What this guide covers
What the h-index is — and why it matters
The h-index is a single number that tries to capture two things at once: how much a researcher has published, and how often that work is cited. It was proposed in 2005 by the physicist Jorge E. Hirsch (hence its other name, the Hirsch index), and it has since become one of the most widely quoted figures in academic careers — appearing in CVs, grant reviews, and promotion files.
The definition is precise:
A researcher has an h-index of h if h of their papers have each been cited at least h times — and their remaining papers have no more than h citations each.
So an h-index of 20 means: 20 papers, each cited 20 or more times. The appeal is that it resists two kinds of distortion. A researcher can’t inflate it with one wildly-cited paper (that lifts the count by only one), and they can’t inflate it by publishing dozens of papers nobody cites (those never reach the threshold). It rewards sustained, cited output — which is exactly why evaluators reached for it. Understanding it matters because, fairly or not, this number will be read as shorthand for your standing as a researcher.
How to calculate it (worked example)
To find an h-index by hand, list every paper’s citation count from highest to lowest, then walk down the list until the citation count drops below the paper’s position number. The last position where citations still meet or beat the rank is the h-index.
Take a researcher with seven papers cited 10, 8, 5, 4, 3, 2, 1 times:
The cut-off falls at position 4: four papers have at least 4 citations each, but there is no fifth paper with at least 5. h-index = 4.
You will almost never need to do this manually — the databases compute it for you (see next section). But doing it once makes the number concrete, and shows why adding a brand-new paper, or picking up a few more citations on your lower-ranked work, is what actually nudges an h-index upward.
Where to find yours — and why the numbers differ
Three major databases each calculate an h-index — and they will give you three different numbers. This is not an error. Each database only counts the publications and citations it indexes, and their coverage differs enormously.
| Source | Access | Coverage & typical h-index |
|---|---|---|
| Google Scholar | Free — public profile | Broadest net (journals, conferences, preprints, theses, repositories, and self-citations) → usually the highest h-index. Also shows the i10-index. |
| Scopus (Elsevier) | Subscription | Curated journal set → intermediate value. |
| Web of Science (Clarivate) | Subscription | Most selective coverage → usually the lowest h-index. |
Studies comparing the three find Google Scholar’s h-index runs, on average, roughly 1.4× a researcher’s Web of Science value, with Scopus a little above Web of Science — though the exact ratio varies by person and field, so treat those as tendencies, not constants. The practical upshot: always state which database an h-index came from. An “h-index of 12” is meaningless without knowing whether it is a (generous) Google Scholar figure or a (conservative) Web of Science one.
You can see — and show — your own h-index for free by creating a Google Scholar Profile. It displays your h-index and your i10-index (the number of your publications with at least 10 citations each, a Google-Scholar-only metric), and updates automatically as citations accrue. For researchers without a Scopus or Web of Science subscription, this is the accessible option.
What the h-index can’t tell you
The h-index is useful, but a single number hides a great deal. Every one of these limitations is well established — and every committee member should know them.
| Limitation | What it means for you |
|---|---|
| It only ever rises | The h-index never decreases, even if a researcher stops publishing. It measures accumulated standing, not current activity. |
| It is field-dependent | Citation cultures differ hugely — biomedicine cites far more than mathematics. An h-index of 15 can be exceptional in one field and average in another. Never compare across disciplines. |
| It favours long careers | Because it can only grow, senior researchers almost always out-score early-career ones. It penalises people with less “academic age.” |
| It ignores your role | It counts a paper the same whether you were sole author or fifteenth of twenty. Author position and contribution vanish. |
| It can be gamed | Excessive self-citation can inflate it, so the raw number isn’t tamper-proof. |
| It flattens your best work | Once a paper passes the h threshold, extra citations on it don’t help. A landmark 2,000-citation paper counts the same as one just over the line. |
Two variants try to patch specific gaps. The m-quotient (Hirsch’s own idea) divides the h-index by the number of years since your first publication, to compare researchers at different career stages. The g-index (Leo Egghe, 2006) gives extra weight to your most-cited papers, addressing the “flattening” problem. Both are refinements — not replacements for judgement.
The African researcher’s context
There is a structural reason African researchers should be especially careful with database h-indexes: African-published journals are severely under-indexed in the subscription databases. A peer-reviewed study found that of Africa’s roughly 2,200 journals, only about 7–8% appear in Web of Science or Scopus, compared with around 46% in Crossref — and sub-Saharan African journals are several times less likely to be indexed than journals elsewhere. Nigeria, which produces a large share of the continent’s journals, is heavily underrepresented.
If much of your work is published in African journals, a Scopus or Web of Science h-index may understate your real output and impact, simply because those databases don’t index where you published. Google Scholar’s broader coverage tends to capture more of it — so for many African researchers it is the comparatively more representative source. State your Google Scholar h-index, and say so.
On promotion: Nigerian academic advancement places heavy weight on publication counts and, increasingly, on publishing in Scopus- or Web-of-Science-indexed journals; some institutions and assessors do reference citation metrics including the h-index. But there is no single documented national h-index threshold — so treat any specific “you need an h-index of X” claim with caution and check your own institution’s actual criteria.
Using it responsibly & how FRELIP helps
The wider research world has moved decisively toward using metrics responsibly rather than abolishing or worshipping them. Two landmark statements lead this:
- DORA (the San Francisco Declaration on Research Assessment) urges evaluators not to use journal-based metrics as a surrogate for the quality of an individual’s work in hiring, promotion or funding. Its main target is the Journal Impact Factor, but the principle — don’t reduce a person to a single number — applies squarely to the h-index.
- The Leiden Manifesto sets out ten principles for responsible metrics; the first is that quantitative evaluation should support, not replace, expert qualitative judgement.
More recent coalitions such as CoARA continue this push. The consensus is not “the h-index is worthless” — it is “one number can inform a conversation, but it cannot be the conversation.”
- Fair visibility. FRELIP indexes open African scholarship so that work published outside the big subscription databases is still findable and citable — helping close the very coverage gap that understates African h-indexes.
- Identity done right. An h-index is only as good as the record it’s computed from. Pair a complete ORCID profile with DOIs on your outputs so your citations attach to you correctly, in every database.
- Context, not just numbers. This explainer is part of a FRELIP series on how research is measured, shared and discovered.
The takeaway: know your h-index, know which database it came from, and know what it leaves out. Use it as one line in a fuller story of your research — alongside the papers themselves, your role in them, and the difference they make — never as the whole verdict.
FRELIP fact-check report
Every factual claim here was checked against Hirsch’s original work, the databases’ own documentation, and the responsible-metrics literature. We separate what is well established from what needs qualification.
Verified Confirmed by primary or authoritative sources
- Definition: h-index = h if h papers are each cited ≥ h times and the rest ≤ h — matches Hirsch’s own wording. (Hirsch, PNAS 2005 / arXiv physics/0508025)
- Proposed by physicist Jorge E. Hirsch in 2005 (PNAS); also called the Hirsch index. (PNAS 10.1073/pnas.0507655102)
- Worked example 10, 8, 5, 4, 3, 2, 1 → h = 4 (verified by direct calculation).
- Purpose: combine productivity and citation impact in one robust number. (Hirsch 2005)
- Scopus, Web of Science and Google Scholar each compute an h-index and give different values; Google Scholar (broadest coverage) is typically highest, Web of Science lowest. (library guides; bibliometrics literature)
- Google Scholar shows h-index and i10-index free on a public profile; Scopus and Web of Science are subscription databases. (Cornell / Google Scholar docs)
- i10-index = number of publications with ≥ 10 citations (a Google Scholar metric). (Cornell)
- Limitations: only rises; field-dependent; career-length dependent; ignores author position/co-authors; inflatable by self-citation; insensitive to citations beyond the threshold. (bibliometrics literature)
- g-index (Egghe, 2006) weights highly-cited papers; m-quotient (Hirsch) = h ÷ years since first publication. (Egghe; Hirsch 2005)
- African journals are under-indexed in Web of Science/Scopus (~7–8%) vs Crossref (~46%); Nigeria underrepresented. (Asubiaro et al., JASIST 2023)
- DORA and the Leiden Manifesto urge responsible, expert-led assessment over single metrics. (sfdora.org; Hicks et al., Nature 2015)
Hedged True but needs qualification
- The “Google Scholar ≈ 1.4× Web of Science” ratio comes from specific studies and varies by person and field — a tendency, not a fixed constant.
- That Google Scholar is “more representative” for African researchers is a reasonable inference from the coverage gap, not a directly measured finding.
- Nigerian promotion weights publication counts and Scopus/WoS indexing, and some assessors cite the h-index — but there is no documented national h-index threshold.
- The m-quotient’s exact benchmarks (m ≈ 1 / 2 / 3) rest on secondary summaries of Hirsch’s paper.
Excluded Claims we deliberately did not make
- That DORA or the Leiden Manifesto attacks the h-index by name — both target the Journal Impact Factor and single-metric reliance broadly; we frame them at that level.
- Any specific national Nigerian h-index promotion requirement — no authoritative source supports one.
Sources
2. PNAS 102(46):16569–16572 — the original h-index paper ↗
3. Cornell University Library — h-index & i10-index ↗
4. UT Dallas — h-index across Web of Science, Scopus & Google Scholar ↗
5. Cornell — Google Scholar Citations profiles ↗
6. g-index (Egghe, 2006) ↗
7. Asubiaro et al. (JASIST 2023) — under-indexing of African journals ↗
8. DORA — San Francisco Declaration on Research Assessment ↗
9. Hicks et al. — The Leiden Manifesto for research metrics (Nature 2015) ↗
10. CoARA — Coalition for Advancing Research Assessment ↗
This is an educational guide to the h-index; it is not affiliated with any citation database. Spotted something to correct? Tell us.
📂 More Open Access & Research Practice explainers