Everyday Scientist

new citation index: the h-bar value

October 9, 2008 at 6:43 pm | sam | literature, science community

I commented a while ago about the Hirsch h index. In fact, it was one of my first posts 2.5 years ago. Since then, the h index has become the standard in reporting the extent of an author citations. ISI and ResearcherID report h index along with total citations, and scientists have started to include their h value on their CVs and resumés.

But there’s still the problem of self-citation: citing ones own papers in later papers or reviews can increase ones own h index.

PI	h
Pande	28	4.5
Moerner	43	6.8
Andersen	48	7.6
Fayer	51	8.1
Boxer	52	8.3
Zare	94	15.0

I now introduce a modification of the Hirsch h index to solve the self-citation problem. The “h-bar index.” The math is very simple: because each researcher cites himself or herself multiple times in each paper, total citations must be divided by a normalizing constant to account for these self-citations. Therefore an author’s  index is his or her h index divided by 2π.

The table above compares the Hirsch h index and the Lord  index of the same scientists I listed in my original post. I argue that the Lord  index is a more accurate measure of an individual’s true citation record.