To the Editor: One must hope that the findings of Dr Ioannidis1 that 16% of highly cited research studies were contradicted by subsequent studies will itself be among the 44% of studies whose findings are replicated. The possibility that it will be among the 16% whose findings are refuted is almost too frightening to contemplate.
If this study is among the 16% that are subsequently found to be false, then its conclusion—that these studies are false—will itself be false, implying that these studies (and, pari passu, the Ioannidis study) are true. But then, if the Ioannidis study is true, these studies (of which it is a member) must be false (since those were its findings), so once again its conclusions will be false. This paradox cycle of “if false, then true; if true, then false” will continue ad infinitum, until one recognizes that under these circumstances, the findings of the Ioannidis study become undecidable.2
Contradictions in Highly Cited Clinical Research. JAMA. 2005;294(21):2695-2696. doi:10.1001/jama.294.21.2695-b