A new test that was substantially different from an earlier test would have to be compared against it to make sure it was measuring the same stuff, and I have seen a few studies checking how well different tests were correlated.