Which grants get funded at single-digit funding rates?

Throughout the last decade, funding rates at most US funding bodies have kept declining, and they have now reached awfully low levels. These days, single-digit success rates are pretty common. At the NSF, in some competitions I’ve seen success rates of 5% or less. As a result, most US scientists are struggling to figure out how to make the best of this abysmal situation. One could view the entire granting process as a lottery, and say that at a funding rate of 5% it takes an average of 20 submissions to get one project funded. And certainly some scientists operate this way and just write grant after grant after grant. However, I’m not convinced that that’s a viable strategy. While there is certainly an aspect of randomness to grant review, and sometimes a mediocre grant gets rated much higher than it should while an excellent grant gets triaged, I don’t think that this randomness matters much at the top end. If funding rates were around 30-40% then yes, one could probably write 3-4 grants and expect one to be funded just by chance. But I will argue that the top 5-10% of grants, the ones in the currently fundable range, are fundamentally different. If your grant isn’t like one of them, it probably has a near-zero chance of being funded. And if it is, it may well have a 50% or higher chance. The trick to being funded is hence to only write grants that can score in the top 10%. Easy. (Yes, I’m being sarcastic here.)

The good news is that I think it’s relatively straightforward to recognize grants that can score in the top 10%. The bad news is that I don’t know how to write them. Or rather, I don’t know how make all the necessary things come together such that I can reliably write grants that will score in the top 10%. All I can do is write a grant and then evaluate whether it’ll be good enough. Usually—but not always—the answer is “no.” In those cases, I probably might just as well not even submit. For the small number of grants where the answer has been “yes,” my success rate hasn’t been all that bad in recent years [1].

So how do 10th-percentile-and-up grants differ from lower-scoring grants? I like to think of them as magical. They have all the required pieces coming together in just the right way. They address an important question, use a sexy study system, have just the right combination of theory and experiment, and have extensive preliminary results demonstrating feasibility. They also are well written and easy to understand even by people outside the field.

As you can see, some of these elements are not under your control. For example, if you’re a theoretician like me, you may not be able to add experiments to your proposal. You could gang up with an experimentalist, but then the two of you may not have any convincing joint preliminary data. More generally, you may have a great idea, but it may require three years of work to demonstrate feasibility and you won’t get it funded until you’ve put in those three years. Also, the study system you chose to work on, for whatever reason, may just not be as sexy as other systems. If you’re studying virus evolution using bacteriophages and somebody else is doing more or less the same work with some deadly human virus, chances are your proposal will be considered less competitive.

On the other hand, some things are under your control. You definitely can improve your writing and your grantmanship. If you’re not regularly scoring in at least the top 30%, then either the science you’re doing is absolutely no good (possible, but not probable) or you just don’t know how to write a competitive grant (more likely). In the latter case, there are resources that you can take advantage of. For example, the seminars and books offered by these people are pretty good [2]. The NIH has a page with annotated examples of high-scoring grants. I will probably also write a few blog posts about grant writing in the future, so stay tuned [3].

Nevertheless, as an individual investigator, in particular if you’re a junior one, it will always be difficult to jump over the bar separating the top 10% from the bottom 90%. Unless you’re literally the best in the world at doing whatever it is you are doing, your individual track record and preliminary results may not be good enough to compete at the highest level. In general, I think it’s easier to be successful as a team, competing for larger pots of money. If a funding body wants to establish a center or make a large collaborative grant on a certain topic, and you pull together the best people in the nation working on that topic, then your proposal will be hard to beat. This doesn’t mean that the single investigator grant is dead. It just means that you should evaluate carefully (and realistically!) where your work falls. And don’t waste your time on proposals that are not going to be fundable. If, however, you can see that you have all the right pieces in place, that you can write that magical proposal where everything comes together beautifully, then by all means go for it!

Notes

[1] Over the last 5 years, about 2-3 grants that I thought were good enough to be funded did actually get funded. That’s good enough to run a lab on. Over that same time period, I’ve probably written only 2-3 grants that I thought had a realistic chance of being funded. So that’s a near 100% success rate among the really good grants. I’ve of course also written many grants that didn’t get funded. But for all of those, by submission time I could have listed reasons why they likely wouldn’t make it into the top 10%, and they didn’t.

[2] I have no affiliation with them. I just took their training once and found it helpful.

[3] Yes, I just put myself on the same level with the leading grant coaches and the NIH. I’m still trying to figure out how this modesty things works.