The ASA Committee on Funded Research (CFR) hosted a webinar September 14 about the National Science Foundation (NSF) fellowships supporting graduate and postdoctoral study. Moderated by CFR member Veera Baladandayuthapani of the University of Michigan, the webinar featured Huixia Judy Wang of NSF, Claire McKay Bowen of the Urban Institute, Elizabeth Chase of the University of Michigan, and Hannah Director of the Colorado School of Mines. The webinar was timed to inform applicants for the 2021 round of awards; however, the advice offered here will also apply for the next round of applicants.

Wang, NSF statistics program officer, was the first speaker. She introduced the NSF Graduate Research Fellowship Program (GRFP) and Mathematical Science Postdoctoral Research Fellowship (MSPRF). She explained that the goals of the GRFP are two-fold:

• Support outstanding early-career individuals with the demonstrated potential to be high-achieving scientists and engineers
• Broaden participation in STEM of underrepresented groups: women; minorities; persons with disabilities; and veterans

Awardees are provided three years of support, which includes an annual stipend of $34,000 and a $12,000 cost of education allowance for tuition and fees that is paid to the institution.

To be eligible, applicants must be a US citizen or a national or permanent resident who either will be prepared to attend graduate school in the fall of the award year or has not completed more than one academic year of a graduate program in an eligible field of study. Undergraduate seniors and bachelor’s degree holders can apply more than once for the fellowship, but graduate students are only eligible to apply once.

The subfields of mathematical sciences supported include statistics, biostatistics, probability, computational and data-enabled science, and computational statistics. For 2021, NSF expects to give approximately 1,600 awards and has named three high-priority areas: artificial intelligence; quantum information science; and computationally intensive research.

The application components include a three-page statement on one’s personal and relevant background and future goals, a two-page graduate research plan statement, at least two letters of reference, and academic transcripts. The statements should have sections on intellectual merit and broader impacts.

The intellectual merit criterion, also supported by the academic records and letters, should establish the potential of the applicant to advance knowledge. Intellectual merit is assessed through balanced consideration of experiences, attributes, and academic achievements and—when considered in combination—how the applicant has demonstrated potential for significant research achievements in science, technology, engineering, and mathematics (STEM) fields and STEM education.

The broader impacts criterion is the potential of the applicant to benefit society and contribute to the achievement of specific, desired societal outcomes based on a holistic analysis of the complete application, including by personal experiences, professional experiences, educational experiences, and future plans.

Wang then turned to the MSPRF, which the NSF Division of Mathematical Sciences initiated in 1979 to support future leaders in mathematics and statistics—and research—“by facilitating their participation in postdoctoral research environments that will have maximal impact on their future scientific development.” US citizens or national or permanent residents who have not held a doctoral degree for more than two years as of January 1 of the year of the award are eligible to apply for one of two options: research fellowship and research instructorship. The awards provide $150,000 either for two years of full-time research (research fellowship) or one year of full-time research and two years of half-time research supplemented by teaching (research instructorship). Approximately 40 awards are given each year.

The MSPRF has many application components. The main components include a project summary, project description, biographical sketch, and 3–4 letters of reference. A key distinction from the GRFP is that the applicant must identify a host institution and sponsoring scientist in the proposal. This sponsoring scientist must submit a statement that addresses a mentoring plan for the fellow and how the host institution will support the proposed research and the fellow’s development in their research career.

Last, Wang discussed the NSF Mathematical Sciences Graduate Internship Program, which facilitates the participation of mathematical sciences doctoral students in internships at federal national laboratories, industry, and other approved facilities. Each year, 40 internships are funded with stipends of $1,200/week for 10 weeks in addition to travel reimbursement. US citizenship is not required for this program.

Bowen, Chase, and Director led the second half of the webinar with advice for applicants from their perspective as previous and current recipients of the GRFP and MSPRF. Bowen also shared insights from a past reviewer’s perspective. Director was supported by the GRFP and is currently a MSPRF. Chase is a current recipient of GRFP. They grouped their advice into three categories, one for both fellowships and ones specific to each program.

All Fellowships

Start early. Although the deadline may seem far away, NSF application packets take a long time to assemble. Ideally, you’ll want to finish your personal and research statements early enough to share them with your letter-writers (and others you want to review them). This means you’ll need to complete them at least a month before the actual deadline. Leave yourself plenty of time to review your application for mistakes and upload it to the NSF portal—you may encounter technological glitches or formatting requirements you forgot.

Check eligibility criteria. The program solicitation gives extensive information about eligibility for the fellowships, and it’s easy to get confused by it all. Read through the eligibility criteria carefully to make sure you actually qualify for the fellowship. In particular, pay attention to requirements on citizenship, year of study, discipline, and previous degrees.

Read all instructions and the program solicitation carefully. The NSF provides detailed instructions in the program solicitation. Although the document is extensive, it provides critical information about the submission requirements and application review criteria.

Review other online fellowship resources. Many past applicants (both successful and unsuccessful) have written about their experiences on blogs. It’s often easier to read than the program solicitation and dives into specific behind-the-scenes details. However, note that these blogs are not a replacement for reading the program solicitation itself.

Be clear with your letter writers. A recommendation letter for an NSF fellowship should differ from a recommendation letter for graduate school or a job. Your letter writers should explicitly address the NSF’s two big criteria (intellectual merits and broader impacts) and focus on your specific research project and potential as a future leader in STEM and scientific research. Provide your recommenders with copies of your personal statement, research statement, CV, and transcript and go over the NSF criteria with them to ensure they understand what their letters should focus on. You may even give them suggestions of what you’d like them to emphasize in their letter.

Audience analysis. For any funding opportunity, you should answer the following questions to help shape your application:

• What is the agency’s mission?
• How does this fellowship function within that mission?
• Who is eligible for this fellowship (i.e., Who is your competition)?

In the context of the NSF GRFP, these questions will help you figure out how you are the ideal candidate to win this opportunity.

Reviewer process. Understanding the reviewer process will help you avoid common pitfalls or other mistakes in your application. Keep in mind that reviewers are looking for losers, not winners. They have to read through a large stack of applications and the easiest way to thin it down is to find mistakes. Did the applicant follow guidelines when applying? Did the applicant address the essay prompts?

Make each sentence count. With a short page limit, you need to consider your format when telling your story. For example, when describing a research experience, consider the following essay construction:

• Thesis sentence
• Set up the experience (research goal/motivation)
• Describe what the applicant accomplished during the experience
• State the broader impact of the experience
• Explain what was learned and what the research products were from the experience (wrap it up)

For the GRFP

Focus on your potential, rather than your project. The GRFP differs from other NSF funding sources in that it prioritizes the applicant’s potential, rather than the applicant’s project. In particular, the solicitation states, “The GRFP … provides fellowships to individuals selected early in their graduate careers based on their demonstrated potential for significant research achievements in STEM or in STEM education.”

Unlike most scientific grants, you need to develop a holistic picture of yourself as an individual. Ideally, your application should present an overarching narrative of who you are as a person, your academic career, and how they tie together with your proposed research. Your proposed research takes a backseat here and should serve as one component of the story you’re telling about who you are. (There’s a reason the personal statement is longer than the research statement for the GRFP.)

If this seems daunting, it is! Start by looking over your CV to see if you can identify any recurring themes. Remember to think about service and achievements outside of professional settings and nonprofessional hobbies that may be relevant. Next, consider how these themes connect to your proposed research, and think of ways you can highlight these connections in your statements. If you’ve done this correctly, your research and personal statements should fit together like puzzle pieces, with each statement emphasizing the broader impacts and intellectual merit of you and your work.

Note that this aspect of the GRFP may make professors experienced in grant-writing less qualified to help you—they might not realize the importance of your sweeping narrative or your discussion of your volunteerism at a homeless shelter. They’d be correct … if this application weren’t for the GRFP.

Emphasize broader impacts. In our anecdotal experience, GRFP applicants who do not emphasize broader impacts will not be successful. From the solicitation: “The Broader Impacts criterion encompasses the potential to benefit society and contribute to the achievement of specific, desired societal outcomes.” This should be present in your personal statement, your research statement, your CV, and your letters of recommendation.

Focus on science, not medicine. For biostatistics applicants, the solicitation explicitly says that “clinical practice, counseling, social work, patient-oriented research, epidemiological and medical behavioral studies, outcomes research, and health services research; interventions for disease or disorder prevention, prophylaxis, diagnosis, therapy, or treatment; community, public, or global health” and more are not eligible for funding. Although it’s alright to mention some of the medical applications of your work, they should not be the center of your proposal and you should include broader scientific applications, as well.

For the MSPRF

Align your proposal with your goals. Unlike the GRFP, the MSPRF is funding specific research. After a fellowship is awarded, changes to the proposal are typically not allowed. So, take some time to design the best possible postdoc for you. Consider what skills and experiences you want to get in the next 2–3 years, what research topics might be a good launching point for your future research program, and which advisers and institutions could best support your goals.

Craft a detailed research plan. Clearly identify the main statistical and scientific questions your research will answer. Then, break your research into smaller milestones that come together to answer your main questions. Describe precisely how each milestone will be reached. Use detail to show reviewers you’ve thought about how to successfully carry out your proposal.

Write to explain, not to impress. The reviewers of your proposal are unlikely to all be statisticians, so be sure to explain the big picture. Describe the problem you’re solving and why solving it matters. A solid literature review is needed to convey how your research fits into the field and how your proposal will expand statistical knowledge. Including some equations is a good idea. However, notation takes up a lot of space, so weigh how much information each equation provides against the value of additional text or figures that could be used in its place.

Make clear why NSF should fund you and not just your research. The purpose of the MSPRF is to “support future leaders in mathematics and statistics.” So, describe yourself as someone who is on the path to a career as an NSF-funded researcher. Use the Career Plans section to describe your career goals and how receiving the fellowship will best prepare you to be an independent researcher. Use the Biographical Sketch and Past Accomplishments sections to highlight what you’ve already accomplished. Tangible research products are especially important to highlight here. Examples include published papers and preprints, produced software, and presentations at scientific meetings.

Contact the ASA CFR with any questions or suggestions for webinars or articles. View the webinar slides or recording for free.

Additional Resources

Claire Bowen’s Advice Page
Mallory Ladd’s Advice Page
Alex Lang’s Advice Page
University of Missouri’s Advice Page
Fellowships for Graduate Study
Graduate Fellows Discuss Fellowship Experiences
Post-doc Fellowships, Programs, and Opportunities