It actually does mean that in most cases. Having false confidence in a hypothesis based on non-rigorous studies is much worse (and in some cases, a lot more dangerous) than withholding judgment until the data (all the data, not some cherry-picked slice of it) really is conclusive.