continuous growth is needed, even when high levels of excellence have been achieved. There is no fixed and final state of expertise, nor of excellence. In teaching, as in other fields, experts must always continue to develop. In sports, when Michael Jordan develops new capabilities, his environment soon adapts to him, so he must engage in another round of development The situation is similar in mathematics teaching. The ways that teachers think about mathematics, teaching, learning, and real-life problem solving strongly influence what goes on in their classrooms; but what goes on in their classrooms also requires teachers to develop more powerful and sophisticated understandings about mathematics, teaching, and learning. The cycle is never-ending, and teachers who fail to get better risk being not very good at all.
These principles should be taken into account in assessment programs for teachers. Just as in the assessment for K-12 students, the assessment of teaching should focus on activities that are meaningful and important in their own right. These activities should enable teachers to both develop and document their development without interrupting their instructional activities. And, as in assessments of students or of programs, assessment activities for teachers are more than indicators of progress. They are interventions that can induce either positive or negative changes in the systems and individuals they describe. Therefore, care must be taken to ensure that these influences are positive. This "progress-focused assessment" is aimed at documenting progress in directions that are increasingly "better" without necessarily beginning with a fixed and final definition of "best" and without labeling individuals as good or poor relative to one another.
In the Teachers project, teachers collaboratively write performance assessment problems for their students and analyze students' responses to such items. Then, in weekly meetings over a ten-week period, these teachers produce a library of useful problems and response analysis procedures while refining their collective conceptions about the nature of good problems and good responses. During the process, many participants have developed new insights about the nature of their discipline, of its applications, and of students' understandings and capabilities. And the problems that were written and responses that were analyzed produced a trace of the teachers' own progress that was very impressive to school administrators.
Enormous progress has been made in clarifying future-oriented instructional objectives in mathematics. Now efforts are being made to create new types of tests and test items consistent with these instructional objectives. Yet, precisely because significant progress has been made at the