How relevant is the predictive power of the h-index? A case study of the time-dependent Hirsch index

The h-index has been shown to have predictive power. Here I report results of an empirical study showing that the increase of the h-index with time often depends for a long time on citations to rather old publications. This inert behavior of the h-index means that it is difficult to use it as a measure for predicting future scientific output. 1. Introduction The Hirsch index or h-index of a researcher has been defined by Hirsch (2005) as the largest number h of publications of a researcher which have obtained at least h citations each. It has become popular and is frequently used as a measure of the scientific achievement of a researcher. But it is controversial whether the h-index is indeed a representative measure) has shown that the h-index has predictive power in the sense, that there is a high correlation between the Hirsch index values after 12 years and after 24 years of the career of researchers. In short, " the h-index is a better predictor of itself than any of the other indicators " investigated in that study and " the h-index has the highest ability to predict future scientific achievement ". Hirsch concludes that the h-index " can be profitably used in academic appointment processes and to allocate research resources ". It is the purpose of the present investigation to raise doubts about this conclusion. Using empirical data, I demonstrate that the increase of the h-index with time after a given point of time (e.g. the time of appointment or the time of allocating resources) is not necessarily related to the scientific achievements after this date. Specifically, I show examples where the growth of the h-index is the same, irrespective of whether the investigated researcher had performed as he or she did or whether he (she) had not published any further work. Of course, the above mentioned consideration that the h-index is used in academic hiring processes and research fund allocation processes is only one aspect of looking at the evolution of the h-index with time. The other aspect is that a substantial increase of the h-index even after the researcher has stopped publishing bears testimony of the significance of this researcher's achievements. Recently, Acuna, Allesina, and Kording (2012) have fitted the time-dependent h-index of neuroscientists using a large number of 18 parameters and were able to predict the future h-index rather accurately for several years. Even …