Workshop Results
That the workshop was timely and addressed a significant need was reflected in the passion with which many participants expressed their views. Workshop participants were able to identify several key areas that deserve attention and why they are important, to set goals that NSF can help advance, and to attempt to foresee potential trade-offs and unintended consequences that might result from changes in NSF policies or practices. Although participants noted possible arenas of influence for NSF, they did not recommend specific actions. Rather, by identifying the goals to be pursued and the places where NSF has leverage, the workshop developed options for NSF’s consideration. The consensus was that NSF professional staff knows best what actions are feasible and most likely to succeed in different institutional environments.
Several core themes permeated all the discussions. The most prominent was that the postdoctoral appointment must be designed to develop the recipient into an independent professional with all the skills needed to succeed and take leadership in a variety of possible careers. The second was that a postdoc has valuable advanced training and should receive compensation and benefits for it. The third was that to build the most talented and creative scientific and engineering workforce, institutions, programs, and PIs must do a better job of recruiting, supporting, and mentoring women, ethnic minorities, and persons with disabilities. The final theme was that the postdoctoral experience needs to be better documented and analyzed: Analysis should include such topics as the origins of postdocs (who they are, where they come from, and what their expectations are), details of their postdoctoral experiences, and their subsequent career paths. These data are especially necessary for several reasons:
- They are critical for evaluating the impact of any postdoctoral program.
- They are important for mounting effective strategies for recruitment, creating development plans, and fostering emergence into independent careers, especially for underrepresented populations whose needs require specific attention.
- They are essential to address the needs and expectations of foreign postdocs, who compose a substantial percentage of postdocs in the United States and the majority of postdocs in many disciplines.
The Purpose of the Postdoctorate
One theme of discussion throughout the workshop was the relationship of the training dimension of postdoctoral appointments to the other facets of preparation for their professional lives. Postdocs who are appointed by principal investigators make critical direct contributions to grant-supported research. However, the purpose of the postdoctoral experience should be broader in order to prepare the postdocs for professional careers that are not solely as faculty members at research-intensive universities. Much of the discussion therefore centered on what skills the postdoc should be acquiring and what responsibilities PIs and institutions have to ensure that postdocs acquire them.
The participants all agreed that in addition to acquiring research experience and proficiency, postdocs need to learn how to write grant proposals, manage people, develop research topics, strengthen communication skills, and develop other abilities that will help them in academic, government, industrial, or other careers. Although PIs have an obligation to serve as mentors, they often do not have the knowledge and experience necessary to guide their postdocs in all areas. Indeed, some participants felt that individuals other than the research advisor should serve as mentors. The participants agreed, however, that PIs, with the help of their institutions, need to develop the means to foster a wide range of necessary skills in their postdocs.
In seeking ways to improve the postdoctoral experience, several participants noted that it is necessary to begin by defining the postdoctorate. The agencies that provide the funds that support postdocs must ensure that program managers understand that a postdoctoral appointment is fundamentally about preparing a PhD holder to be successful in any of a variety of scientific careers. Principal investigators who hire postdocs must accept that career development is an essential component of the postdoctoral experience, and ensure that postdocs have access to mentoring and the time to pursue necessary professional development activities. If attention to professional development of the postdocs is to become a more prominent feature of NSF-funded research, proposal reviewers also will have to be instructed to examine critically the plans for PIs and their institutions to meet these expectations.
Postdoctoral Development Plans
One potentially effective way to ensure that PIs and postdocs have a clear and shared understanding of the nature of the position is to have a signed individual development plan for each postdoc. The postdoctoral experience should then be more satisfying and productive for both parties. Furthermore, grant applications requesting postdoctoral support could be expected to include an outline for these plans, including institutional resources available for career development.
The nature and culture of postdoctoral training differs among disciplines, and extensive variation can be found even within disciplines. The percentage of PhDs who become postdocs, the average length of a postdoctoral appointment, the source of funding, and where most postdocs hope to work when they finish also vary by discipline. Likewise, within disciplines there are differences between those who have fellowships and those who are supported by PI grants, between those who want to work in academia and those who seek a career in government or industry. Therefore, attempting to establish a single detailed template for what a postdoctoral appointment should entail is counterproductive. Rather, every postdoc needs to develop an individual plan with a PI, mentor, or sponsor. The resulting clarity would benefit everyone.
The workshop identified several existing models for such plans, and participants suggested that NSF could refer grant applicants to them. Development plans can specify length of the appointment; outline duties and responsibilities of the postdoc and the PI; clarify issues of authorship and intellectual property; and delineate procedures for evaluation, renewal, or termination. While details would need to be tailored to individual settings, plans can be based on a discussion between the postdoc and the PI of what should occur during the term of the appointment to meet the postdoc’s stated goals for the future. Having this discussion and agreeing on a development plan should reduce the potential for conflict, facilitate more effective time management, and result in an experience that is more satisfying and productive for postdoc and PI alike. An added benefit is that postdocs will be able to understand the management and mentoring activities that will later become their responsibility, regardless of where they work.
Participants acknowledged that development plans would have to allow for such conditions as an unexpected occurrence in the research or a change in the postdoc’s career goals. Furthermore, a system for monitoring outcomes is needed. One mechanism could be to include information on the development plan in the annual reports on the progress of their research that NSF requires from PIs, including the actual length of each postdoctoral appointment and the subsequent professional positions that postdocs assume. To gain a more complete picture, NSF could also consider requiring postdocs to report separately on their progress and satisfaction. Any additional reporting requirements must, however, be considered thoughtfully, especially the consequences if a required report is not submitted, and how the reports would be used. The last relates directly to the issue of data collection and program evaluation discussed earlier in this report.
Because diversity is a vital goal in recruiting postdocs, individual development plans need to speak to the special training, mentoring, and career goals of women, minorities, and persons with disabilities who are underrepresented in science and engineering. National needs require that the pool from which participants in the scientific and engineering enterprise are drawn embraces and attracts the most highly capable people. The NSF Engineering Research Center (ERC) and Science and Technology Center (STC) programs already require plans for broadening participation, which are subject to critical review internal and external to NSF. The NSF Minority Postdoctoral Fellowship program run by the Biological Sciences and the Social, Behavioral and Economic Sciences Directorates has a wealth of information on special minority training and mentoring needs and issues provided by minority postdocs and their mentors in proposals. These programs could be seen as an increasingly important reference for the development of plans that respond to postdocs who are U.S. citizens and members of groups underrepresented in the science and engineering enterprise.
In an extension of this concept, research proposals could be required to include a description of the institution’s plan for promoting diversity and broadening access to NSF-supported disciplines, and how postdoctoral positions will contribute to this objective (as already seen in NSF STC and ERC programs). Applicants for research support would be obliged to participate in their institutions’ diversity goals, and institutions will have to require PIs to do so. NSF could also identify and publicize those plans that are most successful.
NSF must also keep in mind the importance of postdocs who are not U.S. citizens. While their important role in sustaining the research enterprise and propelling an understanding of the global nature of science and engineering was acknowledged, their circumstances were not addressed. Nonetheless, many workshop participants did note that the needs of this population must be acknowledged along with the importance of broadening participation by U.S. citizens.
Transitions Into and Out of the Postdoctorate
Many participants observed that attention is also required both to the transition into a postdoctoral appointment from graduate study, and out of a postdoctoral appointment into an independent career. They understand that the postdoctorate must be closely integrated with graduate education and professional careers; this temporary position is, in many disciplines, an essential stepping-stone to an independent professional career.
Access to postdoctoral positions needs to be better facilitated, especially because of the decline in the number of U.S. citizens and permanent residents seeking them. It is particularly important to ease the entry of women, underrepresented minorities, and persons with disabilities into this critical transitional appointment. Participants suggested that NSF could announce all its supported postdoctoral positions on the Internet, either alone or in collaboration with professional societies that already have mechanisms for that purpose. These announcements could be complemented by linkages with other existing information sources, such as Science’s NextWave.
Once into the position, the postdoc must have the time to develop skills and investigate opportunities that will be critical when the postdoctoral appointment ends. Because preparing for this transition has often been neglected, and postdocs increasingly choose among alternative careers, several participants suggested that NSF experiment with new types of supplemental awards for which postdocs would apply directly that would permit them to spend time to acquire additional skills necessary for a professional career. Others suggested that NSF make grants directly to institutions to enable them to develop effective new mechanisms for facilitating the movement of postdocs into independent careers.
Independent of how the postdoctoral period proceeds in detail, workshop participants unanimously agreed that the cumulative term must have a fixed limit. Indeed, at least one university system was reported to have already implemented such a rule. While some that felt differences among fields may prevent establishment of a single term limit applicable to all fields, all agreed that the postdoctoral period must not be open ended. For scientists and engineers who choose to remain affiliated with a PI’s research enterprise beyond a normal postdoctoral period, a category of long-term research-support personnel could be defined that does not incorporate the educational function of the postdoctorate.
Institutions that appoint postdocs have to take responsibility to provide support services for them. Because postdocs in academic settings are usually considered to be neither students nor employees, they often are ineligible for university services. Several participants reported that a number of universities have created an office of postdoctoral services or other formal structures for addressing the needs of postdocs. Offices of this type help in securing medical benefits, provide advice and assistance to noncitizens in obtaining visas, provide information on career development, and identify individuals who provide information and guidance for PIs lacking a particular expertise or experience in guiding postdocs. The offices also serve as a resource for PIs who want to understand how best to meet their responsibilities to their postdocs, and for graduate students who have questions about what to look for and to expect in a postdoctoral appointment. One participant suggested that small institutions with only a few postdocs could collaborate to form a single office that would serve them all. A few workshop attendees proposed that NSF could facilitate creation of institutional postdoctoral services by providing seed/matching funding for new and existing postdoctoral offices and related organizational support. Clearly, the cost-benefit ratio would have to be analyzed, and sustaining institutional support committed. Nonetheless, NSF participation would strengthen the roles of these offices or services, create local advocates for postdocs, and because the effort would be leveraged by university funds, could prove to be a cost-effective way to improve the postdoctoral experience.
For postdocs working with PI support, and even those on individual fellowships, effective mentoring by the PI is often the most important resource that an institution provides. But no mentor is likely to be able to meet all of a postdoc’s needs, and many PIs lack effective mentoring skills. While a postdoctoral services office could assist PIs in addressing mentoring issues, it would be intended to strengthen the PIs’ mentoring responsibilities, not to supplant them.
New Structures and Opportunities
Acknowledgement that postdoctoral training needs to serve a more diverse purpose than the preparation of academic faculty members led the participants to consider whether a system designed primarily for a narrow purpose could be successful at achieving a broader goal. They concluded that there was good reason at least to experiment with new funding mechanisms and other structural changes to the traditional postdoctoral experience.
In the 1970s, the number of postdocs receiving fellowships was about equal to the number supported by a PI’s research grant; today, those with fellowships are outnumbered by at least ten to one. More specifically, of the approximately 6,000 postdocs supported by NSF funds, less than three percent have fellowships. Workshop participants felt that that this distribution bears reconsideration.
Fellowships have several advantages:
- They transfer more responsibility to the postdoc for directing independent research.
- They can facilitate more rapid maturation of postdocs.
- They can increase diversity because they are more open to competition.
- They can reduce conflicts with an advisor.
- They can provide portability that enables a postdoc to move to a different setting if circumstances require.
At the same time, several possible disadvantages could accrue to fellowship programs:
Fellowships tend to provide lower institutional overhead and can limit institutional flexibility.
- They may lead to a concentration of fellowship holders in only a small number of well-known laboratories, unless a cap is set on the number of fellows allowed at any single institution.
- Foreign nationals are excluded from eligibility.
- Mentoring could be less adequate because the mentor connection may be tenuous.
- Progress could be slowed if research settings are changed.
Nonetheless, as one component of a mix of postdoctoral support mechanisms, postdoctoral fellowships deserve closer NSF scrutiny. NSF should study the difference in outcomes for postdocs with fellowships compared with those who are supported on research grants. It should also explore different types of fellowships, some of which could embrace career-development components explicitly. There was general agreement that the number of fellowships should be increased, but no agreement on how much. Workshop participants recognized, of course, that the implications of creating new programs for the ability to sustain and expand existing programs needs to be thoroughly explored, especially with current and projected limitations on Federal budgets and staffing.
To broaden preparation of postdocs for careers in nonacademic settings, NSF could explore formation of partnerships with other settings such as industrial or governmental laboratories, or with college or precollege teaching institutions. Existing programs such as the Integrative Graduate Education and Research Traineeship (IGERT) program or the newly launched Discovery Corps may serve as useful models. New endeavors would have to be flexible enough to adapt to the differing needs of each discipline.
Postdoctoral Research-Initiation Grants
Because one of the goals of the postdoctorate is to foster professional independence, workshop participants suggested creating a program of grants for postdoc-initiated research. Such a program could give postdocs the experience of developing an independent research project, writing a proposal, and (if successful) completing a small research project different from that of their mentor’s funded research. For those planning a research-oriented career, the project award would be invaluable in demonstrating ability to secure research funding.
Participants noted possible downsides to such a program. The application process could require a large investment of time for a small grant. It could divert attention away from the core research project that is the focus of the postdoc’s program, and it could result in extending the postdoctoral appointment. Institutional barriers could make the program difficult to implement. Nonetheless, NSF might want to explore a pilot activity.
An alternative could be joint postdoc/mentor proposals, similar to those existing at the National Institutes of Health (NIH), which could provide the postdoc with the same experience but more guidance. This concept has similar potential downsides, and adds the risk that other pressures may limit the extent of PI mentoring or encourage the PI’s appropriation of the project. In either case, available time to carry out an independent project could be a problem for postdocs with short-term appointments.
Current and recent postdocs at the workshop emphatically stressed the need to improve standards of living. For people with doctorates, who are often older than 30, the salaries and benefits are usually significantly less than those of their peers in other professions. The problem derives in part from the postdoc’s ambiguous institutional status. Because grants are to institutions, not to individuals, grant-supported postdocs are technically institutional employees. However, in practice, such appointees are not usually afforded the same status as ordinary employees. In effect, they are unofficial employees of PIs and are often not covered by institutional salary guidelines, health benefits, family leave, paid vacation, and many other policies that govern treatment of employees. All participants agreed that if the Nation expects to attract the most capable young people to science and engineering, the postdoctorate cannot require undue sacrifice by the postdoc. Openly stated compensation policies will allow postdocs to make informed decisions about where to work and could lead to increased transparency in compensation across institutions.
Prescribing a specific salary scale for all postdocs, while a conceptually simple means to guarantee adequate compensation, is unlikely to work when market forces determine different salary scales for different institutions and disciplines. Moreover, participants acknowledged that in times of budgetary constraints, one potential result of postdoctoral salary increases could be a reduction in the number of available research grants and fellowships. Workshop participants were unable to identify straightforward means to this end, but they did discuss a number of existing models. For example, the NIH has set minimum compensation standards for postdocs who are working in NIH laboratories. Some universities, such as the University of California, have set salary and benefits standards for postdocs. Some workshop participants recommended that universities set a minimum postdoctoral salary that corresponds to a specific percentage of what a newly hired assistant professor in that field would earn. Participants did suggest that all institutions should be required to have a stated policy about postdoctoral compensation to be eligible for grant support while NSF continues to explore how best to ensure that postdocs receive fair compensation.
In addition to salaries, participants agreed that postdocs need access to health insurance and to other standard employee benefits, such as family leave. However, workshop participants were not able to specify the level of benefits or mechanisms for paying for them.
For this effort to succeed, NSF needs to avoid becoming too prescriptive or formulaic in its requirements. Institutions need the flexibility to experiment with new approaches and to capitalize on what works for their particular circumstances while still attending to the needs of postdocs.
One reason that some uncertainty exists about the postdoctorate being a springboard to a successful professional career is that the career pathways of postdocs have not been effectively tracked after their departure. Knowing what has become of those who completed postdoctorates is essential to determining the characteristics of a good postdoctoral program. The Association of American Universities has collected some data on postdocs from its member institutions, and Sigma Xi is conducting a survey of current postdocs in various academic institutions, but a comprehensive, longitudinal study has not been carried out. Collecting information on postdocs clearly falls within NSF’s mandate to gather information on the science and engineering enterprise at large, and participants noted that NSF is already doing so to a limited degree. Workshop participants felt strongly that NSF needs to take responsibility for collecting and analyzing more data about the postdoctorate.
Tracking postdocs will require substantial investment and may be difficult, particularly for noncitizens who return to their own countries. Nonetheless, attendees thought this to be an important investment because it can inform the entire postdoctoral experience and ultimately contribute to more effective use of Federal funds. Participants warned against the danger of interpreting the results simplistically by defining success narrowly as moving into a faculty position at a doctoral institution. As participants emphasized throughout, although doctoral scientists in some disciplines often begin postdoctoral training with that intention, a variety of scientific and engineering career pathways exist: among them are industry or government research and development, teaching in an undergraduate or high school setting, or doing policy analysis for a government, think tank, or professional organization. The challenge is to determine the impact of the postdoctoral period in influencing career choices. Several participants made the point that collecting data is only the first step; sophisticated analysis will be necessary to make effective use of the data.
As more information about effective postdoctoral programs and about the subsequent career paths of postdocs is acquired, it needs to be made available to NSF and other granting agencies, grantees, host institutions and, most importantly, graduate students who are considering postdoctoral training. Participants reported that information on specific topics could be found at scattered locations. Alumni networks can provide information on an individual institution; some university websites offer guidance on selecting mentors and projects, while others provide information for mentors; and various professional societies post listings of postdoctoral openings in specific disciplines.
Potential postdocs could benefit significantly if a single place to find information existed. Workshop attendees felt that NSF could ensure that at least all its own information about the postdoctorate is easy to find. If it is impractical for NSF to do so itself, the Foundation could provide support to an independent organization to maintain a postdoctoral information portal. A central site for general information on the postdoctorate could also help universities to find available information and thus allow their own postdoctoral offices to focus on information specific to their own needs. More complete information will enable potential postdocs to make better informed choices, which should lead to more satisfactory and productive postdoctoral experiences.
A general information portal could also describe evaluation of successful innovative programs in other contexts that could serve as models for more effective postdoctoral experiences. Some have been identified earlier in this report. Others include the Fellowship Program in Biological Informatics, and the ADVANCE Program for women in science and engineering.
Workshop participants agreed that enough is known about the current state of postdoctoral training to identify action in several key areas. A coordinated and sustained endeavor to improve the postdoctoral experience is required, but it must take into account the substantial differences that exist across and within disciplines. Participants identified the following possible ways that NSF can be effective, while also recognizing the need for NSF to have flexibility in developing specific changes to current policies and practices:
- Modify research proposal application and review procedures to make more explicit the expectations of the postdoctoral period, including adequate training and mentoring of the postdoc for an independent career, and overt delineation of the responsibilities of the postdoc, the research advisor, and the institution, including issues of compensation and benefits.
- Collect more information about what occurs during and after a postdoctoral appointment for the analysis of the outcomes of postdoctoral policies and practices. These results can clarify how the long-term scientific and technological interests of postdocs and the Nation are served, and help create mechanisms that address the transitions from graduate education to the postdoctorate and to independent careers.
- Disseminate information about postdoctoral openings, expectations and models for adequate training, and the array of available funding mechanisms to provide choice and flexibility in postdoctoral training and career path development.
- Create mechanisms and pathways to broaden postdoctoral opportunities for underserved populations—women, minorities, and persons with disabilities. NSF can provide incentives for shaping professional development plans that lead to the professional accomplishment and advancement of the members of these communities.
Workshop participants noted that additional issues remain, some of which were discussed earlier in this report. They also acknowledged that many of the workshop’s themes link to graduate education and to early-career professional development issues. As important as these topics are, however, they were not the focus of the workshop: that purpose was to identify positive ways for ensuring that the postdoctorate better meets the needs of today’s new PhDs and of a 21st century scientific, technological, and educational enterprise. The challenge may be daunting, but the goal is vital.
