I need to find time for more reading! One of the observations that Miraca Gross makes in her book is that the children in her study who had great multi-potentiality would generally be extended in the area that was easiest for their educators (math/science). So what they ended up pursing to the highest level did not necessarily reflect their area of highest potential in terms of childhood IQ. I seem to recall reading an article not that long ago that suggested almost the opposite, possibly a refletive article about SMPY and related studies, I must try to find that again, and I will be interested to read these articles you have linked.
The comments you make about sample size and identifying more verbal children make sense.