from that stuff (other than checking to make sure come clinical study used the right test) can see the problem then a certified actuary certainly would whether they were a high school drop out or had a PhD.

Not that I think a PhD is a bad thing or useless, but certification is definitely better cost and time wise. If employers learned this and dropped the formal BS/MS/PhD education hoops they make everyone jump through our society would be a lot better off. I sure hope you developed some training for those engineers to learn what statistical tests they should use. It isn't that they can't learn it (and in your scenario they certainly should have figured that out) it's just that the teaching in the on the job training was terrible. In fact, it doesn't seem there was any on the job training.