What is h-index and how to calculate it

The h-index is one number that looks at two things together: how many papers you have published, and how often they have been cited. The physicist Jorge Hirsch proposed it in 2005, in a paper in PNAS. He defined it as "the number of papers with citation number higher or equal to h".

In plain words: your h-index is h if h of your papers have each been cited at least h times. An h-index of 6 means you have six papers with at least six citations each, but not seven papers with at least seven.

h-indexnoun

the largest number h such that h of a researcher's papers have each received at least h citations.

Hirsch wanted a number that rewards steady work. A long list of papers nobody cites does not raise it much. Neither does one famous paper on its own. To reach an h-index of 10, you need ten papers that other researchers have each cited at least ten times.

Dr. Pradhan took a sheet of paper and asked Tenzin to read out the "cited by" number for each of her eight papers from Google Scholar. Then he showed her the method. You can do it with any list of citation counts.

1. List your papers.Write down every paper with the number of times it has been cited.
2. Sort them.Put the most cited paper first and the least cited last.
3. Number them.Give each paper a rank: 1, 2, 3 and so on, down the list.
4. Find the crossing point.Go down the list while the citations are still equal to or greater than the rank. The last rank where that is true is your h-index.

Paper six has six citations, so the condition holds at rank 6. Paper seven has three citations, which is less than 7, so it stops there. Tenzin's h-index in Google Scholar is 6.

Notice what does not matter. If her top paper had 240 citations instead of 24, her h-index would still be 6. If she published a ninth paper tomorrow with no citations, it would still be 6. It rises only when another paper crosses the line: here, when paper seven reaches seven citations and paper six still has at least seven.

Because the h-index needs both many papers and many citations, it smooths out the extremes. That is its strength, and also why it hides things.

• A single famous paper counts once. A researcher with one paper cited 500 times has an h-index of 1.
• Many uncited papers count for nothing. Forty papers with two citations each give an h-index of 2.
• It rarely goes down. Citations only add up, so the number usually stays the same or rises, even after you stop publishing.
• It grows with time. Hirsch himself observed that h should increase roughly in step with the years of a career, so a senior professor will almost always have a higher h-index than a new PhD scholar.
• It ignores your role. First author, last author or tenth author: every paper you are on counts the same.

For an early researcher like Tenzin, the last two points matter most. Her h-index of 6 after five years says very little next to a professor's h-index of 30 after thirty years. Hirsch suggested dividing the h-index by the number of years since a researcher's first paper, which he called m, as a fairer way to compare people at different stages.

You rarely need to calculate the number by hand. Three databases work it out for you, each from the papers and citations it has indexed.

Google Scholar's help page says it computes "two versions, All and Recent, of three metrics": the h-index, the i10-index and the total number of citations. The i10-index is simply the number of your papers with at least ten citations. Tenzin's i10-index was 2.

If your profile is meant to be found, make it public and add a verified email address at your university. Google Scholar says both are needed for a profile to appear in its search results.

Tenzin's 6 and 4 were both correct. Each database is likely to produce a different h-index for the same person, because each one covers different publications and counts different citations.

The University of Texas at Dallas library guide explains that Google Scholar's h-index "tends to be higher" than the ones from Scopus and Web of Science, because Google Scholar also includes other publications such as books, while the other two count citations in journals and conference proceedings. It adds that Google Scholar has no filter to remove self-citations, and is less good at removing duplicate entries.

Dr. Pradhan found Tenzin's gap quickly. Two of her citations came from a book chapter and a master's thesis that Scopus does not cover, and one of her papers appeared twice in Google Scholar. She merged the duplicate, which Google Scholar's help page explains how to do: select both versions and click "Merge". The h-index stayed at 6, but now it was honest.

These three numbers are often mixed up on forms and in conversations. They measure different things.

Our post on what is impact factor explains the journal number in detail. The key point is that a high impact factor belongs to the journal, not to you or to your paper. The San Francisco Declaration on Research Assessment, known as DORA, asks that journal-based metrics not be used "to assess an individual scientist's contributions, or in hiring, promotion, or funding decisions."

The h-index at least looks at your own papers and their own citations. That makes it more personal than an impact factor, but it is still only a count.

There is no single good h-index. It depends on your field and on how long you have been publishing. Hirsch noted that typical values differ between fields, and that high h-indices in the life sciences are much higher than in physics. Fields where papers carry long reference lists, or where many people publish often, produce higher numbers. So compare yourself only with people in the same field and at a similar stage, and better still, with your own number from last year.

As for raising it, there are no shortcuts worth taking. The honest ways are the slow ones.

Our guide on how to publish a research paper covers the steps that make good work easy to find and cite.

Tenzin filled in the box with both numbers, the database names and the date. In the section for her research statement, she wrote about what her orchid surveys had shown and who had used them. "The committee will see the number," Dr. Pradhan said, "but they will remember the orchids."