Common Pitfalls in Classical Testing Models