Like high-stakes student assessment, high-stakes teacher evaluation threatens to be an occasional event that is disconnected from day-to-day teaching and learning, producing results that do not help teachers improve their performance and placing teachers in a passive role as recipients of external judgment. For several years, the authors have worked with teams of teachers, helping them develop their capacity for self-assessment using two prominent, comprehensive frameworks for effective practice: Robert Marzano's Art and Science of Teaching model (2007) and Charlotte Danielson's Framework for Teaching (2007). Both of these models systematically describe various components of research-based effective classroom practices. Through this work, the authors have come to understand that as valuable as such comprehensive frameworks can be, it's not enough to simply put them in place as rubrics that supervisors use to rate teachers' effectiveness. Just as students need to be actively involved and empowered as partners in classroom assessment (Stiggins, 2004), teachers need to be actively involved and empowered as leaders in the formative use of the tools that will be the basis for their own summative evaluation. The most effective supervision and evaluation systems empower teachers to accurately assess their own practice and self-diagnose areas for growth. In such systems, teachers use comprehensive frameworks throughout the school year to collect data related to their teaching, reflect on their practice, and identify specific instructional strategies they can work on to improve their repertoire of skills. The school culture in such systems supports teachers by recognizing the need for improvement as an asset rather than a liability. Some of the beliefs that are central to such supervision and evaluation systems are highlighted.