I saw this tweet yesterday and it's been on my mind ever since. Some thoughts:

1. "Too much social science research" is an ambiguous phrase. Too many papers being published? Too many studies being run? Too many human-hours being devoted to social science? 1/n https://twitter.com/ModeledBehavior/status/1305150891205459968

2. I believe social science research is valuable. If I didn't think that, I think I would quit.

I think social science has contributed a great deal to our understanding of how the social, political, and economic world works, and how we can make it work better. 2/n

3. The problem with social science -- and other sciences too -- outlined in the OT comes from the structure of career incentives. It is not fundamental.

We are in an equilibrium where there are few jobs, many capable scientists, and historical norms/power brokers... 3/n

... As a consequence, we are in an equilibrium where the quantity of discrete scholarly outputs (of specific _types_) is excessively up-weighted at the cost of quality (of specific _type_) in determining career success.

This is VERY bad. 4/n

3. That problem is compounded by our mentorship and publication models which could be a lot better at encouraging/facilitating "knowledge accumulation" rather than "boundary pushing."

The latter is good and important, but the former is very fundamental part of science... 5/n

... It would be neat to see journals encouraging dialogues where authors get publication credit (in that journal) for replicating/extending a study already published there.

It would be neat if we could also move away from using "novelty" and "puzzles" as our only frames. 6/n

4. So the answer is definitely NOT "do less social science research" (at least, if you mean "devote fewer human-hours to social science").

Why would that help? It would probably just make things worse.

A better answer is: incentivize and teach a different type of output. 7/7