the presentation of conference pa- 

 pers, which nearly quadrupleci over 

 the period; and in the output of tech- 

 nical reports, which tripled (chart 29). 

 Although there was a nearly two- 

 fold increase in U.S. conference pa- 

 pers, the U.S. share of this literature 

 declined from 45 percent to 41 per- 

 cent, and there was an absolute as 

 well as relative decline in the num- 

 ber of U.S. -authored technical re- 

 ports. 



According to the Science Citation 

 Index (SCI), Japanese researchers 

 produced 5.1 percent of the world's 

 S/T journal articles in 1973 and 7.3 

 percent in 1982; U.S. shares were 39.2 

 percent and 37.2 percent, respec- 

 tively (table 5). By 1984, the Japa- 

 nese share had increased to 7.6 

 percent, compared to 36.8 percent 

 for the United States.^^ The MSK se- 

 ries (see footnote 36) shows a larger 

 Japanese share — 10.3 percent in 1973 

 and 13.1 percent in 1982 — but this 

 difference may be partially ac- 



Table 5. Share of world's science and technology articles written by Japanese 

 and U.S. researchers, by field: 1972-73 and 1982 



[Percent] 



Field 



Agriculture 



Biology 



Biomedicine 



Chemistry 



Clinical medicine 

 Earth and space . 



Engineering 



Mathematics 



Physics 



All fieldss^ 



SCI' 



Japan 



1973 



NA 

 5.3 

 4.0 

 9.4 

 3.5 

 2.0 

 5.4 

 3.9 

 6.5 

 5.1 



1982 



NA 

 6.6 



6.5 

 11.6 

 5.6 

 2.1 

 7.9 

 6.0 

 9.0 

 7.3 



United States 



1973 



NA 

 46.4 

 39.2 

 23.3 

 42.8 

 46.7 

 41.8 

 47.9 

 32.7 

 39.2 



1982 



NA 

 44.0 

 41.1 

 21.9 

 42.1 

 43.1 

 41.5 

 38.7 

 29.6 

 37.2 



MSK 



Japan 



1972 



11.5 



10.4 



10.7 



8.3 



NA 



5.0 



13.4 



7.9 



11.3 



10.3 



1982 



11.0 

 10.4 

 12.1 

 17.9 

 NA 

 4.3 

 13.4 

 8.0 

 13.4 

 13.1 



United States 



1972 



31.3 

 36.4 

 32.2 

 38.3 

 NA 

 43.0 

 36.4 

 37.6 

 32.4 

 35.7 



1982 



33.6 

 36.4 

 34.5 

 31.6 

 NA 

 45.1 

 36.4 

 37.4 

 24.1 

 33.8 



^Based on articles, notes, and reviews in a (ixed set ot about 2.300 lournals covered by the Science Citation Index since 1973 



^Includes psychology 



NOTE NA Not separately available 



SOURCES SCI data Irorr Computer Horizons. Inc , unpublished findings: MSK from Mitsubishi Sogo Kenkyusho, Kagaku Gijilsu Mo no 



Kokusai-teh Ryutsu Arikala ni Kan-suru (Tokyo, Japan MSK. March 1984), p 12 



"These SCI figures are based on a constant 1973 

 journal set which includes psychology. 



counted for by the more select range 

 of journals covered by SCI, which 

 has fewer journals in the fields of 

 engineering and agriculture. 



Although the MSK data show 

 considerably higher Japanese and 

 slightly lower U.S. shares of the sci- 

 entific literature, both data sets gen- 



erally reflect similar trends and 

 distributions among fields. Both 

 sources indicate that Japanese re- 

 searchers' shares of publications are 

 largest in the fields of chemistry, en- 

 gineering, and physics; and lowest 

 in the earth and space sciences. Both 

 also indicate that the U.S. share of 

 the world's scientific literature 

 slightly declined from 1972 to 1982 

 by 2 percentage points. 



According to the SCI data, the 

 Japanese share of S/T articles during 

 the 1973-82 period was greatest in 

 the pharmacy subfield, with Japa- 

 nese publications accounting for 25 

 percent of the articles written in that 

 subfield (table 6). The next largest 

 shares were in polymers (18 per- 

 cent) and marine biology and hy- 

 drobiology (14 percent). 



Journal citation rates are one mea- 

 sure of the quality of articles being 

 published. On this indicator, Japa- 

 nese authors have done slightly less 

 well than expected overall, but have 

 improved significantly over the 1973- 

 83 period and performed strongly in 

 the engineering subfield. Using share 

 of total citations to share of total pa- 

 pers as a relative citation indicator, 

 a value of 1.0 reflects expected per- 

 formance: articles are being cited at 

 the same rates at which they are 



32 



