Meta-Analysis of Teacher Coaching

A new study authored by researchers from Brown and Harvard universities points to one promising strategy for helping current teachers improve: one-on-one coaching.

Teacher coaching has emerged as a promising alternative to traditional models of professional development. Authors reviewed the empirical literature on teacher coaching and conducted meta-analyses to estimate the mean effect of coaching on teachers’ instructional practice and students’ academic achievement. Combining results across 37 studies that employ causal research designs, researchers found pooled effect sizes of .57 standard deviations (SD) on instruction and .11 SD on achievement. Much of this evidence comes from studies of literacy coaching, which have an effect of .14 SD on reading achievement.

Although these findings affirm the effectiveness of coaching as a development tool, further analyses illustrate the challenges of taking coaching programs to scale while maintaining effectiveness. Coaching effects in large-scale effectiveness trials with 100 teachers or more are roughly half as large as effects in small-scale efficacy trials. In addition, coaching is an expensive intervention.

One way to make coaching more affordable and specialized would be to do it virtually rather than in-person. Encouragingly, the researchers did not find a major difference in effectiveness between the two approaches, suggesting that online coaching might be a good alternative. Another approach, the study highlights, would be to focus coaching on the subset of teachers who most need to improve.


For more commentary, see:

To read the report, see: