Applying for an NSERC Discovery Grant

> français

SSCThis is a reprint of an article by Duncan Murdoch, Paul Gustafson, and Mary Thompson in SSC’s Liaison Volume 28 No. 3 in August of 2014.

Discovery Grants since 2009

Most academic statisticians in Canada receive research funding from NSERC, CIHR, SSHRC or their home university. In this article we will discuss the NSERC Discovery Grant program and describe resources to help you write a successful NSERC Discovery Grant.

There have been major changes in the Discovery Grant application process in recent years. Over the 2009/2010 competition years (i.e. for applicants who applied in the fall of 2008 or 2009), NSERC merged 28 grant selection committees into 12 evaluation groups. The statistics and probability GSC 14 was merged with two mathematics groups into the present EG 1508 in Mathematics and Statistics. Within that group there are 3 sections, corresponding to the original GSCs, but group members may move between sections for particular cases, or visit other evaluation groups if their expertise is needed. The evaluators assign scores to each applicant to place them in a “bin”, and the final amount of funding is determined by that bin, almost always at a common level across all three sections within EG 1508. (The 2011 competition year was an exception: because the scores in mathematics were consistently higher than those in statistics, the sections were split.)

For the 2014 competition year, NSERC left the evaluation process unchanged, but changed the application procedures, doing away with the free-form Form 100 that described past contributions, and adopting a variant of the Canadian Common CV. The Form 180 Notice of Intent to Apply for a grant was also modified significantly, and renamed as the Letter of Intent (LOI).

One of us (Murdoch) was a member of the Statistics Section of EG 1508 for the 2011-2013 competition years. It was my observation that while all group members make an honest attempt to consistently apply calibrated scores, there continue to be differences between the sections in “culture,” with consequent differences in the distributions of scores (albeit smaller differences than in 2011). The SSC Research Committee has had a continuing discussion with NSERC about these issues, and NSERC has made efforts to address them, though they are not completely successful: in the results reported by NSERC at the SSC annual meeting in Toronto, there was still a noticeable difference in distributions between mathematics and statistics.

The Problem of Biostatistics

There have been expressions of concern within the Canadian biostatistical community that biostatistical research is prone to falling in a gap between NSERC and CIHR. Of course biostatistics is but one of many research disciplines that crosses over between the natural sciences and engineering (NSE) on one hand, and the health sciences on the other. The federal legislation that established NSERC in 1978 specifically excludes health research from its mandate. However, particularly in today’s landscape where interdisciplinary research is so prevalent, the boundary between the NSE and health research is necessarily blurred.

NSERC is actively involved in clarifying its message concerning this boundary. Late last year they posted a summary document titled “How NSERC determines whether a Discovery Grant application fits its mandate” (available at http://www.nserc-crsng.gc.ca). And earlier this year they convened an External Advisory Committee on Mandate and Subject Matter Eligibility. This committee is comprised of researchers from various disciplines where crossover between the NSE and the health sciences arises. The biostatistical community (and the statistical community more generally) are represented on this committee by Paul Gustafson from the University of British Columbia. The committee’s deliberations are still ongoing, with the aim to help NSERC sharpen its guidance to the research community concerning subject-matter eligibility.

As for the current situation, please be reminded that the document “Selecting the Appropriate Federal Granting Agency” (www.science.gc.ca) explicitly lists “Research whose major challenges lie in the NSE (materials science, engineering, computer science, chemistry, etc.) which could eventually lead, among other applications, to the treatment and prevention of human disease” as eligible for NSERC support. While it would be nice to see statistical science mentioned explicitly, we can take it as read to be included via the et cetera. So it does behoove applicants with biostatistical proposals to stress first and foremost the statistical science challenges in their proposed programs of research.

CANSSI

At the CANSSI annual meeting on May 24, 2014 in Toronto, several in attendance suggested that CANSSI too should play a role in supporting NSERC applicants for Discovery Grants and other more targeted programs. CANSSI will be posting on its website articles which have appeared in Liaison over the past few years on programs of NSERC and other funding agencies, particularly articles with advice to applicants. CANSSI also plans to post notices of post-doctoral fellowships and collaborative opportunities. Look for these under the Research and Training Opportunities tab on the CANSSI website http://www.crm.umontreal.ca/CANSSI/.

Writing a Successful Grant Proposal

NSERC’s aim in the 2009 changes was to make sure that grants were awarded on the basis of the merit shown in the proposals. What this means is that all researchers need to write their grant proposals carefully: your reputation in the community will likely play a smaller part than it did in the past. In particular, we would emphasize the following three points.

  1. Read NSERC’s Peer Review Manual (available online at http://www.nserc-crsng.gc.ca). It contains the criteria that the evaluation group will be using. Don’t assume that the evaluators will fill in any blanks: if your proposal misses something valuable that you have done or propose to do, you will likely receive a lower score than you could.
  2. Prepare your first draft early. The deadline for submission is late October. For most of us, that is a busy time of year with teaching duties. We would suggest that you should aim to have a first draft of your proposal at least six weeks earlier, with most of the work done before teaching starts. This will give you time to think about your proposal, to have others read it and offer suggestions, and to polish it before final submission.
  3. Ask others to read your proposal and to comment on it. If a colleague is a member of EG 1508, s/he will not be allowed to participate in the evaluation of your proposal at NSERC due to a conflict of interest, but may be an invaluable resource to you to help you to write it so that it will be received well there. SSC members at smaller institutions and first-time applicants who have difficulty finding colleagues to comment on their proposals are invited to contact Duncan Murdoch (murdoch@stats.uwo.ca) who will try to coordinate volunteer readers at other institutions. Experienced researchers who are willing to volunteer to help in this way should also contact him.

Besides these three short points, readers are referred to longer articles in past issues of Liaison. The August 2012 issue (v 26, no 3, p 57) contains questions and answers about preparing a successful grant proposal, and Dave Stephens’ article in August 2011 (v 25, no 3, p 52) contains a detailed discussion of the evaluation criteria. The only significant change that we are aware of is described in the 2012 article: the rules for determining eligibility for NSERC funding were made more stringent after the 2011 article was written.

DuncanMurdoch

Duncan Murdoch

PaulGustafson

Paul Gustafson

Mary Thompson

Mary Thompson

Comments are closed.