Instructional Coaching and Mentoring: What Actually Moves Teacher Practice

Quick answer

Instructional coaching is one of the most rigorously evidenced forms of professional learning available. A landmark meta-analysis of 60 causal studies found effect sizes of 0.49 standard deviations on teaching practice and 0.18 on student achievement (Kraft, Blazar & Hogan, 2018), a meaningful achievement gain in the range of what several years of classroom experience typically adds. But the same research is equally clear that scale is the enemy: coaching programmes lose most of their power once they're rolled out broadly, and the number of coaching hours matters far less than what happens inside them.

Key facts

- A meta-analysis of 60 causal studies found pooled effect sizes of 0.49 SD on instruction and 0.18 SD on achievement from instructional coaching (Kraft, Blazar & Hogan, The Effect of Teacher Coaching on Instruction and Achievement, Review of Educational Research, 2018).

- Practice opportunities during coaching sessions were a statistically significant predictor of impact; modelling by the coach was not (Glover, Reddy & Crouse, Journal of School Psychology, 2023).

- Large-scale coaching programmes are meaningfully less effective than smaller ones. Effects from effectiveness trials of larger programmes were only a fraction of those found in smaller efficacy trials of the same approach (Kraft, Blazar & Hogan, 2018).

- More coaching hours were not clearly associated with bigger effects, suggesting the quality and focus of coaching matters more than raw contact time (Kraft, Blazar & Hogan, 2018).

Why coaching outperforms most one-off PD

The effect sizes above are large by education-research standards, comparable to some of the most effective interventions studied in the field. The mechanism isn't mysterious: coaching is job-embedded, individualised, and includes real feedback on real practice, three things a workshop day structurally cannot offer. It's the same logic behind why we differentiate and personalise learning for students, applied to adults (see why professional learning should look more like great teaching: https://eduspark.world/blog/future-of-teacher-professional-learning).

What predicts whether coaching actually works

Not all coaching is equal, and the research is specific about what separates effective coaching from expensive shadowing:

Practice opportunities matter more than watching. A 2023 analysis of coaching actions (Glover, Reddy & Crouse, Journal of School Psychology) found that structured opportunities to practise a new strategy were a statistically significant predictor of both implementation fidelity and student achievement, while modelling the strategy for a teacher, on its own, was not.

Feedback loops matter, but not uniformly. In the same study, feedback during coaching predicted stronger English language arts achievement but did not predict maths achievement, a reminder that "more feedback" isn't a universal fix independent of subject and context.

Smaller is usually more effective than bigger. This is the counter-intuitive finding school leaders most need to hear: coaching programmes that work well as a pilot with 10 teachers often underperform once scaled to 200, because the individualisation and fidelity that made it work in the first place is harder to preserve at scale. Plan for this before you scale, don't assume it will hold.

Hours matter less than focus. There's no clear evidence that simply adding more coaching hours produces proportionally bigger gains. A tightly focused coaching cycle on one specific practice will often outperform a loosely structured programme with more total hours.

Coaching and mentoring aren't quite the same thing

Instructional coaching (structured, often skill-specific, sometimes short-term) and mentoring (broader, relational, often longer-term, especially for new or newly-arrived staff) overlap heavily but solve slightly different problems. In international schools specifically, where a meaningful share of staff are new to the school every August, a standing mentoring structure for new arrivals and a more targeted instructional coaching cycle for specific practice goals are both worth running, not one instead of the other.

Building coaching into a whole-school PD plan

Coaching works best as one deliberate tier in a wider PD plan, not a replacement for everything else. We cover how to sequence coaching alongside self-paced courses and collaborative sessions in how to plan a whole-school PD programme (https://eduspark.world/blog/plan-whole-school-professional-development), and how coaching specifically supports new-staff retention in how PD improves teacher retention (https://eduspark.world/blog/pd-teacher-retention-international-schools).

Frequently asked questions

Does instructional coaching actually improve student achievement, or just teacher confidence?

Both, according to the evidence, but the achievement effect (0.18 SD) is smaller than the instructional-practice effect (0.49 SD). Coaching reliably changes what teachers do in the classroom; the downstream effect on measured achievement is real but more modest.

Should we run coaching for every teacher or target it?

The research on scale suggests targeting matters. A focused coaching cycle for teachers working on a specific, well-defined practice tends to outperform a diluted, whole-staff rollout of the same coaching model.

How is coaching different from a mentoring programme for new teachers?

Coaching tends to be skill-specific and structured around practice and feedback on a defined area. Mentoring is broader and relational, and is especially valuable for new or internationally-relocated staff who need more than a skills cycle in their first term.

Sources

- Matthew Kraft, David Blazar & Dylan Hogan, The Effect of Teacher Coaching on Instruction and Achievement: A Meta-Analysis of the Causal Evidence, Review of Educational Research 88(4) (2018): annenberg.brown.edu

- Todd A. Glover, Linda A. Reddy & Kristen Crouse, Instructional coaching actions that predict teacher classroom practices and student achievement, Journal of School Psychology (2023): pubmed.ncbi.nlm.nih.gov/36641220

See how EduSpark builds coaching and mentoring into a whole-school growth plan. Book a demo (https://eduspark.world/book-a-demo).