FUND THE PEOPLE!!!!!!!

Feb 24 2014 Published by under [Education&Careers]

I have to admit, anytime someone says that we need to start funding people not projects, my suspicion meter just about craps itself. As with all "solutions to the status quo" it always turns out that the person espousing this view sees themselves as a major beneficiary of this system. The will cite HHMI is a shining example of success, of course, and conveniently forget that those HHMI peeps had to become independent before they were picked up. I may be mistaken, but I don't see a lot of Assistant Profs at the HHMI table.

I am all in favor of figuring out ways to fund good science and change the fact that I spent most of my time chasing money. ALL for it. But it's a damn waste of time to just toss out quarter-baked ideas without stopping for half a second to flex the mental muscle and see what the logical conclusion is. Why does HHMI work? Well, it identifies promising scientists and gives them the resources to get shit done. Cool! Why doesn't HHMI fund the scientific enterprise in the US? Well, for one, that's not their mission and for another, they don't have the money. And oddly, neither does NIH if it wanted to go to this model.

It's simple math. If the goal is to provide long-term stable funding so that certain people can be as creative as they like, the cost is going to be large in terms of the science that doesn't get done. Let's just take the NHLBI data that DrugMonkey is all wrought up about as an illustration of a point. These data show pretty convincingly that the science happening at the 1%ile is just as impactful as the science that gets scored at the 35%ile. Unfortunately we don't have data beyond that point, but one could hazard a guess that a tailing off doesn't happen right after the 35%ile. Let's just suppose for a second that the drop off is somewhere in the neighborhood of the 50%ile.

Current funding ranges at NIH are pretty much at the 15%ile and below at the moment. The expanse between 15 and 50 is, well, massive. So let's say we had enough NIH funds to keep the top 20% (however one would measure that) of labs comfortable. What happens to all that science that would have had the same impact but isn't happening in the right labs?

Now I realize there's no perfect correlation between the Proposal %ile and number of labs, but you get the idea. We would be forced to either give a much larger number of labs much less money or effectively cut off an enormous amount of productive impactful work. Indeed, if you're at a top institution and all you care about is your own situation, it's NBD, right? And if you're too dumb to figure out how to make it work, well then you're probably not the type of scientist that should be funded anyhow.

12 responses so far

  • DrugMonkey says:

    No, no, the solutions are obvious and simple! Clearly you are just an apologist for the status quo.

  • tom selleck says:

    I think you're missing the primary point of most of the "fund the people" proposals, which is that the investigator-based budgets would be substantially smaller than the typical R01 budget (though they could also have multiple tiers, as in Michael Eisen's proposal). The core idea is to get a subset of PIs in the pool to give up some money in return for not having to fight nearly as hard for funding. Speaking personally, I would be happy to give up a large chunk of my current R01 in return for a guarantee of future funding--and I don't doubt that many other PIs would too. The fundamental problem with the current system is not really that there is too little money, it's that the incentives are structured in such a way as to promote a winner-takes-all approach to science. Of course there are many PIs who work in fields where you simply have to apply for big-money grants because the science is expensive; but there are also many PIs who apply for big money because the incentives frankly don't support a more modest approach (e.g., as an NIH-funded PI, I have almost zero incentive to apply for anything other than R01s, even though I could still get a lot done on less money).

    It's true of course that there's no panacea, and that there would still be many problems under a partly investigator-driven funding system. But I think it's a bit uncharitable for you to imply that the people proposing people-driven science funding are simpletons who haven't thought about the math involved. If you ask around I think you'll find a pretty substantial proportion of PIs who would be happy to give up a large chunk of change in return for much more time to do research. And my (admittedly anecdotal) experience is that many of these people are actually PIs doing quite well under the current system, so it's not like they're just arguing for whatever system they think will provide the biggest pot of money.

  • proflikesubstance says:

    Okay, let's think numbers.

    NIH budget ~$30b

    Subtract 11% for the intramural peeps = $26.7b

    Assuming that all gets paid out, we'll need to lop off 1/3ish for overhead

    $17.35b

    If we were to push that entire sum into lab awards and ignore all the other mechs (which is unreasonable, but let's play) and fund labs at less than the modular limit, say $200k, that gives us almost 87,000 labs we can support at a modest level. In reality, once the non-R mechs are added in, we're probably closer to 50,000.

    Just for perspective, Harvard Med lists nearly 12,000 faculty. Yale, Texas A&M and their ilk list closer to 2000 per. So, which 20-30 university are going to have their labs modestly funded in this new system? I'm just curious.

  • tom selleck says:

    If we're going to talk numbers, let's first get the right numbers. NIH funded under 5,000 R01-equivalent awards in 2013, and under 10,000 RPGs in total. I don't know how many faculty Harvard Med lists, but I can assure you that there aren't 12,000 active NIH PIs at HMS. I would imagine the vast majority of those are clinicians with part-time affiliations.

    If your assumption is that 1 R01 = 1 active lab, it would seem that the current system is able to support much less than 50,000 labs, no?

  • proflikesubstance says:

    5000 Ro1s in 2013, but those are new awards. That's not continuing. And I'm sure that there are not 12000 PI-eligible faculty at Harvard med, but apparently there are 17,500 faculty at Texas Medical Center, so you get the idea.

  • tom selleck says:

    No, I don't get the idea. Even if we're generous and assume 5 years per new R01 award (and many are for less), that takes us to ~25,000 awards. And many of those are held by multi-R01 labs. So in practice the number of labs currently supported by NIH is considerably smaller than 50,000. If I take your numbers at face value, the implication I should presumably draw is that the current NIH system is already grossly unable to support the current crop of PIs, since there are more faculty at HMS and TMC alone than there are R01s nation-wide. But clearly, what matters is not the number of people listed on hospital websites, it's the number of actual labs depending on NIH support. That number currently appears to be substantially less than 50,000, so if anything it seems like I should take your proposed reallocation as a resounding success!

    Setting that aside though, my point was not about specific numbers, it was that it's uncharitable to suggest that people who favor a PI-based funding track haven't thought deeply about the issues (including crunching the numbers). For one thing, virtually no one that I've seen is seriously proposing completely redistributing NIH funds on a purely PI-based basis, which is what your exercise above seems to assume. All of the proposals I've seen (e.g., Michael Eisen's) are in favor of allocating a chunk of the NIH budget to a separate PI-based track. The idea is not to do away with project-based funding, but to alleviate some of the pressure by giving PIs who don't need a ton of money to opt out of that track, thereby making life easier for both themselves and the people who do want to pursue project-based funding. And it's not like this system can't demonstrably work in the real world, since many countries already have such programs to varying degrees (e.g., NSERC's Discovery program, the NWO's Veni/Vidi/Vici programs, etc.).

    Moreover, it's not even as though the tradeoff between smaller budgets and higher success rates is an essential part of the equation, because you could, for example, simply start allocating R01s based purely on past record, without consideration for proposed project, while retaining the same size budgets. This would presumably give you roughly the same distribution of awards as we currently have, while necessarily reducing the amount of work PIs have to do (since one couldn't apply for more than one award simultaneously). Note that I'm not saying that this would be a good system (it wouldn't), but just making the point that it's important to separate the basic idea that funding PIs rather than projects is a good idea from the details of any proposed implementation (which vary widely). It's not "simple math" to suppose that there isn't enough funding to go around with PI-based funding unless you make a number of strong assumptions that most people who've proposed such ideas don't actually subscribe to.

  • proflikesubstance says:

    Actually, now that I have a few minutes to hunt down some numbers, it looks like I way over estimated, according to the 2013 numbers (http://nexus.od.nih.gov/all/2014/01/10/fy2013-by-the-numbers/), so the pot would be considerably smaller.

    But that's not really the point. The idea of "track record" almost always come back around to the major issue with NIH - the enormous age skew in their funding scheme. It's already an issue, but when you start shifting to funding the people, what do you think will happen?

    Also, I love when people bring up NSERC but don't acknowledge that the average award has never exceeded $35k/yr.

    Like I said, I'm not saying we shouldn't explore options, but this is neither a simple issue, nor an easy fix, as many often seem to think. Well thought out proposals, Like Eisen's are a far better start point than the conversation that spurred the post.

  • DrugMonkey says:

    You are deranged if you think partitioning off a big chunk for the small time plodders makes it easier for the labs with higher budget demands.

  • tom selleck says:

    Right, PLS, but that was my point: it's a complicated issue, and your post made it sound as though it was simply a crazy idea to fund people over (or in addition to) projects. It's not.

    As far as NSERC goes, $35k/year goes much farther in Canada than it does in the US; PIs at Canadian universities don't need to pay for a lot of things that come out of their direct costs in the US. But even if we take the amount at face value, I'm not sure what your point really is. Are you saying that $35k/year is so little that NSERC may as well get rid of the Discovery grants? To a PI who doesn't have any other source of funding, that $35k/year can easily be difference between doing research and publishing at a slow pace and not being able to do any research at all. It doesn't seem to me like this is a bad idea at all; I suspect if NIH had a comparable program that made it relatively easy to get $50k/year, it would be pretty well received (while consuming a pretty small chunk of the budget). And note that in Canada PIs can still apply for the bigger grants from CIHR, so again, it's not like anyone has to make a career-altering choice when applying for one kind of funding vs. the other.

    DM, I'm just going to assume you're wearing your troll hat this evening. It suits you.

  • proflikesubstance says:

    Almost every time it is brought up, it is absolutely a crazy idea that someone sees at the perfect system, mostly because it will benefit them. Outside of Eisen's proposal, the only people who EVER propose this idea is the Eat The Rich fan club members who believe all science can be done on the cheap because that's how they do it. If you want a tremendous example, see if you can find the NSF is Broken forum, if it isn't closed down.

    As far as NSERC goes, $35k/year goes much farther in Canada than it does in the US; PIs at Canadian universities don't need to pay for a lot of things that come out of their direct costs in the US.

    Free health care, reagents, supplies and postdocs, right? A technician in every lab! Yes, there are a few things that Canadian scientists do not have to pay for, but the basic costs remain the same. You hiring a postdoc or tech on $35/yr?

    I suspect if NIH had a comparable program that made it relatively easy to get $50k/year, it would be pretty well received (while consuming a pretty small chunk of the budget).

    And this is where you go off the rails. Easy to get $50k/yr without consuming much budget, eh? Considering that the system currently can, in no way, support the scientists who are applying (see: 14% funding rate), I assume you think that we'll just take care of, what, half the people applying with this $50/yr business (of course, who's going to pass up easy money, right?)? NIH had over 28000 R01 applications last year, so are we writing off $700m/yr in direct costs to this program?.

    And then what? You don't think more people are going to want in if the barrier is lower? I promise you every institution has a large number of people who have tried for NIH funding in the past to little success who would jump at that new program. And of course Deans will adjust their expectations to: One base award and at least one NewR01 for tenure? So you'll have a bunch of people puttering along and, by default, significantly reduce the number of R01 awards available.

    Somehow that doesn't sound like the way to launch US science competitiveness into the next generation.

  • tom selleck says:

    Almost every time it is brought up, it is absolutely a crazy idea that someone sees at the perfect system, mostly because it will benefit them. Outside of Eisen's proposal, the only people who EVER propose this idea is the Eat The Rich fan club members who believe all science can be done on the cheap because that's how they do it. If you want a tremendous example, see if you can find the NSF is Broken forum, if it isn't closed down.

    Well, again, it's a model that works quite well in many countries, so I don't see how it's a crazy idea. I'm sorry to say that the Twitter exchange that prompted your tirade is hardly "crazy" on my reading; but perhaps you're expecting more number-crunching in 140 characters than is reasonable. Not that you gave that dude much time to express his views before jumping down his throat. I mean, what expression of this same basic idea would you not have considered crazy in two tweets? Serious question.

    Free health care, reagents, supplies and postdocs, right? A technician in every lab! Yes, there are a few things that Canadian scientists do not have to pay for, but the basic costs remain the same. You hiring a postdoc or tech on $35/yr?

    No, you're not hiring a postdoc. You're buying supplies and equipment for your graduate students (and, if you're lucky, NSERC- or university-supported postdoc). If you think it's crazy to do any kind of science on a $35k/year budget, you're going to have to explain how so many Canadian academics manage to do science and publish at a steady rate with NSERC as their primary source of support. (You might also have to explain to the non-negligible portion of American scientists who work on small NSF grants why they're crazy, especially when they're repeatedly writing much larger 15-page proposals.)

    And this is where you go off the rails. Easy to get $50k/yr without consuming much budget, eh? Considering that the system currently can, in no way, support the scientists who are applying (see: 14% funding rate), I assume you think that we'll just take care of, what, half the people applying with this $50/yr business (of course, who's going to pass up easy money, right?)? NIH had over 28000 R01 applications last year, so are we writing off $700m/yr in direct costs to this program?.

    Right, right, I've gone off the rails. Never mind that applications to NSERC have been stable for the past few years, with success rates also stable at ~33% (and > 50% for early stage academics). Or that, as I pointed out above, there's no necessary relationship between high funding likelihood and investigator-funding (worried about too many people applying? Fine: adopt the Dutch system, with PI grants but large budgets and relatively low success rates). Or that there are a range of policies one could implement to prevent people from behaving the way you suggest is inevitable (e.g., an N-month moratorium on applying for project-based grants from the same agency once you accept the PI funding, which forces people to make tough but necessary choices about what their research agenda is). Right. It's simply crazy to suggest a program like this.

    And then what? You don't think more people are going to want in if the barrier is lower? I promise you every institution has a large number of people who have tried for NIH funding in the past to little success who would jump at that new program. And of course Deans will adjust their expectations to: One base award and at least one NewR01 for tenure? So you'll have a bunch of people puttering along and, by default, significantly reduce the number of R01 awards available.

    Come on. If there's a silly statement here, this is it. Everything you say here already applies to our current system, except with absurdly lower success rates. What, you think it isn't already a race to the bottom? Where's the disincentive for PIs to apply for R01s as opposed to smaller R03s? Have you noticed that the way NIH has managed to maintain any semblance of sanity in success rates is to lower the mean award size dramatically over the last two decades, effectively already moving in exactly the same direction? Do you think deans' current expectations are sustainable given the spiraling success rates? Why is it so crazy to think that some proportion of scientists applying for R01 funding don't really need anywhere near that much money to operate, and would be happy to take much smaller amounts in return for much higher success rates, even if it meant opting out of the big-money rat race?

    The reality is that the current system is broken in many ways, and something is going to have to change. That doesn't have to be the introduction of a PI funding track, but once again, it's not in any way crazy to suggest that a PI track might be part of the solution, given that it already is a common mechanism in many highly productive countries. In the worst case, if we do nothing, that "something" is going to be the forced exile of a large portion of the current scientific work force. But hey, maybe that doesn't bother you, because after all, you already got yours, and if all those people left, there'd be more $$$ for you! How dare those other people argue from their self-interest in favor of solutions other than "my job is going to go away unless we change something!"

  • DrugMonkey says:

    Like it or not, what is considered the very hottest and most impactful science requires a lot of research funding. So why woudl funding bodies move to support more and more small time labs, instead of a smaller number of high-fliers? Does that benefit science?

Leave a Reply