Determining gender in academic authorship: a comprehensive and methodical approach

Purpose This study introduces an advanced method for ascertaining the gender identity of scientific authors of academic papers indexed in Web of Science or Scopus. This study aims to improve gender information retrieval, thereby benefiting gender-based bibliometric studies. Design/methodology/approach A sample of 187 authors was examined using two identification methods: manual examination and R scripts performing queries in different databases. Gender verification involved consulting databases such as Scopus, ORCID, Virtual International Authority File (VIAF), Wikidata and Gender-API. Findings Findings revealed that manual checks identified the gender of 50.8% of authors in Wikidata and 37.43% in VIAF, achieving high reliability but varying success across datasets. The R script identified 42.24% of authors in Wikidata and 31.01% in VIAF, underscoring the value of automated tools despite limitations. In addition, this study identifies and addresses biases in automated gender identification practices, which have contributed to the underrepresentation of women in bibliometric studies. Originality/value The study emphasizes the need for more accessible data, including linked open data, to refine author gender identification within the scientific domain. By highlighting and mitigating biases in automated methods, this dual-method approach not only enhances gender identification accuracy but also contributes to more equitable representation in academic bibliometric analyses.