The h-index and what counts as good

Work out your h-index

Type or paste one citation count for each paper. Separate the counts with commas, spaces or new lines, so a spreadsheet column pastes as it is. Take every count from one source, such as your Google Scholar profile or an iCite export, so they are counted the same way.

How to work it out by hand

J. E. Hirsch, a physicist, proposed the index in a 2005 paper as “the number of papers with citation number ≥h”. Four steps find it.

1. 1

   List each paper’s citation count.

2. 2

   Sort the papers from most cited to least cited.

3. 3

   Number them down the list: 1, 2, 3 and so on.

4. 4

   Find the last paper whose citations are at least its number. That number is the h-index.

Take Ada Lane, a made-up researcher with eight papers, cited 41, 22, 15, 9, 7, 6, 3 and 0 times.

The sixth paper has 6 citations, and the seventh has 3. Ada Lane’s h-index is 6.

The i10-index

Google Scholar introduced the i10-index in 2011 as “the number of articles with at least ten citations”. Its profiles show it beside the h-index and total citations, in two versions, All and Recent. Ada Lane’s i10-index is 3: her papers cited 41, 22 and 15 times. The cutoff is 10 citations in every field and at every career stage, so the i10-index adjusts for neither.

What counts as a good h-index

No single number counts as good everywhere. Three things move it.

• Field: Hirsch expected typical values to differ by field, set in part by how many references a paper holds, how many papers each scientist writes and how many scientists work in the field. He expected fields with large collaborations to show larger values.
• Career length: Hirsch expected h to “increase approximately linearly with time”. To compare scientists of different seniority, he used m, the h-index divided by the years since the first paper.
• The database: each one counts citations from its own sources. Google Scholar’s values differ from OpenAlex’s, and NIH’s iCite leaves out citations from books and conference posters.

Hirsch gave these figures in 2005, from research that “concentrated on physicists”. He offered the tenure and full professor values “with large error bars”.

ScienceDex’s public directory covers “every medical school, department, and investigator funded by the NIH”. Its h-index page counts “ScienceDex researchers with papers in OpenAlex, an open index of published research, as of September 23, 2026”: 50,136 of them.

These figures mix every field and career stage, and they count OpenAlex’s citations. An h-index from iCite, as Impact shows it, can run lower, since iCite leaves out some citations. Compare an h-index with people in the same field, at the same career stage, counted in the same database.

How ScienceDex shows your h-index

Impact works out the h-index from the papers on your own CV, with citation counts from iCite, NIH’s public citation database. It reads the PubMed IDs in your CV, and one click gets iCite’s counts. ScienceDex sends iCite your papers’ PubMed IDs and nothing else.

Its Citations group shows four figures, each with a sentence on what it means.

• Citations: the times other papers cited yours, as iCite counts them.
• h-index: in words such as “18 of these papers have at least 18 citations each.”
• h-index per career year: the h-index divided by your career years, counted from the first paper in your CV, much as Hirsch’s m is.
• Citations per career year: your citations divided by the same years.

iCite finds papers by their PubMed ID, so the figures cover the papers in your CV that have one. Once your CV holds iCite’s counts, the faculty dashboard shows your h-index too, and the Impact report prints the figures for a promotion packet. Faculty privacy sets out who can open your CV.