So I actually wrote about this in an earlier post. https://medium.com/franklopenscience/sciences-two-reproducibility-problems-6eb634dc32cc
There are lots of genuine reasons for replication failure that don’t mean the original finding is wrong. A “failed” replication might lead to the discovery of hidden moderators or boundary conditions for the effect. I think it’s also completely appropriate to ask whether the people attempting to conduct replications are sufficiently competent.
But Begley’s point is that in many cases, the original researchers were unable to replicate their own experiments in their own labs. And from his other writings it seems that the critical difference between the original finding and the self-replication failure is that the second time around they were blinded to condition.
Anyway, I’m hoping to catch up with Glenn again soon and will report back!