<p>
Frey and Detterman found a correlation coefficient of .82, and .87 when corrected for nonlinearity. This is definitely not "statistically low.</p>
<p>
Okay, so let’s say student A and student B take the PSAT with no prep. Student A is about average and scores 140. Student B is the OP and scores 222. Your argument is that within finite time both students can raise their scores to perfect. I can buy this, but I guarantee you it will be much, much harder for Student A. </p>
<p>Just because it is theoretically possible for student A to score a 2400 doesn’t mean that the SAT has no power to differentiate between intelligence levels. The fact that A did significantly worse on the test than B when both had no preparation signifies that certain skills are lacking. You may argue that these skills are unrelated to intelligence, but not on the basis that a 2400 is theoretically possible for anyone. </p>
<p>An IQ of 150 (as measured by a specific, fixed test like the Wechsler) is theoretically possible for anyone. If Weschler tests were as freely available as practice SAT exams and students were incentivized to score well, I guarantee you that people who practiced the specific question types would see their scores rise. This doesn’t mean that they are getting more naturally intelligent, simply that they are getting better at tests designed to measure intelligence. But just because people can prepare for the Wechsler and raise their scores does not imply that the score you receive upon first seeing the test is irrelevant to intelligence, in the same way that an initial SAT score (with no prep) is not irrelevant to intelligence.</p>