Nature 396, 210 (19 November 1998) | doi:10.1038/24249

Region-based citation bias in science

Gianmarco Paris1,3, Giulio De Leo1, Paolo Menozzi2 & Marino Gatto1

  1. Dipartimento di Elettronica e Informazione, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy
  2. Dipartimento di Scienze Ambientali, Universit|[eacute]| degli Studi, Viale delle Scienze, 43100 Parma, Italy
  3. Fondazione Lombardia per l'Ambiente, Foro Bonaparte 12, 20121 MilanoM, Italy


R. M. May1 ranked countries according to their share of articles and the share of citations the articles received. As the ratio of the two indicators is less than unity in France, Germany, Italy and Japan, he claimed a lower-than-average quality of their publications. This might be the case if most of their papers were published in journals with low impact factor (IF); instead, however, we think that most papers receive less citations than they deserve even if they appear in journals with good IF.

To this end we have considered 206 international journals in 14 environment-related categories of the Institute for Scientific Information database and have analysed the citations received by Italian scientists. For each category, we have computed the Italian papers' IF as the weighted sum of the journal IFs according to the number of Italian papers in each journal. In 12 cases out of the 14 (94% of the journals), the Italian papers' IF is either equal to, or significantly higher than, the mean IF of the journals in the respective category. This shows that Italian scientists tend to publish in high-quality journals.

We then compared the number of citations received by each Italian article with the expected citation rate (XCR), that is, the average citation per paper based on the journal title, year of publication and type of document. As suggested by Barreto2 and May3 himself, if there were no bias, the number of citations per article would not be significantly higher or lower than the XCR. However, this is not so (see Table 1).

A further clue to bias comes from analysing the scientific productivity at Ispra (Italy), the Joint Research Centre of the European Commission comprising mostly non-Italian environmental scientists. Although the IF is not higher than average, the number of citations significantly exceeds the XCR (Wilcoxon matched-pairs test, P <0.05).

It is difficult to pinpoint the causes of the bias highlighted by our analysis. There are many motivations behind the citation process |[mdash]| for example, acknowledging pioneers, giving credit to related work and providing background reading. We believe that, while the process of accepting a paper on a journal is reasonably objective, with several referees commenting on the quality and merits of the paper regardless of its country of origin, that of citing a paper is more subjective and certainly more open to considerations of career and funding opportunities. As English-speaking countries produce 49.2% of the world's papers1, they have a dominant position in the scientific world. Not citing their scientists' work would produce negative feedback, while the converse would not. In other words, awareness of the citation game can influence the choice of citations.



  1. May, R. M. Science 275, 793 (1997). | Article | ISI | ChemPort |
  2. Barreto, G. R. Science 276, 882–883 (1997).
  3. May, R. M. Science 276, 885 (1997).