> By ETS's own admissions over the years, the best they can say about > ANY test that they supply is that it is RELIABLE, that is, they can > predict the margin for any student retaking the test within a certain > probability (e.g., a retake will be within 50 points of the initial > score with a 95% probability). As much as ETS experts have tried over > the years, they have not come up with a measure of VALIDITY, which is > what Lou is talking about. There is a drastic difference between a > laboratory test which measures a determinate quantity, such as the > amount of a chemical agent present in a blood sample, and a > standardized educational test for which even its authors cannot agree > on what it measures. Again, ETS jumps back and forth between aptitude, > achievement and IQ measures on most of their tests, usually without > even defining what any of these alleged quantities are.
Indeed, this is the central issue. Standardized tests do have a (small) role they can play in things like college admissions. (Though a growing number of colleges are learning just how small that role is, and, in consequence are abandoning them. And some of these colleges are highly selective ones, too.)
But Wayne has repeatedly told us that no curriculum should be put into place in a school (let alone a district or a state) until the results of implementing that curriculum have been validated by standardized tests. This is a ridiculous statement on its face, because the standardized tests themselves are not validated--except by being tested against other standardized tests.