Hey, I agree with you that Behe looked bad on the business of the reviews of his book. I’m not sure what to make of it, but I am willing to give him the benefit of the doubt and assume it was just carelessness on his part. He may have left the review process in the hands of the publisher and assumed that they did a better job than they actually did.
What you need to understand is that the peer review process is not always absolutely rigorous. In engineering, research papers usually get three reviews, and sometimes the reviewers are not intimately familiar with the exact micro-topic addressed in the paper. Also, sometimes the mathematics is too difficult for some of the reviewers to follow in detail. Think you know mathematics? Try reading a few papers in the IEEE Transactions and see if you can follow the mathematics in detail.
The other thing that sometimes happens is that a particular journal has a group of “insiders” who scratch each other’s backs. So you give me a favorable review, and I’ll return the favor. Yes, reviews are supposed to be anonymous, but the reviewer can always violate the anonymity by simply telling the author.
And the most important thing to understand is that certain taboo topics are often simply prohibited by the review process. ID is one, for example. The powers that be in much of the scientific community has simply decided that ID is off limits, and it cannot get past the review process. If a particular reviewer signs off on it, the editor rejects it. And there is a strong feedback effect: if reviewers know they will be “punished” for accepting claims of ID, most will refuse to do so. Ditto for editors. How could a reviewer be punished? By making it harder to get his own papers published.
The bottom line is that peer review, while it may be the only alternative, is far from foolproof.
But this should hardly be surprising here on FR. After all, imagine trying to get a paper advocating school vouchers published in some Leftist sociology, education, or political science journal. It won’t happen.
I'll trust you on the engineering peer review process. It seems that's not significantly different from economics peer review, which I'm slightly more familiar with. (My dad is an economics professor.) I have no doubt that the mathematics can be very difficult in engineering - it's already pretty difficult in economics and finance, but I do have some doubts that the reviewers wouldn't be able to follow the math in detail. Unless I'm mistaken, the reviewers of a paper should be very familiar with the subject area of that paper (i.e. should know how to at least follow the steps of the math.)
It's unfair to ask an engineer to think of a proof on the spot. It's fair, though, to ask him to follow the steps of a proof.