More, smaller grants

[Aureliano Buendia]

I have been told that the Canadian system funds well over 50% of applicants for relatively (to NSF) small amounts, say $20-30k a year. Is this true? (Apparently yes, see reply below.)

This seems to be a much more humane system to me. Less time is spent on grant writing and there is more predictability for researchers planning their programs.

Certainly some of us need huge grants. But I know from experience that grants are often inflated. I've been in labs that were desperate to spend large sums of money as grant expiration dates approached.

[Aureliano Buendia]

The flagship grant program of the Natural Sciences and Engineering Research Council of Canada (NSERC) consists of "Discovery Grants." The overall funding rate for these grants in the 2010 competition was 58%, a decline from the 2009 rate of 64%. Average grant size was about $30,000 Canadian per year, with a five year grant term.

A pdf summary of the 2010 Discovery Grants competition is available from this website: www.nserc-crsng.gc.ca/Professors-Professeurs/DiscoveryGrants-SubventionsDecouverte/Index_eng.asp

I have no direct experience with this program, but it strikes me as humane and respectful.

Here is the first paragraph of the program description from the NSERC website:

Discovery Grants (DG) Program supports ongoing programs of research (with long-term goals) rather than a single short-term project or collection of projects. These grants recognize the creativity and innovation that are at the heart of all research advances, whether made individually or in teams. Researchers are free to work in the mode most appropriate for the research area.

The rest of the description includes the following highlights:

• If applicants don't want five years of funding, they have to provide a strong justification for a shorter award period, and be prepared to get the full five years anyway
  
• Researchers are free to change their research program without losing funding
  
• Discovery grants include a minimum amount designed to support one graduate student or two undergraduates
  
• The application forms appear to be quite simple, and easily completed
  

I don't understand this program well enough to endorse it. But it certainly appears to be a sensible alternative to the NSF all-or-nothing approach of writing enormous grants and getting funded on about one out of eight attempts. Imagine what you could do with your time if you didn't have to spend months each year writing grants that were probably destined for the "Not Competitive" stamp.

How do other countries handle science grants?

Source: www.nserc-crsng.gc.ca/Professors-Professeurs/Grants-Subs/DGIGP-PSIGP_eng.asp

[Anon]

While funding rates are high in Canada, the small size of the grants acts to handcuff early researchers. Yes, it is nice to have $30K of support in year one, but by the time year 5 rolls around, that same amount is stifling progress. The high success rates also means that Canadian institutions do not provide much for start-up, and investigators instead rely on CFI grants for equipment.

The bottom line is that without the ability to apply to CIHR (Canada's NIH), most researchers find it very difficult to ramp up their research programs in the first 5 years, when it is critical to do so. There are many successful programs in Canada, but almost none of them can afford to rely on NSERC alone.

[Anon2]

No! No! NO!!! The funding system in Canada is absolutely broken! As the previous poster indicated, it is very difficult for a lab to be competitive when you are receiving such small amounts of money each year. To compensate, there are other funding avenues in Canada, but they are very focussed and generally require a business angle (i.e., Genome Canada). If we want US science to regress 50 years then by all means go to a 50% funding rate with $100,000 over 5 years.

The only problem with funding rate is the lack of funds from Congress. You want to fix the system, lobby congress for more funding. Otherwise, write better grants. Trust me, many of those 75% that I have reviewed are not competitive, especially on their first submission. Step up your science or step off your soapbox. You have been peer reviewed.

[Aureliano Buendia]

Bof. Yes, I've been peer reviewed and received overwhelmingly positive evalutations. But no money. That's because NSF is broken. You can write pithy smart comments, but the fact remains, most of us who do good science aren't getting our NSF grants. Period. 10% funding rates? That's a joke.

[eclecticos]

While funding rates are high in Canada, the small size of the grants acts to handcuff early researchers. Yes, it is nice to have $30K of support in year one, but by the time year 5 rolls around, that same amount is stifling progress.

I think the Canadian system might work okay in more theoretical fields. In computer science, for example, a grant budget in the U.S. goes mainly toward

If I understand correctly, Canadian professors do not have to raise money for 1. and 2. because they are paid for by other sources (although perhaps there are fewer paid-for RAs per professor than in the U.S.?). I can imagine that $30K/year might be able to cover 3. and 4.

Can the above posters comment on this? I haven't checked my thinking with my Canadian friends.

[eclecticos]

Also: The previous poster says that average size of a Canadian Discovery Grant is $30K * 5 years. How much variation is there, and can one investigator get several grants to support different research directions?

[eclecticos]

Here is an idea for improving the process.

I like reviewing small & medium grants, because they make specific and clear proposals. Reading such a proposal is like reading the introductory sections of a few related journal articles. It is also valuable for the researchers to write these proposals to work out their plans -- writing the proposal is the first stage of research. In general, I believe that my own field (I can't speak for other fields) benefits most from small science -- smart people exploring good ideas in a curiosity-driven way.

I don't so much like reviewing large grants, because in my own field (again, I can't speak for others) they tend to be grab bags. A large grant proposal contains lots of ideas of varying merit, tossed in by researchers of varying quality, some of whom are just hangers-on who are at the same institution as the good people. There is insufficient description of the individual ideas and not enough of a story about how the researchers will work together.

Thus, I would rather require separate small/medium sub-proposals for the individual efforts within a large grant. Each would be written by 1-2 researchers who are justifying that effort. Perhaps these sub-proposals should be considered in the context of an overview statement for the big grant, so that overall motivation and cross-project synergies could be taken into account. But each sub-proposal should be rated separately, and it should have to compete with ordinary small/medium proposals that were not submitted under the aegis of any big grant. NSF should be able to decide which ones to fund, like a line-item veto.

Presumably, each sub-proposal would have its own PI who would largely control the money allocated to that sub-proposal, although there might be a role as well for the PI of the whole grant. Some graduate students might divide time among 2 or more sub-proposals since they are related.

Note: In my directorate, one must usually designate one's proposal as "small," "medium," or "large" according to the budget size. The different sizes are paneled separately to allow fiar apples-to-apples comparison within a given size. But I am not sure how the competition between the different sizes is currently handled. My impression is that the program managers allocate separate pots of money for the different sizes, but that they may adjust based on the quality of reviews.

Request: Can someone (e.g., the thread initiator) please post information about how much of each NSF directorate's budget goes to small, medium, and large grants? And what the funding rates for each category are? I am sympathetic to the argument that too much money is going to large grants, but I would like more data to be sure.