I think it's more of a problem of misaligned incentives rather than the method itself. There's a lot of room for politics, ass-covering, etc. Studies that fail to replicate something are shelved, even though they're useful information. Studies that indicate a negative, ditto. Studies are "nudged" to show something so the author can show that it wasn't a complete waste of time, etc.

They're understandable, but they're bad for science.