I’m in the middle of applying to have my NSERC Discovery Grant renewed. I’d almost forgotten how much time and energy the process can consume. I’m not really in a position to take it lightly, either, given that the funding for my graduate students depends on the outcome of the competition. (Yes, this is one of the reasons for the slow blogging rate. That, and I’m feeling a bit burnt out.)

This is the first year that the new formatting requirements for Form 101 are in place. According to the NSERC instructions (we also received an email confirming this), we have to divide our 5-page proposal into sections, using the following headings:

Recent Progress

Describe your recent progress in research activities related to the proposal and, in addition for renewals, the progress attributable to your previous Discovery Grant.

Objectives

Define the short- and long-term objectives of your research program.

Literature Review

Discuss the literature pertinent to the proposal, placing the proposed research in the context of the state-of-the-art.

Methodology

Describe the methods and proposed approach, providing sufficient details to allow the reviewers to assess the feasibility of the research activities.

Impact

Explain the anticipated significance of the work.

This has me wondering: is it really just a formatting change? Or does it signify a change in NSERC’s policy on research funding, a shift from funding research *programs* to research *projects*?

The required format would fit me just fine if I were, say, a graduate student submitting a thesis proposal. In describing the tentative project, I would of course have to explain the objectives, the progress so far, and everything else on NSERC’s list; and I might as well do it in the suggested order.

As a senior researcher, though, I’m finding this format less than helpful. A senior researcher’s research program typically encompasses several areas or subareas of work. In each of these areas, we have one or more specific projects at any given time, in various stages of completion. These projects are often interrelated or even feeding directly into each other, but they’re still separate projects, each with its own objectives, literature and methodology. My proposal will probably end up looking like this:

Recent Progress

Recent progress in Area 1. Recent progress in Area 2. Recent progress in Area 3.

Objectives

Objectives in Area 1. Objectives in Area 2. Objectives in Area 3.

And so on, in each of the 5 categories. The criteria above are perfectly reasonable and it makes sense to require that each of them be addressed in the proposal. It would make more sense, though, to drop the rigid formatting requirement and allow applicants to talk about their work in Area 1 first – goals, literature, methodology, everything else – then in Area 2, then in Area 3. As it is, each part gets chopped up unnecessarily and artificially, and, given the 5-page limit, you never get to say more than 1-2 paragraphs about any one thing at a time.

Now, I’ve seen similar formatting requirements for junior researchers, but then they do often only have one research project at a time. (Except when they don’t.) I’ve also seen language on NSF webpages that indicates support for “research projects” rather than “programs”, but in actual practice as I know it (I have submitted applications to the NSF and have served on an NSF panel), project descriptions at the NSF are not much different from what research program descriptions used to look like at NSERC. Everyone understands that senior scientists in general, and mathematicians in particular, usually work on a number of interconnected projects and that the total sum of these projects constitutes the activity that should be supported.

Indeed, a case could be made that senior mathematicians **must** propose to work on more than one thing. The duration of an NSERC grant is 5 years. If a senior mathematician produces 1 paper in 5 years, maybe 2 papers on the same problem, and these papers don’t happen to prove the Riemann hypothesis or at least the twin primes conjecture, there’s precious little chance that the grant will be renewed, and rightly so. There’s a long list of reasons to have at least a few projects going on. For example, we don’t know in advance what will work out and what will not, so having several projects insures us against putting all eggs in one basket and then breaking them all; ideas brought from one project can help with another one; and it’s easier that way to generate projects for students and postdocs, an activity that NSERC has been putting much weight on lately.

That, and it’s not always clear how to chop up each part of the proposal. In mathematics, the *objective* of our research is often to develop a *method*. PDE come to mind as a good example. If you find some differential equation that hasn’t been solved before and solve it by using well known methods, that’s a student project, possibly a Master’s thesis or something like that. If you solve a differential equation that’s beyond the reach of the currently known methods, thereby introducing new methods in the area, you might have a major paper on your hands.

Another example that might be even better (I’m borrowing it from Tim Gowers’s excellent Two Cultures paper) is Paul Erdös’s discovery of the probabilistic method in graph theory. The actual result (a lower bound on the Ramsey number of a complete graph on N vertices) is almost incidental to the real discovery here, which lie in introducing a new point of view on this type of problems and paving the way for the now-ubiquitous probabilistic arguments in combinatorics. With this in mind, would the introduction of the probabilistic method belong under “methodology” or “objectives”?

Because in hindsight, it should really count as “impact”.

My understanding is that funding research projects, rather than programs, is the norm in much of the sciences. As you say, math grants don’t really work that way, but mathematicians are a very small percentage of scientists. So my theory would be that they weren’t thinking about mathematicians at all when changing the format.

One interesting problem caused by the project/program issue, is that you’d expect very different rates of success in the two cases. If you’re funding projects, that means that everyone submits several grant applications each and only some of them get accepted. Whereas for funding programs you have many more people getting 100% of their proposals funded (because they only made one proposal). Apparently this is a problem in the US because math has “too high” a percentage of proposals funded relative to other fields.

I’m also applying for a DG this year, and am reading the rules to mean: describe

‘Area 1 first – goals, literature, methodology, everything else – then in Area 2, then in Area 3.’

Maybe this is a flawed interpretation, but I doubt that a proposal will get thrown out for such a rearrangement. Otherwise, as you say, the proposal is a very choppy read indeed.

In correspondence with M. Bastien of NSERC, I received the following advice. (Sorry it is posted here too late for 2012 applicants…..)

“In response to the question you asked me on the phone yesterday with respect to the methodology section of an application, I contacted the Executive Committee for 1508 (Group and Section Chairs) and here is a composite of their answers. Hopefully this will help.

I get this question all the time, and my reply is basically this. When I give a problem to a PhD or postdoc, I don’t know how to solve it. However, I always have some sort of idea of the type of approach that might work, and I describe this to the student to get them started. This is exactly the sort of thing that we are looking for under Methodology, particularly for those projects that fall under short term objectives. It is not enough simply to say that you will prove Theorem A: you must be able to convince the reader that you have some sort of handle on the problem. As far as the longer term objectives are concerned, this isn’t so obvious. Often it is only through solving the initial problems that you get a handle on the ones that follow, and so we don’t expect to see much detail here.

The applicant needs to convince the panel that he/she has a good appreciation of the complexity of the various problems posed and some solid ideas on how to tackle them.

Any mathematical expert is presumably an expert in some technique(s) of proof, solution, or analysis. And if they claim to be doing research on a mathematical problem, then they have to have some plan for attacking it and some method(s) that are suitable for doing that.

Methodology would be the bit where you describe how precisely you are going to achieve your Objectives.

I would also like to point you to the Peer-Review Manual (the 2011-2012 version has been posted on our web site last week), to section 6.6.1.2 which describe what the Evaluation Group looks for when evaluating the criteria ‘Merit of the proposal’.

http://www.nserc-crsng.gc.ca/NSERC-CRSNG/Reviewers-Examinateurs/IntroPRManual-IntroManuelEP_eng.asp “

I certainly agree with that. That’s how I have been writing proposals all along, and I might have given very similar advice myself to a first-time applicant. My point was that the required headings are at odds with communicating the content clearly and efficiently, especially when the total length is restricted to 5 pages (NSF has 15) and when there are several different groups of problems to cover.

It’s NSERC’s general practice not to fund SSS rated application.

Lev Idels ( Math Department, VIU, Nanaimo BC)