Count tables: A single-model census of 40 million PubMed articles (seniority, discipline, and team composition in the gender gap in biomedical authorship, 2002-2025)

Aggregated count tables reproducing the figures and tables of the study 'A single-model census of 40 million PubMed articles: seniority, discipline, and team composition in the gender gap in biomedical authorship, 2002-2025'. These CSVs are the aggregated outputs of the analysis and reproduce the published figures and tables without the underlying PubMed mirror (the coverage table over 1945-2025 is derived from the full mirror and is not included). Gender is name-inferred (perceived) gender, a binary proxy; discipline labels are model-assigned. Inference is restricted to 2002-2025. Single-author articles are counted as last authors only; homophily excludes the team leader. Analysis and classification code: https://github.com/sheppard94g/biomedical-gender-gap-pubmed . Gender validation uses the World Gender Name Dictionary (WGND 2.0, doi:10.7910/DVN/MSEGSJ); bibliographic metadata are from PubMed/MEDLINE; neither third-party dataset is redistributed here.