It should be nigh impossible to overcome a strong prima facie case with evidence of secondary considerations.
I agree with that, in theory. But, as Einstein (or Yogi Berra?) said, "In theory there is no difference between theory and practice.
In practice there is."

The standard test of obviousness is hard to make objective, there is always a subjective element to
it. I think in practice it has an element of contest of wit and eloquence, which is the lawyer's expertise. Which means a
good lawyer is capable of patenting stuff that ideally shouldn't be. That's why I think a confident lawyer would consider the
secondary evidence, well, secondary. It also means that a pro se inventor like myself faced with an examiner who is more
experienced in such matters, has a chance to lose cases that are solid patents, simply due to inexperience, even if the patent
is properly built (right claims, good spec). So if in those cases such evidence may very well be "secondary" to a lawyer, as he/she
will tear apart the examiners objections and go have a lunch, a pro se inventor will go, "uh, umm, lookie here, nobody thought
cold fusion was possible and I did it! Can I have a patent? Pretty please?". So it really depends on the point of view. It seems
that the courts allow for both possibilities, the lawyers can go after prima facie, and bring the secondary evidence when it comes
to that. But nothing prevents somebody from laying down everything in front of the examiner and hope he/she sees the
light.
EDIT: I should add that I am guessing that in the majority of cases that pass through the USPTO and the lawyers hands the
secondary evidence is not available/not very persuasive. That also affects the lawyer's point of view I think...