Abstract:
This paper presents two new graphical devices to describe the lower and the upper tail of a citation distribution, as well as a novel methodology to compare the research performance of two sets of scientists. The key to these contributions is the identification of a citation distribution in any scientific field with an income distribution. Then the approach to poverty measurement developed by economists since Sen (1976) serves to evaluate the lower tail or low impact sector of the citation distribution. The upper tail, or the high impact sector of that distribution is evaluated by means of, say, an affluence measurement approach that is seen to be the reverse of the poverty one. The paper illustrates this methodology comparing the performance of U.S. and European researchers in the nineteen natural sciences, the two social sciences and the Arts and Humanities category distinguished by Thomson Scientific. The critical value of the citation distribution, or the number of citations that separates the low from the high impact articles, is taken to be the one corresponding to the 70th percentile of the world citation distribution. For all values below or equal to the critical one, in 19 fields the low impact of the citation distribution is found to be larger in Europe than in the U.S. according to any low impact indicator in a large class of admissible measures. For all values above the critical one, in 20 fields the high impact of the citation distribution is found to be larger in the U.S. than in Europe according to any high impact indicator in a similarly large admissible class.