I'm looking for a book that explains why r-squared is invalid for nonlinear regression.
While there are some articles online about that subject, I'm specifically looking for a book as a reference.
Reason: I'm in the process of trying to convince our consultant that r-squared isn't valid for nonlinear regression. And in this case, I think referencing an actual book would be most effective.
- It's easy to say, "Well, you can find anything on the internet". Whereas it would be harder to discount a book.
- There are academic papers on the subject too. But I suspect they're over the party-in-question's head. A book might put things in simpler terms.
With that said, I suspect a question about a book recommendation would be shot down pretty quickly on Cross Validated SE. But I'm wondering, is there a way to rephrase the question that would be acceptable?