Responding to federal policy and recent research, states and districts have developed and begun implementing multiple-measure teacher evaluation systems. These systems generally include observational and/or student survey measures of instructional quality alongside measures of teachers' contributions to student learning (e.g., value-added models [VAMs]). While the research base on VAMs is large and growing, less is known about the measures of instructional quality. This study focuses on the year-to-year stability of observational and student survey measures, drawing on data from the Bill and Melinda Gates Foundation's Measures of Effective Teaching (MET) study. Results suggest that the observational and student survey measures used in MET are somewhat more stable than measures of teachers' contributions to student learning, but the stability is lower than that found in comparable measures in higher education. Nevertheless, reclassification rates based on these measures are high, particularly when based on criterion-referenced cut scores.