1.     The Research Services Group surveyed all papers in the journals indexed by ISI-Thomson Scientific from 1981 to 1998.

2.     The papers were divided into 22 broad fields, based on the journals in which the papers appeared and ISI's system of assigning journals to specific fields. The fields were broadly defined, such as chemistry, engineering, neuroscience, clinical medicine, materials science, astronomy/astrophysics, agriculture, immunology, molecular biology and genetics, microbiology, etc. Note that papers published in the journals Science, Nature, and Proceedings of the National Academy of Sciences of the USA—all multidisciplinary journals—were assigned to one of the specific 22 broad fields.

3.     The 200 most-cited papers of each year (1981 to 1998) in each of the 22 broad fields were identified. The citation counts were from year of publication through 1999. Thus, for each field, 3,600 top-cited papers were selected. For example, the 200 most-cited engineering papers of 1981 based on their citation counts from 1981-1999, plus the 200 most-cited engineering papers of 1982 based on their citation counts through 1999, plus the 200 most-cited engineering papers of 1983 based on their citation counts through 1999...and so on to the 200 most-cited engineering papers of 1998 based on their citation counts through 1999. In this way, ISI selected 3,600 top-cited, or high-impact, papers for each of the 22 fields (18 years X 200 papers = 3,600 papers). The 22 fields times 3,600 papers = 79,200 papers. When duplicates were eliminated (because some papers were assigned to more than one broad field), the final count of high-impact papers was 76,998. By selecting the most-cited papers of each year and by each field, the Research Services Group was able to control for the effect of age and for field-specific citation averages: older papers have more time to collect citations than younger papers, and different fields exhibit different average rates of citations (sometimes on the order of five to one, such as molecular biology compared to mathematics).

4.     From this database of combined high-impact papers, any paper with at least one Japan author address was extracted. These amounted to 2,992 papers, or 3.9% of the total of 76,998 high-impact papers.

5.     The author names on the 2,992 papers were then summarized and ranked, and any name appearing on more than 12 high-impact papers during the period 1981-98 was identified. ISI found that there were 30 Japanese scientists with more than 12 high-impact papers each. Papers were attributed to the author's credit no matter what position the author name had on the paper. Since Japanese surnames are often common and only one initial is typically used, the Research Services Group reviewed each paper for each Japanese author name to ensure all papers counted were by a specific individual and not another person with the same surname and initial (such false matches were eliminated from our counts).

6.     By this procedure—covering nearly two decades and weighted by the age of papers and by the fields of the papers—ISI was able to recognize individual Japanese researchers whose work has been consistently recognized at a very high level by peer researchers worldwide. Through this systematic, analytical procedure, ISI was able to identify Japan's Citation Laureates, 1981-98.

7.     For further information or details of this analysis, please contact David A. Pendlebury, Manager of Contract Research, ISI-Thomson Scientific, 3501 Market Street, Philadelphia, PA USA 19104. Tel 215-386-0100, ext 1411; fax 215-387-1266.

Back to previous page | Return to Article: Japan's Citation Laureates, 1981-98