Could it be that there is nothing really wrong with the tests we have, only our procedure for affirming type? Consider, what if we were to first have someone take a type test, then get a score, and then have them take a test that "validates" the type?