How to reform scientific grants?

Using the economic theory of contests, Gross and Bergstrom modeled a controversial alternative: awarding grants instead by partial lottery. Under a partial lottery system, funds are awarded by random draw among a pool of high-ranking grants — the top 40 percent, for example. Since applicants would be aiming to clear a lower bar for a smaller prize — a shot at the lottery instead of a guaranteed payout for winning proposals — the contest theory model predicts that applicants would spend less time trying to perfect their applications, Bergstrom said.

Here is more from James Urton, and more here, via Charles Klingman and also Michelle Dawson.


Interesting idea. The current system leaves a lot to be desired. However, I think the equilibrium would be many more applications submitted. Instead of working a lot on their best idea, people might submit several and have their third best idea funded. I'm not sure it would end up better.

You could limit applicants to one submission per call, but this would still happen by applicants making sure they have something in every call.

Don't reform them, end them. It is nothing more than welfare for professors and graduate students.

I'm inclined to agree, but there is useful science that government may and should fund. Perhaps such grants should go only to research that contributes to an important government function rather than science for science sake.

Why is it that, well I'm not sure whether I should call you a modern conservative or a modern populist, but why is it that you do not consider scientific progress and economic growth good?

Because clearly if you considered growth good you would not blinker your view with "government shouldn't help."

You would look at actual results, he said on the internet. Invented with that bad old government money.

Economic growth is good. Using tax money to hand out free stuff is bad and is NOT economic growth. In fact taxing by itself restrains economic growth and should be done with great restraint not with great excess.

It depends on what you think the shape of the quality curve looks like among those top 40 per cent, and whether "perfecting" is really wasteful.

As with many proposals for reform such a system would also diminish the opportunities for favoritism, so very unlikely that those in control of the current system would be in favor of the change.

What if academia is full of people who have specialized in perfecting applications, who have been reinforced to think their success is based solely on merit, who believe a natural hierarchy exists within their field and is justified, would they not revolt if a partial-lottery is implemented?

Does the application process itself improve the quality of research, by eliminating false starts or spotting future problems, that would have happened if the grant had been granted?

Looking at it only from the application stage ignores what happens after the grant has been given, unless you are also relying on the goddess Fortune to allocate future success.

A thorough application process may be a quality control mechanism, and protect even the applicant from wasting time and resources.

Given the professed beliefs in academia, one would think the obvious solution would be to divide the available funding equally among all applicants.

There’s so many other places where this partial lottery would be a good idea. I think it’d be great for hiring for many highly competitive jobs. If you have 200 applicants for a faculty job, narrowing down to the top 25% won’t be too costly, but narrowing it down to the top 0.5% takes immense amounts of time. So the marginal cost of narrowing it down in the later stages is huge, but the marginal benefit is probably trivial, since the top 25% are probably all well qualified. Not to mention that whatever criteria the hiring committee use to narrow candidates down in the last stages are probably not reliable and thus little better than (if not worse than) a coin flip.

I imagine this last point also applies here. There’s no reason to think that the people deciding on grant funding are so good at their jobs that they could genuinely figure out the best grant application with enough effort, though it seems reasonable that they could pick out the top 40%.

I think much the same proposal has been made for admissions to highly selective universities - all applicants who meet the minimum requirements (which could be quite high) go into a lottery pool and are chosen at random.

I have been proposing lotteries, from within the qualified, in these pages for some time. Cheaters.

It’s interesting to note how situational the academic response is - in some cases the very idea of merit is rejected, in others they claim the ability to discern the finest gradations

Why do the grant-makers not rely simply on the scientific purport of grant applications? Haven't these people been TRAINED to make grant assessments on some kind of analytic basis?

Scientific enterprise at the mercy of lotteries (new tests for fatalism and determinism vs. something like pure chance?) and of science-generated algorithms to determine research paths? How utterly scientific.

I don't think you remove measurement error by moving criteria to "hypothetical research result."

That too would likely suffer all sorts of error and benefit from randomization.

As I come to think and believe that measurement errors are NEVER removed (or compensated for), I advocate equipping our scientific grant-making bodies with the grant-making assessment instruments of their choice, whether a dart board, a roulette wheel, or a simple pair of dice (subject to multiple rolls, naturally).

Contemporary epistemological accounts continue to exhibit deficiencies all their own, an undoubted consequence of our having entered the quantum realm almost a century ago.

To hear proffered accounts for how to conduct scientific inquiry, archaeologists should defer to the decision-making prowess of their shovels and pick-axes (unless of course they wander desert wastes equipped with divining rods), et cetera.

The key to "partial" lottery is that you think you do have some idea about qualification, if not absolute ordering.

To avert those epistemic errors, perhaps possibly maybe curators of scientific enterprise could rely solely on the criteria of digitized chronometry: first-come, first-served.

Might yield outcomes at least as startling as anything else we've tried.

Comments for this post are closed