The September issue of Amstat News featured Q&As with recipients of National Science Foundation (NSF) and National Institutes of Health (NIH) postdoctoral fellowships and program officers. This month, we focus on graduate fellowships.

The NSF, NIH, Department of Energy (DOE), Department of Defense (DOD), and other agencies offer graduate fellowships. See the more comprehensive list, Internship and Fellowship Opportunities in Science.

The National Science Foundation’s Graduate Research Fellowship Program (GRFP) has an extensive website, with tips for applying. An October 2012 article, “NSF Graduate Research Fellowship” by Meredith Berthelson and Jennifer Slimowitz Pearl, extensively discussed the GRFP. The 2014 GRFP applications are due in early November.

Because of the NSF GRFP page and the recent Amstat News article, we will focus this introduction on the NIH F31. For DOE and DOD, the primary graduate fellow programs are the following:

• Computational Science Graduate Fellowship (CSGF) (DOE)
• National Defense Science and Engineering Graduate Fellowship (DOD)
• Office of Science Graduate Fellowship Program (DOE)

The NIH graduate fellowship program is the Ruth L. Kirschstein National Research Service Awards for Individual Predoctoral Fellows (F31). Twelve institutes participate in the F31 program. Amstat News staff emailed the following four questions to the staff contact for each of the 12 programs seeking advice for potential applicants from the statistical science community:

• What advice do you have for people considering applying for the predoctoral fellowships through your institute? Do you have any advice specific to statisticians?
• Should a potential applicant discuss their proposal with you, the program officer? If so, at what stage of the process?
• What are the proposal funding rates for this program? How many are awarded annually?
• Any other advice or comments you’d like to add?

The following program directors or officers responded to our questions: Tina Gatlin, National Human Genome Research Institute (NHGRI); Dennis A. Twombly, National Institute of Child Health and Human Development (NICHD); Mark Damico, National Cancer Institute (NCI); Shawn R. Gaillard, National Institute of General Medical Sciences (NIGMS); Michael-David Kerns and Chyren Hunter, National Institute on Aging (NIA); Leslie A. Frieden, National Institute of Dental and Craniofacial Research.

We note that much of the advice received also could be applied to the NIH F32 postdoctoral research program. The F31 Diversity Fellowship, whose purpose is “to improve the diversity of the health-related research workforce,” was also mentioned often.

Based on an August 27 query of the NIH Research Portfolio Online Reporting Tools, there were 357 F31 fellowships awarded in FY12, totaling $12.5 million. The top eight sponsors were NINDS (75), NCI (70), NIMH (59), NIA (43), NIDA (33), NIAAA (32), NIDCD (27), and NIDCR (14). For the F31 diversity fellowships, 154 were awarded in FY12, totaling $5.4 million. The top sponsors were NCI (27), NIGMS (18), NINDS (17), NIAID (12), NHLBI (12), and NIMH (10). The overall F31 and F31 diversity combined proposal funding rate for FY12 was 28.5%. The next deadline is December 8 (and then April 8 and August 8).

The following are summaries and selected excerpts capturing the range of answers to the four questions.

Advice for Applying for the F31

In general, applicants are encouraged to read the F31 program announcement carefully, particularly the review criteria and eligibility requirements (U.S. citizens and permanent residents). Most institutes welcome applications from statisticians and point to the institute-specific guidance at the F31 staff contact page. For each institute, it was stressed that the application must be specific to the research mission/interests of that institute. It also was noted that all F31 applications are first processed by the NIH Center for Scientific Review and may not necessarily be directed to the targeted institute. One program officer emphasized the importance of recognizing this fellowship as a training grant with significant career development, which means the fellow must advance both intellectually and technically and will often take classes as part of the fellowship.

Other responses from the various program officers/directors follow:

• NHGRI: Priority is given to applicants who are developing methods, technologies, etc. that can be broadly applied, rather than applying existing methods to specific biomedical problems.
• NICHD: We have only supported the Diversity F31 until this year. For FY2014, we are signed onto the Parent F31, but limit the topic areas to obstetric and pediatric pharmacology and therapeutics; intellectual and developmental disabilities; fertility, infertility, and contraceptive development; medical rehabilitation research; and biology of the placenta. F31s under those subjects would have to be primarily biomedical in nature with a statistical component.
• NCI: First, read the program announcement carefully, and then read it again. For NCI, it is important that not only the research be cancer focused, but also the sponsor must be funded with cancer-focused grants. Write with the understanding that these are training awards. The training plan is as important as the research plan. Make the case, not only for you, but also for how your institution will enable your training plan. What do they provide? Make sure your sponsor(s) have a track record of training students and postdocs. Make sure your collaborators’ roles are specifically described, as well. Make sure your training plan provides professional development and exposure to facets of cancer research you haven’t had in the past.
• NIA: I [MDK] am particularly attuned to the urgent need to identify and train graduate students in the rigors of statistical science. In my experience, fellowship applicants with a manifest interest in quantitative methodology who succeed in winning research-training support from NIH prepare and submit fellowship applications in much the same way as do other successful fellowship applicants. By that, I mean they propose a highly detailed research-training plan under the supervision of a recognized expert in (this case) quantitative methodology. For NIA, that interest must be further situated within a training program and career trajectory that focuses on aging as a primary variable or process (or as an adjunctive but central interest, as in intergenerational relations, demography of aging, health economics of aging, retirement, etc.).
• NIDCR: NIDCR is focused on training individuals for future careers in oral, dental, or craniofacial health research. We welcome all research topics directly relevant to the NIDCR scientific mission and provide training for a wide range of research disciplines, including statisticians. Applicants should articulate their long-term and short-term career goals within the application.
• The primary mentor for the applicant does not necessarily need expertise in oral, dental, or craniofacial research, but there should be a co-mentor or consultant who can provide training in these areas. Potential applicants should discuss an F31 application with their mentor(s), because the training experience the applicant will receive is an important component of the F31 review.

Contact with NIH F31 Staff

The consensus here is a resounding “yes” and as early as possible. The following institutes provided this additional information:

• NHGRI: While not required, it is strongly encouraged that all potential applicants discuss their proposed project with an NHGRI staff member in the concept development stage of their application, and as early as possible. It is preferable that the initial contact be through email, with the following information: (1) three of four paragraphs about the proposed research; (2) CV; (3) name(s) of proposed mentor(s); and (4) NIH institute/center that supports proposed mentor(s) research. In that way, the initial discussion can be more informative.
• NIDCR: Program officers can answer general questions about the application process and provide specific feedback regarding whether a topic is appropriate for assignment to their NIH institute or center. It is always helpful to find this information out in advance, rather than invest time and effort in an application that is a poor fit for the sponsoring institute/center funding opportunity.
• NCI: Absolutely! You should talk with your PD early and often, if you need it. The PD is in charge of your application process until the application is made. Then, the contact person shifts to the SRO for the review. Ask the PD questions and for advice. It’s our job to help.

F31 Funding Rates

The general response here is “that depends” and, for many institutes, the number of F31s to statisticians has been small. Three institutes had additional information:

• NHGRI: [The F31 proposal funding rate] is higher than the average NIH funding rate of traditional RO1s. NHGRI recently signed on to the parent F31 (before it was just the diversity parent F31), so we do not have an average of how many we award annually. The number of new awards per year, however, we anticipate will be four or fewer.
• NCI: We award as many grants as we can, dependent on the amount of money we have to spend and the number of applications we receive in all mechanisms at NCI. We normally fund somewhere around 25% of fellowship (F) applications.
• NIDCR: The funding rates for the NIDCR F31 vary from year to year and depend on the number of applications reviewed, priority scores, and the budget in a particular fiscal year. The abstracts and specific aims of all NIH grants, including NIDCR F31s, are publically available on the NIDCR RePorter Database.

Additional F31 Advice

• NHGRI: Biostatistics can be applied in many areas, so priority is given to applicants whose mentors are supported by NHGRI.
• NIDCR: NIDCR strongly encourages students to apply for individual F31 fellowships. This provides an invaluable exposure to the NIH grant system and peer-review process.