Submission to the LRP steering committee

Below (under the cut) is the text of the submission I am about to send to the Long Range Plan steering committee. I missed the April 18 deadline for submission of discussion papers, basically because I was too busy and exhausted at the end of the semester, but the committee web page states that “comments and ideas are welcome at any time”, so here are mine. There’s very little here that I haven’t already said on this blog in much more detail (the relevant posts are linked below) and it’s possible that some of the committee members have seen those posts already; this is just a short summary. (A PDF version is also available.)

1. According to the just announced 2011 NSERC Discovery Grants statistics, the Discovery Grants in mathematics are by far the lowest among all disciplines funded by NSERC. Average grants in every other disciplines are 70% to 200% higher than in mathematics. This is not acceptable.

(The equipment and materials cost in other disciplines may be higher than in mathematics, but on the other hand, such expenses are often covered by other grant programs. Anecdotal evidence available to me suggests instead that the inequities between the DG grant levels translate into similar inequities between HQP salaries in mathematics and other disciplines. I would be very interested to see this studied systematically.)

2. The government should restore basic research to its central role in science and fund it accordingly, rather than divert funds to narrowly targeted “innovation” programs of little long-term benefit.

3. The allocation of research funding within mathematics should be based on the quality of current and proposed research, not on politics. This means that the current trend of supporting large research groups based more on their size than on scientific merit (“critical mass”, “priority areas”, etc.) needs to be reversed. Such practices reward success in departmental politics, especially hiring battles, and penalize those researchers who actually prefer to focus on their research. Specific examples of this include the following:

(a) The current emphasis on HQP training in the DG competition puts isolated researchers at an unfair disadvantage. (This includes most faculty in small departments and a smaller but significant number in large ones.) A large group is more likely to have established recruitment pipelines, stable funding, advanced graduate courses offered on a regular basis, etc. It is much harder for an isolated researcher to rack up a comparable HQP record.

(b) The PIMS Collaborative Research Groups program is based explicitly on politics. Only sufficiently large groups (usually at least 10 faculty) are eligible to apply, and the geographical and institutional constraints are not compatible with the natural research-driven process of forming collaborations. The program is tailored perfectly to support large and politically influential groups within departments (such as were formed at UBC in accordance with the 2001 academic plan), but others are either excluded from the opportunity or else forced to form artificial and unproductive alliances based on the need for funding.

4. I do not support any community plan that involves naming (and singling out for support) “priority research areas”, in any form, under any circumstances. As good and merit-based as it may sound in theory, in practice it will only lead to a further politicization of mathematics, with the largest groups taking the lion’s share of every available form of funding.

Canadian mathematics is not a world unto itself. “Isolated” researchers need to be supported, not out of charity, but because they’re not necessarily isolated in the greater scheme of things. In addition to having strong research programs, they’re often part of large and vibrant international networks of collaborators, opening up Canadian mathematics to new developments worldwide. Modern mathematics relies increasingly on creative interplay between many diverse areas of research. Focusing on only a few areas (and all but locking everyone else out of the funding system) will lead to stagnation, inertia, increasing isolation from the worldwide mainstream, and ultimately long-term damage to Canadian mathematics.

5. I do not support the “envelope” option where, apparently, individual researchers will have to compete against the institutes for their research grants. The Discovery Grants budget for mathematics should be (a) higher and (b) actually dedicated to Discovery Grants, rather than subject to clawbacks for the institutes or for any other purposes. The institutes, unlike us, have significant political clout and substantial resources (human and financial) that they can apply towards campaigning. We cannot compete against that. The institutes and the Discovery Grants serve different purposes and are not interchangeable; both are valuable and should be funded, but not at each other’s expense.



Filed under academic politics, research funding

10 responses to “Submission to the LRP steering committee

  1. Ryan Budney

    I’ve often wondered if the Discovery Grant program is a bad fit for mathematics. As far as I can tell they’re reasonable in Ontario and Quebec where there is provincial funding for grad students. But in BC, and even moreso at smaller universities like Victoria (you people at UBC seem to have funds for grad students that we lack), it’s almost impossible to support a single Ph.D student on one discovery grant + TA money unless your student has other sources of income. For starters, you wouldn’t be able to spend any of your NSERC money on travel, visitors, conference expenses, etc. I believe to support a Ph.D student at Victoria who has a full TAship we have to find an additional $12,000 per year. Perhaps a model where funding is more stable and decisions made more locally would be an improvement — say for roughly every 5 to 6 researchers in a departments there would be a “group grant” of about $100,000/year for which the group would make spending decisions for. But perhaps that would introduce too much friction inside departments. Or perhaps a funding model where base grant support is separated from grad student support — so if you found a good student to support the money would be there, but if there aren’t any appropriate students at the moment, you don’t feel like your future ability to support students would be compromised by not having any students for a year or two.

    The standard solution to all this would be to have a university system where the university itself values the support of grad students. U.Vic does not have any systematic support of grad students — just a few grants here and there from the Faculty of Grad Studies. The quantity of the grants and whom to award them to is out of our control as far as I know. This makes long-term planning difficult.

  2. The Discovery Grants would be perfectly adequate to support graduate students if they were brought up to the same level as in other NSERC disciplines: 20-30K per year for early career researchers and 30-40K for established researchers, on average.

    At UBC, the standard research amount for graduate students with a full TA-ship is 8K per year. It has not been raised since time immemorial. It should be, but we can’t afford to do so, given the current situation.

    I’m not sure about the shared “group” grants. Maybe it should be allowed for certain groups that actually work and supervise HQP together, just like NSF allows such researchers to apply together under their individual grant program. (One person acts as the principal investigator and the others are co-PIs. There are famous examples of that involving very well known mathematicians.) But it would not be a good idea at all to force *everyone* into such structures, against their own judgement. That’s kind of the point I was making.

  3. Ryan Budney

    I suppose the point I was trying to make about group grants is they wouldn’t be on specific subject areas. They’d be administrative groups, as in “you five people are responsible for this $100k per year”. If a small subset of the group is particularly successful with grad students they could be encouraged to apply for individual grants. And I do have a few NSF models in mind when I’m thinking about this. It seems like this model can work quite well, but maybe I haven’t seen enough of the landscape to really have a feel for potential pitfalls.

    One of my main concerns is the emphasis on production of “HQP” in our granting process. There does not appear to be a serious criterion for success, other than the number of Ph.D’s and M.Sc’s granted. I mean, it seems like any outcome other than persistent unemployment can be spun as “success” for a graduate.

    I’m still something of a newbie when it comes to NSERC. And I have to admit I’ve been skeptical of the institution for some time. When I was an undergrad (at U.Alberta) it seemed like NSERC’s only criterion for undergrad and graduate awards was GPA, even if high GPAs were achieved by deliberately taking soft courses rather than challenging yourself.

  4. The pitfall in having a grant shared by people who do *not* work together is the fights over how the money is spent. I’ve seen this in my (limited) experience with PIMS CRGs, which do have a nominal subject, and purely administrative groups would be worse. There’s no way that everyone would want to hire the same postdocs or students, and what if several group members have excellent candidates they want to support and there isn’t enough money for all of them? How do you make the bargains? How do you enforce the other ends of those bargains? Isn’t going to work. “Tensions” wouldn’t even begin to describe it.

    I agree that the “production” of HQP measured by numbers is a concern.

  5. Ryan Budney

    Regarding having grants shared by people who do not work together, I suppose the main criterion would be to get grants with people you know you can cooperate with. One does not need to work on a common topic to do that — having an understanding of each other’s work might be enough for most.

    I haven’t witnessed the CRG cycle myself. I’ll likely be putting in an application soon(ish) so I’m curious to see how it goes.

  6. Cooperation is difficult when your goals and interests are mutually exclusive, and impossible when a power imbalance is also thrown in the mix. As a bare minimum, such groups would need to be voluntary *and* more or less in the same area, and even then it doesn’t always work out, not for all of us anyway. Right now we have a group of 4 harmonic analysts at UBC. Some of us work on common topics, some don’t, but we do sponsor postdocs and students together and have had no major clashes about that. But what if you’re isolated in the department, as I was for a long time, and you’re assigned administratively to a group of 4 people in a different area who are not willing to accommodate you? Your funding is lumped together with theirs, but you’re outvoted every single time a decision is made, so you never get a student or postdoc? And you only find out that about them after all the paperwork is signed and the funding is released to the group?

    As for CRGs, I’m planning a separate post about that.

  7. Ryan Budney

    I suppose if I was in the hypothetical situation you describe, I’d move to another university. I’m the only topologist at U.Vic (we have operator algebraists, graph theorists, dynamicists, some partial differential equators, and one number theorist here) and I feel like people have been quite helpful when it comes time to scrounging-up resources for my grad students and visitors.

    Moreover, I can’t imagine ever applying for a grant with someone whose goals are mutually exclusive to my own — certainly nobody in my department fits that description. I suspect the only people in math that have goals close to being exclusive to my own would be people who I collaborate with: me trying to prove certain theorems, them trying to show that the “theorem” is false, etc. But maybe I didn’t understand the context of your comment.

  8. By “mutually exclusive” I mean a very simple calculation: you’d like to spend the grant money on X, the rest of the group would like to spend it on Y, and there isn’t enough money for both. Right now, the arbitration takes place at the level of the Discovery Grants competition and is based on merit. (Or should be, but that’s a whole new discussion.) On the other hand, the funding decisions within the groups you suggest would be based entirely on politics. Unless everyone in the group actually wants that sort of arrangement (and presumably can agree on the postdocs they want etc), I can’t see how that could possibly be a good thing.

    This is a very different situation than, for example, colleagues helping you write a better proposal or making *voluntary* joint sponsoring arrangements, which I suppose is what you are talking about. And as I said before, if several people actually want to apply together, I have nothing against that. I just don’t think that it should be mandatory.

    It would be really nice if everyone, everywhere, was friendly and helpful. That’s not the planet I live on. The point of having arbitration systems (such as grant evaluation panels) is to allocate resources fairly between people who might not be naturally inclined to pool their money together, and thus preventing conflicts that might arise without such arbitration. You end up in a certain negotiating position from which you can then make deals if you wish, or not make deals if that supports your research program more efficiently. That’s better than having to move across the country every time you don’t get along with your colleagues, especially with a funding system that goes out of its way to create opportunities for colleagues to not get along.

  9. Pingback: Weekly Picks « — the Blog

  10. Nilima

    Nice post, Izabella!

    I’d go further and say: the ambiguous nature of the HQP criteria, and the large variations in funding, will make Canadian science more risk-averse than ever before. This is disastrous if we’re to have serious impact.

    Let me clarify. Some projects, particularly in emerging fields or topics at the interface between fields, are ‘high risk’. One doesn’t know if they will work out spectacularly, or produce a damp squib. But one will not know unless one tries. So one takes a professional gamble. To involve a student in such an enterprise may be unprofessional – there is an emphasis on ‘getting them out the door’ in universities, and working on a truly risky project is a luxury. However, unless one involves a student to boost the HQP thingy, one likely will not get funded for the project.

    Large variations in funding are bad.

    Take the following hypothetical: in year Y, I get an NSERC grant of $X. In year Y+2, I take on a graduate student, who (on average) is slated to graduate in year Y+7.5. However, I re-enter the NSERC competition in year Y+5. Even if I’ve moved up a bin or two, it’s entirely possible my funding is now $X-10000. So what happens to the student? S/he still has 2.5 years to go!

    So maybe I should never take on a student in year Y+2? How about in Y+1? I can’t take on a student starting in year Y, because I don’t know the results of that competition.

    In other words, one can be intellectually brave, one can be professional responsible to one’s students, or one can be successful in the current DG system. These are Hobbsian choices.