Posted January 10, 2007
Authorship is a way of making explicit both credit and responsibility for the contents of published articles. Credit and responsibility are inseparable. The guiding principle for authorship decisions is to present an honest account of what took place. Criteria for authorship apply to all intellectual products, including print and electronic publications of words, data, and images. Journals should make their own policies on authorship transparent and accessible.
Criteria for Authorship. Everyone who has made substantial intellectual contributions to the study on which the article is based (for example, to the research question, design, analysis, interpretation, and written description) should be an author. It is dishonest to omit mention of someone who has participated in writing the manuscript (“ghost authorship”) and unfair to omit investigator who have had important engagement with other aspects of the work. (See the WAME policy statement, “Ghost Writing Initiated by Commercial Companies”)
Only an individual who has made substantial intellectual contributions should be an author. Performing technical services, translating text, identifying patients for study, supplying materials, and providing funding or administrative oversight over facilities where the work was done are not, in themselves, sufficient for authorship, although these contributions may be acknowledged in the manuscript, as described below. It is dishonest to include authors only because of their reputation, position of authority, or friendship (“guest authorship”).
Many journals publish the names and contributions of everyone who has participated in the work (“contributors”). Not all contributors necessarily qualify for authorship. The nature of each contributors’ participation can be made transparent by a statement, published with the article, of their names and contributions and WAME encourages this practice.
One author (a “guarantor”) should take responsibility for the integrity of the work as a whole. Often this is the corresponding author, the one who sends in the manuscript and receives reviews, but other authors can have this role. All authors should approve the final version of the manuscript.
It is preferable that all authors be familiar with all aspects of the work. However, modern research is often done in teams with complementary expertise so that every author may not be equally familiar with all aspects of the work. For example, a biostatistician may have greater mastery of statistical aspects of the manuscript than other authors, but have somewhat less understanding of clinical variables or laboratory measurements. Therefore, some authors’ contributions may be limited to specific aspects of the work as a whole.
All authors should comply with the journals’ policies on conflict of interest.
Number of Authors. Editors should not arbitrarily limit the number of authors. There are legitimate reasons for multiple authors in some kinds of research, such as multi-center, randomized controlled trials. In these situations, a subset of authors may be listed with the title, with the notation that they have prepared the manuscript on behalf of all contributors, who are then listed in an appendix to the published article. Alternatively, a “corporate” author (e.g., a “Group” name) representing all authors in a named study may be listed, as long as one investigator takes responsibility for the work as a whole. In either case, all individuals listed as authors should meet criteria for authorship whether or not they are listed explicitly on the byline. If editors believe the number of authors is unusually large, relative to the scope and complexity of the work, they can ask for a detailed description of each author’s contributions to the work. If some do not meet criteria for authorship, editors can require that their names be removed as a condition of publication.
Order of Authorship. The authors themselves should decide the order in which authors are listed in an article. No one else knows as well as they do their respective contributions and the agreements they have made among themselves. Many different criteria are used to decide order of authorship. Among these are relative contributions to the work and, in situations where all authors have contributed equally, alphabetical or random order. Readers cannot know, and should not assume, the meaning of order of authorship unless the approach to assigning order has been described by the authors. Authors may want to include with their manuscript a description of how order was decided. If so, editors should welcome this information and publish it with the manuscript.
Authorship Disputes. Disputes about authorship are best settled at the local level, before journals review the manuscript. However, at their discretion editors may become involved in resolving authorship disputes. Changes in authorship at any stage of manuscript review, revision, or acceptance should be accompanied by a written request and explanation from all of the original authors.