I may have not articulated this very well: Both groups in my metaphor (those who naturally understand the SAT questions from learning mathematical concepts, and those who are tutored on known SAT questions) were meant to be among the high scorers on the test.

This would imply that SAT scores will not predict real performance, and the same implications would apply to job interview questions that are gamed in a similar fashion.

"Those schools that attract students with a wide range of SAT scores generally have higher correlations between the scores and first-year grades." --http://abcnews.go.com/Technology/WhosCounting/story?id=98373...