Thursday, January 20, 2011

A New Index from Jorge Hirsch: h-bar

Jorge Hirsch is a physicist at the University of California, San Diego who is mostly famous (outside physics at least) for inventing the h-index. A scholar's h-index is the number of his or her academic papers that have received at least h citations. It has become popular in large I think because it is easy to compute using Google Scholar or the ISI citation index. The main obstacles to computation are:

1. Distinguishing between different authors with the same name (a real challenge for Chinese and Korean authors especially as ISI only index papers using the initials of first names - there are a lot of S Liu's (my wife's name) out there for example).

2. Tracking down all the citations to a paper which were incorrectly referenced. A big problem. But to compute the h-index you only need to do this for the papers that have a a citation number close to the author's h-index.

In a paper in Scientometrics Hirsch proposes a new index that is supposed to address the issue of coauthorship. This index is called h-bar and is defined as:

"the number of papers of an individual that have citation count larger than or equal to the h-bar of all coauthors of each paper"

Getting your head around what this means is much harder and in order to compute it you need to get the data for computing the h-index for all of an author's coauthors. So, I don't think this will be as popular as the h-index until the two problems I note above are solved satisfactorily.

An example will help explain what h-bar means. Here are some of Shuang's more cited papers. The top four of these papers would contribute to her Google Scholar h-index. But none of these papers contribute to her h-bar index because they all have fewer citations than Costanza's h-index, which is somewhere around 70. Hirsch notes that the h-bar index for junior researchers tends to be much lower than their h-index while for senior stars they tend to be much closer to each other. He also suggests averaging the two to get a number that is fairer to junior researchers.

My paper coauthored with Costanza has received fewer citations than my h-index. So coauthoring with Costanza doesn't reduce my h-bar-index. I think my h-index and h-bar index are quite close. A couple of papers coauthored with Charles Perrings and Robert Kaufmann might drop out of my "h-bar core" because they got enough citations to contribute to my h-index but not to my co-authors' h-indices.