Confessions of a NSF Program Director

Active INFORMS member recounts shift from traditional academic career to government administrator.

By Sheldon Jacobson

Sheldon Jacobson

In the spring of 2012, I was invited to interview for the Operations Research (O.R.) program director position at the National Science Foundation (NSF). The interview (which lasted one day) consisted of meetings with other program directors and staff, as well as a presentation of my research, my vision for the program and what may be accomplished during my tenure at NSF. A few weeks later, an e-mail arrived indicating I had been chosen for the position.

NSF logo

The long line of O.R. colleagues who preceded me in this position helped to traverse the subtleties that go into working for the federal government. Of course, experiencing the government shutdown from Oct. 1-16, 2013, provided a new and unique experience that even the most seasoned NSF employees had never experienced.

Nonetheless, the preparation in shifting from a traditional academic career (of 24 years) into a government administrator appeared daunting. Some of the bookkeeping issues involved letting my department head know of the opportunity (this also including cancelling an approved sabbatical), assigning a substitute principal investigator (PI) for my NSF grants (one cannot be a PI on an NSF grant while serving at NSF) and creating a feasible plan of keeping my research group (five Ph.D. students and one former Ph.D. student visitor) moving forward. Fortunately, I have been blessed with superb colleagues, students and collaborators over the years, and this group of individuals was no exception.

NSF Rotator

As an NSF program director rotator (how we are referred to at NSF), we trade our university teaching and service responsibilities to running panels (and using such insights to recommend which proposals will be funded) and providing service within our NSF division and NSF as a whole. Given the growing number of proposals, and the shrinking amount of funds available, funding rates are typically in the low teens (dipping near single digits at times). The 2013 sequestration further exacerbated the situation.

The Operations Research program attracts an unusually large number of superb proposals, with upward of 40 percent at funding quality. However, the economic reality typically cuts this number by nearly two-thirds.

NSF rotators are given the opportunity to continue to be research active by creating an independent research/development plan (or IR/D). This allows for attendance at professional meetings (such as the INFORMS Annual Meeting), visits back to our home institution (to work with graduate students) or visits to other institutions (to collaborate on joint research activities). The funds for such travel, as well as the funds to bring people in for panels, are all taken out of the same pool of money used to support grants.

The workload at NSF centers on when proposals come in for your core program. For proposals in the Division of Civil, Mechanical and Manufacturing Innovations (CMMI), where the operations research program resides, deadlines are late July (for career proposals), Sept. 15 (for unsolicited proposals) and Feb. 15 (also for unsolicited proposals). Given that panels are typically scheduled eight to 12 weeks after these dates, life becomes very busy reading and sorting through all the proposals so as to create and run panels. Identifying an appropriate mix of panelists is a delicate balance of capturing the right areas of expertise, minimizing the number of conflicts of interest (NSF has well-defined rules for what constitutes such events), keeping everyone’s workload manageable, and holding the cost of running the panel down (every dollar saved on a panel ends up in someone’s grant at the end of the fiscal year).

Congenial Organization

In between such peak panel times, NSF is a very congenial organization at which to work. There are numerous seminars, meetings discussing new initiatives across the Foundation, and opportunities to interact with colleagues across a broad swath of fields. The intricate mix of permanent staff and rotators provides a delicate balance of maintaining institutional memory and injecting new ideas into the funding process.

Given that NSF is a federal government agency, funded by tax dollars, it is imperative that NSF be able to establish the value provided by the investment of such money. Each NSF grant awarded is an investment to identify new knowledge that brings some value to society at large. As such, a steadily increasing focus on the broader impact of each grant is of interest. In simple terms, each panelist would be asked to evaluate proposals using the question, “How will the world be better as a result of this research?”

In addition, all the funded Operations Research program grantees were asked to consider this question when describing the broader impacts of their research. To make a case for why operations research deserves a growing share of the NSF funding pie requires such information to be clearly and forcefully communicated to a broad audience. Indeed, communication is something that academics as a whole do not do particularly well. We often view our research in a very myopic manner, focusing on the narrow impact of our work within our field. PIs were encouraged to create their own “30-Second Elevator Speech,” accessible to anyone, which highlights why our field is as exciting as we know it to be.

Even the most esoteric and fundamental concepts can be communicated in a manner that makes people willing to invest their tax dollars in our research. Communication should be viewed as an important component of service to the profession. If we do not take this initiative, it will become more challenging for future generations of academics in our field to benefit from the resources available for research and discovery available from NSF and other funding agencies.

Research Risks

NSF is interested in taking research risks and investing in research ideas that can revolutionize a field. These ideas, labeled as transformative research, propose giant leaps in thinking, rather than incremental jumps in thoughts. At its core, transformative research ideas and concepts are designed to make today’s practices and modes of thinking obsolete. NSF welcomes such bold and revolutionary ideas.

Chasing “hot topics,” that often become cold by the time most people begin to work on them, is not something that NSF can afford to invest in. Although difficult to formally define, transformative research is often easy to recognize, and NSF is interested in rewarding PIs for such initiatives. Program directors are positioned to encourage and engender such bold thinking in their research community.

NSF invests in human capital. Research grants are awarded to academic institutions, and academic institutions are comprised of people (i.e., professors, students, post-docs). By investing in the creation of new knowledge by such individuals, it not only solves the problems that exist in society and the world at large, but also creates new opportunities for future research, as new problems emerge. By accelerating the resulting cycle of knowledge creation, NSF hopes to improve the nation’s economic and societal well-being, by defining new industries, catalyze economic growth and enhance global competiveness. This provides a basis for justifying the use of tax dollars to support basic and applied research, and it allows the academic research engine to continue to contribute to the well-being of society and the nation as a whole.

Great Research Proposals

Writing a research proposal is easy; writing a good proposal can be a challenge. I came across an excellent book, “From Good to Great” by Jim Collins, which chronicled common characteristics and decisions that allowed certain companies to transform into truly great organizations. These lessons apply to proposals that tend to review well at NSF. Three points are particularly noteworthy.

  1. When the PI is proposing an idea for which they are the best or second best person in the field positioned to tackle it, panels often reward such proposals with positive evaluations.
  2. When proposals are highly focused on a particular problem, with a clearly defined objective, a well communicated plan and a clear statement on the value of this research, panels often understand what the proposal aims to accomplish and rewards such proposals with positive evaluations. Collins refers to this as the hedgehog principle.
  3. Lastly, PIs who are invested in a problem such that they have established credibility in their work are also viewed favorably by panelists; Collins refers to this as the flywheel principle.

Proposals that scored well across these three metrics consistently ranked highly, while those lacking in one or more of these dimensions often went unfunded.

Taking the time to serve our research community as an NSF rotator is a sacrifice. Given that one cannot submit NSF grants while serving in this capacity, any funding that you have when beginning your assignment will likely be depleted before the end of your rotation at NSF. In addition, recruiting new graduate students is a challenge, since your visibility with such individuals is essentially nil.

On the other hand, the knowledge and insights gained by seeing the NSF funding mechanism from the inside is something you will take back with you to your home institution, to help you and your colleagues better compete for future funding. There is never a good time to take this leap, yet it is a leap definitely worth taking.

Sheldon H. Jacobson (shj@illinois.edu) is a professor of computer science in the College of Engineering at the University of Illinois at Urbana-Champaign. He served as the program director for Operations Research at the National Science Foundation from September 2012 through August 2014. He wishes to thank the program directors before him who set an example for him, and provided much information and support.

He also wishes to thank the many program directors he worked with during his time at NSF (both permanent and rotators). He particularly wants to thank Dr. Edwin Romeijn, who joined NSF at the same time as a rotator, and who helped make this experience most enjoyable. A number of other people provided feedback on earlier drafts of this article; he thanks them for their comments and input.

He also wishes to thank the many NSF support staff members who patiently provided guidance on the intricacies of the mechanics of running an NSF program. All the successes achieved were due to their efforts, while any problems that occurred were due to my own limitations.

Lastly, the thoughts expressed here are mine and mine alone, and do not reflect the views of the National Science Foundation nor the United States federal government