Individuals should be included as authors on a work submitted for publication if they have provided significant contributions affecting its direction, scope, or depth. These contributions may take many different forms. Generally, the principal author will have designed many of the experiments, performed much of the work, analyzed most of the data, and written the manuscript. In some cases, a senior author or mentor may have provided much of the work involved in the development of a project and, after it was initiated, provided long-term guidance to its completion. Other advisors may have provided the creative spark or the idea that was carried forward in the work. Some mentors may have developed and performed methodologies without which the project may not have reached a timely completion. It is imperative, as noted below, that this methodological input extend beyond the performance of routine assays by a technician, sometimes for a prearranged fee. The contribution of other authors to a manuscript may be in the analysis or interpretation of the data. The conclusions of some projects might not have been reached without this vital input.

While it may be difficult in some instances to decide whether specific contributions warrant authorship, there are clear circumstances under which individuals should not be included as authors. The simple provision of resources such as a laboratory space, instrumentation, or even research funding without direct involvement in a project should not of themselves be grounds for authorship. If a "collaborator" provides services, consulting, or materials for a fee or reimbursement under a contractual arrangement, he might not be considered as an author on a scientific project. This principle should also extend to the provision of routine technical work, as may be provided by any paid technician in a laboratory without significant input into the design or conduct of a study. In clinical areas, contributions limited to involvement in the care of a patient or to the provision of specimens from a patient should not be grounds for inclusion as an author on a manuscript. Occasionally, supervisors, section heads, or departmental chairpersons insist upon inclusion as authors simply in recognition of their status, but this is inappropriate unless there are other grounds that warrant such recognition. Simple proofreading or editing of manuscripts should provide no basis for inclusion as an author. Occasionally, a principal investigator on a project may seek advice on narrowly defined problems or on problems unrelated to the project's objective. Provision of such advice should not provide grounds for authorship.

In addition to the benefits of prestige, authorship carries with it the burdens of responsibilities. The responsibilities of authorship should apply not only to written and published documents but also to verbal communications in public fora including the press. The principal author must be responsible for establishing the list and order of authors. He must be able to identify the specific contributions of each author and understand the significance of each contribution to the conclusion of the project. The principal author, representing all of the authors, must be willing to share details of the methodologies and data used in the course of investigation. Currently, it is the policy of many journals that publication also implies a willingness to share regents such as antisera and recombinant clones, thus it is important that the authors recognized the specific policies of a journal before submitting their work to it for publication. In any case, the unselfish exchange of information and reagents is a basic assumption of science and every effort should be made to adhere to it provided that it does not compromise an individual scientist's research efforts. Each author should be able to defend the methods and data pertinent to his specific contribution. On a larger scale, each author has the responsibility to read the manuscript in which his work is included and understand it sufficiently to be able to agree with the general conclusions and interpretations of the paper. Any disagreements should be resolved prior to submission of the work for review. Ultimately, any individual author has the right and the responsibility to remove his name from a manuscript if he has substantial concerns with its conclusions.

Authors have additional responsibilities regarding the content of their manuscripts. Above all, the manuscript must represent an accurate and complete reflection of the methods utilized and the data obtained. Sketchy outlines of methodology make it impossible for others to duplicate important experiments and may lead to unwarranted controversy over the results obtained. It is of importance to report data that are both supportive and unsupportive of the general conclusions of the paper. Withholding unsupportive data may suggest selection bias in reporting the results of an experiment. Despite the academic pressure, real or imagined, to demonstrate excellence with quantity rather than quality of publications, every effort should be made to avoid fragmentation of a complete body of work into separate publications. Moreover, the practice of publishing the same material in more than one manuscript is inappropriate except in clearly identified review articles with citations of the original work. When ideas, concepts or text of others are used in a manuscript, appropriate citations should be made. Furthermore, prior work that served as the basis for a manuscript must be cited. In their citations, authors must strive to acknowledge data that conflict with their own theories as well as data that are generally supportive. It is important to acknowledge the sources of funding for a publication to ensure that the funding agencies are appropriately credited and, moreover, that any potential conflict of interest is identified. In general, abstracts may be somewhat less detailed because of their brevity; however, they must be considered as scientific publications and, as such, are subject to the same considerations regarding responsibility of authorship as full-length manuscripts.