3 ms·
Who says I'm defending the status quo? I have zero desire to defend the status quo: I don't think it maximizes either scientific productivity or my own persona
by eli_gottlieb 12y ago
Who says I'm defending the status quo? I have zero desire to defend the status quo: I don't think it maximizes either scientific productivity or my own personal opportunities.
>Great. What exactly is the plan here?
We can start listing from incremental reforms and work our way up to radical ones.
1) Shift from post-docs to "perma-docs". This will immediately address one of the biggest hidden costs (in the sense that the accounting methods of grant funding conceal the cost entirely) in science: by the time you've got a grad-student trained to do decent research, they graduate, and by the time you've got a post-doc running experiments at a decent pace, they too leave. The only people who stay in a single lab long-term are the professors, and they've become glorified managers. The result is that science is actually losing a lot of productivity to skilled-labor turnover. The solution? Just stop pretending everyone will be a professor someday, and allocate money for paying permanent staff scientists with PhDs to actually run experiments and publish results rather than shepherding graduate students.
2) Lengthen the grant cycle. This does mean more money per grant, but it doesn't cost the government anything to allocate the same dollars-per-grant-year in longer blocks of years. Use this reform to commit to letting scientists do long-term work: more like five years than two years, or even, ideally, something like 10 years. Since whole sciences are not routinely totally revolutionized every decade or two, funding periods should look more like substantial fractions of one's whole career than like the time it takes to train up a new MS or PhD student.
3) Build and use a framework for allocating grants based on value-of-information calculations that reward doing experiments whose outcomes are uncertain, rather than experiments that will probably work (which really means: experiments that don't teach us anything new). Find a way of actually quantifying the uncertainty a subfield carries about its open problems before and after new publications, and reward the scientists who do best at surprising their peers.
And we can get more radical from there if I keep wasting my time thinking of these things instead of doing useful work.
- dnautics 12y agoIt is distressing how every single one of your suggestions basically throws more money and puts more power in the hands of professors. 1) with little advancement and negotiating power, perma-docs will become basically overeducated and exploited. I mean, your suggestion has already happened with the creation of "staff scientist" position, with marginal improvement in quality of science. The phenomenon of "perma-doc" is basically railed against anyways. Why bother putting someone through 5-6 years of economic and advancement distress as a grad student when basically, they should have had on-the-job training as a tech to do what you are suggesting. 2) Lock fundraisers (professors) in to money for long periods of time with limited for accountability or checks. Yup. No way that could go wrong. 3) The more seasoned professors with grantmanship will then gravitate towards uncertain outcomes. Of course this is no guarantee that these professors will actually have the skill to execute, and you know, how do you judge the reasonableness of an experiment with a highly uncertain outcome? Actually, this already happens to a certain extent - have you been to any DARPA jamborees?