It's still science even with a low p value! It's just a negative result instead of a positive one. Such results really ought to be published, even if they aren't usually.

Well, not all of them deserve to be published. But I understand your meaning.

True, but not all positive results deserve to be published either!

No, because p < 0.05 can be achieved by simply dropping inconvenient variables from the estimation model.

The results dont have to be conclusive for science, they just have to be documented.