the highest total link strength among the seven clusters and contains a stable set of foundational works published mainly between 2000 and 2010. The argumentation theory cluster is the smallest in size but has the highest mean link strength, indicating that it is cited across clusters. Keywords emerging in the final two years of the dataset are associated with generative artificial intelligence. Two implications are drawn, one for the design of search strategies in review studies and one for research on argumentation in school science.