This is a pretty big problem. I've given it some thought, but haven't cone across a good answer. Incentivizing this is hard right now. The best idea I've had i

This is a pretty big problem. I've given it some thought, but haven't cone across a good answer. Incentivizing this is hard right now. The best idea I've had is to minimize effort for negative experiments. Like arxiv, but perhaps a even less formal. Allow it to be easily searchable. It's still not incentivized, but it's a fuckton easier to put it out there. I could also see dedicated data analysts publishing negative data of open data sets.