The post makes the point that a large amount of science produced in the US is not trustworthy. Fraudulence implies intentional deception, but it is more likely that it is simply a response to incentives. I don't see why you can't understand what I'm saying, but in the hope that it's easier I'll put it in bullet points and try to be direct
* The post is a sample of papers across many fields
* It describes a method to determine whether widespread p-hacking is occurring in a field
* Widespread p-hacking in a field will lead to unreproducible results (because the results are an artifact of the data rather than reflecting any underlying cause)
* You said: The replication crisis stuff is mostly a message board fixation. The problem we have is that people unfamiliar with the field don't have a sense of what the denominator is, only the numerator.
* The approach in the post uses a statistical method across many papers across many fields. It does not rely on the existence of some number of non-replicable results (in fact, one should expect some percentage of results to be non-replicable).
* Therefore, it doesn't rely on "the numerator" so to speak, but a sample of all papers in a field to characterize the field's replicability.
You're a smart guy. I follow your posts here and on Twitter. So I can only assume that either I've flipped some bozo bit or something in my communication style is abstruse. I don't understand why you're saying that p-hacking has nothing to do with the replication crisis (most would expect the two to be quite related), or that saying "the replication crisis is real" means I am claiming that individual papers exist that are non-replicable (which is fine, but not a replication crisis). In either case, we're clearly getting nothing out of talking to each other so I might as well leave off here.