When student ratings first started being widely used and the research associated with them was prolific, many faculty objected to the focus on teaching effectiveness. “It shouldn't be about me! I'm a good teacher if my students are learning. Measure how much my students have learned and then make judgments about my teaching.” The problem has always been those intervening variables between the teacher and the students that make it hard to tie the learning to the teaching. Is the student learning lots because she's highly motivated and wants to learn? Is the student not learning because his life is full of work and family without time to study? Variables like these affect how much a student learns regardless of the quality of the instruction. And there's a second persistent problem: End-of-course ratings are anonymous, which makes it impossible to tie the ratings of the instructor to the student's performance in the course. Are the high teacher ratings being given by the students who learned the most or by students who liked the course because it was easy?