What evidence do you have that suggests that the cost of doing a video audit (an expert carefully watching hundreds of hours of experimentation) and a replicati

What evidence do you have that suggests that the cost of doing a video audit (an expert carefully watching hundreds of hours of experimentation) and a replication study for every submitted article would be cheaper than the occasional paper that leads to a second experiment that doesn't pan out?