François CholletVerified account

I think this problem has a lot to do with how scientific papers are distributed, and how the ideas in those papers are communicated.

Having been on many review panels over the years, my vote is always for readable ---and thus, replicable.

How about a stats standards group certificate that a paper could get, to prove that the conclusion is supported by correct stats techniques.

If you want impact, its important to find the people your results affect, make the pitch to them directly, and start a collaboration.

It would help most if novel findings got funded. That's not what happens. #incremental #inbred

BETTER summaries would help....with an eye to framing " what this does mean/not mean" to stop bad journo takes.

The internet could change sharing research radically. Journal paper model benefits few people w strong track-records and publishing co's.

Both? 95% would have no impact but the 5% would have much more impact

Would science progress faster if scientific journals practiced affirmative action?