Tests ought to be run on multiple thinking/reasoning levels and then scores reported for each level they were run on.