Researchers argue when to publish null results — and a Bayesian prior may answer it
RexDouglass · x · 2026-09-14
Ryan Briggs revisits the perennial question of when null results should be published, arguing in line with his paper "It must be very hard to publish null results" that most reasons for rejecting null results apply equally to significant ones.
Rex Douglass adds that this is only a puzzle for frequentists: if you actually synthesize a meaningful SOTA prior, the answer to where to spend effort is simply whatever generates more useful posteriors.
More from Research
- RL luminary Szepesvári: researchers actually want AI models that generate insight — CsabaSzepesvari · 2026-09-14
- Ted Nelson's 1965 Thoughts on Hypertext Resurface — dreamwieber · 2026-09-14
- Meme nails AI research's oldest gripe: papers without reproducible code — CSProfKGD · 2026-09-14
- Csaba Szepesvari: Math's Love of Infinity May Be a Fiction — CsabaSzepesvari · 2026-09-14
- TCS researcher reflects on problem-solving vs theory-building in the AI era — jasondeanlee · 2026-09-14
- Study: dropping 0.003% of LMArena votes can flip the top model — rishabh16_ · 2026-09-14