Success rates continue downward spiral with latest NSERC Discovery Grants competition

New selection process implemented

Greater numbers of returning applicants, higher quality research and stagnant funding have resulted in a further decline in the success rate in the latest competition for the core research grants supporting scientists represented by the Natural Sciences and Engineering Research Council (NSERC). The recently announced results of the Discovery Grants Program (DGP) competition produced a 58% success rate for the 3,355 researchers that applied, down from 63.5% in 2009, 71% in 2008 and a high of 87% in 2002 when substantial new funding was being injected into the program.

Just six universities scored a success rate of 70% or better: HEC Montreal (80%), École polytechnique (78%), Queen's Univ (73%), Univ of Waterloo (73%), Univ of Toronto (71%) and McMaster Univ (70%)

The average grant size in 2010 also declined from $38,142 last year to $35,143. Established researchers who already held a DGP grant achieved a success rate of 72% and received an average of $36,333, while only 28% of those who did not previously hold a grant were successful for an average of $27,552. Early career researchers achieved the same success rate as the overall competition (58%) but with a much smaller average grant size ($23,930).

returning applicants staying in system

Returning applicants numbered 921 and comprised 27% of the competition, up significantly from the 2003-2005 timeframe when they accounted for between 9% and 11% of the overall competition. The result is less money being freed up for other applicants.

Overall, the DGP allocated $350 million for the 2010 competition, roughly the same as 2009. That amount could rise slightly if NSERC council decides to apply the $8 million it received for advanced research in Budget 2010 to the recent competition — a decision that should be made by June.

"It is worrisome. There's going to be a new equilibrium that sets in. At some point, unsuccessful applicants will try three- four or five times and then cease to try," says NSERC Isabelle Blain, VP of NSERC's research grants and scholarships directorate. "At the same time, the quality bar is going up … The quality of research is just remarkable and those researchers are becoming the norm for the others, producing astounding results and becoming the comparator in the competition."

conference model implemented

The DGP competition marks the first time NSERC has fully implemented the so-called Conference Model to review applications, reducing the number of grant selection committees from 28 to 12. The Conference Model was recommended by the 2008 DGP Structure Review panel, chaired by Dr Adel Sedra, dean of engineering at the Univ of Waterloo (R$, June 20/08). It involves panels with at least 30 members organized along thematic lines that can be split thematically into topic groups that can decide on specific areas of research. It is considered ideal for assessing research proposals that are multi-disciplinary or cross traditional boundaries, bringing together optimum expertise when and where required.

"The committee members were pleasantly surprised at how smoothly it went. They moved from room to room and exchanged between disciplines with lots of up and down elevators," says Blain. "There are much larger groups now but parallel streams based the topics of proposals. We assign people according to expertise ... The model helps bring experts from other disciplines and increases the sense that you have the right expertise to assess the applicants."

In addition to the Sedra report, the DGP was also subject to an international review panel, which deemed the program to be an exceptionally productive investment for Canada that fostered a high degree of research excellence and afforded researchers a base to lever internationally competitive levels of funding from other sources (R$ May 20/08). Chaired by Dr Peter Nicholson, the panel recommended that NSERC switch to a so-called bin system to rank proposals according research excellence before determining how much each successful researcher will receive. The two-step process has been implemented and appears to be working well. The international panel also made two key recommendations that have yet to be implemented.

unfunded recommendations

One is to increase the funds available to the DSP so that average grant sizes do not shrink further. This recommendation has proved impossible as the federal government has not increased non-targeted allocations to the granting councils.

The other recommendation that remains unimplemented is a doubling of accelerator supplement grants to 200, using new funds. Accelerators are viewed by NSERC as "an effective tool to quickly elevate funding levels of meritorious individuals and accelerate the progress of their research".

NSERC council may consider implementing this recommendation when it convenes in June.

Last year the mathematics research community was critical of the DGP success rates, prompting the Mathematics Committee for Liaison with NSERC to issue a letter urging the government to make "crucially needed investment in Canadian research via NSERC and the other federal granting agencies". The success rates for that community fell from 77% in 2008 to 64% in 2009. This year, the success rate fell again to 62%, although early-career researchers achieved a 73.2% success rate.

R$