The classical method for computing the impact of one's research is based on ISI's Science Citation Index. This has two problems. The first is that Science Citation Index does a notoriously bad job of indexing computer science publications. The second is that the computation of an individual's research impact requires sifting through a number of figures, including the total number of citœations, the impact factor of the venue in which the papers have appeared, etc. H-index is a single parameter measure that has been proposed by J. E. Hirsch, who is a physicist at UCSD, as a simpler method of computing one's research impact. The original paper is here. The computation is as follows:
"A scientist has index h if h of his/her Np papers have at least h citations each, and the other (Np − h) papers have no more than h citations each."
H-index is interesting, but, as with all of these measures, it has its problems. The following are what I can think of, but these are generic to any of performance measure, not only h-index:
- It looks at the publications of a researcher over a lifetime. So, there is a danger that it may not properly reflect the impact of those who have only written a few publications that are very important. Hirsh recognizes this problem in the first paragraph of his paper:
- It is as accurate as the database over which it is computed. Google Scholar is not the most accurate database of one's publications (but this is a perennial problem in computer science).
- As other citation-based evaluation indexes, this one also assumes that the quality of the paper is equivalent to it being highly cited. I am not sure I entirely believe that this is true in computer science (I don't know about other disciplines).
- There is no normalization for various factors. If I have an h-index of x after n years of being in this profession, I can tell I am not doing as well as another colleague who has an h-index of x after m years if m<n, and I can compare h-indexes with others at my seniority level. However, I have no sense of how close to the mean or median I am at my seniority level or how close I am to the mean or median within my area of research.
- It does not account for the impact of people who have published few papers which are very influential (this is related to the second point above). The author recognizes this problem, however, and states the following in the original paper: "For the few scientists that earn a Nobel prize, the impact and relevance of their research work is unquestionable. Among the rest of us, how does one quantify the cumulative impact and relevance of an individual’s scientific research output?" Thus, the measure is not intended to measure the impact of these types of people, but "the rest of us."
"For the few scientists who earn a Nobel prize, the impact and relevance of their research is unquestionable. Among the rest of us, how does one quantify the cumulative impact and relevance of an individual’s scientific research output?"
On this issue , here is what Wikipedia, the current ultimate authority on all things, says:
"It is not difficult to come up with situations in which h may provide misleading information about a scientist's output. Most importantly the fact that h is bounded by the total number of publications means that scientists with a short career are at an inherent disadvantage, regardless of the importance of their discoveries. For example, Evariste Galois' h-index is 2, and will remain so forever. Had Albert Einstein died in early 1906, his h index would be stuck at 4 or 5, despite his wide acknowledgement as one of the greatest of physicists. " (Wikipedia lists a number of other criticisms.)
Attempts to modify h-index to address these issues can be found here. A bunch of papers addressing various issues related to h-index can be found here.
A list of computer scientists with an h-index of at least 40 is given here.
Back to Vita