I've noticed that people rarely seem to see much value in testing procedures against questions with known answers.