Transcript Document

HOW TO WRITE A SUCCESSFUL PROPOSAL
Daniel J. Jacob
THE FLOW OF RESEARCH MONEY
Federal :NSF, NASA,
EPA, DOE, NOAA, DoD…
State
Industry
Foundations
Research institutes
(NCAR, EPRI. SAO…)
Research teams
(TES, GMAO…)
Individual investigators
Arrow on right is most important for university investigators; arrows on left
may be more important for scientists in national labs (or other large labs)
THE DIFFERENT POTS OF MONEY
•
•
•
Federal government is by far the largest pot of research money
– NSF’s mandate is curiosity-driven research
• Half (!) of its atm sci budget goes to NCAR
• Atm chem pot available for universty investigators is 10-15M/y, widely
distributed through individual proposals
– Other agencies sponsor research directed at their program needs
• Some of that research is done in-house (national labs)
• Some is competed to the outside
State funding is variable. Great opportunities in CA (CARB), NY (NYSERDA),
zilch in MA
Industry, foundations are very directed in the research they fund and have in
general less $ than the feds; but they are much less bureaucratic and will
throw serious $$ at problems they consider important
HOW TO COMPETE FOR MONEY
•
•
•
•
Unsolicited proposals: that’s the way NSF mainly functions. Send in your
proposal anytime, to the program that seems to fit best
Solicited proposals: that’s the way other agencies mainly function. They
issue a ‘call for proposals’ in a certain area, with a certain pot of $ to be
competed with a deadline. Programs within agencies generally issue yearly
calls in order to maintain a stable community of investigators
Fellowship proposals: encourage young investigators at the PhD, postdoc,
or early faculty level. Selection is mostly decided by your record, reference
letters, vision expressed in proposal. They generally have low overhead
rates.
Industry, foundation: funding is more ad hoc, often uncompeted; need to
develop relationships
HOW PROPOSALS GET REVIEWED
•
•
•
•
•
Proposals are sent to program manager; at least for all govt funding the
proposals go for peer review
NSF solely uses mail-in reviews (3-6 reviewers); decision to fund is largely based
on reviewers’ ratings.
Other agencies and fellowship programs use review panels, sometimes in
combination with mail-in reviews; idea is to have a cohesive program, calibrate
across reviewers, make the best of a fixed pot of money.
Review panels typically have primary/secondary reviewers assigned to individual
proposals;
– These primary/secondary reviewers (+ any mail-in reviewers) are the ones
who will actually read your proposal; others on the panel may only read
abstract and check out your CV for evidence of quality.
– First round of discussion categorizes proposals as ‘definitely fund’, ‘try to
fund’, ‘don’t fund’.
– Add up costs of ‘definitely fund’ proposals, compare to total pot of money
available.
– If extra money is available, bring in ‘try-to-fund’ proposals; consider breadth
of program, cost-cutting opportunities…This is where most of the discussion
takes place
Typically 20-30% of submitted proposals are funded; 10-20% of fellowship
proposals are selected.
RESPONDING TO A SOLICITATION FOR PROPOSALS
•
•
•
•
Read very carefully the technical solicitation. Every word in it has been weighed. Your proposal
must fit centrally into the program. Peripheral proposals will not get funded. You have to deliver
on what the program wants.
You might think, ‘what if I want to do something else than is solicited?’ You can contact the
program manager with your idea and whether it fits in their program – it’s their job to respond. Or
you can submit your proposal to NSF, which provides no directives.
NSF proposals: important to contact the program manager before writing the proposal to
describe your idea and ask whether it fits in their program. They will generally be encouraging,
but may also point you to another program that could provide a better opportunity
Don’t be afraid to go at it alone
– You may feel that you need the gravitas of senior scientists as co-Is or collaborators in your
proposal but that’s a mistake. A small tight proposal by a new investigator with a good
pedigree enjoys favorable prejudice. A proposal with senior co-Is bloats the budget and even
if it doesn’t it suggests that the investigator can’t stand on his/her own two feet.
– Your senior colleagues can in fact help most by having you as co-I on their proposals.
– Good proposals from meritorious young investigators generally fall in the ‘try to fund’
category. Increase your odds by being low-cost so you can squeeze under the funding bar.
You can afford to be low cost because (1) you do work yourself and your salary is not too
high, (2) you have the benefit of your start-up package.
GENERAL FEATURES OF A GOOD PROPOSAL
•
•
•
Has clear and itemized objectives: shows clear vision
– Generally need a bullet list in bold font of 1-sentence objectives (1 to 4) early in the
proposal. Can be expressed as objectives, questions, or hypotheses
– Some solicitations require you to deliver a specific product; this ‘deliverable’ is then
an objective
– Explicitly state how your objectives respond to the solicitation (if appropriate & nonobvious)
Has clear and itemized description of tasks, i.e., the work to be done to achieve these
objectives
– There should normally be a ‘Task’ section mapped to each objective
– Describe your research tools and how you will apply them. Reviewers want to know
what work will actually be done to justify the budget
– Have enough technical detail to convey competence, not so much as to lose
reviewer or invite pickiness.
– Mention tools that you may use, options if things don’t work, etc.
Oozes with competence: you need to convince reviewer that you’re the right person for
the job
– Proposal should include mini-review of literature on research topic, including
discussion of your own previous accomplishments; this is usually the material that
the reviewer enjoys reading most
– Nothing beats having relevant publications to show (1) that you know what you’re
doing, (2) that you can publish. Advertise your publications.
– Most proposals are tied to general professional development and reviewers will want
to see this. Talk about how the proposed work fits into your larger-scale research
plans and interests, your career development, your teaching/outreach objectives,
etc. A proposal should be intensely personal.
TYPICAL PROPOSAL STRUCTURE
0.
Abstract
•
Should include clear statements of objectives and tasks. I generally use numbered lists of
objectives and tasks; a proposal isn’t a literary work. It is most important to be clear. Program
manager and broader review panel may only read your abstract.
1. Introduction/background
•
Start from big picture (show that you have perspective) but quickly focus on motivating what
you want to do; stay focused
•
Review of relevant literature should be exhaustive; show that you know what’s been done and
how your proposed work builds on it
2. Objectives
•
1-sentence bullet list, followed by some explanatory text if needed; this should be at the end of
intro or as a separate section right after the intro.
3. Previous accomplishments
•
Very important; advertise what you’ve done that demonstrates your expertise, and show that
money spent on you is well spent. NSF requires such a section. Include it in proposals to other
agencies too.
4. Research tools
5. Tasks
•
One task section/subsection per objective
6. Expected outcome
•
State how knowledge will be increased and the world will be a better place as a result of your
research. Why should one invest in your proposal?
•
This may include education/outreach
7. Schedule
Solicitations may require a certain proposal structure (which you should follow
scrupulously) which generally is some variant of the above.
TARGETING THE REVIEWER
•
•
•
•
•
•
As with a presentation, it is crucial that you know your audience and pitch your message accordingly
Expect the reviewer to be a specialist in your field, so write to a very high level of expertise (you must
convince the expert that what you’re proposing is new and worthwhile); if reviewer can’t follow the
technical content they will recognize it as their problem.
At the same time, you need to put your work in perspective and show broad vision
– A non-expert reviewer will expect to at least understand why you want to do your work, how it fits
in the broader future of the field, and the general lines of what you want to do
– An expert reviewer will also expect that you place your work in perspective
Have enough technical detail to convince reviewer that you know what you’re doing and that your
tasks make sense.
– Avoid controversy; don’t give rope to hang yourself.
– Avoid vague statements not backed up by evidence; reviewer will conclude you don’t know what
you’re talking about. Proposal should ooze expertise!
– Your tasks should be clear about the first steps of your proposed research. Beyond that, it’s
perfectly OK to be intelligently tentative. If plan A doesn’t work, you will try Plan B, etc. This is
research after all.
Try to have a mix of objectives/tasks on which you can clearly deliver and others that you recognize
as risky
– Program managers have to account for the dollars they dole out – a proposal with risk of zero
return has little chance of funding
– Reviewers put a lot of stock on whether the proposer can actually deliver on the work, which
tends to discourage risk-taking (this is a known problem). Fellowship and foundation funding are
more amenable to risk-taking
– It may however be a good idea for the proposal to have some element of risk and identified as
such.
Repetitions of important points, bullet and number lists, bold text are OK and may be important to
ensure that the reviewer gets the message you want to convey. Don’t assume that the reviewer will
read every word in your proposal. Make sure that the important points are in their face!
SOME MYTHS ABOUT PROPOSALS
1. You can only get funded for work you’ve already done. Would be a little
ridiculous (not to mention unethical) and reviewers could see through this.
But there is some truth in that you generally can’t get funded unless you
demonstrate previous expertise, i.e., that you can do the work. And
showing preliminary results can be very valuable in that regard.
2. Modeling proposals are cheaper than experimental proposals. In either
case, most of the cost is personnel.
3. It’s nearly impossible for a young investigator to break in. Most agencies
will try to fund young investigators but you need to be low-cost
• Agencies try to maintain some stability in their programs and
successful past investigators therefore enjoy some favorable
prejudice (these will typically occupy the ‘must-fund’ category). But
this leaves room for low-cost proposals from young investigators,
which also enjoy some favorable prejudice.
4. Personal contacts with program managers are key. The funding system is
in fact highly meritocratic. What is really ‘key’ in getting funded are prior
relevant publications.