Reply #4: Is there *anything* to support that model besides the *assumption* that it works? [View All]

[eppur_se_muova]

There seems to be a built-in assumption that all these factors can, and have been, successfully deconvoluted from one another, and that no factors have been overlooked. Furthermore, they are reporting data derived from test scores to three significant figures. No one in their right mind considers any written test to have that sort of confidence level. The SAT has a reliability of ~0.9, and it's been studied and refined more than just about any test out there.