My understanding is that the SAT was a very good predictor of collegiate success up to approximately 1994. Prior to 1994, not only did the test score track closely with college success, it also tracked closely to IQ, which was why Mensa accepted SAT scores for membership. But then it had an overhaul that introduced subjectivity into the verbal scoring (open blank questions instead of multiple choice) and students were allowed a calculator. Afterwards, even more subjectivity has been added, with the introduction of an essay portion, and the analogy section (which was the part of the test most influenced by intelligence, not achievement) has been eliminated.
So I'd say that the effectiveness of today's SAT as a predictor of college success is an open question that needs to be researched.
