The original reason for publication--to communicate results--has become secondary to a scientist's professional advancement. Scientists need authorship on papers to survive. Recent years have seen many misleading--and some downright dishonest--authorship decisions, and some of them have gotten a lot of attention. There is a movement afoot to reestablish the accountability of authors.

Most authorship controversies arise because standards are fuzzy and because scientists often fail to understand those standards that do exist. If your name appears as an author on a manuscript, you are responsible for the manuscript's veracity; but to what extent, and to whom, isn't clear. Still-evolving signals from different organizations, journals, and laboratories concerning the ethics and even the definition of authorship confuse the issue further.

So what is a young scientist to do? Learn the rules that do exist and do your best to observe them.

Many journals are lax in specifying authorship principles, but some journals--especially biomedical journals--have specific rules. Most existing specifications in the biomedical science journals are variations of the recommendations of the International Committee of Medical Journal Editors, in "Uniform Requirements for Manuscripts Submitted to Biomedical Journals." In a nutshell, the committee recommends that authorship be based on:

  • substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data;

  • drafting the article or revising it critically for important intellectual content; and

  • final approval of the version to be published.

  • All three conditions must be met. A description of the contribution of each author must be provided, and, even for large groups, each author must meet all criteria for authorship. Acquisition of funding, the collection of data, or general supervision of the research group, by themselves, do not justify authorship.

    Some journals require authors to detail their contributions. The Journal of the American Medical Association ( JAMA), which has for years tried to emphasize authorship responsibility, requires each author to delineate his or her contribution in each of three areas: intellectual and experimental; manuscript writing and editing; supervisory and funding.

    Most journals are far less specific. But there are other standards to consider. ...

    Institutions, including universities, may have expectations of authors

    Increasing numbers of institutions require all authors to sign a statement indicating the contribution of each author. Some institutions require papers to earn the institution's endorsement before a paper may be submitted in the institution's name. The documentation needed for institutional approval or endorsement encourages accountability and enables the institution to legally deal with a person who has not kept up standards--although smaller institutions, especially, still worry that they might be held liable for an author's fraud.

    Every author is responsible for fulfilling the policies of their institution. So wipe the dust off that huge binder and read your institution's policies on authorship. Ask faculty administration or the university's legal department about anything that isn't clear.

    But your responsibility doesn't end with the university. Some departments and divisions have more exacting standards than the institution. All new faculty members ought to make sure they're familiar with university-level, division-level, and department-level authorship rules, but sometimes they don't ask or aren't told. If you don't know the rules of your institution, division, and department--find out.

    Authorship Resources

    International Committee of Medical Journal Editors, "Uniform Requirements for Manuscripts Submitted to Biomedical Journals," Ann Intern Med. 126, 36 (1997) ( http://www.icmje.org/index.html)

    D. Rennie and A. Flanagin, "The Contributions of Authors," Editorial, JAMA 28, 1 (2000).

    D. Rennie, V. Yank, and L. Emanuel, "When authorship fails: A proposal to make contributors accountable", JAMA 278, 579 (1997).

    Authorship Responsibility, Financial Disclosure, Copyright Transfer, and Acknowledgment, JAMA (2002). http://jama.ama-assn.org/issues/v288n1/fpdf/jin0703_crit.pdf

    Guidelines for the Conduct of Research in the Intramural Research Programs at NIH. ( http://www.nih.gov/campus/irnews/guidelines.htm).

    Can one person be responsible for the content of the entire paper?

    As specialization in science increases, it gets harder for one author to understand in depth all facets of a paper. But as a practical matter, at least one author needs to be willing and able to take responsibility and to explain all of a paper's data and conclusions.

    Corresponding authors are responsible for the form and content of the paper. Corresponding authors are not, as many assume, merely in charge of sending reprints. The corresponding author is expected to do everything from collecting compliance paperwork from the other authors to assuming ultimate responsibility for the paper as a whole. While each author bears some responsibility for the paper's contents, the corresponding author has the additional obligation of making sure that all the other authors have met their obligations. Quizzing your collaborators on their work carries risks, but failing to do so also carries risks, which, though of lower probability, are graver.

    With the lack of a universal policy about authorship, what should you consider your responsibilities to be?

    Here's a checklist:

    • Have all authors contributed to the work?

    • Is the work true and unpublished?

    • Are the contributions of those who helped with the paper, but whose contributions don't rise to the level of authorship, listed in the Acknowledgments section?

    • Have all people described in the Acknowledgements, and all people listed in Personal Communications, been notified?

    • Have all authors read the paper prior to submission, and been given the opportunity to comment? Have their comments been dealt with to their satisfaction?

    • Have all potential conflicts of interest been disclosed?

    • Have all protocols for human subjects and animal use been observed?

    • Is the literature review balanced? Does it cite publications of rival groups and those that present opposing theories?

    • Have you retained all the raw data associated with the paper? (There is no legal limit on the amount of time data must be kept, although 5 to 7 years are suggested.)

    • Are you, personally, convinced that the data are sound?

    • (Keep all correspondence with journals for at least 6 months after the paper is accepted for publication.)

    If you submit a paper to a journal with exacting specifications, answer those specifications honestly. Don't sign any piece of paper without reading it and without agreeing to its contents.

    Most decisions about authorship responsibility are done at the level of the laboratory.

    Journal, disciplinary, institutional, divisional, and departmental expectations are all well and good, but as a practical matter most decisions about authorship are made locally, and that's where breakdowns often occur. PIs who make the lab's authorship customs clear to all people coming into the lab can avoid conflicts about ethics, accountability, and the accompanying bruised egos. So, before you begin writing, ask your PI how authorship is determined and what responsibilities will accompany your authorship. If you are a PI, make sure everyone in your lab knows the rules. If you don't know, ask.

    If you have conflicts, what do you do?

    The most common problem junior scientists encounter is exclusion from primary authorship. Try to handle all disagreements about authorship within the group of authors and with the PI. Protesting after the fact and beyond the group carries serious professional risks. Working out percentages of contributions is a subjective but effective way to fairly determine the order of authorship. Seek a consensus that's agreeable to all, and do your best to resolve all authorship disputes with your colleagues, in a professional manner.

    Only if you are unsuccessful in discussions with the other authors, and only if the issues involved are really important, should you begin to maneuver up the ladder of checks and balances, seeking subtle mediation. Your goal should still be to reach consensus, not to get anyone in trouble.

    First stop after lab head is department or division head, or perhaps an ombudsman in the dean's office. Some institutions also have postdoc and student organizations that will open a discussion. These offices can only mediate, but an intelligent interception is often all that is needed to get an authorship discussion to move more fairly. Work behind the scenes if you can; try hard not to alienate anyone.

    It is impossible to overstate the importance of distinguishing between honest disagreements over authorship issues like the allocation of credit, and issues of scientific misconduct. It is natural for people to have an inaccurate sense of their own contributions. You probably do, too. Disagreements over credit are not evidence of misconduct.

    If you have evidence of plagiarism or fabrication or falsification of data, you shouldn't seek mediation; you should, instead, notify the chairperson, as well as your human resources or legal department--or, if you are uncomfortable with this, you should find a new lab to work in. But either way be sure your evidence is strong.

    The best situation for students and postdocs is to be affiliated with an institution and lab that promotes accountability and workplace ethics. The authorship-policy pendulum will probably swing further toward greater accountability before credit is again the driving force in decisions. Standards of authorship will continue to be refined. But no matter, if you are in a lab where authorship decisions aren't transparent--especially if transparency is deliberately obstructed--you may be in the wrong place. Transparency breeds accountability.