What is the h-index? How it is computed and what it tells you
The h-index tries to answer “how much work, with how much impact” with one number. The definition is short: a researcher has index h if h of their papers have at least h citations each. Everything else — which database, which field, journal or author — is the part people skip when reading it.
Definition and a worked example
Sort the publications by citation count, highest first. Walk down the list until a paper’s citations fall below its rank; the last rank that still holds is h. Example: seven papers cited 40, 22, 15, 9, 6, 3, 1 times. Paper 1: 40 ≥ 1, paper 2: 22 ≥ 2, paper 3: 15 ≥ 3, paper 4: 9 ≥ 4, paper 5: 6 ≥ 5, paper 6: 3 < 6 → h = 5. Five papers with at least five citations each.
Two properties follow immediately: one very highly cited paper cannot raise h on its own (4,000 citations instead of 40 still gives h = 5), and uncited papers do not lower it. That is by design — the physicist Jorge Hirsch proposed it in 2005 to capture the intersection of productivity and impact.
One person, three different h-indexes
The h-index depends on the database because citation counts do. Google Scholar counts books, theses, preprints and conference papers, so it usually gives the highest h. Scopus and Web of Science count only citations from the journals they index, and their journal lists differ from each other. Seeing Scholar 28, Scopus 19, WoS 16 for the same researcher is normal; none is wrong, they count different universes. If an application asks for “the h-index”, read which source it means.
Why comparing across fields misleads
Citation habits differ by field several times over. A molecular biology paper carries 60–80 references and the field publishes fast; in mathematics, 15–20 references and multi-year publication cycles are ordinary. The same productivity and the same relative impact therefore yield very different h values in the two fields. The h-index is only a meaningful comparison within one field, at similar career age, in the same database. Across fields, field-normalised measures (paper-level FWCI) are fairer — we explained how to read them in quality signals.
Journal h-index: the number you see on Pofolia
Apply the same computation to a journal’s papers and you get the journal h-index: the journal has h papers cited at least h times. SCImago publishes it yearly for every journal, and the “h-index” shown next to SJR and quartile on Pofolia’s journal pages is that number. A journal h-index depends heavily on age and volume — a 100-year-old journal publishing 3,000 papers a year will almost always beat a young, selective journal publishing 60. So when comparing journals, look at quartile and SJR first; read the h-index as “how much accumulated work” rather than “how selective”.
What the h-index leaves out
- Author position: first and tenth author receive the same credit. Large consortium papers raise h quickly.
- Self-citation: counted in most computations. Scopus and WoS also give h without self-citations; a large gap is worth noticing.
- Career age: h only ever rises. A 12 after thirty years and a 12 after five are not the same thing; some institutions divide h by career years (the m-quotient).
- A single landmark paper: one paper with 5,000 citations makes h just 1. Total citations and the most-cited paper fill that gap.
- Methodological soundness, data sharing, replication: none of it counts. The number helps you rank, not skip reading.
Where to look it up
- Google Scholar profile: the researcher’s own profile shows h-index and i10-index (papers with at least 10 citations); free.
- Scopus: the author page, with a self-citation-excluded option; through a university subscription.
- Web of Science: via Author Search; Core Collection citations only.
- For a journal: the SCImago Journal Rank (scimagojr.com) journal page; Pofolia’s field pages and journal guide list the same number next to quartile and SJR.