Mentorship and the Research Training Experience
David H. Guston
Contemporary interest in the role of mentorship in the current research environment has two parts: one, there are concerns that, at a minimum, a trainee1 is not abused or exploited during the mentorship experience; and two, there are desires to affirm the role of the mentor in transferring both the technical and ethical aspects of good research standards and practices. Abuse of trainees may, in its extreme, constitute a form of misconduct. The absence of sound training does not in itself constitute misconduct but, over time, it may lead to an erosion of research standards and thus compromise the integrity of the research process.
This paper describes the role of mentorship in the contemporary research environment and distinguishes it from other important relationships in the training of new researchers. The paper also discusses efforts by universities to improve mentorship practices.
Central to the research training experience is the duration of the training period. The period of training usually consists of graduate school and postdoctoral training, although many scientists now begin their research careers during undergraduate and sometimes even secondary education, and most continue to learn from their colleagues throughout their careers. The formal period of graduate and postdoctoral training varies considerably from one field of study to another. In 1988, the median time to doctorate for recipients of the Ph.D. was 6.5 years. The disciplinary median varied: 5.5 years in chemistry; 5.9 years in engineering; 7.1 years in health sciences and in earth, atmospheric, and marine sciences; and 9.0 years in anthropology
David H. Guston is a research fellow at the Center for Science and International Affairs at the John F. Kennedy School of Government at Harvard University. He was the research assistant for the Panel on Scientific Responsibility and the Conduct of Research.
and sociology. The length of time to the doctorate has increased for many disciplines in recent years.2
In some fields the duration of postdoctoral training has also increased. These increases reflect the increasing technical complexity of science and engineering and the continually expanding body of knowledge that the trainee must master. Other factors may include the lack of faculty positions and the need in laboratories for cheap labor.
Formal course work required for training also varies considerably among fields and institutions. Course work can involve from 1 to 3 or 4 years of formal courses. The duration of formal course work is important in that, in addition to instruction in a particular field of science or engineering, formal courses can address specific issues in the conduct of scientific research, such as statistics and research practices. Where course work includes formal classes in statistics and allows for discussion of the appropriate use of statistical methods, training reinforces good research practice through instilling concepts of research design, formal hypothesis testing, and the application of appropriate statistical analysis. Formal courses in the ethics of professional and research conduct are now quite common in law and medical schools and are becoming common in business schools. But formal course work can, at best, merely complement the actual substance of the trainee's work, for it is on the job—in the laboratory or the field—where most of research training takes place. To a great extent, research training depends on the mentor-trainee relationship discussed below, and it takes place in the context of the research work itself.
THE ROLE OF THE RESEARCH MENTOR
Defining the Mentoring Relationship
A mentor is defined as that person directly responsible for the professional development of a research trainee. Professional development includes both the explicit conduct of scientific research (e.g., instrument use, research design, observational technique, and theoretical or cognitive frameworks) and the implicit development of scientific standards (e.g., selection of research questions and data, authorship practices, and norms of communication, interpretation, and judgment). It applies to all levels of professional development, from undergraduate work to junior faculty positions, although the focus here is on graduate students and postdoctoral fellows.
It is important to realize what the mentor role is not. A mentor is not simply a patron who provides financial and other material support
(such as laboratory space and equipment, reagents, and so on), because the mentor should provide personal and professional support beyond patronage. Nor is a mentor simply an advisor who provides formal links between a student and the department or institution. The mentor is not just a supervisor who oversees the student's dissertation, because the mentor is responsible for professional development of the trainee in areas not immediately pertinent to the curriculum or dissertation. Furthermore, the mentor is not merely a role model, because the latter can influence a trainee indirectly, unknowingly, or from a distance.3 Although trainees may frequently have mentors in their patrons, advisors, and supervisors (and sometimes one person fills all four roles), there is no necessary connection among them.
The Importance of the Mentor
The research literature generally supports the conventional wisdom that the mentoring relationship is a valuable one. Although the empirical evidence is ambiguous and contradictory in places, research studies on mentorship suggest that a mentor is an asset to the professional life of the young scientist or engineer.4 For example, the productivity of graduate students with mentors may be greater than the productivity of those without mentors.5 Scientists with mentors may be more "self-actualized" than those without.6 Junior faculty with mentors may publish more books, receive more grants, and serve as leaders in more organizations than those without mentors. Academics who find mentors earlier in their careers tend to outperform their colleagues who find mentors later.7 The prestige of mentors also influences the prestige of academic appointments for trainees,8 but the mentors' teaching and promotion of the trainees are not as important as a record of collaboration.9
The potential contributions and harms of mentorship become clear when it is realized that mentoring is a replicative phenomenon; what happens in one relationship between mentor and trainee may be reproduced when that trainee becomes a mentor.10
Characterizing the Ideal Mentoring Relationship
The mentoring relationship is a unique and important one in academia, combining elements of other relationships, such as parenting, coaching, and guildmastering. One mentor has written that his "research group is like an extended family or small tribe, dependent on one
another, but led by the mentor, who acts as their consultant, critic, judge, advisor, and scientific father.''11 Another mentor speaks of trainees who have lost their mentors by death, job changes, or in other manners as "orphaned graduate students."12 Others see "[g]raduate students … as apprentices [who] begin to work gradually over time … [and] become journeymen and eventually graduate and become masters themselves."13
Research studies on mentoring in science and engineering are sparse compared to those in professions that have a major social component, such as business, nursing, and education.14 It is generally recognized that mentors transmit both technical and professional skills, regardless of their field.15 The continuity and community of practice provided by mentoring is vital for good science, particularly in a profession whose authority is traditional and whose decision making is highly individualized.
The research mentor is believed to exercise a fundamental role in shaping the career development of the trainee. If the trainee shows promise of excellence, he or she may become a successful protege to the senior figure. If a trainee is struggling, the personal care of a mentor may rescue a productive scientific career that could otherwise have been lost.
Another generalized role for a mentor is to assure that work conducted under his or her supervision is completed not only in a sound and honest manner, but in a timely manner as well. This role is especially important if the supervised work is progressing toward a Ph.D. Given the trend of increasing time to the doctorate,16 the role of the research mentor in assisting a trainee to select and complete a challenging yet manageable dissertation is among the most important a mentor can play.17
What is not certain, however, is the set of practices that distinguish good mentorship practices from those that are inappropriate or unacceptable. Because of the complexity and diversity of roles assigned to the mentor, some individuals may excel in providing certain kinds of guidance while neglecting others. For example, some mentors may be extremely resourceful in providing sources of patronage or assistance in securing professional opportunities while failing to maintain personal supervision or regular review of the work of their trainees. Others may provide more immediate guidance and be more accessible for their students, but are limited in their abilities to provide the economic resources or professional advancement that may be critical for young investigators.
Ideally, one would expect to have mentors who successfully perform in a variety of categories, and indeed many do. There are concerns,
however, about situations in which the mentor abuses a relationship with the trainee in a manner that violates fundamental standards of professional integrity, and situations that prevent the mentor from providing adequate training and guidance. These situations may result from personal factors such as emotional stresses, substance abuse, or discriminatory practices. Or they may result from environmental factors that foster a climate in which mentoring becomes a secondary or tertiary responsibility.
Whatever the source, damage can result from poor mentorship practices, whether abusive or neglectful. Such inappropriate practices need to be identified and corrected at the earliest possible moment, with a regard for the privacy of the involved parties. Institutional officers should make efforts to establish an appropriate climate within the research setting that encourages research collaboration and educational training and fosters ties between mentors and trainees that go beyond a formal employer-employee relationship. This climate should also encourage the identification of poor practices at an early stage. The climate fostered by the institution should also encourage a broader mentor-trainee environment that becomes important if some unanticipated event—such as the death of a mentor, or an instance of abuse or misconduct—disturbs the relationship. A broader environment would provide necessary support, both emotional and material, to the trainee under such circumstances from the resources of the department or institution.
Student descriptions of the characteristics of good mentors reinforce the idea of mentoring as a complex social relationship.18 And mentors tend to confirm beliefs about the central importance of their social and personal characteristics.19 Even with respect to practices within the laboratory, rather than within the departmental, professional, or extracurricular lives of their students, good mentors contribute to the personal, social, and creative decisions of their students. Snyder, for example, emphasizes the transmission of creativity in the laboratory in experimental design and choice of research direction—rather than experimental technique and instrument competence—as the primary focus of good mentoring.20
The ideal mentor will assist the trainee in pursuit of career goals and in the acquisition of the requisite technical, professional, and social skills for conducting research in a particular field. The ideal mentor challenges the trainee, spurring the trainee to higher scientific achievement. The ideal mentor helps the trainee navigate the difficult course of doctoral and postdoctoral education and helps socialize the trainee into the community of scientists, without taking advantage of a position of institutional superiority.
The ideal trainee reciprocates by adding continuity to the mentor's work and providing the creativity and impetus to learning inspired by the young mind. The ideal trainee performs in the laboratory and classroom, and labors to establish scientific credentials for the benefit of the mentor and the scientific community, as well as for him or herself. The ideal trainee does not abuse the extension of trust from the mentor and does not undermine the mentor's legitimate authority.
THE RESEARCH TRAINING EXPERIENCE
Despite the attractive qualities of good mentoring, the realities may not always incorporate the ideals. Mentoring is "a complicated relationship. … It can be very good or it cannot work so well"; the question is, "are the problems systemic or are the problemsidiosyncratic?"21 This section addresses the research training experience and identifies points of conflict between ideal and real mentoring with a view to determining whether the problems are the result of systemic flaws or individual and idiosyncratic faults.
Market of Mentors
In the world of ideal mentorship, mentors and trainees might find each other through some open market in which each could select the other with an eye toward scientific merit, intellectual and personal compatibility, and other relevant criteria. An open market of mentorship would reduce the abuse of inequalities in the relationship, because trainees could reject unfair or exploitative mentors in favor of others available on the market. Likewise, mentors could select graduate students who best fulfill their obligations and perform their research.
Opinions vary over the extent to which the mentorship market actually works in science and engineering training. Perfect markets operate only under conditions including, among others, clear and available information and unconstrained expression of preference. Some observers believe students select mentors on the basis of informed choices. Others see the choice of mentors as a "random" event predicated on the difficulties of assessing faculty credentials, talents, and reputations in a new department.22
Expression of preferences may be constrained as well by access to information among the faculty. Faculty must usually choose from the pool of graduate students available in their department and therefore rely on their department's admissions criteria. Rivalry for graduate
positions, particularly research fellowships and assistantships, is highly competitive. Competition for postdoctoral positions is similar.
From the trainee's perspective, however, the situation may be more constrained. Students may be forced to choose a mentor very early in their professional training because of funding considerations. Thus, because the trainee is constrained by funding, he or she may choose to work with a faculty member solely on criteria of patronage. Because not all potential mentors have the resources to support graduate students or postdocs, the market for mentors may not be perfect. Furthermore, in departments or disciplines where faculty advisors are assigned to graduate students, there are often expectations that the advisor will be that student's mentor. Without careful consideration of common research interests and perhaps even personal characteristics in the assignment of advisors to trainees (or vice versa), and without clarifying the expectation among all parties that the advisor should do more than merely sign course schedules, the expectation that the formal advisor will also be a mentor may be dashed.
The financial and administrative factors that affect the mobility of graduate students and postdocs therefore become important issues.23 One possible method to perfect the mentorship market is to increase the share of portable fellowships, assigned to students themselves, or training grants, assigned to departments, rather than research assistantships assigned to laboratories. Training grants may be preferable because they relieve junior graduate students of the burden of setting a research agenda and applying for funds, just at the time the student should be freer to consider research options. Such a recommendation does not necessarily mean increasing spending, but merely changing current funding practices to offset what has been a trend toward nonfungible support in the form of research assistantships.24
Conflict of Roles
A second issue in the reality of mentorship is the potential conflict of roles or disincentives for researchers to perform well in the mentoring role. Although the strength of the U.S. system in linking education and research has been recognized and admired for nearly three-quarters of a century,25 the conflicts between teaching and performing research are becoming more apparent. In particular, faculty members at research universities that train most graduate students are usually not directly rewarded in their career advancement for their graduate teaching or training skills. They may receive indirect rewards from the contribution
of well-trained graduate students to their research and enhancement of reputation owing to their students excelling elsewhere in the system, but these rewards may not be significant in tenure or promotion decisions.
Another conflict of roles in mentoring may occur between the demands of maintaining a laboratory in the contemporary research environment and the need to provide appropriate attention to trainees. It has been suggested that the rise of "scientific managers" heading large laboratories and "riding the circuit" in order to fulfill other professional obligations and solicit necessary funding has fostered an environment where trainees may not receive the full benefit of experienced and personal mentorship at the laboratory bench. Although laboratory heads may fail to participate in the everyday workings of the laboratory for the most beneficent of reasons—finding funds to support graduate students, for example—their inattention or benign neglect may not serve their trainees' education. Sometimes non-faculty postdocs or researchers fill the hierarchical gap left by the faculty member's absence from the laboratory (a phenomenon described as "surrogate mentorship" or ''mentor displacement"). This substitution may be problematic in that non-faculty, outside the tenure system and loosely connected with their institution, have even less incentive than faculty to address the educational and other aspects of mentoring not immediately connected to research productivity.
Size of Research Groups
The development of big science may create a laboratory atmosphere that requires more consistent attention to good mentorship practices. As the size of research laboratories expands, even for the beneficent cause of providing for trainees, the quality of the training environment may decline.26 In the highly competitive contemporary environment, laboratory heads may be tempted to make research decisions for the good of the team, rather than for the best educational interests of the trainees, and to use trainees for the instrumental pursuit of a predetermined research goal. "The advisor does harm to the student if he uses him in the laboratory as a pair of hands on a fixed piece of equipment or as a computer algorithm for a theoretical thesis."27 Under current circumstances, graduate students risk becoming "indentured servants."28
In experimental sciences such as high-energy physics, the research agenda has fostered an increasing size of research groups, with a concomitant impact on questions related to graduate student training. The same may be true of some aspects of biology, where broad,
interdisciplinary studies—occasionally involving large groups of 40 to 100 or more—are commonly carried out by collaborative arrangement under a single investigator. As only one member of a large team, working on a complex instrument, the graduate student's work may become ever more specialized. Such changes raise questions about expectations of appropriate doctoral research; for example, does machine design or improvement qualify as dissertation-level research? Size and specialization in the laboratory also raise such issues as how varied an experience mentors are obliged to provide their students, and other issues—such as who counts as an author and who is entitled to access to research data—that are especially sensitive where trainees are concerned.
It may even be possible for the research goals themselves to be in conflict with the best educational and training interests of the trainee. Graduate students and postdocs, although generally receptive to funding from industry, may recognize the risks involved to free and open communication of basic research. Although industrial ties for trainees may help them adapt to the realities of the contemporary research environment, such ties may bind them prematurely to an industrial culture not completely appropriate for an educational environment.
Personalities and Gender
Regardless of environmental pressures from big science, industrial science, or competitive science, mentorship is still a relationship between two individuals, and much relies on the personalities and compatibility of the two. Just as parent-child relationships can turn sour, so too can mentor-trainee relationships.
One way in which the mentoring relationship can turn bad is that instead of training an apprentice to be independent in the pursuit of a unique research question, a mentor may engage in "cloning," trying to reproduce an exact copy, in the scholarly sense, of the mentor in the trainee.29 A mentor's attempts at cloning him or herself in the trainee can only serve to block the novelty and stifle the innovation that come from researchers entering a field of inquiry.
The scholarship on mentoring and sex differences emphasizes the personal and social factors that influence mentoring. There are disputes about whether women in academia receive the same attention and benefits as their male colleagues from a (traditionally male) mentor.30 The most solid conclusion in the literature seems to be that male students avoid female mentors, while "female students neither gravitate toward nor avoid" them.31
Some serious problems do occur in opposite-sex mentoring relationships with sexual harassment, misunderstandings, or envy from coworkers or spouses, and with the lack of ability on the part of some male participants to treat the female mentor or trainee professionally.32 Romantic liaisons between mentors and trainees, especially graduate students, are generally discouraged without being prohibited. It is not surprising that shared interests and close contact should foster such relationships. Some participants abuse these relationships, causing emotional stress as well as damaging careers. The power imbalance in such a relationship is a serious factor that deserves attention.
Because mentorship involves a relationship between a senior and junior figure, the inequalities of power and institutional standing are important features to consider in seeking to foster responsible practices. Good mentorship requires a great deal of trust on behalf of both participants. Trainees and mentors share their vocations, their theories, their aspirations, and their reputations with each other. When the relationship is mutually rewarding and supportive, there are often no reasons to dispute the allocation of credit for new discoveries, even if the credit appears to be uneven.33
When conflicts arise, however, the expectations and assumptions that govern authorship practices, ownership of intellectual property, and references and recommendations are exposed for professional—and even legal—scrutiny. Mentorship and collaborative research practices rely heavily upon implicit standards and practices that have been shaped by customs and traditions over several decades. It is only recently that these standards and practices have been called into question by individuals who feel that they have been betrayed or wronged in their professional work. In some cases, these complaints reveal an individual's unwillingness to tolerate behaviors that others may have endured. Some cases may reflect a broader social trend toward litigation and relying on the courts to resolve interpersonal disputes. In others, mentors or trainees may have felt compelled to betray the other's trust in response to new competitive pressures in the research environment.
Whatever the cause, the inequalities between mentor and trainee can exacerbate ordinary conflicts such as the distribution of credit or blame for the quality of research.34 Abuse of the relationship could even include the suppression of or retribution against whistle-blowing activities on the part of the trainee.35 The mentor need not take overt action against a challenge from a trainee for basic inequalities of the
relationship to become apparent. Mentors have warned graduate students in no uncertain terms about the power politics of science when writing that "any suggestion of malpractice becomes immediately a most serious matter … [and] it is often expedient for the establishment to find a scapegoat, and the low person in the organizational hierarchy (who else but the graduate student?) is a prime candidate."36
If difficulty between a mentor and trainee occurs, then the institutionally insecure position of the trainee can become a great liability. The trainee often lacks the prestige and the ability to move about in the department, institution, or professional community that may mediate the difficulty. Trainees are dependent on mentors for research problems, laboratory space, and references. Although students and postdocs have some procedural rights (more at public institutions), their mentors, who are likely to be tenured faculty, are far more secure in the institution. Graduate students and postdoctoral fellows frequently exist in a netherworld in which institutional obligations are unclear or not forthcoming. At some institutions, the unionization of graduate students is indicative of this uncertain status.
Naturally, disjunctures exist between the ideal mentoring relationship and actual relationships. Because the relationship is exceedingly personal, many of the deviations from the ideal are liable to be of a personal and idiosyncratic nature. But because of the institutional settings common to most mentoring relationships, systemic problems may also occur. More intense scientific competition, a proliferation of roles for researchers in academia and industry, the instrumentation requirements for scientific education and research—these are all part of the new research environment and must be confronted in the training of every new researcher.
MAKING MENTORSHIP BETTER
Recognizing that there is a disjunction between the ideal mentorship experience and the reality of training in research, many groups and institutions have begun to encourage good mentorship practices. These efforts range from issuing training guidelines and definitions of responsibilities to establishing formal evaluation programs and course work. These efforts all acknowledge the importance of explicit guidance in the ethical and social aspects of training, in addition to the technical aspects.
Some universities have written guidelines for the supervision or mentorship of trainees as part of their institutional research policy guidelines.37 Other groups or institutions have written "guidelines"38 or "checklists,"39 or have suggested "areas of concern" and "devices."40 The Harvard University, University of Michigan, Institute of Medicine, and National Institutes of Health guidelines all affirm the need for regular, personal interaction between the mentor and the trainee. They indicate that mentors may need to limit the size of their laboratories so that they are able to interact directly and frequently with all of their trainees. Although there are many ways to ensure responsible mentorship, methods that provide "continuous feedback,'' whether through formal or informal mechanisms, are apt to be most successful.41 Such practices as departmental mentorship awards (comparable to teaching or research prizes) can recognize, encourage, and enhance the mentoring relationship.
The principles outlined in the 1989 Institute of Medicine guidelines suggest "that the university has a responsibility to ensure that the size of a research unit does not outstrip the mentor's ability to maintain adequate supervision."42 Three of these four guidelines are explicit about the mentors' responsibility to provide a research atmosphere conductive to responsible and fruitful research, to ensure that the trainee is not just a technical worker, to socialize the trainee into the appropriate standards of scientific research, and to offer appraisals, assistance, and advice on research strategies, problem choice, and career prospects. Two of the four specify that the mentor should accept responsibility for all work done under his or her supervision. One of the four emphasizes thoughtfulness in managing the complexity of matching mentors with trainees and considers conflicts of interest as a potential problem area in research training.
In a recent document on the role and nature of the dissertation, the Council of Graduate Schools makes several recommendations about the graduate training and the mentoring relationship, including reaching prior, written agreement about access to data and intellectual property rights in collaborative research between mentor and trainee; increasing the availability of information about other graduate students and their faculty advisors and mentors to new graduate students to aid in their selection process; preparing handbooks for faculty and students with guidelines to clarify expectations and mutual obligations in graduate education and dissertation research; and monitoring graduate student progress more closely by departments.43
Formal Programs and Requirements
Recognizing that the mentoring relationship is important for research training, some universities or departments have inaugurated formal educational efforts to promote good mentoring.44 The National Institutes of Health and the Alcohol, Drug Abuse, and Mental Health Administration require that "all competing National Research Service Award institutional training grant applications must include a description of the formal or informal activities related to the instruction about the responsible conduct of research that will be incorporated into the proposed research training program."45 Already, such institutions as UCLA and the University of Chicago have established their responses to the new requirements.46
The UCLA Medical Science Training Program held a two-day retreat that included a total of four hours of discussions on scientific responsibility and good research practices. Students responded to questions about their experiences with regard to research conduct, listened to lectures on ethics in research, and examined case studies of questions in responsible research. The University of Chicago program presents a series of seven lectures and discussions during the course of the year addressing such topics as the government concerns over scientific integrity, human and animal subjects research, and the university's academic fraud procedures.
This review is not an exhaustive one of how institutions can and have encouraged good mentoring and the transfer of professional standards and practices from one generation of scientists to the next. But it does suggest that the closer one looks at graduate education and training in science and engineering, the more importance one attaches to the mentoring relationship. Although the relationship is a personal one between faculty and student, some institutional support is appropriate and necessary.