Testing for alphabetical bias in the c-score: Implications for evaluating scientific excellence

This study investigates whether the initial letter of a researcher’s last name is associated with the likelihood of being included among the world’s top 2% scientists listed in the widely used Ioannidis Top Scientists List (ITSL), based on the c-score. Building on prior evidence of alphabetical biases in citation practices, authorship conventions, and evaluative metrics, we test whether scholars whose surnames begin with early-alphabet letters are over-represented relative to expected national distributions derived from a global dataset of Scopus-indexed authors. Comparing observed and expected surname-initial distributions, we detect substantial deviations both when surnames are grouped into four alphabetical tiers and when treated as a 26-point index. A Poisson regression framework with country fixed effects reveals a negative and statistically significant association between alphabetical position and the probability of appearing in the ITSL. Country-level estimates show considerable heterogeneity across major scientific systems, while field-level regressions indicate that alphabetical bias is strongest in humanities, economics, and social sciences, and negligible or reversed in engineering and computing. The findings suggest that the c-score may incorporate structural distortions linked to authorship conventions and surname ordering, raising broader concerns about fairness and interpretation in high-stakes bibliometric evaluation.