The ones you wouldn't remember probably weren't sufficiently different from the mean to warrant a positive or negative score in the first place. Meanwhile, the ones that you do remember probably were.
Just make opting out of each individual question easy (add a N/A or "don't remember" option). Sure, you'll collect less data, but I imagine that you'll get at least partial cancellation after multiplication by the SNR, possibly much more than partial cancellation.
But you remember the great ones, and you remember the god damned awful ones. The great leave an emotional mark on you, they inspired you to go into their field, they showed you something that wowed you and has stuck with you over those 10 years. And the damned awful ones leave emotional scars that don't fade, and you'll talk to old veterans/classmates about them like you did a tour of 'Nam together...
You might be surprised. The teachers that taught you relevant things that you end up using over and over tend to be remembered in my experience.
Of course some courses lend themselves to imparting this information better than others, and if you end up in fields other than your major there may be less chances to find your learning useful.
But then you are evaluating class he was assigned to teach and its relevance to your career, not his teaching. I did not used whatever I learned about networking or cryptography much, but that does not mean those teachers were bad. Nor does it mean CS major should not contain those classes.
