Wednesday, February 16, 2011

Who's on Your Proposal?

In the past few months, I have written quite a few proposals. I am in a physical sciences field, and the agencies that can fund my work are the NSF, DOE, and several DOD agencies. These agencies vary greatly in how much (a panel of) external reviewers vs your program manager influence the selection of a proposal for funding.

I currently have 5 pending proposals, of which 4 were submitted in January and February. It occurred to me that, in terms of the proposal structure and PI manpower, my current batch of pending proposals does span pretty much all the different types that I have encountered so far. All of these come with different time commitment for the PI, as well as different levels of frustration and types of challenges in terms of preparation.

1) The single-investigator proposal. Obviously, by far the most work for the PI, but you have all the control over the idea, the timeline for execution, and every word and figure that makes it in. If a single-author proposal gets funded, the money is all yours, and usually there is a fair bit of flexibility in how you distribute it (again, this depends on the agency). If you get bad reviews, it's only you who sucks and there's no other way to go about it.

But, I love single investigator proposals mostly because I don't have to compromise. I compromise all the time, in all aspects of my personal and professional life, and it's nice to sometimes just have 100% my way. Ideally, I come up with an idea that excites me and then write a proposal. Excitement shows in the writing and I think it's important to convey this seductive element, as it helps your proposal find a champion on the panel. I think the best single-investigator proposals are those where you are literally in love with the project. I find they are certainly the most enjoyable ones to write, but it does sting twice as bad if the panel says negative things about your precious.

2) A collaborative proposal with another PI. I recently struck up a new collaboration with an experimental colleague from a nearby institution, and writing this proposal was the first concrete collaborative effort we have had together. I think it went well, despite the fact that we didn't have a lot of time and didn't really know each other before this. I think these can be very enjoyable to write, as you do get to share the workload, there is built-in feedback, but you presumably still have a fair bit of control over the outcome. That is, unless the other person is slightly more egomaniacal than the average academic.

For instance, in this experiment(collaborator)/theory(mois) proposal, I think we will get criticized that the theory is not a second, but more like a sixth fiddle to experiment, whereas this is supposed to be a collaborative proposal, with each institution getting the same amount of money. But, my collaborator really wanted things his way, so, since he's the lead, all I can do is defer. It's funny, I sent the collaborator a fair amount of text and a fair number of comments on what he had written; he said he had taken some of my comments into account, but it took me a long time to find which ones. I honestly thought at some point he had sent me the previous version of the text. It doesn't have to always be my way, especially in a collaboration, but when I tell you something seems unclear or is worded in a cumbersome way, trust me that I am actually saying it out of the best of intentions: I actually want our proposal to get funded. But I guess some people really like themselves and their writing. The proposal did turn our well, I have to say.

3) The small-group proposal (typically 3-6 people): I have had 2 of those in the latest batch of proposals, and these are, in my experience, the most common type of multi-investigator proposals. You end up having 2-3 people who do a lion's share of the writing, and then the rest contribute significantly less. I think the composition of the team is key for the success -- I have had successful previous proposals of this type, and it's always with people whom I know well and whose expertises complement one another. One of my currently pending proposals is a continuing grant to a DOD agency, with long-term collaborators, and I have high hopes for its renewal, as the program manager seems very favorably disposed. The second proposal turned out OK, but I cannot say I have high hopes for its funding; it was submitted to the NSF and I only know one of the co-PI's well. The PI did a very good job tying everything together, but I think the proposal does feel a touch half-baked...

4) The large-center, multi-investigator proposal (more than 10 and typically a few tens of PI's often organized into smaller subgroups of less than 10 people). These are a pain overall and really hard to get funded. The amount of text each PI contributes is fairly small, but it's really hard to make these into a coherent effort. People who are the subgroup leads are really screwed, because, in addition to herding cats, they are the ones befallen by the task of trying to get something coherent out of the hodge-podge of writeups from many various people, most of whom barely know each other.

These proposals are very difficult to put together -- months of meeting, brainstorming, inviting and disinviting PIs, trying to create a team from a group of near strangers (often several off-site people), where the people have complementary expertises, but where also people's preferences for alternative collaborators may have to be accommodated. These massive efforts are usually led by the Big Cheezes. You can actually spot a future Big Cheez (or a Big Cheez wannabe) as a person under the age of 50 who wants to take on such a monstrous task.

I am having high hopes for this batch of proposals, which I guess means I am not an entirely crusty, disillusioned, burned-out PI after all.

Happy proposal-writing trails to all!


pika said...

I am currently in a situtation where a grant resulting from a type 4 proposal has gone horribly wrong - due to internal politics, incompatibility of different fields and misinterpretations of coPIs of what funding agency is actually funding. I have not been involved in this during the writing stage, as I only got my current job when the grant was already in place (so I didn't really have a choice if to join or not, I had to as part of my employment). But after this, all I can say is, never again, thank you! I'll stick to proposals/grants of type 1-3 from now on.

Grumpy said...

Interesting post, thanks! Can you comment on the success rate of single PI vs 2-4 PI grant proposals? I understand that there are limited number of calls one can apply for, but still I'm surprised that my advisors haven't tried for more single PI grants from a $$ maximizng standpoint.

prodigal academic said...

Even though there is potentially more money in type 3 and 4 proposals, I have had a few multi-PI projects go horrifically wrong when I was at National Lab. I greatly prefer type 1 and 2 proposals. Less stress over the life of the project. I deal better with scientific stress than interpersonal stress.

GMP said...

Thanks for the comments!

Pika and Prodigal, my experiences with No 4 grants have also been largely negative. In addition to personality conflicts, these grants come with huge administrative hurdles -- very frequent reviews, lots of anxiety about constantly having to report something to someone. I think they are beneficial to the institution as a whole, as they come with lots of funds for facilities, but I have always found their direct monetary worth to any one PI a bit questionable (especially if, like me, you don't use the fancy facilities). One upside is that these big center grants do serve as an incubator for other grants -- there is usually money for seed projects, so you can gather preliminary data and then, with a smaller group of collaborators, get independent funding and spin off. Still, I know that most people find these efforts quite frustrating.

Grumpy, it's hard to talk about success rates solely in terms of the number of co-PI's. I think it makes more sense to think in terms of agency (e.g. NSF panel-based decisions versus DOD where knowing the program manager and his programmatic interests gives you a significant edge), and whether we are talking specific solicitations or unsolicited proposals. For intance, single PI grants for junior PI's may have relatively high funding rates, but once you are out of the junior status I don't think single-PI grants have success rates different than other grants. I think often it makes sense to collaborate since the team likely has a broader expertise than individual PI's. But if it's between a hastily cobbled up collaboration versus going out on your own, you are probably better off alone. It's about having an idea and finding an appropriate place to sell it; that's what one has to appreciate about the NSF -- there is really room for all sorts of grants on all sorts of topics. In terms of money, NSF (at least my directorate) funds at a rate of about $80K per PI per year (1 student plus a month of summer), but even for collaborative unsolicited grants you cannot get much more than $400K per 3 years. I think DOE funds typically at about $100K-150K per year per PI. DOD agencies vary and the funding amounts are usually higher. Also, for all agencies, the funding rates and annual funding amounts for specific solicitations can vary greatly.