I had an Education professor in college who had done time at UT-Austin, and his research subject was standardized testing. His conclusion was that standardized tests at ANY level--elementary, high school, university, post-doc, whatever--are really only good for one thing: Predicting how the student would do on other standardized tests.
The more I administer these things as part of my job, and the more I have to take them, the more convinced I become that he's dead right.