Reader Comments

Post a new comment on this article

A single number?

Posted by hermensf on 28 Dec 2008 at 17:13 GMT

Creating a single number that expresses the performance of a researcher remains difficult.

The existing measures seem to have an important problem: They do not take into account the research field in which the researcher is active.

Fewer researchers work in mathematics than in cancer research. Impact factors and such do not seem to take the size of the community into account. This has the consequence that if one is working in mathematics, it will be much more difficult to achieve a high score as a scientist.

Moreover, different research field value different kinds of publications differently. Researchers in biology and psychology will more often publish in journals, whereas in technology, conference proceedings play and important role. In some research fields, patents are important.

Because of these reasons, I would propose different scoring mechanisms for different research fields.

RE: A single number?

clementkent replied to hermensf on 30 Dec 2008 at 16:58 GMT

I agree with hermensf that different fields may need different weightings. That is why I suggest that a vector of rankings be returned for each researcher. Then a discipline or use-specific function can be applied to that vector. hermensf has said why discipline-specific functions are needed; additionally use-specific functions might differ between members of hiring committees, tenure or promotion committees, and grant review boards.