Cricket, I'll comment that the notion of whether profile analysis is a reasonable thing to do is a highly contentious one within the professional world -- it's one of the few topics that can reliably provoke a flame war on the professional mailing lists. A few very vocal folks are certain that nothing matters other than the overall score (g), while most others take the approach that as you fly the helicopter lower, you see more detail but need to be mindful of the lowered reliability of basing interpretations on smaller samples of behavior. (Full disclosure: I'm in the second camp.) Don't assume that just because you found one article making a case in one direction that this is the way the whole field thinks.
In this case, the mediated > automatic split is *huge* and I cannot imagine defending the idea that all we care about is the FSIQ. I agree that there can be a broad range of possible causes for this finding, which Grinity has laid out the biggest possibilities... can't determine what might be going on from the data we have thus far.
Unfortunately, testing is a field in which you typically get what you pay for (or rather, if you don't pay for it, you don't get it; it's certainly possible to pay a lot for worthless testing... sigh). In the States, the skilled and thoughful testers are getting hit very very hard by insurance companies slashing reimbursements (I mean *slashing,* as in, "Oh, by the way, starting in a week, you're getting 40% less than you thought you were getting.") and denying authorizations for testing in the first place (they have never been willing to approve anything related to educational issues for kids). Increasingly, most of the good folks are refusing to take insurance or reducing the slots available for insurance-based testing, because we literally cannot make a living doing the quality of work we need to do for the rate of pay we can get from insurance. (The folks who are less professional or perhaps more ethically challenged grind out awful reports that are cut-and-paste jobs from the automatic reports generated by the scoring software.)
Sigh.
