In part 3 of this series , our panel of audacious scientists shared their strategies for funding risky research. They had two main suggestions: First, maintain a balanced portfolio combining high-risk, high-reward projects with safer, more incremental work. Second, when proposing ambitious work, focus on the incremental next step, de-emphasize the risk in the proposal, and do the bolder parts on the side.
The first suggestion is fairly obvious -- a sound strategy for pretty much anything: career management, financial investments, or whatever. The second is less obvious -- and possibly even troubling for scientists and, especially, the funding agencies. Vilayanur Ramachandran, a behavioral neurologist and director of the Center for Brain and Cognition at the University of California (UC), San Diego, advised spending as much as 20% of grant money "on other exciting stuff as long as you tell them it's vaguely related." Johns Hopkins neuroscientist Solomon Snyder's preferred technique is to write, "We discovered x, and now we're going to do the boring follow-up of x" -- and then use the grant to do riskier work that really excites him.
That's fine if you're a scientific legend, but what if you're just getting started? Is this still good advice?
The question gets at the nature of the relationship between scientists and the sources of their funding -- a relationship that isn't always as clear as it should be. Scientists and institutions funded by the National Institutes of Health (NIH) and the National Science Foundation (NSF) are bound to specific -- if not widely known and understood -- terms and conditions. Many scientists we've heard from believe they're expected to stick closely to the work described in their proposal. Others -- like our audacious panel -- obviously believe otherwise.
Which view is correct? Might there be appreciable leeway, as our luminaries suggest? What are the expectations of funding agencies? Is a grant proposal a contract, or is it more like an audition, an opportunity to show that you've got what it takes to do good work? Are you bound to the work you outlined, or can you deviate substantially from your stated goals?
When we asked NIH whether it was okay to deviate from the research plan, and by how much, Sally Rockey, NIH's acting deputy director for extramural research, directed us immediately to NIH's Grants Policy Statement. A grant, Rockey writes by e-mail, is neither contract nor audition. Instead, it "is a Federal Assistance mechanism." She continues:
The NIH Grants Policy Statement indicates that in general the PD/PI may make changes in the methodology, approach, or other aspects of the project objectives. However, the grantee must obtain prior written approval from NIH for a change in the direction, type of research or training, or other areas that constitute a significant change from the aims, objectives, or purposes of the proposed project (hereafter "change in scope"). Potential indicators of a change in scope (that require NIH prior approval) include, but are not limited to:
- Change in the specific aims approved at time of award.
- Substitution of one animal model for another.
- Any change from the approved use of animals or human subjects.
- Shift of the research emphasis from one disease area to another.
- Application of a new technology, e.g., changing assays from those approved to a different type of assay.
"The requirement for prior approval allows NIH to provide investigators the flexibility needed to maximize the discovery potential of the project while making sure there aren't ethical issues to the new direction (with new animal or human subjects research); it remains within the scope of the peer-reviewed project, and NIH can assure that the new plan is fully thought out and feasible within the remaining grant period," Rockey writes. "There is quite broad latitude to allow science to progress in a direction unanticipated at the onset of the project and without NIH approval."
Here's the key section of NSF's official policy:
Neither the phenomena under study nor the objectives of the project stated in the proposal or agreed modifications thereto should be changed without prior NSF approval. Such changes should be proposed to the cognizant NSF Program Officer by the PI/PD. If approved by NSF, the Grants Officer will amend the grant.
311.2 Changes in Methodology
NSF believes that the PI/PD, operating within the established policies of the grantee organization, should feel free to pursue interesting and important leads that may arise during the conduct of a research (or other grant-supported) project or to adopt an alternative approach which appears to be a more promising means of achieving the objectives of the project. Significant changes in methods or procedures should be reported to appropriate grantee official(s) and the cognizant NSF Program Officer.
Although those rules sound constraining, Jeff Nesbit, NSF's director of the Office of Legislative and Public Affairs, emphasizes flexibility. "This concept of taking money and then doing something else entirely -- there actually is some latitude in NSF grants to do that," Nesbit says. "Up to a certain extent, ... you're not going to be penalized for doing it. There is some accountability in that you've got to work with your program officer. But as long as you're straightforward in what you're doing, then, generally speaking, it's fine.
"Certain goals and measurements (like when and how money must be spent and accounted for) are fixed," Nesbit continues. "But if, in the course of their research, they need to change direction to pursue an area of discovery shown by the research, a researcher obviously needs to have the flexibility to do so. It wouldn't make sense -- for the agency or the researcher -- not to pursue a promising area of discovery when research indicates it.
"If it is a substantial change, far beyond what was originally proposed, then they should discuss it with their program officer," Nesbit says. "But, in the end, it's the researcher and the university that provides a final project report outlining how the research was conducted, and what resulted."
Even if it's okay to propose one thing and do another -- is it a good idea? After all, why would anyone even consider emphasizing the boring stuff when there's exiting work to be done? The reason, say our audacious scientists, is that NIH's culture is conservative. Committees are charged with making sure poorly conceived science does not get funded -- and they do a good job of that, says Steven McKnight, a biochemist at the University of Texas Southwestern Medical Center at Dallas and a National Academies member. "But committees also tend to delete the really bold and crazy ideas that might provide huge, incremental jumps. Because all it takes is one person on a committee to say, 'Oh, that won't work,' and the grant doesn't get funded."
"I know of several good scientists, with outstanding track records, who have submitted proposals recently [to NIH] with several mundane aims and one really novel and exciting aim. In every case, they were told to take the novel aim out, that it was too speculative, or too ambitious, or too unproven," biochemist Gregory Petsko of Brandeis University in Waltham, Massachusetts, writes in an e-mail to Science Careers. Petsko wrote a column last July in Genome Biology exhorting colleagues to stop "trashing research proposals from beginning investigators as 'overly ambitious' '' and accusing senior colleagues that do of being "nit-picking, supercilious jerks."
"I think many program officers actually recognize novel and exciting science when they see it but are reluctant to go against the recommendations of the reviewing panels," Petsko writes. "This isn't somebody else's problem; we are the reviewers and we are the ones who can change the system, by rewarding boldness instead of shying away from it."
Keith Yamamoto, a molecular biologist at UC San Francisco who co-chairs NIH's TR01 study section, agrees that there's a problem at NIH. When it comes to securing grants, the best advice is to "write conservatively and act boldly," he says. "It's terrible, but it's the reality of the day." And that conservatism, he says, feeds on itself: "The message has gotten through not to include bold or novel ideas in applications," he says. "So, there really aren't [many bold ideas] to fund."
Some people think it's getting worse. "There's no question that the whole system has become much more conservative over my lifetime," says David Botstein, a geneticist at Princeton University who co-chairs NIH's TR01 study section with Yamamoto. "The regular study section is risk averse for a very simple, perfectly human reason: It's just easier to see how an incremental improvement in existing knowledge and an experiment which is very, very obviously feasible is a better use of the money to the average study section member than something which requires any kind of leap of faith or which is insufficiently justified, in their view. It's just conservatism, and it's real. When people feel there isn't enough money, the first thing they want to do is make sure that it isn't wasted."
A senior NIH official, who didn't want to be identified, expresses a similar view, speculating that tighter paylines increase the focus on outputs and deliverables. He points, too, to the increased linkage of research funding to jobs and subsistence as factors that promote conservatism -- not just among reviewers, but in the community as a whole. Over the years, the move toward fulfilling promises and ensuring deliverables has sacrificed the willingness to explore, he says.
That tendency to safeguard funds is alive and well at NSF, says Richard Behnke, section head and acting division director for NSF's Division of Atmospheric and Geospace Sciences. "When your funding rate is low, I think everybody kind of retreats a little bit and is less willing to take risks. And that maybe is a problem."
Yet "NSF truly realizes that there is a perception out there that we don't do innovative research, or we don't do enough of it," Behnke says. "We don't want people having to game the system -- 'we'll do the incremental, non-transformative research on your dime, and we'll do the really interesting stuff when nobody's looking, because you guys aren't capable of realizing it.' That's just not what we want to happen."
Indeed, in recent years, NSF seems to have embraced risk in a big way, at least in their rhetoric. "If it's 'safe science,' NSF should not fund it," Director Arden Bement Jr. said 3 years ago in a speech  at The Academy of Medicine, Engineering and Science of Texas, continuing, "If there are no big unanswered questions in a proposal, NSF should pass it up." At other venues, Bement has talked often about "dogging the frontiers."
NSF's cites its own anecdotal evidence -- which, ironically, takes the form of rejected proposals -- suggesting that the shift toward greater risk-taking is real. Nesbit says he's heard stories about experienced researchers who "thought that because of their track record they could just make incremental changes and propose round two of what they're doing." Those proposals were declined because, he says, "they weren't taking enough risks, weren't pushing the envelope enough." It could be just rhetoric, but if it is it seems calculated to send a clear message: Send us your bold ideas.
Behnke has seen a lot of those rejections, which NSF calls 'declines.' "I probably have looked at a thousand declines in the last few years. Read them all. And the number of declines that say 'Too risky -- has a high payoff but is too risky' -- I don't see any of those, maybe one or two. The number of declines that say, 'This is just incremental -- not exciting. Good, solid proposal, but it looks like more of the same.' Tons of those declines -- lots and lots of those."
"Those guys who are out there thinking they're going to do some safe research, I think they're on the wrong track," Behnke continues. "We're not the incremental science agency."
So, should you downplay risk in your proposals? If the people we spoke to are correct, the answer is 'yes' at NIH and 'no' at NSF. In the end, it's your decision. Evaluate the evidence, assess your personal tolerance for risk, and design your research portfolio accordingly.
NIH's Rockey says there's latitude to move beyond the stated research goals but remains vague about just how much. Meanwhile, the organization's policy statement clearly states that written authorization is needed for any substantive change. Our recommendation: downplay the risk in your proposal and then cover your butt by vetting all changes with your program officer.
NSF's official rules are no less confining, but in practice, NSF-funded scientists appear to be granted much greater flexibility in following changing directions. Our recommendation: At NSF, include the new and risky directions in your proposal and when unexpected opportunities arise, call your program officer. And, of course, if you need to do incremental work, propose it. Just make sure to balance it with a bit of innovation.
Anne Sasso is a freelance writer and may be reached at