If the research on professional development has a winner, it is instructional coaching. Kraft, Blazar, and Hogan's 2018 meta-analysis of 60 causal studies — one of the strongest evidence bases for any single PD approach — pooled the results of coaching programs that had been tested with rigorous designs.
What the Numbers Say
- 0.49 standard deviations improvement in instructional quality — large by education standards, equivalent to moving a teacher from the 50th to the 69th percentile of practice.
- 0.18 standard deviations improvement in student achievement — comparable to (or better than) many whole-school reform interventions.
Coaching works because it is PD's only truly individualized model: the feedback loop runs through each teacher's own classroom, with their own students, on their own goals.
What Good Coaching Looks Like
The studies behind the meta-analysis share a pattern: cycles of observation and feedback, a trusted relationship, and a focus on one or two concrete practices at a time rather than everything at once. Coaches who model lessons and co-plan with teachers showed some of the strongest effects. The Learning Forward analysis of the study highlights a practical caveat: effects were strongest when coaches worked with individuals or small groups — coaching "spray" across an entire faculty diluted the impact.
If your school has a coach, protect that time. If it does not, peer coaching with a trusted colleague using the same cycle — observe, debrief, retry — captures much of the same mechanism.
