News FocusScience and the Stimulus

NSF Boosts Success Rates, But at What Price?

See allHide authors and affiliations

Science  27 Nov 2009:
Vol. 326, Issue 5957, pp. 1181-1182
DOI: 10.1126/science.326.5957.1181

Dipping into a backlog of highly rated proposals is a safe bet. But are there enough high-risk, high-reward ideas in the pot?

Hilairy Hartnett got the bad news in January. The National Science Foundation (NSF) had turned down, for the third time, her application for a CAREER grant, a prestigious 5-year award for tenure-track faculty members who want to combine their research interests with their role as educators. Hartnett's idea to study how the Colorado River takes in and transforms carbon as it flows through the region on its way to Mexico had gotten high marks from reviewers. But her scores were never quite good enough to beat out the very stiff competition.

So Hartnett, a geobiochemist up for tenure this year at Arizona State University, Tempe, decided to rework the idea and resubmit it as a regular proposal to NSF's geosciences directorate. She knew the odds weren't much better, but she had no choice: Researchers can apply only three times for a CAREER award. She also regretted losing the educational component, which would give undergraduates a chance to carry out complex environmental field studies and learn about an important natural resource in their state.

But before Hartnett had done anything, national politics came to her rescue. In May, she got a call from her program manager at NSF, asking if she was still interested in carrying out her CAREER project. Of course, she stammered. But why did it matter? What had changed?

The answer, in short, was the $787 billion American Recovery and Reinvestment Act (ARRA), which was signed into law on 17 February. NSF's slice of the so-called stimulus bill added $3 billion to its regular $6.5 billion budget for 2009: $2 billion for research activities, $900 million for three different infrastructure programs, and $100 million for a handful of specific education and training activities (see top graphic). That huge increase was based on the repeated complaint from NSF Director Arden Bement in congressional testimony that the agency received billions of dollars'worth of good research proposals each year that it couldn't fund. Although Congress specified how NSF should spend the two smaller pots of money, it bowed to NSF's reputation for quality peer review by giving it great leeway in disbursing the largest component.

Bement decided to use the recovery funds to whittle down NSF's huge backlog of highly rated proposals and raise success rates—improving the odds for applicants like Hartnett. Bement also told program officers to give priority to young investigators and to “high-risk, high-reward research.” The latter phrase is code for two important policy debates swirling around NSF. “High-risk” addresses concerns from Congress and the scientific community that NSF is too conservative in making its granting decisions. “High-reward” signifies that the agency's portfolio is relevant to important national needs, such as research on climate and energy, that are top priorities for the new Obama Administration.

Bement's priorities were a trifecta of good news for Hartnett. She is a young (albeit at age 41 hardly inexperienced) investigator. Her CAREER proposal was technically still in play. And her project explores the link between terrestrial and aquatic ecosystems, important components in understanding the effects of climate change. “I was stunned, and I'm infinitely grateful,” says Hartnett, one of many researchers whose rejection letters were turned into grants once NSF received the recovery money. Coming on the heels of a 5-year, $389,000 NSF grant that expires in March 2010, the 5-year, $573,548 CAREER award will also sustain her lab. “It's dramatically changed my situation.”

All told, NSF staffers spent an additional $165 million on the CAREER program, boosting the number of awards by 50%. (Each of NSF's seven directorates participates in the program, with the two largest, engineering and math & physical sciences, making the most awards.) What's more, some 45% of the grants funded with recovery money included at least one first-time grantee, a much higher percentage than normal and another perennial goal of NSF program managers. NSF has also calculated that 10% of the stimulus money (489 awards) has gone to energy-related research, with another 17% (800 awards) supporting climate-related research.

Cashing in.

Adding $2 billion to a $5 billion research account has given scientists a better chance of winning an award from any one of NSF's seven directorates.


Bement's policies have had a spectacular impact on success rates. The NSF-wide average for competitive awards jumped from 25% in 2008 to 32% in 2009. The geosciences directorate tops the list, with a 44% success rate (up from 31% in 2008), and the rate for NSF's largest directorate, mathematics & physical sciences, climbed to 40% from 29%. (All numbers are for the fiscal year that ends on 30 September.)

But those higher success rates may not last for long, warn some science lobbyists, if Congress fails to deliver on a promised doubling of NSF's budget over 10 years. “It's happened so many times in the past that the scientific community has come up short,” says Michael Lubell of the American Physical Society. Bement says he has faith in Congress and the Administration, pointing to what is likely to be an increase of almost 7% in 2010 and what he hopes will be a healthy 2011 request.

The “summer from hell”

Agencies are required to keep their recovery money separate from their regular appropriations so that the spending can be tracked more easily. But that doesn't mean the funds have to be spent on different programs. Some 307 of the 694 CAREER awards this year, for example, are funded with ARRA money. (That's up from 456 CAREER awards in 2008.) And NSF's engineering directorate was able to boost the number of $2 million grants it made this year for a signature program called Emerging Frontiers in Research and Innovation to 20, from 12 in 2008, by using ARRA money to finance seven of the awards.

Mindful of the rules governing the use of recovery money, Bement decreed that all the projects it supported would be fully funded from the start, rather than receiving the money over several years. But he says that his real focus was on “the fact that we had four and a half months to spend $9.5 billion.” Pushing nearly 50% more money out the door before the 30 September end of the 2009 fiscal year translated into 15-hour days for program managers and budget officers across the foundation.

On the money.

Arden Bement made the case for boosting success rates.


“It was the summer from hell,” recalls Jarvis Moyers, head of the division of atmospheric and geophysical sciences within the geosciences directorate. And despite all the extra work, Bement notes that “we didn't get a single nickel to cover our additional administrative costs.” One saving grace: NSF's decision not to hold any new competitions (except for two infrastructure programs) meant that the community didn't need to devote additional time to writing grant proposals or serving on review panels.

NSF's bigger budget has increased political expectations that the agency will place more bets on the type of high-risk, high-reward ideas that have the potential to transform the scientific landscape—and eventually pay off big for the country. One program designed to do exactly that is called EArly-concept Grants for Exploratory Research (EAGER, which until January were called Small Grants for Exploratory Research, or SGER). The grants, which allow investigators to collect preliminary data on a promising idea, are green-lighted by a program officer and don't require external review. Bement has urged all directorates to spend up to 5% of their budgets on such awards.

Program officers traditionally have resisted that suggestion to avoid further depressing already low success rates for their regular programs. And the influx of recovery money doesn't seem to have changed that mindset. Last year, for example, NSF funded 430 SGER awards, representing 0.6% of its research budget. (They are typically smaller than the usual NSF grant.) This year the combined total of the two grants rose to 502. But that's still less than 1% of the budget and far below Bement's target.

Last month a House of Representatives science panel heard suggestions on how NSF and other federal agencies can do more to foster high-risk, high-return research. Richard McCullough, vice president for research at Carnegie Mellon University in Pittsburgh, Pennsylvania, suggested that NSF create a second program to supplement EAGER “so that, once a transformative discovery occurs, we can accelerate its development.” And he pointed out that his university, despite receiving $26 million in stimulus funding and millions more overall in fiscal year 2009 from NSF, has exactly one EAGER award.

A fount of knowledge.

Hilairy Hartnett shows students how to take a water sample from an acidic hot spring in Yellowstone National Park.


Gerald Rubin, vice president for research of the Howard Hughes Medical Institution's Janelia Farms Research Campus in Ashburn, Virginia, wants NSF to go even further. “We fund people, not projects” at Hughes, Rubin explained, predicting that federal agencies would get a much bigger bang for their buck by doing likewise with a small slice of their budgets. As an example, he suggested giving $10 million over 5 years to each of the finalists for NSF's prestigious Alan T. Waterman Prize, awarded to the best U.S. scientist under 35.

Bement doesn't think much of Rubin's suggestion. “As a federal agency, we have to be a bit more prudent,” he says. “That's not to say we don't fund people. But sometimes a person who has done outstanding work gets stuck in their own paradigm. … We are always on the lookout for stars, but only those who continue to twinkle.”

Holding a new competition, like the one that National Institutes of Health conducted, would have been one way to attract fresh ideas. But that approach had two strikes against it as far as Bement is concerned: It would have depressed success rates and slowed down the process. “We felt that it was important to get as many grants out the door as possible,” he says. “So we took full advantage of what we had on hand.”

As for whether NSF's overall research portfolio is benefiting society, Jeannette Wing, head of NSF's computer and information science and engineering directorate, has a ready answer. “We funded a project at Stanford University more than a decade ago to create a digital library,” she explains. “And on the last page of the final project report, it says, ‘and we founded a company called’”

She hardly needs to finish the story. “An investment in one project at one university turned into a completely new industry that transformed society,” says Wing. “And you can't predict that.”

Navigate This Article