It's true that SAT scores correlate poorly with college performance -- but that's because someone can be admitted with, e.g., a bad SAT and a great writing samp

It's true that SAT scores correlate poorly with college performance -- but that's because someone can be admitted with, e.g., a bad SAT and a great writing sample, or a great SAT and a diffident high school record. Take a middle-tier state school: if someone has a Harvard-worthy GPA and chose the mid-tier school, it's probably because of other factors that kept them out of Harvard -- and those other factors will probably lead to a lower GPA. If another student's SAT's would imply community college, you might expect them to have some kind of focus or determination that would let them perform at the state school level (and you could detect that determination in their extracurriculars or in student interviews). So this is only surprising data if you try hard not to think about it. What's amazing is that a four-hour test has almost as much predictive value as four years of GPA data -- is four years of your life and tens of thousands of dollars really a good investment, when the result is a ~20% more accurate measure of your abilities? Doubtful.