but this is probably the best treatment on manuscript reliablity testing:
http://www.leaderu.com/orgs/prob…