Tuesday, August 9, 2016

Should science funding be random?

Some studies looking at the current system of science funding find it quite useless.

Instead of having endless people writing grants, reviewing grants and the whole torture, we will fare better by simply allocating the grants randomly based on super simplified and effortless criteria.

This kind of insight - while intuitively offensive to many - is a recurring robust finding in many fields.

Money management: Index funds outperform managed funds. Very robust long term data. Most mutual funds just cheat you by taking the fees for useless "active managing"

Interviews and student assessments: Simple numerical models outperform most personal thinking on how a student will succeed henceforth. Those time wasting procedures not only are wasteful, they actually lead to worse results than a simple formula.

The list goes on. ....

BUT.... one thought occurred to me.
At the tails, it is possible that the current system is superior.

When an idea is quite clear to have very good chances to succeed, the assessment system might almost guarantee it funding. Which a random system will not.

So, for some aspects a managed system might help

Of course, we can adjust for it. Allocating 75% of the funding randomly, and 25% to "must do" projects. 

No comments: