A review of ranking problems in scientometrics and informetric

Abstract

The paper highlights problems with the ranking of average scores of scientific groups. These can be considered in a wide sense - e.g. scientific disciplines formed by journals or authors and where one studies e.g publication or citation scores. Other scores as aging or prices (for journals) are also possible as well as other groups (e.g. a country, studied w.r.t. a scientific discipline). It is so that, depending on the used scoring device, the relative importance of such groups can increase or decrease. It is even so that rankings can increase (i.e. the group becoming less important) while their relative score increases. The use of geometric averages, to overcome this problem, is explained and it is strongly advised to use it.