摘要：InmanyEUcountries thereisarequirement to count research, i.e., to measureand proveits value. These numbers, often produced automatically based on theimpact ofjournals,are used to rank universities, to determinefund distribution, to evaluateresearch proposals,and to determinethe scientific merit ofeach researcher. Whilethereal value ofresearchmay be difficult to measure, oneavoids this problemby counting papersand citations inwell-known journals. That is, the measured impact ofa paper (and thescientificcontribution) is defined to beequalto theimpact ofthe journalthat publishes it. Thejournalimpact (and its scientific value) is then based on thereferences to papers in this journal. This ignores thefact that there may be huge differences between papers in thesamejournal; that therearesignificant discrepancies between impact values of different scientificareas; that research results may be offered outsidethejournals;and thatcitations may not bea good index for value. Sinceresearch isa collaborativeactivity, itmay also be difficult to measurethecontributions ofeach individualscientist. However, thereal danger is not that the contributions may becounted wrongly, but that the measuring systems willalso haveastrong influence on the waywe performresearch.
