Sunday, January 24, 2010

How I use the h-index

I am a big fan of the h-index, when used appropriately.
I review numerous research grant applications and job applications and I find the h-index is a good filter to consider how seriously to take an application.

In comparing theory to experiment any discrepancy by an order of magnitude is usually a pretty good indicator that a theory is in trouble. Factors of 2 to 5 can also be useful.

What I find interesting and useful about the h-index is that there are often differences by factors of as much as 2 to 3 between individuals at the same career stage and applying for the same job or grant. I find it surprising that there are such large variations. Furthermore, there appears to be no correlation between the h-index and how much noise an individual makes about the significance of their work.

Some things I do keep in mind are:
-the younger the person the less reliable and useful the h-index
-differences of less than fifty per cent (e.g., 9 vs. 12 or 20 vs. 24) are of no significance
-different sub-fields have different traditions of citing each others works

One thing I never do is to look up an individual's h-index just out of curiousity. I only look it up if I have to actually evaluate the person.

1 comment:

  1. It really has to be compared only for people in the same field. for instance, if someone belong to a large experimental community, he can have a big h index without actually being one of the main authors of the papers. However, if you write a paper in nuclear theory, without comparing to experiment, your chance of being cited are small just because the community in your specific topic is small.