The h‑index, impact factors, and similar absurdities are increasingly divorced from the original purpose that spawned them. Citations were intended for forward‑referencing, to be able to follow the subsequent development of a topic or field. Just as our publications carry references to past papers to place them in context, citations allow us to see how the work develops into the future.
Originators of science citations may never have meant them to be an evaluation of an author’s worth. Yet it seems that awards and promotions are increasingly based on the h‑index and other concocted indices. No studies have justified such use, and there is no reason to claim that two people with the same h value are somehow of equal worth. Deans and other administrators, in moments of reflection, will concede that an h‑index cannot be used for promotion decisions, but the tempting simplicity of a single number—as with IQ and indicators in the past—is sometimes hard to resist, especially when one is not familiar with the work itself.
The innumeracy is compounded when h is divided by some time unit, such as years ranked as assistant or associate professor or years since PhD. How is that meaningful? We as physicists should know better than to invest meaning simply because we can multiply or divide two numerical quantities. The sole purpose of computing such indices seems to be to make some case for timeliness of promotion. If such a case could be reasonably made, we could do away with any promotion deliberations and just replace them with some automatic gate or threshold.
Even as the amount of literature on- and off-line explodes and purportedly quantitative measures of abilities proliferate, people seem to actually read the literature less and less. Referencing to previous work is often grossly inadequate, especially with authors using electronic search sites and limiting their searches to recent publications. I’ve also heard of gaming the system, with friends agreeing to cite each other’s papers, regardless of relevance, simply to boost their h‑index.
In the light of all this, I propose a new index, the r‑index. A primary meaning of the “r” is as one of the original three r’s, reading. The r‑index is the fraction or ratio of the references cited that the author has actually read in full at least once.
My tone should make it clear that I do not want or expect anyone to start computing the r‑index, or the h‑index, or the latest “E‑index” with further rococo embellishments (PHYSICS TODAY, November 2010, page 12). Physicists spawned the h‑index. It is for physicists now to banish it from any rational discourse.
