American Scientist

The postdoctoral experience has become integral to building a career in science. The number of postdocs in science, engineering, and mathematics in the United States has grown from fewer than 20,000 in 1980 to upward of 60,000. At the same time, the number of years a newly minted PhD seeking a tenure-track job spends in a postdoc has increased—in many fields to well over three years. The importance of the postdoc phase to a mathematician’s or scientist’s career has, on the whole, become much greater. Even though 80 percent of postdocs are at academic institutions, only one out of five landed a tenure-track job in 2012, according to a recent poll by Science’s blog Careers. The unsettling nature of this statistic resonates with my own experience as a postdoc in mathematics at Duke University. In particular, I remember facing the exhaustion of a recent PhD-writing adventure coupled with the stress of an uncertain future.

After years of intensive study, the job market is tough for recent PhDs, especially those with geographic constraints or who seek a particular type of institution. Sometimes they must take a string of postdocs before landing a permanent position. Yet, postdoctoral positions offer freedom to pursue research and practice teaching. Postdocs are fully certified to fulfill the obligations of an academic or industry position, but they are still navigating new territory, even if they have significant teaching and research experience. At this crucial juncture, it can be challenging for a postdoc’s mentor to know what kind of guidance will be most appropriate.

Most grants supporting postdocs require a written mentoring plan. The America Creating Opportunities to Meaningfully Promote Excellence in Technology, Education, and Science (COMPETES) Act, legislation passed by Congress in 2007 and meant to increase U.S. competitiveness in research, was designed to address mentoring concerns. It required that institutions applying for funding from the National Science Foundation (NSF) for a postdoc salary had to include a mentorship plan in their grant proposals. The plan must summarize the career development opportunities that a postdoc will receive in a program. I have written these plans and reviewed them for NSF as part of a panel making funding recommendations. Although it is certainly true that a proposal without a mentoring plan would not be funded, the details of the mentoring plan are not necessarily a major focus of the funding consideration. It is not clear that there would be serious consequences if a yearly grant report failed to indicate follow-through on the mentoring plan.

A lack of accountability regarding postdoctoral mentoring could be especially problematic because postdocs may not be mentored as much as graduate students, who are generally less independent, or new faculty, who are long-term investments for the institution. Mentoring takes energy, and faculty members have many pulls on their time, most of which, if neglected, have more negative repercussions than a postdoc failing to land a dream job. However, the mentoring plan requirement does communicate that grant seekers should consider best practices for postdoctoral advising. Over the course of my career from PhD to tenured faculty, I have gathered a set of such mentorship practices based on both experience and literature.

A Personal Perspective

When I look back on my postdoc at Duke, I remember it as an ideal time. After the stress of completing my PhD thesis, I was able to continue to work with my previous advisor, Michael Shearer, while collaborating with a new research mentor, Tom Witelski. I enjoyed teaching with a cadre of collegial postdoctoral buddies. I had no service obligations, and a luxurious teaching load of one course per semester.

Nevertheless, I remember wondering what level of mentoring a postdoc should expect, and I had a lot of questions about my future. I wasn’t sure whether to apply for jobs right away or wait a year. I didn’t know whether to focus my research on my graduate school topic, start a new project, or try to do both. I didn’t know how to balance my time between teaching and research. I was trying to get a sense of what an academic job might entail, but I did not attend department or faculty meetings, so I had little sense of the departmental administrative and decision-making processes. I didn’t really hear about opportunities in industry and although friends had jobs at national labs, I wasn’t sure how I might fit in at that type of institution. I also didn’t have much insight into the tenure process experienced by the junior faculty. Because we postdocs were short timers, it seemed difficult and perhaps not desirable to fully integrate us into the regular life of the department, as would happen with new tenure-track faculty members.

I soon could reflect on these issues and develop ways to help other postdocs navigate the job search. When I arrived at my new job in the mathematics department at Harvey Mudd College in 2007, my colleague Andrew Bernoff was already dreaming of establishing a teaching and research postdoctoral fellowship. Harvey Mudd College had recently received the inaugural American Mathematical Society Award for an Exemplary Program or Achievement in a Mathematics Department. It seemed like a good time to share the culture and expertise of our department with future faculty. Bernoff invited fellow department members Jon Jacobson and me to work with him on the plans. At the time, the NSF was interested in developing best practices for mentoring and had several programs focused on transition points in careers, such as Enhancing the Mathematical Science Workforce, Vertical Integration of Research and Education, and later Mentoring through Critical Transition Points.

I took on the design of the mentoring program, and turned to a Sigma Xi study led by a friend of mine, Geoff Davis, who is now a quantitative analyst at Google. At the time, this survey was the only one of its kind. Even today, there are not comprehensive data to elucidate the postdoc experience, although the National Postdoctoral Association, National Institutes of Health, American Association for the Advancement of Science, and several other organizations are working to fill the gaps. To try to determine what factors lead to success in early career stages, Sigma Xi surveyed postdocs from 2003 to 2005 at 47 institutions that employed 40 percent of the U.S. postdoc population. They evaluated four metrics of success: satisfaction, quality of the postdoc–advisor relationship, absence of postdoc–advisor conflict, and research productivity.

The analysis revealed that “overall, professional development and structured oversight had the broadest and largest impact on measures of success.” In particular, when postdocs undertook a written plan with their advisors at the outset, they submitted 23 percent more papers to peer-reviewed journals, were first author on 30 percent more papers, and wrote 25 percent more grant proposals, compared to those without a written plan. The study noted, “Teaching experiences, exposure to nonacademic careers, and training in proposal writing and project management were also associated with multiple positive outcomes.” The results resonated with me. A written plan would have helped me raise questions with my mentor at Duke, while giving him a chance to share ideas with me. I resolved to put these recommendations into practice when I had the chance.

The Postdoctoral Fellowship at Harvey Mudd College became a reality with the support of our college and the NSF. Over five years we provided two-year positions to five postdocs, all of whom obtained tenure-track positions at liberal arts colleges. We keep in contact to celebrate successes and transitions with them, as well as support them during challenges. It is exciting to see where their careers are going and to see them put their skills into practice in new environments. I also appreciate the ways their voices played a role in our department while they were part of our faculty.

The Postdoctoral Mentoring Plan

As our grant comes to a close in its fifth year, there are several aspects of our successful program that I would suggest for other faculty mentoring postdocs. Our program, designed to mentor future faculty at teaching- and research-focused undergraduate institutions similar to our own, ensures that the postdoc has an ambitious yet realistic plan, receives advice from a diversity of mentors rather than just one, is involved in faculty life, and is prepared for the job market. I have often asked our postdocs (only semi-jokingly) if they receive too much mentoring, but they don’t seem to think so.

As I mentioned earlier, studies show that a written plan helps orient a postdoc toward clear, feasible goals. Sometimes after the intensity of the thesis process, it can be hard to maintain productivity, even with a light course load. Working together to set short- and long-term research goals and meeting on a regular basis help postdocs maintain momentum. Adjusting the goals regularly can serve as both motivation and reward.

As in most programs, each of our postdocs is assigned a research mentor even before they enter the department. In addition, a teaching mentor and a professional development mentor provide new perspectives, connections, and advice.

Our postdocs teach the equivalent of one course per semester. Limiting teaching duties allows the postdoc to focus on research, while providing the strong teaching experience essential for a position at a teaching-focused college. Postdocs in our program are first assigned to teach an introductory course required for all of the students at our college. We pair the postdoc with an experienced faculty member teaching different sections of the same class. While the teaching pair discuss course planning, lectures, homework, and exams, a different faculty member not involved in the course planning or teaching observes the class lectures and discussions. That way, two faculty can write strong, detailed recommendation letters about the postdoc’s teaching qualities from different perspectives.

After the first semester, our postdocs have the option to teach more advanced courses. Often these courses are team-taught, which provides the postdoc support with the more difficult, less standard course material. The supervision also helps ensure that the course is taught with the usual rigor.

In addition to research and teaching mentors, we assign postdocs a professional development mentor who reviews their application materials, runs mock interviews, gives feedback on job talks, and discusses job negotiations. My colleague Michael Orrison noted that mentoring discussions “often focused on how best to navigate the job search. Having been a member of many search committees over the past decade, I enjoyed being able to share my thoughts about what it might mean to be a strong job candidate in the eyes of a committee composed of several mathematicians with several different measures of strength.” Our postdocs apply for a few dream jobs in their first year and if necessary conduct a more comprehensive search in the second year.

We also involve postdocs in faculty and college life, so that they better understand what it means to be a faculty member. In my postdoc I did not have a good sense of the administration of the department. In contrast, our postdocs attend the yearly department retreat, weekly department meetings, and all-college faculty meetings, at which they participate as members of our faculty. Over their time in the department I have seen them gain a voice in the decision-making process and thus make significant contributions. (Of course, in certain meetings, such as the discussion of reappointment, promotion, and tenure cases, the postdocs’ attendance is not appropriate.)

Another way to enhance mentoring is to invite postdocs to orientation and training activities that were developed for new tenure-track faculty. In a large enough institution, this group might consist of many people in a single department, but our postdocs participate in professional development with new visiting and tenure-track faculty from across the college. Recent postdoc Amanda Ruiz said, “The new faculty lunches made me feel like part of the community. They never made me feel like less than the tenure-track faculty, and I learned a lot about college structure, politics, and surviving the first few years.”

In job applications, it is beneficial for postdocs to talk about their contributions to the institution. Our postdocs have participated in departmental outreach programs and student clubs, such as the Society for the Advancement of Chicanos and Native Americans in Science (SACNAS) and the Society for Industrial and Applied Mathematics student chapter. Erin Byrne, who is now at MathWorks, took students to a rock climbing gym; Ursula Whitcher, now at University of Wisconsin–Eau Claire, took students to see rare math books in our library’s special collection; and Amanda Ruiz, now at the University of San Diego, went with students to the national SACNAS meeting.

To further involve postdocs in our institution, we encourage them to mentor one of our yearlong senior capstone experiences, and they have successfully mentored senior thesis students in one-on-one research collaborations. Our postdocs had funding to mentor two summer students, usually paired with senior faculty members and their students. They have also co-mentored in our clinic program, which provides industrial project experience for groups of about four students with a faculty mentor and industry liaison.

All aspects of our postdoc mentorship ensured that they emerged from the program with top-notch applications—compelling and detailed references, strong research, as well as in-depth teaching, mentoring, and service experiences. But the application process goes even more smoothly if a candidate has a strong network. During my postdoc, attending conferences was crucial to my successful job search, even before I was interviewed formally.

Our postdoc position includes travel funding, and we encourage our postdocs to attend conferences and visit neighboring institutions to present their research. Before they go, faculty watch their talks and provide feedback. Traditionally, this responsibility falls to the research mentor, but it can be helpful to have someone outside the field and unfamiliar with the research as a test audience member. I wonder how many graduate students or postdocs treat conferences as informal job interviews. Some seem much more confident about meeting people than others. Mentors play an important role by making introductions to people in the mathematics community.

Practices and Policies

Although postdocs are not privy to every decision in our department meetings, they begin their next jobs with firsthand experience of how a department functions. They gain perspective on faculty life from many members of the department and from colleagues across the college. There is no question that our multifaceted mentoring requires significant commitment from senior faculty. Although our postdocs were terrific when they arrived, they insist that the mentoring they received at Harvey Mudd College helped them land dream jobs and begin new positions with more confidence.

Required postdoctoral mentoring plans and Internet-based resources have helped focus faculty attention on the needs of postdocs. But ultimately it is a mentor’s responsibility not only to plan well but also to follow through with these plans. Perhaps granting agencies should consider simple mechanisms, such as a checklist in the yearly report, to encourage investigators to assign the work of mentoring carefully and follow up to see that the promises of the mentoring plan have been fulfilled.

One might argue that postdocs can make it through the job search process without faculty mentoring by supporting each other. In a September article in Nature, Paul Smaglik and colleagues report that “the guidance that researchers receive about their work—whether from superiors or coworkers—contributes to their level of satisfaction. But in Nature’s latest salary survey, most respondents gave less than glowing reviews. Just one in four say that they are happy with the amount of guidance they have received in the past year, and half say that it has had little effect.” Evidence such as the Sigma Xi survey indicates that focused faculty attention can improve the job satisfaction of faculty at the postdoc stage and beyond. I hope that further study will provide even better information about the transition between the postdoc and early career. We still do not know if some aspects of mentorship affect whether a postdoc stays or leaves academia, nor what could be done to prevent a postdoc from getting stuck without a potentially permanent position. How much mentoring is enough to be effective?

Our program ensures that the postdoc receives advice from a diversity of mentors rather than just one.

• Facebook
• Twitter

I believe that the growing awareness of postdoctoral needs coupled with policy changes have improved life for postdocs. The National Academy of Sciences, National Academy of Engineering, and Institute of Medicine have produced the book Enhancing the Postdoctoral Experience for Scientists and Engineers as a guide for structuring productive postdoctoral experiences. Its clear recommendations for the institution, postdoc, and faculty mentor resonate with many conclusions from the Sigma Xi study, such as the need to write down goals and provide periodic feedback on performance. Information about best practices is available. Now we need these ideas to be implemented consistently in postdoctoral programs around the country.

I found it tremendously rewarding to work with our postdocs on professional development and watch them navigate their early careers. I was also surprised at how much time and energy it took me to do the job. Thankfully, my department valued and acknowledged the service. I hope that this article will inspire other departments to develop a set of best practices, invest the faculty time to support their postdocs, and aid faculty to design and carry out successful mentorship plans.