Collaboration and Mentoring


Science increasingly depends on collaborations. This is reflected in the increased number of authors per publication [Mussurakis, 1993, Khan et al., 1999]. For example, a random sampling of 20 journal articles published in Science in 1966 had an average of 1.9 authors per article, with a maximum of 3 authors. By the year 2000, the average had more than doubled to 4.3, with a maximum of 12 authors. This increase reflects both larger research groups and collaboration between research groups. The rise in collaborations is a result of many factors. First, no single person has the skills, knowledge, and resources to address all research problems. A judicious choice of collaborators can save considerable time and money. Second, the funding and structure of science tend to favor programs in which recognized authorities are involved from each key area. Third, breakthroughs are often more likely to come from collaboration across disciplines than by adherence to tried and true methods. Fourth, collaboration between the private sector and academia is being encouraged by legislation (e.g., the Bayh-Dole Patent Reform Act of 1980 allowed universities to negotiate patent rights with industrial partners), industry (which recognizes the benefits of the expertise and reputation of academics), and academia (which can benefit from immediate and long-term sources of private funding). Finally, collaborations are easier than they were before. With obvious improvements in communication (phone, Fax, Email), shipping (one-day delivery), and travel (to national and international conferences), potential collaborators are more likely to find each other and are more able to maintain their collaboration. Whatever the reason, collaborations are increasingly beneficial and possible.

Collaborations are a frequent source of problems, and this results in part because collaboration can take many different forms. It certainly implies two or more people having joined together for a common purpose, but this might involve almost any arrangement of shared time, work, resources, unique materials, data, ideas, or money. Once the work is completed, credit and responsibility can then be shared in a number of ways.

In some cases, collaborations may not even begin because of reluctance to share or work together [Cohen, 1995]. Once started, collaborations can be marred by misunderstandings of what is to be provided by each of the participants, unhappiness with a slow collaborator, disagreement about what and when to publish, or conflicts regarding authorship and credit. [Kahn et al., 2000; Wilcox, 1998]. Although there is no single panacea for such problems, it is evident that any solution needs to be based in improved communication.‚Äč

While successful collaborations depend on explicit communication, such communication is often difficult. In some cases, different cultural backgrounds are an impediment to understanding. The culture of the private sector emphasizes discovery and application of profitable products, for instance, while academics may be more interested in mechanisms and new discoveries. In international collaborations, participants may literally speak different languages. Even when a common language is available, participants may have very different styles and understandings of communication as well as different perspectives on sharing and ownership.

Different research disciplines can also be a source of miscommunication. Because of the nature of the work, some disciplines may have very different expectations about hours to be worked (e.g., many biochemical and molecular biological studies require long hours), standards of proof (e.g., different disciplines have developed different views about the need for statistical methods), or the pace of work (e.g., high quality electron microscopy can often be elusive and require many days or weeks of searching for acceptable images long after a study has been otherwise completed). Similarly, communication across disciplines can be impaired by different understandings about the science, vocabulary, or methods.

Different individuals can simply have very different standards and interpersonal styles. Some people consider a verbal agreement to be binding, while others prefer explicit, written contracts. Some favor rapid publication of each new finding; others prefer to amass a body of work for a single large publication. Some are convinced that authorship and credit should be reserved only for those who have made the most substantial contribution to the study; others are much freer in assigning credit. Some readily and clearly speak their minds; others are more withdrawn and will volunteer information only if asked.

Although guidelines or regulations do not explicitly cover all these aspects of collaboration, the goal should be communication that clarifies expectations of all parties involved. Although it may not be necessary to put everything in writing, attempts should be made to explicitly address relevant issues. Finally, it is important to keep in mind that although collaboration is in the best spirit of science, opening a collaboration can leave a scientist vulnerable to the actions, or inaction, of his or her collaborators. Therefore, choosing colleagues should be based not only on the science, but also on the likelihood of an amicable relationship in which lines of communication can be kept open.

The nature of collaborations is so variable that it is difficult to identify a comprehensive set of ethical principles; however, responsible collaborations are defined by openness and communication.

Collaborators should be open about the research. Science is a communal enterprise; both science and society are best served by collegiality and open collaboration. Collaborators should be clear with one another about the research to be undertaken, the methodology, the results, and so on.

Collaborators should be open and clear about the terms of the collaboration. Collaboration is most likely to succeed if expectations are clearly communicated (and perhaps documented) before commitments are made. There should be a shared understanding of what is to be exchanged through the collaboration and how the products of the collaboration will be shared.

The process of collaboration is regulated primarily at the institutional level, not by the public or private funders of research. The presumption is that the community is best served by minimal barriers to free and open collaboration. Nevertheless, the outcomes of collaboration, particularly patents and copyrights, are restricted by both public and private funders of research. Moreover, nearly all institutions have rules and guidelines governing collaboration. For example, most academic institutions have explicit rules governing ownership of the products of work done by employees of the institution, material transfer, and limitations on academic-industrial agreements that might compromise the institution's academic mission. Some institutions also have guidelines for issues such as sharing and ownership of data, assignment of authorship, and credit and responsibilities for authors [Eastwood et al., 2001]. In general, collaboration with someone outside of an institution cannot proceed without involving the institution.

  • Bayh-Dole Patent Reform Act (1980): Section 6, Patent and Trademark Amendment of 1980, PL 96-517; implementation by OMB Circular No. A-124 [superseded by PL 98-62 and 37CFR401, 1987]Cohen J (1995): Share and share alike isn?t always the rule in science. Science 268:1715-1718.
  • Eastwood S, Fike JR, Cogen PH, Rosegay H, Berens M (2001): BTRC Guidelines on Research Data and Manuscripts (Brain Tumor Research Center, University of California San Francisco, 1989). Revised and updated in 2000 and reprinted in (Bulger RE, Heitman3/4E, Reiser SJ, eds.): The Ethical Dimensions of the Biological Sciences, 2nd ed. Cambridge University Press, New York.
  • Khan KS, Nwosu CR, Khan SF, Dwarakanath LS, Chien PF (1999): A controlled analysis of authorship trends over two decades. American Journal of Obstetrics and Gynecology 181:
  • Kahn JO, Cherng DW, Mayer K, Murray H, Lagakos S for the 806 Investigator Team (2000): Evaluation of HIV-1 ImmunoGen, an Immunologic Modifier, Administered to Patients Infected With HIV Having 300 to 549 X 106/L CD4 Cell Counts: A Randomized Controlled Trial. JAMA. 284:2193-2202.
  • Macrina FL et al. (1995): Dynamic Issues in Scientific Integrity: Collaborative Research. American Academy of Microbiology, Washington, D.C.
  • Mussurakis S (1993): Coauthorship trends in the leading radiological journals. Acta Radiologica 34:316-20.
  • Wilcox LJ (1998): Authorship: the coin of the realm, the source of complaints. JAMA 280:216-7.
  • National Academy of Sciences, National Academy of Engineering, Institute of Medicine, and National Research Council (1999): Overcoming Barriers to Collaborative Research. Report of a Workshop. National Academy Press, Washington, D.C.


Mentoring the next generation of scientists is a responsibility for current scientists. A mentor has experience with the challenges that will be faced by a trainee, the ability to communicate that experience, and a willingness to do so. A mentor assists the trainee in understanding and adhering to the standards of conduct within their profession. In this way, mentoring of new researchers by senior investigators passes on the informal and possibly unwritten standards from one generation of scientists to the next. Within a small research group, this mentoring may readily occur, but many current research groups are too large or competitive. Whether or not this has changed the extent to which new scientists become aware of prevailing standards of conduct, it appears that issues of responsible conduct are discussed infrequently.

A mentor teaches responsible conduct explicitly and by example; mentoring involves both what is verbalized and what is demonstrated in practice. For better or worse, the default method of teaching the traditions and standards of science is often by unwitting and serendipitous example. Unfortunately, without discussion of ethical principles and the purposeful assurance that everyone is included, this approach to training is seriously flawed. Principles of decision making are not explicit and are therefore open to interpretation and misinterpretation; moreover, many important roles of scientists, such as peer review and negotiating collaborations, are not observed by the trainee. In their survey of 2000 doctoral students, Anderson et al. found that departmental climate was the strongest predictor for misconduct. Overall, misconduct occurs more often in those departments in which the climate favors competition and discourages collaboration (Anderson et al., 1994). However, research misconduct occurred least often in those cases in which students felt that their advisors, or others, provided useful feedback and evaluation. These findings are consistent with the view that explicit mentoring reduces the risk of research misconduct.

Eastwood et al. (1996) found that nearly 40% of postdoctoral research fellows responding to a survey at the University of California, San Francisco reported having had no guidance in ethical research from a scientific mentor. Brown and Kalichman (1998) found that half of graduate students responding to a survey at the University of California, San Diego reported that the total time spent discussing responsible conduct of research with a major professor or advisor had been one hour or less. In a nationwide survey of doctoral students, Swazey and Anderson (1998) found that for nearly every defined dimension of training in ethics, over half of the respondents reported that faculty members provided little or no help.

In order to comply with Section 7009 of the America COMPETES Act, the National Science Foundation now requires that all proposals requesting funds to support a postdoctoral researcher include a one-page mentoring plan as part of the supplemental information.  Proposals submitted without a mentoring plan will not be considered.  Principal investigators must report on mentoring progress as part of their annual and final reports for funded awards.  Information about the mentoring plan and reporting requirements can be found in the Grant Proposal Guide.

Many resources are now available to guide both mentors and trainees in optimizing mentoring, and the discussion here can only raise some of the issues.

Mentors in the sciences should help trainees in their technical development as capable researchers. Mentors also help trainees prepare for the job market, relating an understanding of the current job market and providing opportunities to make contacts with other researchers. Another focus of mentoring is socialization of trainees, helping them toward an understanding of the political, ethical, economic, and social dynamics within the academic community. This training includes skills for teaching, communication, working in teams, leadership, listening, expressing ideas, administration and planning, and budget management. Often trainees will have special circumstances to be addressed through mentoring; perhaps problems of gender, race, national origin, language, or disability.

Be available. Both mentors and trainees must encourage one another to participate fully, while respecting differences in their commitment to the mentoring relationship.

Although the terms mentor, thesis adviser, and research supervisor are frequently used interchangeably, mentors provide information beyond scientific concepts and laboratory techniques. Hopefully supervisors are mentors, but this is not always the case. Initiating a discussion with a supervisor about authorship criteria, the funding process, or mentoring itself might stimulate the supervisor to become a better mentor. Regardless, whether or not a supervisor is an effective mentor, it is unlikely that one person alone can teach everything a trainee needs to learn.

Allow for differences in personalities. Some mentors will be uncomfortable offering advice or initiating discussions unless first asked by a trainee, while other mentors will readily volunteer information and advice without any clear indication that help would be welcomed. Similarly, some trainees see frequent and probing discussion with a mentor as invasive micromanaging, while other trainees thrive on frequent feedback. Mentoring is most effective when the personalities of the mentor and trainee are a good match.

Mentors teach by words and by example. Modeling good skills and behavior is a necessary element of mentoring. A mentor who argues for rigorous authorship criteria must act on that advice, or trainees will see it as hypocritical posturing. Yet a good example is not always enough; it's important that mentors make explicit the often implicit rationale for their behavior, because a trainee can not learn the policy and philosophy that underlie exemplary behavior by observation alone.

Trainees must make their own decisions. The role of the mentor is to provide advice, help, and encouragement, to guide rather than decide matters for the trainee. In issues that are the trainees' responsibility, they must act based on their own values, goals, and experience.

Mentoring is an ongoing process. Widely ranging needs at different stages of a career are not likely to be met by a solitary mentor, and few established scientists can bring the requisite time, knowledge, and interest to the full range of issues that are likely to confront a trainee. For these reasons, the term mentor is best used broadly to mean any person who helps another with one or more aspects of the latter's personal or professional development. In this sense, trainees are encouraged to seek out multiple mentors, each of whom can provide the expertise and experience to help fulfill the trainee's needs.

A study that analyzed the guidelines for the conduct of research at U.S. medical schools summarized the duties and responsibilities of mentors and trainees as follows (Douglas-Vidas et al., 2001):

Guidelines recommend that mentor responsibilities include:

  • Holding regular meetings with trainees
  • Ensuring that trainees are familiar with academic and non-academic policies
  • Carefully supervising trainee work
  • Treating trainees with professional courtesy
  • Keeping trainees' best interests in mind
  • Involving trainees in small group research unit meetings
  • Offering candid advice
  • Encouraging trainees to view job prospects realistically
  • Being alert to behavioral changes indicating trainee stress
  • Write candid letters of recommendation
  • Assist in career counseling and job placement
  • Schedule career planning sessions to monitor progress and avoid conflict

Guidelines recommend that trainees:

  • Conduct themselves in a mature manner
  • Be mindful of mentor time constraints
  • Avoid overidentification with the mentor
  • Be proactive in terms of career direction

Several principles are relevant to issues of mentoring, but they might be distilled down to one: Mentoring is a shared professional responsibility of all scientists.

Effective mentoring is essential for promoting responsible conduct of research. The importance of mentoring for training in the responsible conduct of research has been recognized in several national reports on the integrity of research. For example, a report from the Institute of Medicine (1989) noted the importance of mentors and specifically recommended that departments and research units should monitor the supervision and training of young scientists to ensure that it is adequate [Committee on Science, Engineering, and Public Policy, 2000]. Although mentoring alone may be insufficient, its importance is clear.

More experienced scientists have a responsibility to be effective mentors. Taking an active role in helping to train the next generation of scientists should not be optional. The enterprise of science depends on effective communication not just about the science, but about the practice of science, standards of conduct, and ethical and social responsibility. This responsibility for communication extends to all members of the community, not just to senior researchers. It is likely that a newly arrived undergraduate student could benefit from the mentoring of a graduate student, technician, or even a more senior undergraduate.

Less experienced scientists have a responsibility to seek effective mentors. Just as scientists have a responsibility to become mentors, scientific trainees have a complementary responsibility to take an active role in their own development and seek mentors. Most scientists are in the best position to know their own career aspirations and worries, but few will be fortunate enough to have good mentors step in to help. The obvious solution is to seek out more senior scientists, and sometimes peers, who have experience that one is lacking. Finding someone who will be an effective mentor is, in part, a responsibility of the trainee.

Effective mentoring is important to promote the responsible conduct of research, but no regulations explicitly require or prescribe standards for mentoring. The lack of absolute rules is appropriate, since the success of mentoring depends on the widely varying skills, needs, and attitudes of different individuals. Nevertheless, federal requirements encourage and sometimes require 'instruction in the responsible conduct of research in National Research Service Award institutional training grants' (NIH, 1989, 1992), and mentors have an important role in delivering instruction in the responsible conduct of research.

A number of professional societies and journals have also published guidelines that address various aspects of collaborations. In 1995, the American Academy of Microbiology published a document summarizing many of the important issues in collaborations plus suggested guidelines for successful collaboration (Macrina et al., 1995). Another report, with a focus on universities and industry, makes a variety of suggestions about how to overcome the existing barriers to collaboration (National Academy of Sciences etc., 1999).

  • Anderson MS, Louis KS, Earle J (1994): Disciplinary and departmental effects on observations of faculty and graduate student misconduct. Journal of Higher Education 65: 331-350.
  • Brown S, Kalichman MW (1998): Effects of training in the responsible conduct of research: A survey of graduate students in experimental science. Science and Engineering Ethics 4: 487-498.
  • Douglas-Vidas J, Ferraro A, Reichman M (2001): Analysis of Guidelines for the Conduct of Research Adopted by Medical Schools or Their Components(PDF). Published on line by the USPHS Office of Research Integrity
  • Committee on Science, Engineering, and Public Policy (2000): Enhancing the Postdoctoral Experience for Scientists and Engineers: A Guide for Postdoctoral Scholars, Advisors, Institutions, Funding Organizations, and Disciplinary Societies, National Academy Press, Washington, DC.
  • Eastwood S, Derish P, Leash E, Ordway S (1996): Ethical issues in biomedical research: Perceptions and practices of postdoctoral research fellows responding to a survey. Science and Engineering Ethics 2: 89-114.
  • Institute of Medicine (1989): The Responsible Conduct of Research in the Health Sciences. National Academy Press, Washington, DC.
  • NIH (1989): Training grant requirement. NIH Guide for Grants and Contracts 18(45).
  • NIH (1992): Reminder and update: Requirement for instruction in the responsible conduct of research in National Research Service Award institutional training grants. NIH Guide for Grants and Contracts 21(43)
  • Swazey JP, Anderson MS (1996): Mentors, advisors, and role models in graduate and professional education. Association of Academic Health Centers, Washington, DC.