Authors:Giovanni Abramo, Tindaro Cicero and Ciriaco Andrea D'Angelo
An increasing number of nations allocate public funds to research institutions on the basis of rankings obtained from national evaluation exercises. Therefore, in non-competitive higher education systems where top scientists are dispersed among all the universities, rather than concentrated among a few, there is a high risk of penalizing those top scientists who work in lower-performance universities. Using a 5 year bibliometric analysis conducted on all Italian universities active in the hard sciences from 2004 to 2008, this work analyzes the distribution of publications and relevant citations by scientists within the universities, measures the research performance of individual scientists, quantifies the intensity of concentration of top scientists at each university, provides performance rankings for the universities, and indicates the effects of selective funding on the top scientists of low-ranked universities.
Authors:Giovanni Abramo, Ciriaco Andrea D'Angelo and Tindaro Cicero
National research assessment exercises are conducted in different nations over varying periods. The choice of the publication period to be observed has to address often contrasting needs: it has to ensure the reliability of the results issuing from the evaluation, but also reach the achievement of frequent assessments. In this work we attempt to identify which is the most appropriate or optimal publication period to be observed. For this, we analyze the variation of individual researchers’ productivity rankings with the length of the publication period within the period 2003–2008, by the over 30,000 Italian university scientists in the hard sciences. First we analyze the variation in rankings referring to pairs of contiguous and overlapping publication periods, and show that the variations reduce markedly with periods above 3 years. Then we will show the strong randomness of performance rankings over publication periods under 3 years. We conclude that the choice of a 3 year publication period would seem reliable, particularly for physics, chemistry, biology and medicine.