Teachers learning together: insights on leadership and collaborative learning in Kenya and Rwanda

Charlotte Jones
Andrea Tiganescu
Donvan Amenya

Teacher collaborative learning, such as in communities of practice, is a highly promising approach to teacher professional development, but there has been a lack of evidence on how these communities can work best in low-income settings – or at scale. In a new study, we examine communities of practice in our programmes in Rwanda and Kenya to gain important insights into the connections between facilitation quality, leadership, and the impact of these interventions.

Teacher professional development (PD) is central to teaching quality – and therefore to student outcomes. It is rightly recognised as a key area for intervention by governments, multilateral organisations and NGOs around the world (two thirds of the World Bank’s education programming includes a teacher PD element), but results too often show that PD interventions are ineffective, especially when applied at scale.  That said, there is some consensus that teacher collaborative learning – teachers learning together as peers, such as in teacher communities of practice (or CoPs) – is a highly promising approach to CPD. However, there has been a lack of evidence on how to adapt such collaborative learning settings for low-income settings or on effective scaling of these interventions.

Education Development Trust already has excellent impact data on successful CoPs from our Girls’ Education Challenge programme, and emerging data from our CoPs in Rwanda. However, as a learning organisation, we are keen to understand more about what makes effective practice. We are therefore investing in a major study which evaluates:

  1.  the impact of large-scale teacher community of practice interventions on teacher outcomes (such as practice, knowledge, motivation and self-efficacy)
  2. the features of effective CoPs (such as timing, content, processes and culture)
  3. the roles of different actors in these communities and the effect that these have on CoP outcomes.

As part of our adaptive programming philosophy, the findings from the study are already being used by our teams to improve CoP interventions.

The study features over 700 teachers and almost 200 schools in total, using a mixed-methods (qualitative and quantitative) approach, which is tailored for specific country contexts. In Kenya, we compare 285 teachers (across 14 cluster-based CoPs) with 185 control group teachers, covering 99 schools in total. In Rwanda, there are no control schools, but we are analysing the impact of school-based CoPs at different phases of a programme rollout, observing 278 teachers across 97 schools, as well as three focus group discussions for head teachers. The comparison of school-based (Rwanda) and cluster-based (Kenya) models makes this study particularly distinctive.

We have now completed baseline analysis from Kenya and Rwanda. A full paper of early findings was shared at the RISE conference in June 2019. Our key findings include:

Teacher participation tends to be more active in school-based, rather than cluster-based models.

The CoP models we observed had different structures. In Kenya, CoPs had a cluster-based model, with teachers meeting from five different schools, along with a subject specialist or coach. In these CoPs, the meetings were predominantly facilitated by the coach. By contrast, in Rwanda, where there were school-based models, with teachers, subject leaders and headteachers, facilitation was typically shared more equally between all actors. Teachers in school-based CoPs had a higher rate of participation and were more active agents in CoP governance. In the future, further consideration could be given to whether cluster-based CoP models undermine teacher ownership of the communities and their capacity to self-organise – and whether this has any implications for the sustainability of the CoPs.

Regardless of the facilitation model, our analysis showed a positive association between the quality of facilitation in the CoPs, and the impact that these CoPs had. This is a promising finding and definitely worthy of further exploration, especially as it accords with the wider international evidence on teacher collaborative learning.

The role of the head teacher is important – but this should evolve as the CoP matures

The role of the head teacher in communities of practice appears to be a significant factor in its impact and outcomes. In our research, we fund a statistically significant positive correlation between the level of engagement of the headteacher and CoP impact, regardless of the facilitation method, demonstrating the importance of a head teacher’s investment in teacher peer learning communities in their schools.

However, the true picture is actually more nuanced and complex than we might first assume: our qualitative findings suggested that an assumption that ‘more headteacher involvement is better’ would be flawed. In our CoPs Rwanda, it became clear that the role of the headteacher was changing as the CoPs matured, moving from heavy involvement in the early stages, to stepping back as the community matured. This allowed the teachers to have more control and agency in their learning.

This led us to consider the relationship between CoP maturity and leadership approaches – and it became clear that teacher CoPs need different things at different times. In Rwanda, we saw that although less mature CoPs benefitted from stronger headteacher engagement and directive leadership, such directive leadership can become a barrier to growth as CoPs mature. Teachers may benefit from headteachers stepping back as their CoP grows in experience, instead adopting a more distributive leadership style in which the head teacher provides less feedback. Meanwhile, the opposite appeared to be true for subject specialists: high-performing CoPs – including mature CoPs – seem to benefit from more directive inputs from subject leaders.

Contextual factors are critical to designing effective CoPs

It is clear that there is no one-size-fits-all approach for the leadership of collaborative teacher learning communities. Contextual factors, such as the maturity of the CoPs and the practical feasibility of carrying these out at school or cluster level must be considered in planning for and facilitating effective communities of practice.

Looking ahead, it is clear that we should pay more attention to leadership practices and governance processes with collaborative learning communities, in order to best enable them to flourish and make a positive impact. This does not just apply to levels of leadership inputs, but also to how far – and how well – collaborative practice can be promoted by deliberate leadership practices, and how such practices may need to change over time. This will have inevitable implications for training for facilitators, as well as for scaling projects, and is this an important issue that we will continue to explore in our future programmes and research.

Moving forward, we will be publishing further outputs from this research, including practical tools for effective CoPs. To find out more about the study, or our expertise on teacher PD, please contact us: