Background
The Malawi government has taken important steps towards improving foundational learning through the National Numeracy Program (NNP) by revising the mathematics curriculum for infant and junior classes. In 2022, the NNP reached out to EdTech Hub through our Helpdesk for advice on improving programme design, including tech-enabled Teacher Learning Circles (TLCs). Through this blog, we explore the journey and the Hub’s contribution to this initiative over the last few years, highlighting the research process and how the TLCs were refined and eventually scaled.
In-depth understanding of current practice: observing the TLC Model in Zomba
The NNP was in its expanded pilot phase when EdTech Hub commenced work. To lay the groundwork for the TLCs, the Hub conducted a series of design-based implementation research cycles. Beginning with four schools in Zomba, EdTech Hub spent two weeks assessing existing practices, their implementation, and identifying opportunities for improvement.
Just from sitting in one of these TLCs, it became clear that the environment was not appealing to teachers who had just finished a long day of instruction. With understandably hungry and tired attendees, the sessions were brief and lacked depth. Typically, two teachers would take turns asking their colleagues about the challenges they faced in their classrooms and what they would do differently. These discussions were happening, however, without any guidance or materials to confirm whether the proposed solutions actually aligned with the new curriculum.
Teachers were looking for something more substantial, something that would reassure them they were on the right track with the new curriculum. They wanted to be more involved and feel a sense of ownership over TLCs. Many of the issues raised by teachers were related to workload rather than teaching techniques. It became clear that teachers needed TLC sessions that were directly connected to what they were teaching in the classroom. To confidently teach and support their students, they first needed to grasp the subject matter themselves.
This realisation led to a new focus in the research, aimed at making the TLCs more practical, hands-on, and directly tied to real classroom challenges. There was also a push for more modelling of teaching strategies and keeping TLCs simple but effective.
Co-creation: Expanding to 20 Schools
With insights from the initial phase, subsequent research was expanded to include 20 schools across Lilongwe and Zomba. This phase focused on co-designing and refining the TLC model by working closely with teachers and school authorities through a combination of TLCs, classroom observations, and focus group discussions. The Hub’s teaching experts were key in shaping the new TLC model. Offering valuable guidance on the types of activities and teaching methods to include in these sessions, and ensuring these were effective and practical for teachers. These interactions were invaluable in fine-tuning the TLC model to meet the needs of teachers and the unique challenges they faced in their classrooms.
A new TLC model would first be given to a school facilitator and a day would be set aside for observation. Afterwards, teachers would participate in a focus group discussion where they were asked what they enjoyed about the new TLC, what they had learned, and how they planned to apply it in their classrooms. The key question was, ‘If given a chance, what would you change in the TLC?’
This part really energised the teachers, as they felt involved and heard. They made significant contributions to shaping the TLC to better match their needs, such as determining how much time they wanted to spend on each section. Teachers also gave input on how they preferred the topics to flow and how they should align with their term work and workbooks. Additionally, they suggested certain topics should be given more emphasis due to their level of difficulty.
This process of co-design and continuous feedback helped create a model that was not only effective but also practical and adaptable for schools across different regions. As the NNP was going towards scaling up.
Validating a new model
Following the co-design phase, we tested the final TLC model in five schools. During this phase, facilitators were coached on how to effectively conduct Teacher Learning Circles, with a focus on fostering collaboration and hands-on learning. Observations were conducted to monitor the progress of these TLC sessions, and officials from the Ministry of Education were invited to witness the model in action. Their presence was crucial, as their endorsement was key to scaling the initiative. Reflecting on our journey, this stage was pivotal because it confirmed the model’s potential for scalability. Watching new facilitators successfully lead the TLCs as intended was the outcome we had been striving for.
The hard work paid off. The final TLC model was officially adopted and scaled to over 1,000 pilot schools across Malawi. The TLCs now hold real value for the teachers. In our consultations with them, they report feeling a sense of ownership over the process, and as a result, they’re no longer focused on the time or eager to leave.
Holistic contributions and impact
Beyond our TLC testing and research, our team in Malawi provided direct contributions to the development and implementation of teacher training at the national level, and helped to establish a new tool for classroom and TLC observations and data collection. Ultimately, the expanded pilot phasewas evaluated through a Randomised Controlled Trial (RCT) evaluating the NNP’s impact in 75 treatment schools across three districts. It found that learners in treatment schools achieved a 0.19 standard deviation improvement in overall numeracy, a significant increase. It also (importantly for our many teacher-based contributions) found significant improvement in teaching quality and teachers’ attitudes towards maths teaching and learning.
Continued learning: low-tech in support of TLC facilitators
While our team was really pleased with the success of the expanded pilot programme and TLCs, we knew that more was needed to support TLC facilitators at scale. As preparations for a full-scale implementation across over 6,000 schools in Malawi began, we began to explore: how might we maintain the quality of TLC facilitation over time and at a national scale?
With this in mind, we decided to pilot the use of videos as a tool for engaging with TLC facilitators and preparing them for this role. We conducted a new DBIR cycle in two districts, Lilongwe and Mwanza, where we tested the potential of TLC-specific videos designed to assist facilitators in delivering these sessions.
In collaboration with the NNP team, two types of videos were produced: “TLC Basics,” which demonstrated the structure of each TLC section for facilitators, and “TLC Topics,” which guided how to introduce and facilitate specific topic-based activities. We also experimented with various methods of video distribution, including WhatsApp and tablet-based approaches, to determine the most effective channels.
Feedback from TLC facilitators and teachers suggests that the videos were valuable: many reported feeling better prepared and, on average, demonstrated increased effectiveness in leading TLC sessions.
Takeaways and lessons learned
Teachers’ own needs and preferences need to be a part of effective TPD design. This was made evident in our co-design process, where teachers actively helped shape their TLC model: their feedback played a key role in deciding how much time to dedicate to each topic and ensuring the TLCs were aligned with their lesson plans. This collaborative approach made the TLCs more practical and truly responsive to what teachers needed to improve their teaching.
This is not only true for TLCs but also across tech-enabled TPD. A core theme throughout our work at Edtech Hub is that designing TPD resources to be easy to understand, access, and implement and ensuring that all teachers are included, are all critical elements to success. We were glad to play a role in involving teachers’ voices in the design of NNP and grateful for the partnership of its implementers, who were willing to iterate and grow the model based on these contributions.
