A new system for assessing student ratings of instruction has been implemented at the University of Washington. Two major changes are in evidence. First, there is an explicit recognition that student ratings can and do serve multiple functions, and the same evaluative questions are not necessarily appropriate for each. Secondly, there is an explicit recognition that adequate diagnostic information cannot be efficiently provided instructors with use of a common set of evaluative questions for all classes. Student Rating Forms A through E are appended. (Author/BJG)