An especially interesting piece of recent research on coaching has given us good news in terms of the link with improved teaching practice. At the same time, this research has opened up more questions about how we upscale school coaching programs. A team from north eastern USA, Matthew Kraft from Brown University, David Blazar from Harvard and Dylan Hogan, also from Brown, published The Effect of Teacher Coaching on Instruction and Achievement: a Meta-Analysis of the Causal Evidence in an updated version late in 2017.

While many coaching studies do not employ rigorous research methods such as randomisation or collecting baseline data, the research team found 60 coaching studies that used a causal research design and examined effects on instruction and achievement. For their purposes, they defined coaching programs as “all in-service PD programs where coaches or peers observe teachers’ instruction and provide feedback to help them improve”. The coaching process itself they characterised as one where instructional experts work with teachers to discuss classroom practice in a way that is: individualised, with one-on-one sessions; intensive, with frequent interaction; sustained, with support extending over a period of time; context-specific, being within the context of a teacher’s classroom and focused on deliberate practice of specific skills.

In relation to instruction, the authors found large positive effects of coaching on teachers’ instructional practice. This included pedagogy, relationships, student engagement and classroom climate.

In relation to student achievement, measured on test results, they found positive effects, on average, across studies, while noting smaller gains on student achievement for general coaching programs than content-specific programs.

The team wanted to find out what were the features of effective coaching programs. While this research question was constrained by the nature of the various research designs, they were able to make some observations.

• Pairing coaching with group trainings gave larger effect sizes on both instruction and student achievement. It may be that teachers benefit from building baseline skills, for example, content knowledge prior to working with a coach.

• Pairing coaching with instructional resources and materials, for example a specific curriculum, is also associated with greater gains..

• There was no difference in outcomes for teachers assigned to an in-person coach versus a coach who met with teachers virtually..

• There was no evidence that coaching has to be ‘high dosage’ (in terms of number of hours) to be effective. This suggests the quality and focus of coaching may be more important than the actual number of contact hours.

The other important issue for school leaders in this study was the relationship between the scale of a coaching program and its effect size. Many school leaders hope to scale up their coaching programs in order to provide wider access to in school coaching and to build a ‘coaching culture’ in the school. Further, the team investigated the relationship between the scale of a coaching program and its effect size. For both instruction and achievement, it appears that there could be diminishing effects as programs are scaled up in size. Smaller coaching programs (of up to 50 teachers and a handful of coaches) had up to twice the benefit for instruction and three times the benefit for achievement of larger programs of 100 or more teachers. One obvious point of comparison is that smaller programs are often tailored specifically for teachers who are motivated to participate and the school contexts under which they work. Larger programs generally require training of a large number of coaches and involve teachers who are more likely to have mixed levels of interest in the program.

The study considered other challenges in scaling up programs:

• Building a body of capable coaches whose expertise is matched to the teachers’ expertise.

• The need for teacher buy-in – if participation in a coaching program is mandatory there are more likely to be weaker outcomes for those not interested in actively participating in coaching.

This study confirms some of the practical implementation challenges that schools face and raises some important contextual considerations to be grappled with.

Key considerations emerging from the Kraft et al study:

• The need for investment in the coaching capacity of the school, and consistency of service to the teachers, through the provision of high quality training for school coaches

• The need for clarity of intent of coaching in your context, and shared language when talking about classroom practice (instructional practices in the terms of this study)

• How to maintain and sustain the internal coaching resource of the school, for example through reflective practice and/or coaching supervision, to ensure quality and consistency of coaching experience for teachers.

As always, research does not necessarily show us ‘what works’ but provides a lens through which to consider our own contexts and to help us make sense of the challenges we face and questions we seek to resolve. In a relatively young and emerging field like coaching in education, we need more research to continue to understand not only the dynamics of good coaching but also the broader contextual challenges of implementing new ways of working in schools.