Accumulating evidence across studies: Consistent methods protect against false findings produced by p-hacking