H-index is the largest number h such that h publications have at least h citations.
For example, an author with 10 papers, where 5 papers were cited 5 times each one, h-index is 5; if 6 papers were cited 6 times each one, h-index is 6,...
If you are interested in indexes and metrics to measure scientific production, you might want to have a look at the "Publish or Perish" website: http://www.harzing.com/pop.htm
It describes a number of methods besides the h-index to assess one's scientific impact and there is a software you can download to calculate all these things using the citation database of google scholar.
Sorry, I could't understand you. If some one has 10 publications, 5 of these publications are cited. Three of these five are cited only one time, the fourth one is cited 8 times, the fifth one is cited 4 times. What is the h-index?
To answer Peshawa Muhammad Ali's question: in calculation of h-index only the cited works play a role. The papers should be sorted by the number of citations, in decreasing order. Paper A:8 times, Paper B:4 times, and Papers C,D and E only once. Since there are only two papers that were cited more the once, the h-index of this researcher is h=2. Only if Paper C (the third paper in the sorted list) gets 2 more citations (must be the same paper) then the h-index will increase to 3, regardless how many new citations Papers A and B get. If Papers C, D, or E get only one citation more (i.e. they would have 2 each), the h-index remains still 2.
The reasoning behind this calculation is that the new papers that the researcher publishes should get always the same good reputation as the previous ones. This would mean a constantly high quality publication activity. The h-index is insensitive to the "one-great-paper" effect, if someone writes (or a co-author of) a one single excellent paper.
However, some criticism can be formulated against the h-index, that older researchers have better chance to achieve higher h-indices. The a-index, a derivative of the h-index (which is the h/y where y is the number of years spent in publishing by the author) corrects (mostly) also for the age of the researcher. E.g., a=1 for a researcher who started to publish more than 5 years ago, it is a very good achievement.
Thank you Marius, Balazs and Hooshyar. In fact there is a kind of unfair in this calculation. I know some one has a paper cited 130 times, but he has no quality paper after this one. unfair.
I don't think h-index would be unfair. This is just an index that measures something else. Maybe this is an off-topic here (probably it belongs rather to Stuart Hardy's recent discussion), anyhow, I try to present an example to understand the potential usage of the h-index.
Assume you are the director of a research institute and you want to hire some new researchers. You have all their data (cv, list of publications, citations, indices) and you can invite only 2 of the 4 candidates for an interview. Assume that all candidates have altogether 150 citations and all of them have published 10 papers (well, it is idealized).
Whom would you rather invite:
Researcher A, co-author of paper AA (8 years ago, cited 141 times), paper AB (recent, 3), paper AC (recent, 2), the rest single authored (various years; they are cited max. once per paper)
Researcher B, (co-)author of papers BA, BB, BC, (published 4,5,6 years ago, respectively; each cited 30 times), papers BD, BE, BF ((sub)recent, 15 times each), papers BG (6 times),BH (5 times) , BI (4 times), BJ (new, not cited)
Researcher C, C's (co-)authored papers are paper CA (6 years ago, 30 times), 8 different papers (2-5 years; cited more than 10 times each) and another new paper that hasn't been cited.
Researcher D, D's papers are single authored, paper DA (10 years old, 70 citations), paper DB (8 years old, 55 citations), paper DC (7 years old, 10 citations), 4 papers (3-4 years old, 3 citations) and the rest (0-3 years) are with one or zero citations.
The h-indices:
Researcher A: h=2
Researcher B: h=6
Researcher C: h=9
Researcher D: h=3
I would consider Researcher B and C because it seems that they can publish papers that are of continuous interest for the community and the expectation is that they can work in team and publish new papers that will be cited.
Researcher A might have had a successful project 10 years ago, but maybe less productive since then; Researcher D likes to publish alone and the interest of the community for D's work seems to decrease with time. (Of course, this is an idealized example.)
As a director I would hire a real team-player who can write interesting papers in the future attracting attention in the future as well, and B and C seem to be closer to that expectation.
I am sorry for the long example, but I hope this explains why h-index could be used effectively in some cases.
Calculating h-index using ResearcherID will take into account the citations indexed by Thomson Reuters (ISI) only. This is typically just a subset of citations, it will not include those e.g. indexed by Scopus (that has a better coverage in my fields of interest). In my case the differences are
h(Scopus)-h(ISI)=2 and
h(my own list)-h(Scopus)=2
A further restriction of the method is its incompleteness concerning conference proceedings and books. These are by far less covered by the aforementioned services. In many engineering sciences peer reviewed conference proceedings are common sources of original research papers (e.g., ISPRS Archives), because their turnaround time is often less than that of the journals. There are a lot of frequently cited seminal papers that appeared in this form; the citations received to these papers will be partly included only.
Apart from ISI and Scopus, there is Google Scholar citations and Google Scholar h index. If you have a gmail id, you can login to google scholar and see the citations and h index of you, your colleagues, your peers and other researchers....
You are surely right. Take computer science for example, quite a lot of papers are published as conference proceedings, much of which are of good quality. I agree with @Lijo Francis. From my experience, the h index caculated by Google Scholar is higher than that by ISI, frequently > 2. A scholar's h index should reflect all the citations of journals, books and conference proceedings. But it's not easy to do the statistical analysis. An international standard should be created to make h index compareable among researchers.
Let me recommend you the discussion of Stuart Hardy's question "Any opinions out there on the various citation indexes, metrics etc - fairness etc.? ", especially the recent post of Craig Jones. A nice contribution on some thoughts tackled above. Kind regards, Balázs
The examples of Balázs are interesting also his consequences on the evaluation of these imaginary researchers. However, h-index does not inform on the contribution of an author to articles co-author of which he/she was. Recently, I have seen the publications of a scientist who has had about 600 articles. In most cases he was a co-author among a lot of other co-authors. His name was difficult to find in his articles because over 98% of these articles the number of co-authors was several hundred. In one case I counted the number of authors which was 1636 exactly. It is hard to assess the amount of his/her contribution to this article of 1636 authors.
The example you presented is of course valid, however, consider that there are research branches where there is hardly any individual contribution possible. Just a few example: research in CERN, SRTM mission. Mars Express (or space missions in general), several branches of astronomy (e.g. radio astronomy), certain widespread medical studies, large scale environmental monitoring, crustal-depth seismic sections, ocean drilling projects, etc. In these areas even a few-people-authored paper is difficult, because there are many people who really contribute to the success of the project with technical improvements, special expertise or providing large-scale geographical coverage (e.g. earthquake monitoring).
The person you cited can be e.g., an engineer at CERN dealing with developing and maintaining the magnetic focusing of the beam, or a developer who contributed to Mars Express HRSC data processing (calculation of DTMs from stereophotogrammetric images), or measuring burial ages of rock samples in a thermochronological lab. (The last two examples represent two of my existing colleagues.)
These people are valuable contributors to other projects, without them the achieved result would not be possible. There are institutions who look for such people. (In the latter case, the person is leading a lab today, previously he worked at another similar lab as a scientist, developing the measurement instruments, publishing a number of papers as co-author.)
On the other hand, you are right (I also tried to emphasize that above) that the CONTENT of the contributions is also important, and the role what the contributing person played in the project. Fortunately (and hopefully) not robots will evaluate the candidates based on simple numeric values, even if these values are useful in evaluation.
Thank you for your comment and kind regards, Balázs
Please read the previous contributions, some of them explain the method in details. (You may also want to read the original paper of Hirsch, a well-written paper.) If this does not help, please do not hesitate to contact me directly, and specify your problem.
The index is based on a list of publications ranked in descending order by the number of citations these publications received. The value of h is equal to the number of papers (N) in the list that have N or more citations.
But as per example quoted by him, an author has 8 papers that have been cited 33, 30, 20, 15, 7, 6, 5 and 4 times. This tells us that the author's h-index is 6.
What does an h-index of 6 mean?
An h-index of 6 means that this author has published at least 6 papers that have each received at least 6 citations.
But that is not fixed h-index 6? in many cases it is taken upto 4 (times citation) and also more then 6 (times citation) then only included in h-index.
h-index also differ from Google and Research gate also..
The index is based on a list of publications ranked in descending order by the number of citations these publications received. The value of h is equal to the number of papers (N) in the list that have N or more citations. Before you can calculate your h-index, you will need a complete publication list.
Number the articles in descending order of citations, from most cited to least cited. The h-index is equal to the number in which the article rank is equal to its number of citations.