Despite growing interest in formative assessment of teacher practices, research on rates of change in teachers' practices is sparse. This is the first study to examine the characteristics of observed change in classroom practices using the Classroom Strategies Assessment System (CSAS) across alternative schedules of data collection during instructional coaching. Our primary objectives included examining (a) the magnitude, variability, and precision of estimates of average rates of change in teacher practices and (b) the impact of data collection duration (i.e., number of weeks of data collection) and density (i.e., the number of classroom observations per week) on the precision of estimates of rates of change over time. A sample of teachers (N = 63) participating in instructional coaching was observed 14 times during coaching using the CSAS. Findings revealed a significant gradual improvement in strategy use, with significant between-teacher variation in rates of change. The frequency of observations was associated with the precision of estimates for average rates of change across teachers and for individual teachers, providing initial guidance on minimum number of observations required to monitor change in practice over time.