Well, I'm lazy, so I'll put in something that I responded with, on my friend Toan's website quantumtunnels.com.

Hey Toan,

I'll throw in my agreement here. You know from our personal talks that I'm not a fan of IQ testing as it relates to Spearman's g and such. But in practical circumstances, I'm all for standardized tests. Given the wide range of possible SAT scores (400-1600), there's a very narrow chance that random scoring would give an accurate measurement. The accuracy of the SATs must give an astounding p value.

The SAT is practical because life is practical. It sounds trivial, but think of it this way. I may argue with Toan that a high score on any number of standardized tests doesn't necessarily point to any underlying "intelligence" (or the corollary about stupidity). But it may very well point to an ability to deal with certain situations, such as, well, taking tests. Sure, real life is nothing like an SAT, but if the statistics bear out a correlation of SAT scores to academic performance, then what more can you ask of an objective measurement?

The true debate shouldn't be whether SAT scores ought to be considered in an application process, but to what extent. Objectivity alone is imprecise and, perhaps by definition, unfair; subjectivity alone is the precursor to discrimination. It's the balance that matters. Completely remove the SATs, and the process is open even wider to discrimination.  ¶ 9:52 PM