This isn't realistic for NIH, but could work at the NSF (which has a much clearer basic science mission):

How's about a single-grant cap of $100,000/year and a prohibition on grant concurrency? $100k / year is enough to keep a small program afloat, and this would insure that the total grant pool was spread out beyond the "rich" programs.

A few more thoughts: It seems like the NSF (and NIH) have been tending to throw large blocks of funding toward interdisciplinary centers that cross departmental boundaries at a single institution, or multi-university research initiatives (MURI) that are more single-topic in scope. The funding for these centers has often been substantial, but after seeing a number of these in action, I'm convinced that they are part of the problem. The centers often have one or two "rich" and very well known PIs. There may be some that really do engage in interdisciplinary research, but most center PIs use them as easy sources of cash without doing anything "interdisciplinary". The problem is that as younger faculty are brought into an institution, the centers very rarely open their doors to new members. This tends to leave younger "poor" faculty out in the cold, and keeps the rich programs well-funded. I've seen a number of center proposals that were funded when the individual PI contributions to that proposal were substantially weaker than what the PIs would propose on their own.

Other ideas: the standard length of grants should be increased to 5 years from 3. The rate at which faculty are now required to write proposals (instead of actually, you know, doing research) has become quite high. The biggest benefit of the CAREER program at the NSF is the length of the grant, and anyone who has had one can tell you that those 2 years are a huge benefit, particularly because they typically postpone the renewal until after the tenure decision. I think longer term grants will lower the number of proposals that are submitted across the board.

I need to look at the funding data, but NSF has been hurting for funding basically since 9/11. Sure, Bush signed a statement saying that we should double funding, but it wasn't a law and isn't binding in anyway. Funding for all of the government science agencies is being held flat for this year, which is a real decrease in funding due to inflation and expectations of even modestly higher funding levels.

One reason for the sudden decline in approval rates that has been often floated around by NSF managers and scientists is that because there is so little money, scientists are submitting MORE proposals to improve their odds with the expectation that only one or two might get funded. Also, new proposals often have kinks that need to be worked out, and scientists have come to expect that their proposal won't be funded the first time around, but will need to reviewed and edited, pretty much like any journal article submission. So they'll sometimes turn in a half-cooked idea for the first round to let the reviewers work on it, then resubmit it with revisions and a higher expectation that it will be funded.

Regarding NSF funding of multiple institutions - this is very attractive to the funding agencies because (a) they get to spread the money around, supporting a lot of groups rather than making one group 'rich', and (b) they're spreading the risk that the research might not get done by betting on multiple capable partners.

So how was the science blog conference?

Let's say you have a flask of bacteria grown to saturation. You pour in an equal amount of fresh LB and put it back in the shaker. What happens?

The NIH's budget was doubled and in roughly one generation time, approval rates went back to where they were before (and may now have overshot)? Is anyone genuinely surprised by this? The current process of developing scientists is a pyramid scheme, and like any pyramid scheme, throwing in more capital can only push the problem out a little further.

Reading Zerhouni's piece in Science, where he expressed complete bewilderment at the results of the budget doubling and called for another batch of LB to be dumped in, I couldn't decide which would be more disturbing: that he was being disingenuous or that he was being sincere! If he and the rest of the NIH leadership truly don't understand that more money equals more faculty positions equals the same approval rate, I'd say that's pretty alarming.

How about limiting the number of submissions from any PI in a period of time? Would this help with the downward spiral and half-cooked idea problems mentioned above? If you knew you only had one chance, wouldn't you work to make it stronger?

BioGuy@dodgeit.com

I may have a long term fix. But for it to make sense, I need to further define the problem. The current dip in funding rates is due to an increase in submissions. The increase in submissions is due to: 1) an increase in the number of submissions per submitter and 2) an increase in the number of submitters. I would argue that the first problem is a non-issue after certain current incentives (below) are removed - the best proposals should be funded and no one ultimately has a monopoly on brilliance. In other words, an increase in submissions per investigator should only go as far as that investigator's ability to do the work.

With each funding raise, we increase the number of graduate student stipends in the U.S. This in turn leads to an increase in the number of investigators a decade or so down the road. As every P.I. turns out multiples of PhDs over the course of his/her career the system is inherently unstable in any funding environment other than significant growth. There are many indicators of the problem: decreased domestic undergraduate interest in the field, increasing degree times, increasing time spent in post-doc, and decreasing percentages of postdocs landing tenure track positions. (This is in spite of a modest increase in total tenure track positions nationwide.) In sum, the larger problem is a glut of PhD scientists at every level – but as it pertains to the current question we will consider only the increase in principal investigators. There are far more applicants than P.I. positions, and universities stand to make mint off of NIH grants with indirect charges ranging from 40%-100%. The result is an eagerness for institutions to hire relatively cheap professors to rake in as much grant money as they can. Of course, the grantees require space which requires a large capital outlay thus blunting the growth of this phenomena. Nonetheless, the end result is the growth in investigator numbers in periods of NIH growth to levels beyond what most would consider comfortable in a zero fiscal growth environment. Thus, I argue that any increase in funding will only result in a short-term solution. The real problem in approval rates are twofold: the creation of fiscal incentives for universities whereby they financially gain from research beyond the true ‘cost of business’ and a simple supply/demand imbalance in the PhD scientist marketplace.

The fix for the first problem is in the way indirect costs are considered. NIH and NSF grant scores should be adjusted by a multiplier of the institution’s overhead. Thus, proposals from institutions charging high overhead would be at a vast disadvantage. Competition for grants would push the marginal gains from indirect funding to zero for the institution, thus removing the incentive to spew the marketplace with applications. I believe that there has been some movement in controlling indirect costs, but the existence of charges in excess of 50% at many institutions indicates it continues to be a problem.

The second solution is a trickier fix. I think the best option in fixing the investigator supply/demand imbalance is to overhaul the system at the base so allow me to digress before returning to the problem at hand. We are producing more PhDs than the market would otherwise demand. On the other hand, there is an ever growing need for more science to be done. Why not solve both by a simple rearrangement of funding? If we were to gradually phase out the funding of graduate stipends and instead fund professional scientist positions we would be subsidizing the demand for professional scientists while allowing the supply to float with the ‘intellectual marketplace’. Graduate studies would need to be self funded, as is the case with the M.D. Initially, the number of students willing to pay their own way would be near zero – this is itself is a good indication that the rewards of the degree are not commensurate with the demands and the reason for the gradual phase in of the shift. But as the demand for new scientists continued to grow through natural means plus the increase in NIH/NSF funded scientist positions the salaries of these individuals would be driven upwards. This increase would eventually justify students in taking out loans from financial institutions for the PhD. Domestic interest in the field would rise while our ability to retain the best foreign scientists (regardless of country of education) would increase through a stable job market. The federal government would get more productivity for it’s dollar as trained scientists working as a team are more productive than ever-rotating and inexperienced graduate students. Academic institutions would see an increase in funding as federal grants became supplemented with private tuition fees. Scientists would see an increase in competition for their services resulting in increased compensation. Sure, the young would be saddled with debt, but this burden seems less problematic when set against increases in salaries. See the market for US physicians for a comparison. I suppose that private industry would be dissatisfied with having to pay higher scientist salaries, but the implication that Pfizer needs federal subsidization in this area seems wrong.

Lastly, the supply/demand fix and its implication to the original question. As the overall market for PhD scientists became more competitive, universities would have to follow with increased salaries. This would naturally result in a decline in the number of positions available. (If this personally seems like a move in an undesirable direction, remember that it has resulted from an increase in the overall scientist market combined with a decrease in supply.) Combined with the decreased incentive from indirect funding, universities would be less likely to gamble on short term gains from the research lottery and instead focus on its vast intrinsic worth. The number of NIH submitters and submissions would drop until success rates were high enough to warrant the investment.

This proposal is incomplete and certainly can be improved. But it’s something different and I believe the basis for righting the biosciences funding and employment problems in the U.S.

Discussion, criticism, and improvements welcome.

Since there seem to be a surplus of scientists, maybe we could cut back on production a bit. NSF could help through requiring grad students and postdocs paid off grant money to be paid above a minimum level - say $25k for a grad student and $40k for a postdoc. This is bound to reduce overproduction and may also lead to better quality both on incoming students and on the training offered since outsourcing more janitorial tasks may be a more attractive option than previously.

Thank you Geoff Davis for inviting discussion and potential solutions to this problem.

As a junior scientist who prudently left academia (and any remote desire to spend 80+ hours/week writing grants to survive) for better pay and better odds in industry, I think it is my duty to plead to anyone who will listen for sensible science policy.

The current academic grant funding crisis is NOT a temporary aberration due to short-term policy swings. Congressional budgets come and go. Funding increases for several years, then there is the predictable cut, and THEN there is the mad scrambling, the dying off of the junior tenure track people, the hard times and gnashing of teeth in the halls of the ivory tower.

While increased federal funding and/or complex changes by granting agencies in how grants are awarded and/or reviewed can mean career life or death in specific situations, no amount of funding or policy changes can undue the tragic consequences of a sustained glut in the academic labor market.

If you want to fix the grant funding crisis, fix the underlying illness plaguing the scientific profession in the first place! Fix the shortsightedness, the mindlessness, that fuels a perpetual overproduction of PhD's. If we want to be a healthy profession, if we want to attract the bright and the young to this career, then we must convince them that this is a noble and rewarding profession. There is nothing noble about a system that rewards rather than punishes universities and PI's for crass exploitation of the youngest and most vulnerable. There is nothing noble about a profession in which the odds are AGAINST a young scientist ever being able to feed children with their PhD, a profession in which a young woman is punished rather than rewarded for wanting children.

For all but the most established and secure scientists, this is an outrage. It is a shame upon every one of us. It is destroying this profession. And it doesn't take a scientist to understand the long-term consequences of eating your young.

The question is not whether funding/policy agencies like the NIH and NSF can make meaningful policy changes; the question is not even "how" to do it. The REAL question is "when will agencies such as the NIH and NSF be willing to make changes?" Was the NSF not responsible in the early 1990's for all those bogus alarms over phantom scientist shortages? Did not Zerhouni himself recently decline to meet face to face with disgruntled junior scientists? There will be no meaningful change until there is willingness to change. And I don't see that on the horizon.

Until then, the best and the brightest of a new scientific generation are LEAVING in droves.

(Sorry if someone mentioned such a scheme above, but I just didn't have time to read all the comments.)

By no means have I given up on coming up with new schemes for funding, but here is the most recent one that has occurred to me. I observed the type of big science that is done in labs with access to millions of dollars or euros of funding, and am of the opinion that this work needs to be facilitated. Unfortunately, it remains a minority of such highly-funded laboratories who actually produce value for money. On the other hand, a lot of really interesting science can be done for very reasonable sums. How many Ph.D.s would kill for a secure corner of lab space, a salary, and 50K a year for reagents? I would therefore propose breaking up at least a part of funding into tiers or leagues.

It would function as follows:
For any overall grants budget, the money would be divided in some ratio appropriate to the goals of the agency or the pool of candidates. Let's say 30% big league, 70% little league.

The big league grants are going to be the very juicy tidbits, the ones that the monster labs want. Grants of $1-3M per year, say. Obviously, with such grant sizes, relatively few such grants are given out, and competition is stiff.

On the other end of the scale are the little league grants. These are of much more limited size. Say $75-150K. Of course, because they are small, there are many more to be had.

Now the trick is, researchers must declare themselves "big league" or "little league", and they will be restricted accordingly. Thus Weinberg and Vogelstein will not be vaccuuming up grants in the little league. At some defined intervals, say once a year or 18 months, researchers can switch leagues if they feel their circumstances have changed.

Along the same lines, I think physical institutes need to reassess their vocations. Many seem to want to be mini pharmas or Celeras. I certainly have no problem with pushing to have big research done, but I think it's important not to neglect the bottom of the pyramid. My institute recently underwent a change in direction that was followed in short order by the expansion of three or four laboratories and the departure of a large number of junior researchers. I'm of the opinion that the productivity of the larger labs does not equal what was lost in the junior labs. (The institute must remain larger than any of its individuals.)

Last point. I like the US army's concept awards. These are 1 page applications, no preliminary data required, 5 references maximum. Most importantly, no name or identifying marks allowed on the application. A great means to foster new ideas.

I have a question...

Would it be feasible to have a permanent position for a nonscientist whose sole job is to deal with the scientific labor market, or perhaps as an advocate for junior scientists? Ideally, a labor market economist would have both the distance from science and the credibility to act in such a role. Importantly, are there any mechanisms by which such a position could be created so that it both A) has power to enact policy changes and B) is independent of other agency scientists?

This last post really is the best. Eric C Seales would want nothing more than a "scientific labor market" czar that would make the decisions. You could hire Adolf; i guess he'll do a great job. What makes you thing you would have better odds?

And why 100K, why not 10K? If we spread the honey wide, let’s make it wide enough....

There are 2 easy ways out:

• The govt. should get out of funding rich schools that can afford to find the cure for baldness by themselves.

• Or it should auction the grants to the highest bidder.

The former will never happen (too much pork involved...); the latter might have a chance.

Let's imagine that you are a brilliant researcher from a small poor school (econ theory says if you really are brilliant you would not be at the poor school long, but anyway, let's assume..), you come to the NSF for the grant. There are n! Other guys like you in there, how do you discriminate amongst so many alternative use of money? You could have a political process..like about now. or you can use the old price system. Proclaim that those are interest-free loans that need to be returned from the proceeds you get for the "thing" you invent (if not enough to cover the initial outlay you know your invention was a WASTE) . Both researcher and school would be responsable for paying back the money. NSF could also require lower returns from research done in preferred fields. I guess nobody that wasted time to write on this page (me included) would no longer apply.

PS If you want to know why it looks like a pyramid see: Lazear, E. P. and S. Rosen, (1981) "Rank-Order Tournaments as Optimum Labor Contracts," Journal of Political Economy, 89, 5, 841-64