The Classroom Strategies Assessment System (CSAS) is a multi-rater, multi-method (direct observation and rating scale methodology) assessment of teachers' use of research-based instructional and behavior management strategies. The present study investigated the association between teacher self-report and school administrator ratings using the CSAS Teacher (CSAS-T) and Observer (CSAS-O) Forms in 15 high-poverty U.S. charter schools. The CSAS-T and CSAS-O were designed to be used concurrently as a valid formative assessment of teacher practice. Findings include small, but statistically significant correlations between the CSAS-T and CSAS-O. Analysis of a multi-trait--multi-method (MTMM) matrix found teachers and observers to be measuring different constructs. No mean score differences were found between teacher self-reported instruction and behavior management strategy use compared with school administrators' observed ratings. Furthermore, school administrators and teachers have similar ratings of overall effectiveness, with the majority of teachers in the sample being rated at or above effective. Overall, findings offer support for using the CSAS-O and CSAS-T for guiding professional development conversations.