Measuring the robustness of the journal h-index with respect to publication and citation values: A Bayesian sensitivity analysis

Braun, Glänzel, and Schubert (2006) recommended using the h-index as an alternative to the journal impact factor (IF) to qualify journals. In this paper, a Bayesian-based sensitivity analysis is performed with the aid of mathematical models to examine the behavior of the journal h-index to changes in the publication/citation counts of journals. Sensitivity of the h-index was most apparent for changes in the number of citations, revealing similar patterns of behavior for almost all models and independently to the field of research. In general, the h-index was found to be robust to changes in citations up to approximately the 25th percentile of the citation distribution, inflating its value afterwards.