Bioethics in Asia

pp. 355-360 in Bioethics in Asia

Editors: Norio Fujiki and Darryl R. J. Macer, Ph.D.
Eubios Ethics Institute

Copyright 2000, Eubios Ethics Institute All commercial rights reserved. This publication may be reproduced for limited educational or academic use, however please enquire with the author.

P8. Multiple Authorship in Japan and the West
Michael D. Fetters^.

Department of Family Medicine, University of Michigan Medical Center, USA

Todd S. Elwyn, Fourth-year student,

University of Michigan Medical School; USA Fulbright Graduate Research Fellow, Department of Law, University of Tokyo.

Summary
Objective: To assess and compare the mean number of authors per article in a Japanese and a Western biomedical publication.

Design: The numbers of articles and authors appearing on original contributions, case studies, or communications in Japanese Circulation Journal (JCJ) and Circulation Research (CR)for the years 1983, 1993, and 1996, were tallied and means calculated.

Main outcome measures: Mean number of authors per each article type.

Results: There were 1,590 original research contributions to JCJ by 246 Japanese authors for a mean number of 6.5, and range of 1-13 authors per article. There were 2,022 original research contributions by 461 non-Japanese authors to CR for a mean number of 4.4, and range of 1-16 authors per article. There were 296 authors on 45 case reports in JCJ for a mean of 6.6 authors per case report with a range of 2-12 authors, and 225 authors on 59 communications in CR for a mean of 3.8 authors per communication with a range of 1-11 authors per article.

Discussion: In the years 1983, 1993, and 1996 there were roughly 2-3 more Japanese authors per article in the Japanese Circulation Journal than there were non-Japanese authors per contribution in Circulation Research. Groupism and the hierarchical structuring of Japanese research groups arguably account for legitimately larger number of Japanese authors per contribution than non-Japanese authors per contribution. These data and analysis illustrate that publication of the results of scientific inquiry is in fact the interface between the scientific method and the culture of the contributing investigators.

Introduction
Strict adherence to objectivity and high ethical standards are two defining virtues of science which have led to the emergence of a culture of science and scientific inquiry respected worldwide. Though scientists share a common language and culture, they each are also members of separate countries and cultures with divergent values and norms which may influence the way they conduct scientific research. For example, some U.S. scientists perceive Western European scientists as more theoretical than researchers in the United States, which may be due to a greater importance placed upon theory (1). The "frontier mentality" of U.S. culture arguably influences physicians and patients to value taking action over doing nothing -- an attitude that likely limits the use of placebo arms in clinical trials (1). The emphasis on groupism at the expense of individuality in Japanese culture (2) may make it more difficult for Japanese scientists to produce original research (3-5).

An area in which the effects of cultural differences upon the conduct of science that has not been well studied is in the authorship of journal articles. The criteria for authorship most widely used are the “Uniform Requirements of the International Committee of Medical Journal Editors” issued by the International Committee of Medical Journal Editors (ICMJE) which are currently employed by over 500 international journals (6). While the first uniform requirements were published in 1979, it was not until the third edition, published in 1988, when the criteria for authorship explicitly required:

...substantial contributions to (a) conception and design, or analysis and interpretation of data; and to (b) drafting the article or revising it critically for important intellectual content; and on (c) final approval of the version to be published. Participation solely in the acquisition of funding or collection of data does not justify authorship. General supervision of the research group is also not sufficient for authorship." (7)

These criteria take a strong position against unjustified authorship and stand virtually unchanged in the fourth and fifth editions (6, 8).

Our perception that there were more Japanese authors than non-Japanese authors listed on papers in the literature prompted us to investigate how culture influences the practice of crediting authorship. To assess this question, we conducted research that sought first to identify a Japanese biomedical journal and match it with a similar Western journal, and second, to compare the number of authors per original contribution in these journals.

Methods
One investigator (MF) identified the peer-reviewed Japanese biomedical journals carried in MEDLINE having a consistent focus on basic or clinical science research, published at least monthly for over 10 years, and available in the Health Sciences Library of the University of North Carolina. MF reviewed several current issues of both candidate Japanese journals and Western journals covering similar content areas, focusing on organizational affiliation, editorial board size, manuscript length, abstract length, primary content of original contributions, and instructions for authors. MF selected the following pair of journals since they were qualitatively the most similar in the above-described characteristics: Circulation Research (CR), published in the United States, and The Japanese Circulation Journal (JCJ).

To compare the number of authors per article and to classify articles as originating within a Japanese or non-Japanese research culture, we then examined the authorship of each contribution and found three combinations of names: 1) non-Japanese names only; 2) Japanese names only; and 3) a combination of Japanese and non-Japanese names. We assumed non-Japanese cultural values influenced the first combination of authors, and Japanese values influenced authorship for the second and third combinations, but with two qualifications. First, we excluded JCJ articles with a combination of Japanese and non-Japanese authors if there were two or more non-Japanese authors or if there were one Japanese author and one non-Japanese author. The rationale for this decision was that JCJ publishes manuscripts contributed principally by members of the Japanese Circulation Society. Moreover, we assumed foreigners working in Japan would have similar or greater pressures to include others from the research group in the paper as a Japanese colleague who had the same professional rank. Second, we also excluded from analysis CR articles that had a combination of Japanese and non-Japanese authors since assigning a Japanese or non-Japanese research culture to these groups would be purely speculative. Second and third generation Japanese researchers who have no affiliation with Japanese research groups cannot be readily distinguished from Japanese nationals collaborating internationally.

We examined the 1983, 1993, and 1996 issues of these journals; this time frame corresponds to five years before, and five and eight years after the ICMJE first published specific authorship criteria (7). We compared the number of peer reviewed articles per issue, the types of articles, and the number of authors per article.

Results
Journal Characteristics
Table 1 illustrates that the journal characteristics of CR and JCJ are very similar. The 1996 JCJ instructions for authors references the Uniform Requirements only in describing the form that is to be used in citing references. By contrast, the 1993 JCJ referenced the third, 1988, version, and required all manuscripts to conform to all of the Uniform Requirements. No other criteria for crediting authorship are stated in either journal. The editorial structures were similar. Six members of the large CR editorial board had full Japanese names, but not a single non-Japanese name appears on the smaller JCJ editorial board.

Table 1: Journal Characteristics of Circulation Research and Japanese Circulation Journal

	Circulation Research	Japanese Circulation J.
Peer review	yes	yes
Research area	clinical and experimental cardiology	clinical and experimental cardiology
Original research only	yes	yes
Issues per year	12	12
Organizational affiliation	American Heart Association	Japanese Circulation Soc.
Size of editorial board (members)	153	27
Abstract length (words)	250	200
Key words	3-5	3-5
Manuscript length (words)	<6000	not specified
ICMJE instructions 1983 1993 1996	none second edition fourth edition	none third edition fourth edition

Table 2: Contribution Types in Circulation Research and Japanese Circulation Journal

	1983			1993			1996		3 Year Total
Type of Article	CR	JCJ	CR	JCJ	CR	JCJ	CR	JCJ
Original Research	135	71	227	108	227	76	589	255
Original articles	135	71	--	--	--	--	135	71
Original contribution	--	--	212	--	222	--	434	--
Expedited publication	--	--	15	--	5	--	20	--
Clinical studies	--	--	--	73	--	51	--	124
Experimental studies	--	--	--	35	--	25	--	60
Case Reports	--	4	--	15	--	31	--	50
Communications	20	--	21	--	19	--	68	--
Professional Meeting	25	86	--	--	--	--	25	86
Review	5	--	7	2	11	5	23	7
Other	--	--	1	--	7	1	8	1

Original Contribution Types
Table 2 illustrates the contribution types in CR and JCJ. There are two to three times as many original research articles in CR than in JCJ during the years investigated. Case reports were only published in JCJ, whereas communications were only published in CR. The contribution types in both journals have evolved over time. For example, in 1983, both JCJ and CR published papers from professional meetings, a practice that ceased in the 1993 and 1996 editions. Both journals currently focus primarily on original research articles.

Authors Per Article
Table 3 compares the mean number of Japanese and non-Japanese authors on original contributions published during the years 1983, 1993, and 1996 in the JCJ and CR. In each of these years there were roughly 2-3 more Japanese authors per article in JCJ than there were non-Japanese authors per contribution in CR. These trends were similar for the other contribution types as well. For example, there were 296 authors on 45 case reports in JCJ for a mean of 6.6 authors per case report with a range of 2-12 authors. Similarly, there were 225 authors on 59 communications in CR for a mean of 3.8 authors per communication with a range of 1-11 authors per article.

Table 3: Comparison of Authors Per Original Contribution in Articles by Journal and Japanese or Non-Japanese Author Groups

	Year	No. Authors	No. Articles	Mean	Range
Circulation Research (non-Japanese authors)	1983	423	115	3.7	1-10
	1993	702	163	4.3	1-13
	1996	849	173	4.9	1-13
Japanese Circulation J. (Japanese authors)	1983	446	69	6.5	1-15
	1993	632	104	6.1	1-14
	1996	512	73	7.0	3-16
Circulation Research (Japanese authors)	1983	24	7	3.4	2-5
	1993	138	23	6.0	2-11
	1996	116	18	6.4	3-11
Japanese Circulation J
(non-Japanese authors)	1983	0	0	--	--
	1993	20	4	5.0	2-9
	1996	19	3	6.3	4-10

Table 4. Contribution Types in Circulation Research

	1983			1993			1996		3 Year Total
	CR	JCJ	CR	JCJ	CR	JCJ	CR	JCJ
Original Research
Total number authors	423	446	729	632	870	512	2,022	1,590
Total number articles	115	69	169	104	177	73	461	246
Authors/article; mean	3.7	6.5	4.3	6.1	4.9	7.0	4.4	6.5
Authors/article; range	1-10	1-15	1-13	1-14	1-13	3-16	1-13	1-16

Discussion
The literature on multiple authorship has largely addressed the increasing numbers of authors per article over time seen in various biomedical journals (9-11). While most attribute part of this increase to a growing complexity of research (12) and increased collaboration (13), the proliferation of authors is primarily seen as resulting from the unjustified addition of authors due to increasing pressures to publish (14, 15).

The data described by this research raises questions about why there are inter-cultural variations in the number of authors per article. In the Japanese journal, the mean number of authors on original contributions and even case reports-- that theoretically should require fewer authors-- was much higher than the mean for original contributions and communications in the U.S. journal. Indeed, the nearly two-fold difference assumes even greater significance since CR accepts contributions from other cultures which might, like Japan, list a large number of authors (16). Is having 2-3 more authors per article a practice to be condemned as unjustified authorship, or is science in Japan published in a systematically different, yet still legitimate, way?

One explanation for the observed differences may be that the Japanese penchant for groupism leads them to involve more people on each research project, all of whom have made the “significant” contributions required by the Uniform Requirements for authorship. Japanese research groups ritually hold frequent and time-consuming lab meetings, a tradition which promotes group solidarity by giving everyone an opportunity to make legitimate intellectual contributions.

Groups in Japan, including laboratories, possess a cohesive sense of unity and mutual reliance on both the group and on the paternalistic leader who runs the group (17). The relationship has been described as one of dependence on the part of the subordinates and indulgence on the part of the leader, much like the relationship between parent and child (18). Each member of the group has a vested interest in its success, so the leader will fulfill his duty of looking after the welfare of his subordinates, which, in the case of a professor, means finding jobs for them after leaving the lab (19). Subordinates, in turn, will be loyal and cooperative (2), resulting in a system of “competitive communism.” (20)

One consequence of this approach is that the head of a lab, who may be busy with numerous other tasks, will have his or her name added to every paper leaving the lab, regardless of his or her contribution. We base this assertion upon both personal experiences and discussions with Japanese colleagues about the issue. Under the Japanese system of group interdependence, a laboratory head or department chairman who has played a critical role in providing funding and keeping the aisles clear for his underlings through consensus building (21) may have some legitimate claim to authorship since this level of contribution is critical for conducting the research and completing the paper. While the ICJME assesses such activity as insufficient to claim authorship, a stronger case may be made in Japan where these leaders have an indispensable political role in obtaining funding.

According to the criteria of the uniform requirements, some of the contributions of lab heads and other lab members would perhaps better be recognized in an acknowledgment rather than an authorship, but such a discrimination is likely to be antithetical to group harmony for two reasons. First, there appears to be a greater value placed on the act of contributing than on the value of the contribution, regardless of its magnitude or impact. That is, the symbolic process of contributing assumes greater significance than the intrinsic value of the contribution, which precludes assignment of authorship to some group members but only acknowledgments to others. Second, successful completion of a project may depend upon contributions made by lab members which may, practically, be difficult and counterproductive to sort out as it may disturb group harmony.

Implications
In addressing the problem of unjustified authorship the ICJME, as an assembly of science editors, has attempted to fashion objective, even scientific, criteria to be uniformly applied to all cultures. We believe this approach presents two problems. First, it is not clear that contributions to a research project can be weighed and measured in any kind of scientific fashion. How would one estimate the value of one hour spent by a senior researcher on a project, versus several hours spent by a junior researcher? Indeed, one can envision a multitude of scenarios in which the uniform requirements, if applied strictly as written, would result in unfairly excluding individuals whose contributions were indispensable to a given project (13).

Second, it can be argued that the uniform requirements themselves are more a reflection of the cultural background of the majority of members of the ICJME than the codification of undisputed scientific principles. For most Western scientists, the conduct of research is seen more as an individualistic endeavor than a group one. The image of a “scientist” Westerners hold, somewhere in the background of consciousness, is that of a solitary research pioneer, bent over a microscope into the early hours of the morning. Not all countries, as we have shown, take this position.

Scientific investigation is neither devoid of its own cultural milieu nor immune to the influence of the cultural values and beliefs of those individuals who use scientific methods. These data and analysis illustrate that publication of the results of scientific inquiry is in fact the interface between the scientific method and the culture of the contributing investigators. The goal of crediting authorship to only those who it is due is a noble one. However, the assessment about what constitutes authorship is a cultural, not a scientific, judgment.

References
1. Payer L. Medicine and culture. New York: Henry Holt and Company, Inc, 1996.

2. Befu H. The group model of Japanese society and an alternative. Rice University Studies. 1980;66:169-87.

3. Watanabe M. The Japanese and Western science. Philadelphia: University of Pennsylvania Press, 1988.

4.Bartholomew JR. The formation of science in Japan. Building a research tradition. New Haven and London: Yale University Press, 1989.

5.Herbig P. Creativity in Japan. In: Innovation Japanese style. Westport, Connecticut and London: Quorum Books, 1995:7-17.

6.International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. JAMA 1997;277:927-34.

7. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. BMJ (Clinical Research Edition) 1988;296:401-5.

8. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. BMJ 1991;302:338-41.

9.Epstein RJ. Six authors in search of a citation: villians or victims of the Vancouver convention? BMJ 1993;306:765-7.

10.Howard MO. Multiple authorship: Trends over 50 years in the Journal of Studies on Alcohol. J Studies Alcohol 1996;105-6.

11.de Villiers F. Publish or perish--the growing trends towards multiple authorship. S African Med J 1984;66:882-3.

12. Huth EJ. Irresponsible authorship and wasteful publication. Ann Intern Med 1986;104:257-9.

13. Grant IWB. Multiple Authorship. BMJ 1989;298:386-7.

14. Shapiro DW, Wenger NS, Shapiro MF. The contributions of authors on multi-authored biomedical research papers. JAMA. 1994;271:438-42.

15. Angell M. Publish or perish: A proposal. Ann Intern Med 1986;104:261-2.

16. Halperin EC, Scott J, George SL. Multiple authorship in two English-Language Journals in Radiation Oncology. Acad Med 1992;67:850-6.

17. Nakane C. Japanese Society. Berkeley and Los Angelos: University of California Press, 1970.

18. Doi T. The Anatomy of Dependence. Tokyo: Kohbundo, 1971.

19. Kinoshita J. System's rigidity reduces lure of science as a career. Science 1996;274:49-52.

20.Kenrick DM. Where Communism Works. The Success of Competitive-Communism in Japan. Rutland, Vermont & Tokyo, Japan: Charles E. Tuttle Company, 1988.

21.Fetters MD. Nemawashi essential for conducting research in Japan. Social Sci Med 1995;41:375-81.

Acknowledgement
We gratefully acknowledge the assistance of Debbie White in manuscript preparation.

Funding
This research was made possible in part by the generous support of the Robert Wood Johnson Clinical Scholars Program (MD Fetters).

Please send comments to Email < asianbioethics@yahoo.co.nz >.

To contents page
To Japanese version
To Eubios book list
To Eubios Ethics Institute home page