Awarding grants by throwing dice

March 30, 2018

Yesterday, Alex Holcome’s tweet drew my attention to Shahar Avin’s paper “Centralised Funding and Epistemic Exploration”, currently in press at The British Journal for the Philosophy of Science. You can read the accepted manuscript on PhilSci Archive.

My colleague Filip Jakobsen asked me to explain in layman’s terms what the paper was saying. Here’s what I told him.

The way our academic system works means that some people spend loads of their time — literally, like half of their work time, in many cases — writing grant applications, so they can get the equipment, reagents and other resources that they need for their actual research. There are of course many more applications for grants than can be funded. So applications get competitively reviewed, and only the “best” get funded.

There is growing evidence that this means money keeps getting funneled to the same people — because they have a “track record” — who do more experiments along the same lines they’ve done before, with the danger that it all gets very stagnant.

What the new paper found, by running simulations, is that in terms of advancing science, you actually get better results if you just choose which grants to fund at random. I guess some of the grants then go to no-hopers who waste them, but others go to teams no-one’s heard of who go on to do great and unexpected things — much more interesting new discoveries than those that might arise from the incremental experiements that the established groups would go on to do given the same funds.

It’s like the way venture capitalists throw money at lots of startups, knowing full well that most will fail, but that some will succeed wildly; and that overall they’ll get a better return than if they invested their money in, say IBM stock. IBM will grow, sure, and they’ll get a safe dividend. But only a small one.

It looks like granting bodies like to play it safe and invest in the science equivalent of IBM. Maybe they need to behave more like venture capitalists.

Advertisements

3 Responses to “Awarding grants by throwing dice”

  1. protohedgehog Says:

    Another new paper that shows the same thing, sort of https://f1000research.com/articles/6-1335/v2


  2. […] to Professor James Wilsden for highlighting the paper. The Sauropod Vertebra Picture of the Week blog has also mentioned […]


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: