Having finished the grading process for my undergraduate students, I have some doctoral students to attend to and a few short writing assignments as well this week. So while I am itching to read the Gates Foundation-funded, Tom Kane-directed report on value-added measures and surveys of student-observed teacher behavior, I haven't yet had an opportunity. From my brief perusal, it is clear that Ron Ferguson's work has produced much, much better student surveys than what our survey-benighted students at USF complete every semester. (We have a legacy survey inherited from the University of Florida many years ago, and while we're in the process of discussing changes, we're not there on the change.) Our students are asked their opinion of the overall effectiveness of "the teacher" whether they have one, two, or many who teach in a class (e.g., lecturer, lab assistant, discussion TA, etc.). Our students are asked whether the instructors know their material well, though presumably the students are there because they don't (yet) know the subject (though of course they can tell if an instructor is out of her or his depth in a general sense).
So this may be damning with faint praise, but the questions Ferguson's survey asks are better than at least a few in operation in higher ed.
Robert Tucker has some good analysis and comments on End-of-Course student surveys, and he was good enough to critique a CC survey for me that had been in use for at least a decade.
http://www.intered.com/higheredbriefing/2010/2/24/end-of-course-surveys-part-i-why-do-them.html
http://www.intered.com/higheredbriefing/2010/3/3/end-of-course-surveys-part-ii-executive-actions.html
I have to admit that I had to struggle (unsuccessfully) with his statistical observations, but maybe it will make sense to others. Dr. Tucker is a vocal supporter of the “measurement science” movement, which has influenced UoP’s culture of assessment.