BACK TO THE FEATURE INDEX

I am pleased to offer my advice to the graduate student at Bigtime State University with the vexing authorship problem. Before I do so, I need to make it clear that my opinions are not the official positions of the National Institutes of Health (NIH), but they do represent 7 years of experience dealing with authorship disputes in the intramural program in my position as deputy director for intramural research. Unlike at most universities, the majority of the trainees at NIH are postdoctoral fellows, but I believe that the same principles that govern resolution of authorship disputes among fellows and principle investigators are relevant to graduate students.

This case raises issues related to two aspects of authorship: First, what are the rules that govern who should be an author on a paper? Second, when prior agreements are made concerning authorship, how should they be formulated and do they take precedence over standard authorship requirements?

Publication is the means by which scientific information is disseminated to the larger community of scientists and the public, so the information can be criticized, reproduced, and used to build on existing concepts. Publication also recognizes scientists who have contributed to the literature on a particular subject and appropriately credits those who have been part of the current work. Thus, authorship of a paper and the relative position of authorship are an important part of the integrity of the scientific process.

Who should be an author of a paper? Without a doubt, someone who has made no contribution to the conceptualization of a project, or who has not been involved in the generation of data or its analysis, should not be an author. This is stated clearly in the "Guidelines for the Conduct of Research in the Intramural Research Program at NIH": "For each individual the privilege of authorship should be based on a significant contribution to the conceptualization, design, execution, and/or interpretation of the research study, as well as a willingness to assume responsibility for the study. Individuals who do not meet these criteria but who have assisted the research by their encouragement and advice or by providing space, financial support, reagents, occasional analyses or patient material should be acknowledged in the text but not be authors." This principle does not allow rewards for working hard if this work does not contribute to the paper.

While the interpretation of "significant" and the definition of "conceptualization" are frequently subject to interpretation, the "senior" graduate student as described in the case study has made no contributions to the work in question and should not be a co-author. If the senior graduate student had indeed made significant contributions (and this may require some soul searching by the graduate student who raised the issue), then this other student should be an author, but not the primary author. As an aside here, different disciplines reward "primary authorship" (i.e., the person who did most of the experiments and wrote the first draft of the paper) by different authorship positions: Whereas the primary author is usually the first author and last place is reserved for the senior author in most biomedical research, epidemiologists traditionally reverse this order. For two primary authors who contributed equally to the paper, it is acceptable to place an asterisk next to both names and indicate this; however, the official attribution in the literature will contain the authors as they are listed when the paper is published.

So how does one deal with prior agreements that specify authorship and the order of authors on a paper? Because science is increasingly done by large groups of people, agreements are made in advance to credit the contributions of members of these groups, including position of authors on a paper. In fact, most science administrators who deal with disagreements among collaborators recommend that agreements be fashioned before work begins so there is no misunderstanding about the relative contributions and responsibilities of the parties. However, such agreements cannot assign authorship to someone who has not made a contribution to the work that is being submitted, nor should they specify order of authorship before it is known who is actually the primary author on the work. A well-crafted collaborative agreement among two or more labs might specify who the senior author will be based on the work that is being done, and it might even indicate who the first author will be based on the expected work, but it cannot require such authorship if circumstances change.

A further problem in this case is that the "agreement" within the lab was unknown to the complaining graduate student before the work for the paper was completed. In my experience, it is not uncommon for senior members of laboratories to make such agreements and fail to inform incoming trainees and other participants in the work that such a prior agreement has been made. Anyone beginning work on a collaborative project should ask whether prior agreements about authorship have been made, and senior investigators are obligated to inform anyone working with them on a project about such agreements. If such an agreement violates the principle of authorship mentioned above from the "Guidelines for the Conduct of Research in the Intramural Research Program at NIH," it is not valid.

Who enforces these rules? Increasingly, journals are specifying rules for authorship in their information to authors, and the authors should read them carefully when they submit papers. First and foremost, however, the institution (government lab, university, or private company) is generally considered responsible for making sure these authorship rules are followed, and it generally falls to the dean or the director of research to enforce them as part of the responsible conduct of research.

Michael M. Gottesman, M.D., is currently a laboratory chief and senior science administrator in Bethesda, Maryland.