I think Arxiv and similar could contribute positively by listing replications/falsifications, with credit to the validating authors. That would be enough of an incentive for aspiring researchers to start making a dent.