One-and-done NIH Grants: A bug or a feature?

Oct 29 2014 Published by under Fixing the NIH, NIH, NIH Careerism

There is an interesting new survival analysis for PIs that have been awarded at least one R01 posted at RockTalk.

Briefly, what you are looking at is as follows:


We chose three cohorts of first-time R01-equivalent awardees — those who received their first R01-equivalent award in 1989, 1997, or 2003....We used data on these three cohorts for a Kaplan-Meier analysis to look at rates of retention. ...So we used this to analyzing the number of years between the first year of R01-equivalent funding, and the last time an individual receives any additional research project grant (RPG) funding – whether it be from the non-competing continuation of their 1st R01 or another RPG award.

The bulk of Sally Rockey's analytical comments focus on the sharp dropout associated with the first (presumed) interval of support, i.e., 3-5 years after the very first R01-equivalent award. And her conclusion seems to be that this is a problem that needs fixing.

These data seem to support the concept that if there is an intervention needed in retaining scientists in research, it would need to come at the renewal stage of the first award, or as some call it the “second” award. Indeed, we are giving increased focus to this stage through some of our new award mechanisms, such as the National Cancer Institute’s Outstanding Investigator award, and will continue to seek ways of keeping our talent from leaking out of the pipeline.

I am not sure that I agree with this general conclusion from the data presented.

Remember, we are talking system-wide statistics here, not the fate of your five closest colleagues, your training mentor or yourself.

The extreme case would be that once anyone manages to land an R01-equivalent award as a PI, that the NIH should move heaven and earth to keep them funded for the duration of their career. That is an arguable position, but I think it is wrong.

It is wrong for two reasons, which are related to each other. The first reason is that if we re-adjust the system to keep everyone in once they have entered, this will sharply reduce the entrants. It will reduce the number of people who get a chance to prove themselves. This may seem fine and dandy once you have passed the first-R01 hurdle yourself, but this is mind bogglingly forgetful of the position one was in before this and mind bogglingly arrogant in assuming you would have been one of the lucky few.

From a system perspective, this cuts down on scientific diversity. It cuts down on the ability to try out a range of scientific ideas and approaches to see which ones stick. It substitutes the limitations of advance prediction for the virtues of empirical testing.

This would also increase stagnation and slow progress. It would. When you have sinecure funding, I'm sorry but the pressures are not as high to be creative, productive and to diversity your scientific thinking. Yes, from our current vantage point of the amount of time spent securing funds versus doing the science, this may look better than the usual. But take the longer view here. Our competitive system has its virtues in terms of clearing out the dead wood and encouraging better efforts from those who are actively participating.

I think the real question here is the appropriate balance. The desired survival rate.

And we should be very clear that we expect there to be some amount of PI dropout.

Personally, I think that having the major reduction in PIs after the first interval of funding makes a lot of sense. Better than at year 11 or 16, for example. A given PI has had a chance to try out her ideas. He has been given the opportunity to show what he can do. Again, on a system-wide basis, some of these individuals are going to fail so badly that they are never funded again. Do recognize that many will suffer intervals of no-R01 and then come back. Datahound addressed that in a post earlier this year. But many will disappear. Wouldn't it make more sense to have it happen before the NIH has wasted another 5 year interval on them?

Now before I get too far down the path, I will recognize that this is only the start of the data analysis. We want to know a few more things about who is shelled out of the system, never to return, who is able to fight back in after a gap and who is able to sail along with continual funding. This will allow us to see what may be undesired effects or categories of PI/applications that we wish to specifically protect. Is this part of the perfect storm that hits women particularly hard? Human subjects research? Physician-scientist PIs? Ecology or sociology?

Nevertheless, when I look at these initial survival curves, I just don't see the problem that seems so obvious to Dr. Rockey. I don't see how this is the next problem that requires special fixes from the highest offices at NIH.

74 responses so far

  • bacillus says:

    Am I misremembering, or were success rates for non-competing continuations around the 50-60% mark back then versus 10-20% now? If so, then doesn't this suggest that we aint seen nothing yet?

  • DrugMonkey says:

    In 1994? That seems high to me.

  • iGrrrl says:

    50-60% counting all three submissions as 1 application.

  • DrugMonkey says:

    As late as 94?

  • Personally, I think that having the major reduction in PIs after the first interval of funding makes a lot of sense. Better than at year 11 or 16, for example.

    KILL THE MIDDLE-AGED AND FEED THEM TO THE YOUNGS AND OLDES!!111!!!1!!!

  • GAATTC says:

    " Our competitive system has its virtues in terms of clearing out the dead wood and encouraging better efforts from those who are actively participating."

    This is exactly correct. I'm in my 8th month of what I hope is a gap in funding and I think my ideas and applications now are better than what was present in my renewal applications. 3 - 5 years from now, however, I might not feel the same way.

  • qaz says:

    For someone who spends lots of bandwidth arguing for special treatment for youngsters, your argument is extremely shortsighted.

    A person who loses funding likely loses tenure - remember this is the year that tenure comes up - and tenure committees hate when someone loses a grant. For the record, in my anecdata, this is where my comrades fell. Not at finding a faculty spot. Not at the first grant. But at the first renewal. When they were going up for tenure. When the tenure committee said "do we really think they can continue to get funding?" If someone takes two years to get funded again, but loses it in year five, they're toast.

    So we should not be shipping someone out of the academic pipeline when they finish grad school and have lots of non-academic opportunities, nor when they finish postdoc and we have to invest a million dollars or more in their salary and startup, but when they've spent five years as a faculty?! Are you serious? Now you have 40 year old professors (not even wanna-be professors, but actual got-a-grant, did-research, taught-classes professors) and you want to throw them out?

    What are you smoking?

    BTW, I notice that once again, we are talking about fixing the system when the children of the baby boomers need help. Funny that.

  • DJMH says:

    Why are so many people dropping out even before year 5? What is with the 10% of people who are gone by year 3 or so?

  • DrugMonkey says:

    Interesting qaz- aren't you the one that believes NIH grants and University employment have nothing to do with each other when it suits you? I.e. In discussions of hard/soft money jobs and percent effort? Or do I have you wrong on that?

  • DrugMonkey says:

    And agreed on the children of Boomers...funny that.

    and qaz, you jogged another thought. The R29/FIRST award. We can discuss whether it "worked" as intended until the cows come home but the fact of the matter is that it was ended almost precisely at the Boomer/Gen X divide. Sure, there were some Boomers being appointed after the R29 was discontinued but it is nearly impossible that any GenXer would have been faculty in time to apply for one of those.

    funny that.

  • drugmonkey says:

    What is with the 10% of people who are gone by year 3 or so?

    Perhaps 3 year R01s were more common, especially for noobs back then. I have certainly seen POs offer brand new PIs "well, let's cut it down to three years (or two) and see how you do" deals. I have also seen continuation applications where it was clear that the first interval of funding was only three years*. So it wasn't an automatic death sentence. But you have to think that for a given cohort, getting cut to three years would have seriously decreased survival odds relative to someone who landed a 5 year award.

    It would be interesting to know the proportions of 2, 3, 4 and 5 year R01s awarded as first grants, and how this may have changed across decades, wouldn't it?

    *no idea whether the conventional wisdom was telling youngsters to only ask for 3 years in the first place back in the late 80s either. I wouldn't be surprised if there was such a meme and it affected detectable numbers of newbies.

  • qaz says:

    "Interesting qaz- aren't you the one that believes NIH grants and University employment have nothing to do with each other when it suits you? I.e. In discussions of hard/soft money jobs and percent effort? Or do I have you wrong on that?"

    I believe that NIH grants and University employment SHOULD have nothing to do with each other. I'm definitely not foolish enough to believe that they don't.

    More importantly, I definitely recognize that research careers are often dependent on continued NIH funding, particularly in this day and age. It seems like a very late time to be booting people out.

    I think that I have consistently argued for increased funding stability. While I recognize the problems of a sinecure, I think the thunderdome is much worse. I think we could achieve stability without sinecures if we wanted. (Things like - your chance of renewal is based on past success. So you know two years in advance if you're going to be renewed or not and you can know if you should keep doing the good science or if you need to go into crisis mode.)

  • Comradde PhysioProffe says:

    "(Things like - your chance of renewal is based on past success. So you know two years in advance if you're going to be renewed or not and you can know if you should keep doing the good science or if you need to go into crisis mode.)"

    How is this more predictable? The judgment of "past success" is just as subjective and open to interpretation as is the judgment of the likely impact of a grant.

  • Morgan Price says:

    I think the judgment of past success far less subjective. And it becomes more reliable if you can wait 2-3 years to see if others in the field are building on the work, or refuting it, or whatever.

  • young person says:

    Given the enormous numbers of young people chomping at the bit with new ideas, the NIH should aim to cull much more than 30-50% of middle-aged or old people. If these scientists haven't done something great, give the money to someone else who might.

    The pyramid scheme of academia yields thousands of new PhDs per year, but there is no money for these young people to pursue their own ideas. The old people should get out of the way.

    http://www.ascb.org/ascbpost/index.php/compass-points/item/285-where-will-a-biology-phd-take-you

  • Crystaldoc says:

    The Rock Talk analysis was designed and presented to address a hypothesis that first renewal drop offs correlate with tighter funding at time of first renewal, but I find myself wondering also about possible relationships to pay lines at the time that the first grant was awarded. The attrition seems particularly bad for the 1989 cohort-- so I wonder, is it possible that the first R01 was easier to get at that time, and "chances were taken" on a greater number of new investigators with more marginal or at least less concrete track records? I could be totally wrong; I'm not sure where to look for the historical pay lines.

    I was not around to remember what was going on with any of these historical cohorts, but my impression is that it was once more common for new investigators to have very recently started their independent programs, and maybe not yet have independent productivity to point to, when winning a first R01. In my recent experiences on study section, the NI and ESI grants that are competitive are coming from investigators with 4-6 or more years of independent productivity; they have established labs and survived a considerable time running on foundation and institutional support, bits and pieces of program funding, maybe they had a kangaroo grant, DoD, and leftover fumes. It is not unusual for NIs to already be associate professors, or at the point already where they might be expected to be coming up for tenure. I suspect that the current cohort of new R01 funded investigators may look little like any of the historical ones, and I wonder what impact that may have on later success rates for ongoing survival.

  • Comradde PhysioProffe says:

    "I think the judgment of past success far less subjective. And it becomes more reliable if you can wait 2-3 years to see if others in the field are building on the work, or refuting it, or whatever."

    That you think this strongly suggests either that you have not engaged substantially in the kinds of peer review processes that would be tasked with assessing "past success" or that, if you have, you have been oblivious to what occurred.

  • qaz says:

    I am not arguing that past success is a more reliable predictor of future success than the current "can you plan a project" scenario, or that study section doesn't actually work this way (basically granting based on past rather than future). What I am arguing is that if we had a system based on past success (say study section evaluates progress in year 3 of the grant), you could know that you were going to be renewed, and you would not have to write 5 new grants with the hope of getting 1.

    So the question, CPP, is do you think that a judgement of past success would be a LESS reliable predictor than the current system?

    I'm a firm believer in safety nets. If you have a safety net to fall back into, then you are more likely to take risks.

  • Cynric says:

    From a system perspective, this cuts down on scientific diversity. It cuts down on the ability to try out a range of scientific ideas and approaches to see which ones stick. It substitutes the limitations of advance prediction for the virtues of empirical testing.

    This argument also crystallizes for me why the occasional call to switch from "fund the project" to "fund the investigator" is fraught with peril.

  • DrugMonkey says:

    Exactly Cynric. I bang the exact same drum in opposition to people-based funding proposals.

  • DrugMonkey says:

    qaz- so should everyone be able to stay in once they get the first R01-equivalent? For ever?

  • You realize that your comment about the nature of the current environment being good for innovation can be directly tested, right? You can directly compare to what your counterparts in other countries with fewer grant application burdens accomplish. I would argue that in many cases, you will find that they do just as well and are as creative as their US colleagues.

  • DrugMonkey says:

    Funny you should mention that, PlS. I am familiar with one nonUS system as it affects my fields of closest interest.

    The opportunity and success of essentially parallel trainees and junior faculty under each system differs dramatically. The U.S. NIH system is better

  • Skeptic says:

    First renewal is harder, because the new PI 4 -5 years of running a lab with a few people have to compete with PIs with more resources. One publication per year is considered NOT productive these days. I am wondering if we divide the publication numbers by total funding dollars, how would the "non-productive" first time renewal PI fair?

  • drugmonkey says:

    the NIH should aim to cull much more than 30-50% of middle-aged or old people.

    The survival curve shows that less than 40 % of those that manage to land a major award are still funded by year 15. That has to be the start of your "middle-aged or old people", yes? So you are suggesting that it should be such that less than 15% or so of the original R01 awardee population should still be funded?

    Interesting. Be sure to check back with us in 15 years or so.

  • drugmonkey says:

    I am wondering if we divide the publication numbers by total funding dollars, how would the "non-productive" first time renewal PI fair?

    I think they would compare quite favorably. As a reviewer I tried to attack the "this PI is so wonderfully productive!" meme with bean counting all of the grants acknowledged in the papers on more than one occasion. If there is a hobby horse of minutia that I ever rode, this one is it.

    My impression is that I got no traction with this whatsoever.

  • drugmonkey says:

    I believe that NIH grants and University employment SHOULD have nothing to do with each other. I'm definitely not foolish enough to believe that they don't.

    More importantly, I definitely recognize that research careers are often dependent on continued NIH funding, particularly in this day and age. It seems like a very late time to be booting people out.

    You are trying to have it both ways. On the one hand you want to stabilize the NIH funding of people who have managed to enter the system in part because of the tenure implications. On the other hand you decry the soft-money, high total %-effort-on-grants job categories that are so intimately tied to NIH funding and pooh-pooh the career-related concerns of such people vis a vis NIH grantgaming.

    With respect to 'a late time to be booting people out', this applies to the soft-money / "skin in the game" debate as well. As it does to overhead rates.

    Look, I get where you are coming from. From my current vantage point, anything that makes the acquisition of lots of grant funding easier for those who already are in the system probably works to my advantage. And I can come up with all sorts of reasons why it is totes justified that I not have to struggle to maintain funding at this stage of my career.

    The problem is, I am (so far) incapable of forgetting what it felt like when I was in Years -5 to +5 of my independent career. Many of the NIH policy changes bandied about that would significantly advantage me now might have entirely prevented me from entering the system in the first place. So I just can't see how it is acceptable to conveniently forget this.

  • Grumble says:

    I don't think these data argue against what DM calls "sinecure funding" at all. Here's why:

    The most striking feature is that there is a huge drop-off at about 5 years, due to lots of PIs being unable to renew their first R01. So, lots of first-timers fail. They have been winnowed.

    The next-most salient feature is the extremely shallow decline over the next 5 to 15 years. This means that once you get your R01 renewed, you have a very high chance of getting renewed at least a few more times. To achieve such a high success rate, two things must be the case: 1. the grants these PIs write are good, and 2. the PI's past research was high quality. If one of these two things are missing, grant reviewers would be less likely to score the grants well. Therefore, on the whole, PIs who manage to renew at least once are both good grant-writers and scientifically productive.

    This means that for experienced, posts-renewal PIs, writing a lot of grants just wastes their time. They are going to get funded anyway, so why force them to write 10 grants to get just 1? It's precisely this group that should be given some form of "sinecure" funding.

    A sinecure funding system for senior PIs in no way precludes younger PIs from competing for grants in the traditional system. In fact, these data actually provide a criterion that NIH could use to decide who is eligible for sinecure funding: only those who have renewed their R01s once would be allowed to apply for it (with an application that asks only for past accomplishments and the briefest summary of future plans).

    Finally, I agree with DM that Rocky's worries about losing PIs at the first renewal are misguided. Managing a lab is a completely different business than being a post-doc. So it is completely unsurprising that there would be a big drop-out at this point. There needs to be some process of weeding out the unfit, and the big winnowing at this point actually means the system is working.

  • qaz says:

    "so should everyone be able to stay in once they get the first R01-equivalent? For ever?"

    For the record, I have *never* argued this. What I want to see is that once you are in, if you continue to produce, then renewal is easy (and known in advance).

    For example, imagine the following process - there is a high bar to get "into the system". (Something like the current funding process for new grants.) Once you are in the system, you are evaluated by a study section in year 3, who basically says whether or not you can have another five years (meaning the rest of your grant [2 years] + 3 more). Approval is completely based on PAST productivity. So, it's basically a biosketch. If you get approved, you know you have another five years of life. If you get denied, then you panic, and start preparing to resubmit a new grant to save your lab. If you get approved you would be reviewed again in three years. A well-producing stable lab should be able to get renewed without difficulty. Basically - send in your X papers in quality Q journals - and committee says "is great". A struggling lab will have two years to get its act together and will have to refight to get back in.

    I suspect this process would require some tweaking. For example, we will need some mechanism to allow movement up and down in costs - so labs can expand and contract. Maybe you can applying for additional money when you go up for review, but that requires a real proposal, not just the biosketch. Or maybe the denial can be partial rather than complete. Maybe 3 years is too short. Maybe once the track record is there, the past productivity should be in longer timescales. I don't know.

    The key is to find a way to make renewal for good, well-producing labs easy, while not providing sinecures for deadwood. Our current thunderdome does a terrible job at that.

  • qaz says:

    For the record, DM, I'm not that old that I have forgotten what it was like to try to get my first grant, how hard it was (without a separate NI/ESI review process), and even more so how bad trying to get that first renewal was, or how the tenure committee almost didn't give me tenure because my first grant had run out. (Now I'm well-funded and have a multi-grant lab that contributes mightily to the scientific literature, so it's a damn good thing I wasn't kicked out after that first grant.)

    Those are events that are burned into my memory. I would argue that the processes I am suggesting would work well for both junior and senior people.

    I don't think I've ever "pooh-poohed the career-related concerns of such people vis a vis NIH grantgaming." Quite the opposite. I have learned tremendously from the discussions at blogs like this one. Grantsmithing is a critical skill that I teach my students. In fact, I would argue that the kind of a system that I'm proposing would be BETTER for people in those non-tenure soft-money positions, because it would provide much-needed stability.

  • drugmonkey says:

    The next-most salient feature is the extremely shallow decline over the next 5 to 15 years. This means that once you get your R01 renewed, you have a very high chance of getting renewed at least a few more times.

    As someone pointed out at RockTalk (and I originally didn't quite grasp), the years of funding off-set may be smoothing the curves here and we'd ideally like to see "after second grant" uncoupled to the start date of the first award.

    Second, the data trends do not speak to a competing renewal, just continued funding. Perhaps a minor point but we should be clear.

    They are going to get funded anyway, so why force them to write 10 grants to get just 1? It's precisely this group that should be given some form of "sinecure" funding.

    For this I think DataHound's post I linked ( and perhaps a figure or two from a closely related post) provides a clearer picture. His analysis shows the cumulative probability of getting funded after losing all NIH funding and describes the size of the ~continually funded subpopulation. It gives us a target (of course I'd like to see NIH extend his analysis back for a few decades). What it doesn't necessarily tell us is how well we would do trying to identify the eventual "continually funded" PIs when they get their first award. Still, if NIH first concentrates on who is continually funded, who drops out never to be seen again and who exists in various twilights of intermittent funding then perhaps there is a chance of identifying some predictive variables.

    A sinecure funding system for senior PIs in no way precludes younger PIs from competing for grants in the traditional system.

    It all depends on the rate. Is it not trivially obvious that if everyone who wins a first R01 remains funded forever that this decreases future odds of entry? It walls off chunks of money that the NI cannot compete for. ...I suppose we can argue over whether ESI policies can be constructed to maintain certain pools but overall this just substitutes top-down affirmative action for the unfettered competition of ideas. Me, I prefer policies that promote fair competition amongst all who are proposing studies within a given IC's domain.

  • Grumble says:

    "It walls off chunks of money that the NI cannot compete for"

    I'd argue that, de facto, the current system has exactly this kind of wall. Newbies are rarely going to write grants that are as good as those of senior investigators (nor do they have similar track records, pretty much by definition). So there is a chunk of money that newbies don't have access to. Really, what is the difference if the system is formalized?

    Remember, also, that sinecure funding shouldn't be non-competitive. Ideally it would be based on a competition - but the criteria would have more to do with your track record than your ability to write exquisitely crafted bullshit describing what you are going to do (and which you will then, typically, not do).

  • DrugMonkey says:

    "Track record " has some implications for new and risky directions, no?

  • The key is to find a way to make renewal for good, well-producing labs easy, while not providing sinecures for deadwood. Our current thunderdome does a terrible job at that.

    You keep saying this, but I don't think you appreciate that it is not particularly more straightforward to identify "good, well-producing labs" than it is to identify "high-impact" R01 proposals. Regardless, at current funding rates, making renewal "easy" for "good, well-producing labs" under any definition that would suit your goal would suck away every NIH dollar available for competing grants, and make it impossible for anyone--young, old, or whatever--to obtain new funds.

  • Joe says:

    As for the linkage of funding success and tenure, I think it should be noted that the current poor funding climate has already removed the grant renewal requirement that many university promotion committees have previously used. I see asst profs not getting their R01 until the 4th year or later and thus not reaching the point of needing to renew the grant until after the tenure decision has been made.

    I agree with the point made by others that the idea of saving these people that can't get the renewal is misguided. Furthermore, this reminds me of the discussions I've heard on cutting grant budgets. People are trying to save as many investigators as possible and so they reduce the budgets to 4 years and modular or even cut modules with the idea that more grants can be awarded, and maybe these PIs can make it until the funding climate improves. It's not going to happen in the foreseeable future, so helping people hang on with not enough money is only hurting the system.

  • datahound says:

    The R01 competing renewal success rate was 51% in 2000 (during the doubling), 42% in 2004, and 31% in 2013. The success rate only counts applications submitted in the same fiscal year as one application. The per application funding rate are probably a few percentage points lower.

  • datahound says:

    The discussion at Rock Talk gets at a very fundamental issue about the relationship between academic performance and funding. While tenure has its traditions in protecting academic freedom, it is now used by many institutions as a way of acknowledging financial viability, at least in the sciences. The timing alignment between tenure decisions at many institutions and competing for the first renewal of an R01 grant increases the importance of this renewal. However, I personally feel that it is more likely that reforms at institutions about tenure and tenure processes will better serve the research enterprise than having NIH manipulate success rates for these renewals substantially.

  • jmz4 says:

    Something interesting in the original post that may bear on this discussion is that the analysis of RO1 recipients that reapplied fell much more precipitously in the 89 cohort, indicating that these people weren't being forced out, but leaving voluntarily (perhaps for a booming biotech and pharma sector). This is echoed by the difference between PI's not getting funding in the first three years (before, presumably, any real funding crunch was encountered) between the 89 and later cohorts. I'd like to see some follow up for what happened, since it might give us some ideas for how to ease our current glut.

    Finally, just eyeballing it, the most dramatic change in loss of funding is clearly the 3-6 year peri0d, which lines up with what most of the YG's have been saying, that the hammer falls during resubmission time. Here are the numbers for % of PI's that lost their funding during that period, adjusted for how many PI's dropped out all together (didn't apply for R level funding).
    Failure rate (lost funding/renewal and submitted an application) for year 3-6:
    33% 2003
    24% 1997
    37.14% in 1989

    As Dr. Rocky notes, it makes sense that the 1997 cohort fared better than everyone else during that period, since it coincided with high paylines during the NIH doubling phase. But apparently this problem has been around for a long time, so it's probably worth taking a look at how to ameliorate it.

    I don't think it would be unreasonable to institute an HHMI like program of 5-10 year commitments to young(ish) investigators with a good track record of grants leading to publications. Alternatively, you could make the Pioneer awards a more regular part of the funding landscape, and set them aside for investigators in the 3-8 year range. A big chunk of unexpirable cash would definitely help some of these labs stay in business and encourage interesting and risky projects at a time in her career when a PI is probably the most likely to be able to pull it off. This also has the advantage of letting the PI have some flexibility on how the money is disbursed (e.g. constant stream vs. big upfront investment). Cause honestly, one of the big advantages the HHMI labs get is their awesome/ludicrous equipment budget requests (250 grand a quarter/lab).

    I think that all ages of PI's have their own contributions to make, and that the NIH has an interest in keeping it balanced. To me, it does look like that means retaining some of the people falling through in the +3-6 year gap. I could easily see a good scientist getting his first R01, after years as a postdoc of itching to run with his own ideas, trying to go for a big flashy project out of the gate and failing. If there was a mechanism to encourage that big flashy project a couple years later, you might be able to keep some people that were good, but just impatient.

  • I really am gobsmacked by the magical thinking that is going on here. It's like the whole denial of the simple rules of arithmetic thing occurring again. If you make it "easier", "more predictable", "more stable" for a particular subset of PIs, then you are inescapably making it "harder", "less predictable", "less stable" for some other subset of PIs.

    So all of these brilliant plans to make things better that don't explicitly address the tradeoffs regarding who is going to have an easier time and who is going to have a harder time getting funding are completely worthless.

  • DrugMonkey says:

    Those of us on the good side of the line will do better and screw all those folks who haven't made it past the line yet, PP. Aren't you paying attention?

  • jmz4 says:

    Sorry, I took it for granted that the money would come from R01's awarded to other levels of experience. I'd prefer to see it come from the 35+ year crowd (the 1989 cohort), since it looks like their levels have stabilized in terms of awards applied for vs. the dropout (retirement rate) meaning they can bear some pressure. Also, given the advanced stage of their research, "forcing" them out a little earlier would be the least disruptive to innovation and progress (because they have a slew of progeny).

    I understand its a tradeoff, and I think, with just what I'm seeing on this graph, that giving a junior faculty a big pile of stable funding would generally promote more innovative science than giving the same batch of money to either a total newb (like me) or a venerated elder.
    This is because, when we look at the 97 cohort, 10% more of the cohort made it through the 3-6 year grinder, presumably due to the deluge of funds under Clinton. However, after that, looking at the next 6 years (2003-2009), as funds plateaued, its not like they suddenly stopped getting grants requests at higher rates than either other cohort. This means that they are perfectly capable scientists, which means that, assuming a similar quality (or better) of future cohorts, that 3-6 year interval looks to be where people that are worthwhile are getting dropped.
    Perhaps this is all very obvious, but it runs counter to the thinking of my peer group (postdocs), where we assume that many of the PI's that got their start under Clinton only did so because the funding levels were so great and they couldn't hack it today.

  • jmz4 says:

    Oops, 25+ year crowd, not 35+.

  • You can't just glibly say "oh, yeah, of course it's a tradeoff. we'll put a little more pressure on the youngs, olds, whoever. they can handle it" YOU'VE GOTTA DO THE MOTHERFUCKEN MATH. Are you prepared to kill the research programs of every single PI who's been in the game for more than 15 years in order to provide "stable" funding to every "reasonably productive" PI who is trying to renew the first R01? YOU'VE GOTTA DO THE MOTHERFUCKEN MATH.

  • qaz says:

    CPP - You keep saying do the *** math. But I don't see any math from your side. What I see is a lot of shouting.

    There is no magical thinking here. There is acknowledgement of the scientific enterprise, which tends to have large startup costs and depend on integrated knowledge within a lab, which means that lost transient funding creates large inefficiencies.

    First of all, NO ONE HAS SAID that every PI gets renewed. What we have said is that (1) current systems are unstable and (2) tend to toss out productive PIs after wasting lots of money (what's the most expensive part of the fighter jet? the pilot.).

    The effect of point 1 is that people send in more grants than they need to do the science (wasting time, costing money and efficiency). Another effect of point 1 is that people have to run the lab at higher levels because they have to build in their own cushion. (Remember that in a competition like this if ten people send in ten grants each - those grants aren't distributed one per person. There's shot noise [some math for you] and some get 0 and some get 4. Both extremes are problems.)

    The effect of point 2 is that we have spent a lot of money training those pilots and then we are throwing them out for a new pilot we're going to be spending lots of more money on. Here's some math for you. The cost of a PI in year 6 is 5 years of graduate school (5*50k) + 4 years of postdoc (we'll take a nice average of 4 yrs, which is 342k adding up all the stipends and 50% fringe costs) + startup ($1M?) + salary (let's be cheap and say $100k*6) = 2.2M. You've just thrown that out. Oh yeah, and an R01 for those five years = 1.8M.

    Here's some more math. If you have a death-wall where people leave if they lose funding, and you have much randomness in the system, then probabilistically you are going to have a lot of people who will randomly die off, whether they should or not. (This is foraging theory, and lots of happy math for you.)

    What NIH needs to do is to create a portfolio to mix the age groups. They started to do this when the baby boomer's kids showed up and said "we need more new younger investigators".

    Recognize that the current system is basically a flooded standard labor market with no labor support structures. This means that the buyer (NIH) can make things a competition. NIH is always getting great science for it's money (because there's far more out there than it can fund anyway). But the system as a whole is highly inefficient because the selection process is based on letting the chips fall where they may.

  • qaz says:

    DM - When you started defending the NI/ESI special review, did you say "Those of us not over the line yet will do better and screw all those old folks who are already past the line, PP. Aren't you paying attention?"

    How about trying to find a system that is equitable for all the groups? That contains a reliable, stable funding system for well-productive labs and a continuous entry of reasonable size for new people?

    We have to realize that's the way the owners keep the labor market in check - by making them fight against each other for scraps. Maybe we shouldn't be doing it to either Julia or me?

  • DrugMonkey says:

    I basically said that qaz, yes. I've written many comments on this blog to the effect I don't feel sorry for us. Where us = those who managed to get to the point of holding an R01.

  • DrugMonkey says:

    Wrt your "math", you want to keep people in who "shouldn't" remain in. And that, in a zero sum budget means newcomers who "should" get an R01 cannot. And that leaves scientific acumen lying in the table. Far better to give more people a chance and then do another cull based on specific performance.

  • Grumble says:

    @CPP: "at current funding rates, making renewal "easy" for "good, well-producing labs" under any definition that would suit your goal would suck away every NIH dollar available for competing grants, and make it impossible for anyone--young, old, or whatever--to obtain new funds."

    That is complete rubbish. First, you are tilting at windmills because the idea is not to make funding "easy". It is to make it efficient, so that an established PI doesn't have to constantly write grants to maintain a reasonable level of funding. Under a track record-based funding system, the hard part for the PI is maintaining a level of productivity that would be judged by reviewers as adequate given the amount of money she had. The hard part for reviewers is making this judgment. Nothing is easy here, but the criteria are different than the utter nonsense of the current system (constant bullshit grants that describe experiments that won't actually be done and are judged on minutia) and the grantwriting load is less for productive PIs.

    Second, as I wrote above, well-established PIs are already sucking down a fairly constant share of the NIH budget. There is no reason to let them suck down the entire budget. The idea is that NIH would allocate funds for the competition-based-on-track-record and for the competition-based-on-prospective-experiments separately. PIs in the track record competition who are judged insufficiently productive get their budgets cut or eliminated.

    I have yet to see a criticism of this proposal that holds any water. DM's wan " 'Track record' has some implications for new and risky directions, no?" doesn't point out any serious flaw. Experienced PIs would need to decide how much risk they should take, keeping in mind that very big risks could lead to a poor track record. How is that at all different from the current system, in which every PI has to decide what experiments to actually do based on their riskiness and potential for payoff? You know very well that if established PIs don't publish when they are well funded, it's the very first thing their renewal (and even their new applications) gets dinged for.

    Track record is ALREADY an important criterion for evaluating grants from experienced PIs. I'm suggesting that a pool of money should be set aside in which track record is the ONLY criterion, greatly reducing time spent on grantwriting for a large number of PIs.

  • DrugMonkey says:

    My "wan" comment is your invitation to actually think it through for yourself.

  • Grumble says:

    I have thought it through. That is why I've written pages of argumentation on the issue, which you prefer not to engage directly with. Maybe because my position is hard to actually argue against when you consider it seriously?

    If you are implying that people with good scientific track records tend to take fewer risks and are less likely to lead their research in new directions, I think that's debatable. It depends on how you define the "good" part of "good track record." Someone who keeps doing the same thing over and over, with minor variations, might have a lot of publications in minor journals, but a review panel tasked with evaluating his track record might call that a poor track record. Or at the very least, a poor track record of innovation.

  • Skeptic says:

    "Track record is ALREADY an important criterion for evaluating grants from experienced PIs."

    I think this is one of the reasons that the first time renewal is harder. After running a lab for 4 years, how many papers can you publish in decent journals? How long does it take from submission to acceptance?

    You cannot evaluate track record just from 4-5 years of works. I often hear on the general news media saying we are closer to curing such such disease, this therapy worked on mice. I am sure those PIs are rewarded with grants and tenure. They have "track record of curing disease". But in 2 years, oh, there is this side effect of dying from something else. My point is that evaluation based on short term "perceived" outcome is not effective.

  • DrugMonkey says:

    We get into bean counting over JIF and author numbers because it is so difficult to define "high productivity" as real scientific advance. It's too subjective*. Your scheme, Grumble, will *necessarily* further increase the premium on 1) doing all the experiments as described and publishing them and 2) hitting quantitative, bean counting measures of productivity. Sure there will still be *some* vertically ascending geniuses *that succeed* but the ones that fail (b/c "risky") will be sent to the back of the line again, no matter how good their next idea may be. While plodding PIs that do steady (publishable) work in crowded fields (lots of cites!) at a predictable rate (consistent track record!) will do well at the sinecure game.

    *bias, anyone?

  • DrugMonkey says:

    And to be clear- *I* am a plodder. I do the experiments that I propose in my grants, for the most part. My ability to publish in journals that study section members on panels I submit grants to find acceptable is limited mostly by writing manuscripts up. Generating the data is not the problem, nor is finding something sufficiently exciting a constraint.

    This sinecure stuff would be very good for current-me.

    I still think it's a bad idea for the NIH to pursue.

  • Grumble says:

    DM, I'm really amazed at your lack of faith in the ability of a review panel of your peers to judge your work.

    First, my scheme will NOT "increase the premium on doing all the experiments as described and publishing them" because these grants will not actually contain any proposed experiments! So this is a complete non-issue.

    Second, the degree to which my scheme puts a premium on "hitting quantitative, bean counting measures of productivity" is really a function of the implementation. It is NOT, as you assert, a *necessary* feature. For instance, just as the NIH berates its reviewers on certain issues now, it could insist that reviewers not use quantitative measures, or at least not use them exclusively. Instead they are to write full-sentence critiques that address the impact that the PI's publications have had and/or are likely to have in the future.

  • DrugMonkey says:

    My views on how grant review goes down is based on being a student of this game and of judgement about scientific quality in non-grant situations. I'd say the burden of proof is on you to show how you aren't proposing a severe version of the Good ol Insider's club.

  • DrugMonkey says:

    How does one enter the system in your scheme Grumble?

  • Grumble says:

    As I wrote extensively above, you enter via a more traditional competition, with grants describing prospective experiments. Then, once you've renewed once and have entered the plateau phase of the graphs you showed in the post, you are eligible to enter a grant competition based only on your track record.

    Your worry is that by removing review criteria, the system becomes more susceptible to reviewer bias. But retrospective review criteria could be just as numerous: past innovation, past significance, validity of past approaches, and putting it all together, what is the overall impact of past work? Of course, a reviewer would judge these criteria by reading the applicant's papers and putting each of these criteria in the context of his/her knowledge and experience of the field. If exactly the same study section format were used to evaluate past performance, the system would be no more susceptible to bias than the current system, as it would be checked by the same factors: the need to appear objective to your fellow reviewers, the fact that there are 2 other reviewers, the open discussion and questions of the other panel members, and finally the vote from the entire panel, not just the reviewers, typically constrained to the reviewers' score range.

  • Grumble says:

    By the way, it's a mistake to imagine that what you and your colleagues informally say over beer regarding other people's work necessarily means that that's how you and your colleagues would judge the same work if tasked to do so with specific criteria and instructions.

    Just like your low opinion of Dr. So And So doesn't necessarily strongly influence your criteria-based evaluation of Dr. So And So's grant. Right?

  • qaz says:

    DM, Your logic depends on the concept that R01s are independent. I think the data (much of it on this blog) is very very clear that that is incorrect. Let's look at this logically:

    If every R01 is independent, then it doesn't matter if we have a brand spankin' new PI who has never run a lab or an established PI with techs that know how to do things. If every R01 is independent, then lab continuity doesn't matter. We might as well take new people in and kill them off after five years.

    If every R01 is independent, then it doesn't matter if a lab loses funding for a year or two and gets rid of all its technicians and loses all of its in-lab knowledge of how to do things. If every R01 is independent, then lab continuity doesn't matter.

    On the other hand, if a lab's scientific contribution is a "research program", of which an R01 is but a component, then continuity matters, and swapping out a productive lab for a new PI is not equivalent. If a lab scientific contribution depends on in-lab knowledge, then killing a lab and then restarting it is a major problem.

    I can't speak to how one would enter Grumble's scheme, but in my scheme, you would write a grant application, much like we do now. Importantly, if you fall out of the renewal cycle (because you had a bad productivity cycle for example), then the way that you get back into it is that you have to write a grant application. The assumption is that the grant application is hard, but doable. Look, we already separate NI/ESI from renewals in review. This really isn't that different, except people who fall out of the renewal-cycle are in review also.

    More importantly, no one is arguing that established PIs get a free pass. But we could make the grant-writing (and reviewing!) burden much less. It is important to have new blood in the system. That's why the NI/ESI renewals got separated out. My argument (and Grumble's, I think) do not change that balance.

  • drugmonkey says:

    As I wrote extensively above, you enter via a more traditional competition, with grants describing prospective experiments. Then, once you've renewed once and have entered the plateau phase of the graphs you showed in the post, you are eligible to enter a grant competition based only on your track record.

    "renewed once". So, gotten past this sharp early cull that is the subject of this post? hoookay. And past tenure and basically you are in year 9+ at this point? But just so we are clear this is for everyone correct? This datahound post reminds us that the in/out churn of the NIH grant is on the order of 5,000 investigators every year. This other datahound post reminds us of the cumulative probability of returning to funding after a gap. The approximately continuously funded population over 8 years that he looked was 13% of investigators- undoubtedly this goes down as we stretch the interval of assessment out across an entire career. How would you change this number? How many unknown-quantity trial-=balloon newb grants do we lose by doing so?

    Your worry is that by removing review criteria, the system becomes more susceptible to reviewer bias. ....If exactly the same study section format were used to evaluate past performance, the system would be no more susceptible to bias than the current system, as it would be checked by the same factors:

    It is, and I don't agree with you. The reason that I do not agree with you is that there is already a tinge of "should we keep Dr. SoandSo funded?" that permeates review. I doubt it is usually expressed as "She's will be out of funding if we don't bite on this one" at study section level but POs certainly have this approach. And there are aspects in which this sneaks in around the edges in study section. The "project based" review system fights back against this tendency by putting the focus away from the PI's lab viability. It says in so many ways that we are to decide on the specific research plan. This, I argue, fights against the GoodOldInsider tendency we have to want to keep funding people who have a body of work that we know or people that we know. I think that when you put the focus on the person you return these subjective tendencies to the middle of the pack, regardless of reviewer instruction or what your imagined application would look like.

    it's a mistake to imagine that what you and your colleagues informally say over beer regarding other people's work necessarily means that that's how you and your colleagues would judge the same work if tasked to do so with specific criteria and instructions.

    People as individuals may not express the exact same biases in every setting in which academic merit is being assessed but they assuredly express biases in every setting as a population.

    Importantly, if you fall out of the renewal cycle (because you had a bad productivity cycle for example), then the way that you get back into it is that you have to write a grant application. The assumption is that the grant application is hard, but doable. Look, we already separate NI/ESI from renewals in review. This really isn't that different, except people who fall out of the renewal-cycle are in review also.

    This is already the case. If this is such a great system, maintain the status quo. The fact that you are arguing for the special-princeling award shows that this
    no one is arguing that established PIs get a free pass.
    is bullshit and this
    we could make the grant-writing (and reviewing!) burden much less.
    continues to be the real goal here. You simply imagine that you'd be the dude who got to get the free pass to the amount of research funds that you want and everyone else clearly deserves to have an even harder time then all of us share on average at present. Or at least a harder time than you do at present.

    This is what mystifies me. Your plans clearly want to make it more "efficient" (aka easier) for some people to remain funded. You are completely bypassing the point that this inherently makes it harder for some other group of people. You are being disingenuous about how those two groups are to be selected, particularly those who are selected for sinecure funding. The most specific proposal yet, from Grumble, argues that the sinecure should come on the third round of a PIs funding arc when the data suggest there is no longer anything but a smooth curve and the population has already undergone severe winnowing. This latter point suggests that the people on the bad side of the annual in/out churn and the 20% of them refunded within a year / 40% refunded in 3-4 years are screwed. If you imagine your scheme significantly reduces the churn then you have to take this pool of money out of the amount that is available for never-PIs or something. And that feeds back to lack of opportunity for new entrants.

  • drugmonkey says:

    On the other hand, if a lab's scientific contribution is a "research program", of which an R01 is but a component, then continuity matters, and swapping out a productive lab for a new PI is not equivalent. If a lab scientific contribution depends on in-lab knowledge, then killing a lab and then restarting it is a major problem.

    I agree with you in this. Absolutely. And I agree on the "efficiency" as well. If we could know, in advance, who "deserves" to stay funded for the duration of their career then it would be an obvious solution to fund them on the basis of noncompetitive review of progress reports...or less.

    My problem here is that this comes with the huge probability of an enormous increase in the inefficiency of keeping labs funded that should not, in some sense, be funded for the duration of the PIs career.

  • I'm starting to come around a little bit on this, with the idea that it is just a different kind of grant mechanism than the R01, and could have certain amounts of money allocated to it. I guess this is sort of what NIGMS is piloting, although it is more of a mega-grant dealio, meant to replace multiple R01s.

  • Grumble says:

    "The approximately continuously funded population over 8 years that he looked was 13% of investigators- undoubtedly this goes down as we stretch the interval of assessment out across an entire career. How would you change this number? How many unknown-quantity trial-=balloon newb grants do we lose by doing so? "

    Datahound was looking at all R mechanisms, including R21s. If you include only R01s and equivalents, the proportion funded for 8+ years would be higher. If you include only those who have made it past the first-renewal cull, the number would be even higher. Your point is that sometimes oldies lose their grants, and that money can go to newbies. Fine. Why can't that continue under the sinecure system? Prof. Graybeard loses his grant or gets his budget cut because he couldn't get anything done in the last 5 years. The idea is that NIH decides how much to allocate to sinecure funding, and that means there's a cut-off so not every single sinecured prof gets his sinecure. (Which means, maybe, that sinecure isn't quite the right word.)

    "I doubt it is usually expressed as "She's will be out of funding if we don't bite on this one" at study section level but POs certainly have this approach."

    Look. If POs are already playing this game, then why not formalize it? Why MAKE experienced PIs jump through the hoops of writing endless grants, only to have all of them score poorly, only to have POs fund one of them anyway? How silly is that? Why not skip all that crap and let the scientists tell the POs who is maintaining a good research program and who isn't, rather than making the POs guess?

    "This latter point suggests that the people on the bad side of the annual in/out churn and the 20% of them refunded within a year / 40% refunded in 3-4 years are screwed. If you imagine your scheme significantly reduces the churn then you have to take this pool of money out of the amount that is available for never-PIs or something. And that feeds back to lack of opportunity for new entrants."

    Again, you are making an assumption that there are a lot of experienced PIs who are losing a lot of grants (outward churn, then refunded within 4 years). So if NIH has to keep covering these guys, that's less money available for newbies. But I'll bet you my next R01 that the outward churn among experienced PIs is far less than it is among less experienced PIs. On top of that, some of that churn is due to 2nd or even 3rd R01s that get lost and then refunded (or not). But in the "sinecure" system, the amount of funding pre PI could be limited to 1 R01 equivalent (or maybe even less).

    I think maybe what you aren't seeing here, DM, is that the NIH can control the level of non-sinecure and sinecure funding. That means NIH can ensure that there is always sufficient funding for the traditional system (to the extent that its budget allows). This will make the sinecure competition tough (but the *process* will be easier because it's a 4 page application: a biosketch, and it happens once and not 10 times over 2-3 years - that is the whole advantage). If a PI fails at it, s/he can apply for funding from the other pool. Or keep going with other support if it exists, and re-apply based on his amazing track record from that funding.

  • jmz4 says:

    " Are you prepared to kill the research programs of every single PI who's been in the game for more than 15 years in order to provide "stable" funding to every "reasonably productive" PI who is trying to renew the first R01? "
    I was basing my proposals on the graphs provided. And yes, as I said, I think it would make sense to take money from older investigators. I specifically mentioned the 1989 cohort, who are hitting their 60s (assuming an R01 age of 35). This is because, by looking at the graphs, they do not currently seem to be facing stark attrition due to grant application failure anywhere close to the younger cohorts. That's where my determination they could "bear pressure" came from.

    The conclusion that this would be productive for science was based on my observation that the "winnowing" that you all seem so keen on is not clearly effective at predicting future grant success, as evidenced by the 2003 cohort that got the boost in R01 funding, and then went on to hold their own.

    I think, as Qaz pointed out, it is kind of silly for the NIH to have a technical policy of R01 as a standalone award, without consideration to how research labs are actually built, and how their sunk costs and returns vary at different points of the scientists career. A brand new scientist has had significant costs sunk with limited return, which justifies the early career awards, but that same math applies to the first time renewers as well, possibly more so since you just put an R01 into the lab and probably haven't gotten much use out of all the new equipment and trained staff. This assumes, of course, that the this is a lab that will, in the future, be productive given the funds (i.e. can actually produce a return on investment). If you keep them going through this bottleneck, these graphs would suggest you can get another 30 years of return on funding (or 21.7 on average, according to the attrition rate of the 1989 cohort).
    If you're funding a 60-65 year old, you're extending the lab's productive years, sure, but its hard to imagine that being much more than 10 years. Also, you've recouped most of the costs of your initial investment in training and staffing, so the opportunity cost of NOT funding that lab are less.

    So I guess I'd have to see your math determining its better to be funding the 60+ crowd rather than retaining another third of the 40+ crowd (as happened for 2003). I'm not saying its implausible, maybe the really established labs are just that much better at cranking out good science.
    Almost 7% of PI's are over 65 years of age (as opposed to 1% in 1980). I haven't seen the data to see what percent of the R-level funding those labs hold, but my guess would be that it is disproportionately concentrated. Does anyone have that data?

    On the topic of NIH treating grants in more career-stage appropriate ways, should there be grants available to properly sunset/wind down an older lab? What would those look like? Do universities ever implement these sort of things?

  • drugmonkey says:

    they do not currently seem to be facing stark attrition due to grant application failure anywhere close to the younger cohorts.

    You don't seem to be grasping the point that this cohort has already suffered the selection effects.
    Those who remain are the elite of their cohort. What you are suggesting is that these 33%ile survivors survivors should be winnowed in favor of the 100%ile newbs. I can't really get behind that logic.

    you're extending the lab's productive years, sure, but its hard to imagine that being much more than 10 years.

    This is an argument I can support (since I've made it on these very pages). Absolutely agree that the anticipated payoff duration should go into the decision about which lab should be 'saved' by exception funding.

  • BugDoc says:

    If we are worried about the pipeline, i.e., making sure that we don't drop a lot of scientists from the begining, end or middle, perhaps it would make sense to percentile within separate pools, e.g., assistant, associate, full professor (or some equivalent scale). NIH can fund the top x % of each pool. That way, each rank would only be reviewed in comparison to each other. I don't think this would really require a huge change in the current process. Regardless of whether NIH is low on funds or high on funds, there will always be regulated movement through the pipeline. And if the older scientists want to stay on longer in their career, e.g., past 70, we don't have to have mandatory retiring age, but that full professor pool will get more competitive without squeezing the assistant or associate pool.

  • jmz4 says:

    "You don't seem to be grasping the point that this cohort has already suffered the selection effects."
    -I get that, I just don't think the selection effects are improving the fitness of the population. Admittedly, I'm basing this on limited data showing a bump in survival of the 2003 cohort that I'm attributing to the NIH doubling, without a subsequent fall off (which may be in the wings, post-2009, I suppose), indicating that at least the higher end of those winnowed in other cohorts are likely to be as qualified as their peers that go on to make it. Of course, as a postdoc, I think all PIs are doddering old fools, but I have a tough time accepting the argument that a 65 year old with 2-3 R01s is, all things considered, going to be as industrious as a 45 year old fighting for a second. There's a reason the HHMI awards their stuff to this crowd of people and tapers it off as they get older. They get the most bang for their buck because the renewal crowd has a good combination of experience and drive (unlike the 100% newbs or the graybeards).

    This is not to mention the graying of the professoriate that has occurred in the last 30 years (http://nexus.od.nih.gov/all/2012/02/13/age-distribution-of-nih-principal-investigators-and-medical-school-faculty/), a trend I think the NIH should push back on by winding down labs when the PI's are pushing 70 (the previously mandated retirement age).

  • drugmonkey says:

    I have a tough time accepting the argument that a 65 year old with 2-3 R01s is, all things considered, going to be as industrious as a 45 year old fighting for a second.

    What about "the lab of a 65 year old" versus "the lab of a 45 year old"? Grant funded efforts are staffed by more than just the PI. Some of the staff in the lab of the highly established 65 year old are going to be 32 year old highly elite postdocs who are even more industrious in fighting for a career than is the 45 year old fighting for her 2nd R01.

  • rxnm says:

    "Grant funded efforts are staffed by more than just the PI. "

    Yeah, this is the pushback I hear from big labs or their defenders...they have all the grants but are enabling research by the best postdocs.

    This might be valid if our goal is short-term maximization of paper production and competition for attention. If our goal is a stable and vibrant research enterprise, it's a fucking disaster.

    "The NIH has decided to kill two generations of scientists," a BSD told me today.

  • DrugMonkey says:

    Did I forget my sarcasm font, rxnm?

  • K99er says:

    I had a conversation this past week with a council member for my IC. I asked him what he thought the main problems were with the current system. He gave me a variation of the riffraff theory that I actually thought made a lot of sense. He said that everyone on the study section is in such a panic about getting their own grants funded that nothing that they read is ever good enough. Minutia overwhelms them and they are unable to tell a good grant from a great grant, and everyone involved ends up frustrated because who gets funded seems arbitrary. (I think the same thing could apply to manuscript review as well.) His solution was an intense computer algorithm that gives a score of past productivity and awards money based on that. He had a lot of ideas about comparing investigators at similar career levels so as to also build in some protection for new investigators. It was interesting to hear that someone who actually has some power is thinking in this way.

  • drugmonkey says:

    It's still the riff raff theory though, K99er. It is still permeated with the belief that review "minutia" or other fixable reasons (reviewer self-concern, flop sweat, panic) are interfering with the ability to properly distinguish the deserving "great grants" from the undeserving "good grants".

    Instead of recognizing the stone cold truth that no matter what one individual thinks, the reality is that we have more deserving grants than the NIH can afford to fund.

    Individuals who think there is a way to beat this reality are blinded by the arrogance of their own opinion. It's a version of "If I was the boss of science, I'd fund this, this and this, but not that..." at work here but people think there are universal reasons backing them. There aren't. And since there aren't, each other peer scientist is going to come up with a different, partially overlapping list of "great grants". Put them together and, guess what?, we're over the payline again.

  • rxnm says:

    "Did I forget my sarcasm font, rxnm?"

    No, I got you, just extending the thought.

  • jmz4 says:

    "Some of the staff in the lab of the highly established 65 year old are going to be 32 year old highly elite postdocs who are even more industrious in fighting for a career than is the 45 year old fighting for her 2nd R01."
    I guess this was sarcasm from you ("highly elite postdocs?"), but its a common enough sentiment. It hasn't been my experience. Years of steady funding breeds a certain amount of complacency, and that transfers down to the postdocs. Also, I'd say the big labs attract a mix of highly driven, career orientated postdocs and lost sheep with good credentials looking for a port in the storm. Easier to weed out the latter in a smaller lab.
    Most big postdoc farms I know have at least 1-2 unproductive people per 3-4 productive members. I've never seen a lab under 10 have someone truly useless for more than a year or two (unless they're a struggling grad student).
    What's your experience been?

Leave a Reply