Citation Analysis Suggests Multinational Coauthorship Has Leveled Off

Coauthorship Has Leveled Off Date: March 4, 1996 (The Scientist, Vol:10, #5, pg.13 & 16, March 4, 1996) (Copyright ©, The Scientist, Inc.) Editor's Note: In its January/February 1996 issue, the newsletter Science Watch -- published by the Philadelphia-based Institute for Scientific Information (ISI) -- took a broad look at international coauthorship of scientific papers (Jan./Feb. 7:1-2, 1996). Examining articles cataloged in ISI databases, analysts tracked the percentage of papers written with collaborators from other nations. They did this analysis by tabulating the percentage of papers in all disciplines of science, engineering, and the arts and humanities published between 1981 and 1994 whose primary author collaborated with at least one researcher from another nation. Because the ISI databases list only the one institution that authors give as their contact address, the analysis could not determine where the researchers did their work-at their home institution or at the institution of their international collaborators. Following is Science Watch's report, reprinted here with permission of the newsletter and ISI. For more information, contact Christopher King, editor of Science Watch, ISI, 3501 Market St., Philadelphia, Pa. 19104; (800) 523-1850, Ext. 1341. Fax: (215) 387-1266. E-mail: cking@isinet.com It is commonplace that research is an increasingly global enterprise. But to what extent is this really the case? What are the actual figures behind the truism? To find out, Science Watch surveyed journal literature published in science, math, engineering, and the humanities over the last 14 years to determine rates of multinational coauthorship. The results provide detailed evidence of a steady increase in the trend toward international collaboration in research. For the world as a whole, for the industrialized nations known as the Group of Seven (G7)-Canada, England, France, Federal Republic of Germany/Germany (Science Watch tracked the Federal Republic of Germany through 1991 and the combined Germany thereafter), Italy, Japan, and the United States-and for seven other representative countries, the percentage of papers written with collaborators from other nations increased sharply from 1981 to 1994. In fact, in many cases the rate of multinational coauthorship more than doubled during the 14-year span. What is less clear, however, is where the trend might be heading. Levels of multinational coauthorship seem to have flattened out during the 1990s, at least for the 14 countries examined here. The top figure at right tracks the G7 nations for each year from 1981 to 1994, indicating the percentage of papers from each country written in collaboration with authors from at least one other country. Also charted are the year-by-year figures for the world, indicating the percentage of the world output of papers on which at least two nations shared authorship. The center chart illustrates rates of international coauthorship for seven other countries. To obtain these figures, Science Watch examined more than 10 million journal articles, notes, and proceedings published between 1981 and 1994. Science Watch broadened its scope beyond science for this survey, assessing all fields-including papers published in the social sciences as well as the arts and humanities, although these together represent only 16 percent of the data. For each country, Science Watch first evaluated the output of papers in each year. For the U.S. in 1981, for example, the total number of published papers in ISI-indexed journals was 285,598. Of those, 13,866 papers listed at least one non-U.S. address (as opposed to the 271,732 papers with all-U.S. addresses). Thus, 4.86 percent of U.S. research papers in 1981 were written in collaboration with an author from at least one other country. The same calculation was performed to obtain the 1981-94 figures for each country and for the world as a whole. Among the G7 nations, France posted the highest level of international collaboration at the end of the span, with 26.65 percent of its papers showing at least one non-French address in 1994. Just behind France was Italy, with 26.58 percent, and Germany with 25.52 percent. The U.S. had the lowest level among the G7 nations in 1994, with 11.68 percent of its papers listing international addresses. This was just above the average mark for the world of 10.38 percent. For all of the G7 nations, as the top chart indicates, the percentages of papers with international coauthorship climbed steadily through the 1980s. However, the chart also shows that in recent years the numbers appear to have leveled off. Indeed, when one compares the percentage of internationally coauthored papers for 1992 against the figure for 1994, some of the countries have actually decreased their levels of international collaboration. In 1992, for example, researchers in France published 27.08 percent of the country's output with collaborators outside France's borders. In 1994, however, after 13 years of increase, the percentage fell to 26.65. In Italy, similarly, the 1994 rate was 26.58 percent, down from 27.27 percent in 1992. And the U.S. showed a small decline as well at the very end of the period, with its 11.68 percent mark in 1994 down from 12.03 percent in 1993. Similar trends, for the most part, are evident on the center figure, in which Science Watch tracks the rates of international coauthorship in a selection of seven additional countries. Switzerland, clearly, exhibited the highest rates of collaboration with foreign authors in 1994; Swiss researchers shared international authorship on more than 40 percent of the nation's papers that year. Poland was next highest, with 35.15 percent. Sweden rose steadily through the 1980s in its rate of foreign collaboration, while Mexico's percentage fluctuated over the same span, but both nations arrived at a similar point in 1994, with just over 33 percent of their respective papers bearing at least one foreign address. Australia and the People's Republic of China vied for the lowest international collaboration in 1994 among the nations shown in this chart -- around 22 percent. Most of the countries appear to have reached a plateau by 1994. Even Switzerland shows a small decline in international collaboration, with its 40.12 percent in 1994 down from 40.81 percent in 1993. Although Sweden's rate did not decrease, by the end of the span it was fairly stagnant, rising from 33.19 percent in 1992 to only 33.27 percent in 1994. For the figure at the bottom of page 13, Science Watch looked at each country and obtained a single average rate of international coauthorship over the entire 14-year span. This number was plotted against the country's citation impact (the number of citations per paper) in all fields over the same period. For example, the impact score for the U.S., in all fields for the period 1981-94, is 11.42 citations per paper. The percentage of U.S. papers that bear at least one non-U.S. address for 1981-94 is 8.16. These points intersect in the upper left of the graph, where the U.S. stands alone. Switzerland posted a comparable mark in citation impact for the 14-year span, but its higher rate of international coauthorship places the country in the graph's upper right. Poland, Mexico, South Korea, and the People's Republic of China, with their comparatively high percentage of papers cowritten with other countries and their comparatively low citations-per-paper averages, fall to the lower right. Grouped more or less in the middle are the remaining G7 nations.

Citation Analysis Suggests Multinational Coauthorship Has Leveled Off

The Scientist Staff

Interested in reading more?

Become a Member of