Joshua Rosenau spends his days defending the teaching of evolution at 
After crunching the numbers on the trends in NIH funding, some interesting patterns emerge.
Clearly, the drop in funding of both R01 (initial grants) and R02 (renewals) are substantial over the Bush years. But intriguingly, the average size of a grant funded has risen. This may seem like a good thing. After all, more money is good, right?
Not necessarily.
The size of a grant will reflect various things, including how ambitious the project is, how technological it is, how many people are involved, etc. But it fundamentally reflects how much anyone is prepared to invest in it. Big grants go to research groups with well-established research programs. They are applying for a first grant on a project, but they are probably applying known techniques and proven strategies. They may be refining a technique that worked in previous studies.
Big grants won't generally go to people who are trying to establish a totally novel approach. Those initial projects will be funded with smaller grants.
The drop in the number of grants funded combined with the rise in average amount funded is actually a bad sign. It suggests that federal science funding has become more risk averse.
These sorts of federal grants are supposed to be funding innovation. They are meant to promote risky work that has big potential benefits to society. As the author of the Science paper points out:
Peer review cannot discriminate among and accurately select only 1 of 11 meritorious applications.The researchers only looked at unamended applications. Some of these will be resubmitted and approved. As the author explains:
each revision of a rejected application delays by close to a year the time required before support can be approved and research initiated. For type-1 applicants, this is a slow, uncertain process that often leads to career reevaluation and change by otherwise successful professional contributors. For an ongoing and previously approved type-2 research activity, rejection casts major doubt on eventual continuation and frequently results in breaking up teams of highly trained personnel. Therefore, success rates for funding initial applications are of primary importance.And their decline bodes ill for American leadership in the biomedical sciences.
Comments
I would like to amplify on two things from your post, based on my experiences as an applicant for and recipient of NIH R01 funding:
"But intriguingly, the average size of a grant funded has risen."
One thing to keep in mind about this is that the majority of R01 direct costs--and usually about 2/3--go to pay personnel: some percentage of the PI's salary (plus fringe benefit expenses), and usually 100% of the salaries (plus fringe) of several post-docs and/or graduate students. The salaries of post-docs and graduate students paid off of R01s are usually tied by the recipient institutions to the NIH fellowship pay scale, which has increased astronomically over the last decade.
So, these increased R01 budgets do not even come close to matching the increased salaries and other inflationary pressures on the cost of doing research. The increases in R01 budgets actually mask a substantial *decrease* in the average scope of research project supportable by an R01.
"The researchers only looked at unamended applications. Some of these will be resubmitted and approved."
This is an important limitation of their analysis, because the proportion of applications that succeed in being funded on their first submission has plummeted in recent years, and the numbers of applications funded only as a first or second resubmissions has skyrocketed. I have generated a PDF graph of this trend based on dated I obtained from NIH's publicly available CRISP database, which I could e-mail to you for posting if you would like. (I would post it myself, but I do not have a convenient venue to do so.)
Posted by: PhysioProf | September 13, 2006 5:00 PM
I've always received my funding in the earth sciences from NSF, which has also struggled since 2000. NSF program managers used to look to NIH as a target for the level of finding they'd like to receive from Congress. The decrease in NIH funding is really just a bellweather for a decline in funding across the sciences and should be a cause of concern for everyone.
Posted by: Jackie at ElementList | September 13, 2006 7:05 PM
FYI, there's no such thing as an R02, as far as I can tell. Extramural individual investigator grants are R01s for their entire life, including multiple renewals. R01's get a suffix for the year of the grant, and another suffix for the number of times its been submitted. Thus R01-{grand ID num}-05A1 would be for a grant being resubmitted for a competitive renewal after 5 years of funding, after it was turned down once.
I'm not sure what numbers you're using. Are you just plotting the numbers from the Science paper?
Posted by: Jim Hu | September 15, 2006 11:16 AM
He's referring to the so-called "Type 1" (new submission) and "Type 2" (competitive renewal) grant applications. These are actually prefixes to the grant numbers you refer to. So, a first submission of a new grant application assigned to the NINDS would be numbered 1R01-NS054678-01, a first resubmission of that application would be 1R01-NS054678-01A1, a second resubmission would be 1R01-NS054678-01A2, a non-competitive renewal ("Type 5") for the, e.g., second year would be 5R01-NS054678-02, and a competitive renewal (assuming the original grant was for five years) would be 2R01-NS054678-06.
There are also other, more obscure, grant application types. For example, if I am not mistaken, Type 7 is an application for a change of institution and Type 3 is an application for a competitive supplement.
Posted by: PhysioProf | September 15, 2006 2:53 PM
Yes, I'm plotting numbers from Science, and I did get my labels wrong.
Posted by: Josh | September 15, 2006 9:08 PM