Perhaps a more human-understandible criterion threshold for publishing would be logarithmic inverse p-value... the closer p gets to 0 (stronger correlation), th

Perhaps a more human-understandible criterion threshold for publishing would be logarithmic inverse p-value... the closer p gets to 0 (stronger correlation), the bigger log(1/p) gets. This of course is just one aspect of greater issues.