Inter-rater reliability of h-index scores calculated by Web of Science and Scopus for clinical epidemiology scientists.

Objective We investigated the inter-rater reliability of Web of Science ( Wo S) and Scopus when calculating the h-index of 25 senior scientists in the Clinical Epidemiology Program of the Ottawa Hospital Research Institute. Materials and methods Bibliometric information and the h-indices for the subjects were computed by four raters using the automatic calculators in Wo S and Scopus. Correlation and agreement between ratings was assessed using Spearman's correlation coefficient and a Bland- Altman plot, respectively. Results Data could not be gathered from Google Scholar due to feasibility constraints. The Spearman's rank correlation between the h-index of scientists calculated with Wo S was 0.81 (95% CI 0.72-0.92) and with Scopus was 0.95 (95% CI 0.92-0.99). The Bland- Altman plot showed no significant rater bias in Wo S and Scopus; however, the agreement between ratings is higher in Scopus compared to Wo S. Conclusion Our results showed a stronger relationship and increased agreement between raters when calculating the h-index of a scientist using Scopus compared to Wo S. The higher inter-rater reliability and simple user interface used in Scopus may render it the more effective database when calculating the h-index of senior scientists in epidemiology. [ABSTRACT FROM AUTHOR]