Richard Smith and Richard Horton have expressed
concern regarding research misconduct and, in particular,
misappropriation of information and ideas and definitions of
authorship.1 2 Issues of data ownership and
authorship
are particularly problematic for multicentre projects, where the need
to "publish or perish" among the increased number of
researchers is
compounded by the requirement for research accountability. Multicentre
collaborations are becoming increasingly common: they give greater
breadth of experience and expertise in developing and implementing
protocols, facilitate recruitment of subjects, and allow a greater
variety of setting to enhance generalisability or results. Despite its
importance, little has been written about this area.3-
5
The National Psychosis Research Framework (currently coordinated
at PRiSM, Institute of Psychiatry) recently established guidelines for
multicentre collaboration.
(1) A publication plan should be constructed early on to avoid
overlap
of contents and ensure fair allocation of authorship. Prior written
agreement must be obtained from the principal investigators of all
centres whose data is to be used in any paper, and policy on
authorship, correspondent, and acknowledgement for each paper should
likewise be agreed.
(2) Individuals interested in data collected at other sites should
circulate their publication proposals to those sites, giving a
reasonable specified time period in which to respond.
(3) Authorship should reflect relative contributions to the writing
of
the paper as well as to the analysis, design, and conduct of the study.
It may be appropriate for some contributions to be recognised by an
acknowledgement.
(4) The final draft of any paper must be circulated to all authors,
whose written agreement regarding content and target journal must be
obtained before each submission for publication. Consent must also be
obtained from those whose contribution has been acknowledged.
(5) When a public presentation of data is to be made and data are
taken from a paper already accepted for publication, collaborating
authors must be acknowledged. Where the data for presentation are taken
from a paper still in preparation, permission must be obtained from the
relevant authors for their findings to be presented and the authors
likewise acknowledged. Agreement may be given in advance for certain
categories of presentation.
(6) All papers which combine data from more than one site should
state
the affiliation to the wider research collaboration. Papers should be
distributed to all members.
Andrew Barker Senior research fellow in old age psychiatry
University of Southampton,
Thornhill Research Unit,
Moorgreen
Hospital,
Southampton SO30 3JB
Richard A Powell Researcher
University of Exeter,
Department of Mental Health,
Wonford
House Hospital,
Exeter EX2 5AF
References
1 Smith R. Time to face up to research
misconduct.
BMJ 1996;312:789-90.
2 Horton R. Time to redefine authorship.
BMJ
1996;312:723.
3 Hogg RJ. Trials and tribulations of multi-
centre studies:
lessons learned from the experiences of the Southwest Pediatric
Nephrology Study Group. Pediatr Nephrol 1991;5:348-51.
4 Kassirer JP, Angell M. On authorship and
acknowledgements. N Engl J Med 1991;325:1510-2.
5 Hanson SMH. Collaborative research and
authorship
credit:beginning guidelines. Nurs Res 1988;37:49-52.
Deciding the order of authors on research papers is a
recognised
problem.1-5 Currently, authorship order cannot be
interpreted by readers and editors. The last position often carries
more status.3 In some papers the senior
investigator is
named last, in others it is the head of the laboratory or department,
and in others it is the person who contributed least.
Burman recommended that the research supervisor should be second
author
and the head of laboratory last,4 while Riesenberg
and
Lundberg proposed that order should reflect contribution.1
Most writers agree that the person most closely associated with the
research should be the first author.1-4 The
Council for
Biology Editors recommend that each journal specifies criteria for
authorship order, but journals tend to leave this to authors. To our
knowledge no practical scheme for allowing group consensus on
authorship order has been published, although there are techniques for
calculating credit.
We undertook a piece of participative research to assign authorship
order to our work5 by means of peer judgments. We
compiled
a list of contributions involved in carrying out our research (see
box). Team members spoke briefly about their contributions and ticked
the list appropriately. Using this list and background knowledge about
the development and execution of the project, each team member compiled
three rankings of colleagues' contributions, excluding themselves. The
first was for the project as a whole, and the others for each of two
papers. A top rank was given eight points, a second seven, and so on.
We collated the rankings to give a total score for each team member,
and all nine authors discussed and agreed the resulting order. Despite
initial trepidation and discomfort, we agreed that the outcome of this
process was fair.
| Authors'
possible contributions to research project |
Impetus or
initiative
Preparation
Planning meetings
Sampling
Questionnaire design
Administering the survey
Interviewing subjects
Coding questionnaire
Qualitative analysis
Quantitative analysis
Drafting paper
Commenting on drafts
Attending meetings
Conference preparation |
Our approach makes each author's involvement explicit
within the team and uses an anonymous democratic process. As the
rankings can be secret, abuse of power, status, and reputation are
minimised. This may help to reduce difficulties experienced by
researchers. Determining authorship in advance may reward those who
promise at the expense of those who deliver, whereas our approach
reflects the team's view of actual contributions. We suggest that
others test this process for themselves.
Authorship order reflects peer group views on overall
academic
and practical contribution to the project and is therefore rank
order.
R S Bhopal Professor of epidemiology
and public health
J M Rankin Senior research associate
Department of Epidemiology and Public Health
E McColl Senior research associate
Centre for Health Services Research
R Stacy Lecturer in medical sociology
P H Pearson Lecturer in primary care
nursing
E F S Kaner Research associate
Department of Primary Health Care
L H Thomas Research associate
Centre for Health Services Research
B G Vernon Lecturer in the ethics of health care
Department of Primary Health Care
H Rodgers Senior lecturer in stroke medicine
Departments of Medicine and of Epidemiology,
School of Health
Sciences, University of Newcastle,
Newcastle upon Tyne NE2 4HH.
References
1 Riesenberg D, Lundberg G. The order of
authorship: who's
on first. JAMA 1990;264:1857.
2 Day RA. How to write and publish a
scientific
paper. 4th ed. Phoenix, AZ: Oryx, 1994.
3 Burman KD. Hanging from the masthead. Ann
Intern
Med 1982;97:602-5.
4 Huth EJ. How to write and publish papers
in the medical
sciences. 2nd ed. Baltimore: Williams and Wilkins, 1990.
5 Bhopal RS, Rankin M, McColl E, Thomas LH,
Kaner EFS, Stacy R,
et al. The vexed question of authorship: views of
researchers in a British medical faculty. BMJ
1997;314:1009-12. (5 April.)
Feedback
