h-index

h-index
Name	h-index
Introduced	2005
Creator	Jorge E. Hirsch
Type	bibliometric index
Measures	citation impact and productivity
Unit	integer

Contents

h-index

The h-index is a bibliometric indicator intended to quantify the combined productivity and citation impact of an individual or entity; it is used in evaluations by universities, funding agencies, and bibliographic databases. It is calculated from citation records maintained by services such as Web of Science, Scopus, Google Scholar, Microsoft Academic, and aggregated by platforms including ResearchGate, ORCID, Dimensions, and Publons. Institutions and prize committees at Harvard University, Stanford University, University of Oxford, Massachusetts Institute of Technology, and Max Planck Society commonly consider the h-index alongside metrics like the Journal Impact Factor, Eigenfactor, and Altmetric scores.

Definition and calculation

The h-index is defined for a researcher, journal, or institution as the maximum integer h such that h publications have each received at least h citations; calculation uses ranked citation lists from sources like Web of Science, Scopus, Google Scholar, CrossRef, and PubMed. Computationally, the h-index can be derived by sorting outputs from databases associated with Clarivate Analytics, Elsevier, Digital Science, Wiley, and Springer Nature and identifying the point where publication rank equals citation count. Variants in calculation arise when using curated profiles at repositories such as arXiv, SSRN, Zenodo, Figshare, and institutional repositories at University of Cambridge, Yale University, Princeton University, Columbia University, and California Institute of Technology.

The metric was proposed in 2005 by physicist Jorge E. Hirsch while at University of California, San Diego and disseminated through citations in venues read by scholars at American Physical Society, Institute of Physics, Nature, Science, and Proceedings of the National Academy of Sciences. Early adoption and critique occurred in debates involving researchers affiliated with Sloan Foundation, Wellcome Trust, National Science Foundation, European Research Council, and editorial offices of journals such as Nature, Science, The Lancet, Cell, and PNAS. Historical predecessors and contemporaneous alternatives were developed by statisticians and bibliometricians at Leiden University, University of Leiden, CWTS, Harzing, Gingras, and Hirsch's contemporaries.

Several modifications and complementary indices were introduced, including the g-index by Leo Egghe (used by scholars at Université de Liège, Ghent University, Vrije Universiteit Brussel), the i10-index popularized by Google Scholar Citations and used by Google, the m-index recommending normalization over career length cited by academics at MIT, Caltech, UCL, and the contemporary a-index, r-index, and h-b index discussed in literature from Clarivate, Elsevier Research Intelligence, Scimago Lab, SCImago Journal Rank, and conference proceedings of International Society for Scientometrics and Informetrics. Composite approaches combine h-index with altmetrics tracked by Altmetric.com, usage data from COUNTER, and collaboration measures employed by consortia at CERN, European Organization for Nuclear Research, NASA, and NIH.

Advocates cite simplicity and resistance to outliers when comparing scholars at Princeton University, University of Chicago, Imperial College London, ETH Zurich, and Johns Hopkins University, while critics at Royal Society, Academia Europaea, American Association for the Advancement of Science, Royal Society of Chemistry, and American Chemical Society point to insensitivity to author order, field differences, and database coverage. The metric can be gamed through self-citation patterns documented in analyses involving Elsevier, Clarivate Analytics, Google Scholar, ResearchGate, and editorial boards of PLOS, Frontiers, MDPI, IEEE, and ACM. Legal and policy implications have been considered by agencies such as European Commission, U.S. Department of Education, UK Research and Innovation, Australian Research Council, and Canadian Institutes of Health Research.

Different fields show characteristic h-index distributions, with higher typical values in areas represented at CERN, Human Genome Project, Large Hadron Collider, International Space Station, and Human Connectome Project, compared with humanities and arts communities at British Library, Library of Congress, Getty Research Institute, Museum of Modern Art, and National Endowment for the Arts. Disciplines indexed by PubMed and MEDLINE (medical sciences), IEEE Xplore (engineering), arXiv (physics, mathematics, computer science), and JSTOR (humanities, social sciences) require field-appropriate normalization; normalization methods have been proposed by researchers at Leiden University, King's College London, University of Toronto, Peking University, and Tsinghua University.