Building teacher capacity: data-driven approaches and state reform

In the wake of the 2021 National Achievement Survey (NAS), the outcome presented insights on the academic landscape of Tamil Nadu’s schools. These were performing below the national average in terms of the learning outcomes. The study became a catalyst for change. It paved the way to establish the Ennum Ezhuthum (EE) Mission (Foundational Literacy and Numeracy Mission) as a collective acknowledgement of the learning gaps that the survey brought to the forefront, particularly in grades 1-5.

The mission has established itself as a narrative of recognition. It has evidenced a commitment to bridge educational disparities in the state. These are systemic and programmatic. Their mission has had its effects in operational areas across different domains in the government machinery. These include curriculum, teacher training, assessments and governance.

Within the Ennum Ezhuthum ecosystem, teacher capacity building is specifically crafted to empower educators. It aims to address the unique challenges posed by multi-grade and multi-level (MGML) elementary classrooms. This has been implemented across 38 districts in the state. The training employs a scaffolded model. It encompasses three layers of building capacity – state level, district level, and cluster level. With participation from ~1 lakh teachers in Tamil Nadu, the blended model is a comprehensive approach to professional development. It is contextual and rooted in pedagogical insights.

A typical Ennum Ezhuthum classroom is a microcosm of policy being translated into practice. The aim is to build a robust learning environment that binds all its stakeholders – the student, the school and the community. Differentiated instruction is a key lever for the teachers to execute level-based instruction through buckets such as – Arumbu (seed), Mottu (bud) and Malar (flower). The teacher has a tool kit of a Teacher Handbook (THB), Student Workbook (SWB) and the Ennum Ezhuthum TLM Kit (collection of sensorial teaching-learning instruments).

The scope for training and capacity building of teachers was not prescriptive, where the teacher was introduced methodology in silos. Instead, it was fractionalized to provide opportunities for reinforcement, transferability of skills and reflective practice during the training. It built off from the prior knowledge of teachers. It also tried to include the rich expertise in context and groundlevel awareness. The key tenets adopted by the state were need-based training method and content, contextualized tools to capture quality and efficacy of teacher trainings at scale, and iterative model of teachers’ capacity building monitored through datadriven training needs.

Teachers’ capacity building at scale did present to the state the challenge of planning for sustainable training structures, which had to be standardized across clusters. The priority of the state was to employ accessible immersive communities of practice for teachers, meeting diverse needs presented in an MGML classroom, and integrating contemporary pedagogical practices through approaches, strategies and tools for effective teaching practice. Our training model has three pillars. We share details about these in the following sections.

Scalable content delivery

Everyone has always appreciated the need for differentiated training or need-based training. One of the biggest challenges in executing such a model is the scale of operations and hence the capacity of the system to deliver it effectively at scale. Hence, one of the key aspects of innovation for teacher training was to use technology to enable effective delivery of content. Facilitation at scale was administered through digital content with one-click access through agenda and guidelines. Access was also provided to all training collaterals such as presentations, video scribes and playlists before the training. This promoted readiness in teachers to engage effectively during the training.

Initially, there was a lot of resistance from the teachers and resource persons (RPs) to adopt the new methodology. For instance, some RPs felt that the structure of the agenda did not allow them to express or share their innovative ideas with the teachers. When we began the training in 2022, the Training Alignment and Effectiveness Level (TAEL) was less than 50%. This meant that the training spaces were not aligned with the design, and there was significant deviation from the SOPs (Standard Operating Procedures) laid down.

We listened to the feedback. For instance, for many RPs the infrastructure in the training venue was not adequate to conduct an effective blended learning experience. We institutionalized a set of training sessions for the RPs particularly around facilitation practices rooted in the principles of andragogy. Over time, we observed that the TAEL went up by 36% in one year. Setting up platforms for delivering content at scale is a great beginning. However, rigorous follow up, along with measuring data at every stage of the user experience, is critical to improve adoption of practices over time.

The TAEL is measured through an observation tool for classroom teaching and training, teacher performance in quizzes and demonstrations. First, a digital training observation tool was developed to be used by administrators during training observations. This tool facilitates the assessment of various aspects. These included facilitation practices, participant engagement, content delivery, infrastructure, and the incorporation of Ennum Ezhuthum-specific practices during the training. Teacher adherence is gauged through attendance, while understanding is assessed through teacher quizzes along with a feedback checklist. This is utilized to provide peer feedback and observations by facilitators during teacherled group demonstrations.

Need-based design and iteration

The classroom observation data serve as a rich foundation for developing teacher training content by offering a nuanced understanding of teaching practices. Data analysis is conducted to identify patterns and Areas of Strength (AOS) and Areas of Development (AOD). The findings are shared with the mission’s steering committee, state subject experts and training coordinators.

Subsequently, targeted training content is tailored to address specific needs identified through the observations. This content is embedded into the training through multiple modes. These include input sessions and tasks, reflective spaces and thought partnering, and planning and teacher demonstrations.

Implementation of the training sessions is followed by ongoing monitoring and iterative adjustments. This ensures that the content remains responsive to the evolving needs of the teachers, and promotes sustained professional growth. Through this comprehensive approach, data from classroom observations become a dynamic source for enhancing teaching practices. This also elevates overall teacher effectiveness.

One example of a success was “Assessment practices in the classroom.” One of the new practices instituted in the EE Mission was conducting one-on-one assessments for young learners. Though a global best practice, the assessment practice was new for the teachers. And this became evident in the data collected as well. Hence, we focused training sessions on conducting effective training assessments. Over time, we saw that the alignment of teacher practices around assessments improved.

One example of such an approach not being effective was with the phonics training we provided to the teachers. The data showed that the teachers struggled with teaching English letter sounds and reading fluency lessons in the classroom. Based on this information, we had planned for the training.

The training resulted in increased stress among the teachers, as phonics and English language teaching was a sensitive topic with them. The teachers being trained in the vernacular language themselves, were not comfortable with the “phonics” approach. This bears an important learning that though data may provide us with insights, the decisions need to be based on context and some qualitative information. After this experience, we currently talk to 3-5 teachers based on the data trend to hear their perspectives before deciding on a course of action.

On the other hand, the TPD framework developed by Madhi Foundation maps a progression of pedagogical components that align with the curriculum, EE teaching methodology, and theoretical modular components such as lesson execution, effective use of Teaching Learning Material, etc. The strategy of using the TPD framework is very similar to the theory of leading and lagging indicators. The data from classroom observations give us the lagging indicators for targeting efforts to improve teacher training. Whereas the TPD framework gives us the leading indicators to structure the TPD course plan.

For instance, this year, the training delved into the cultivation of vital parental engagement strategies. The aim was to establish robust parent-teacher partnerships. This is a focus area for the 15th to 24th months of execution of the mission. Ensuring strong parentteacher communication is vital to improving student learning outcomes.

Simultaneously, based on the data collected, the training placed a significant emphasis on equipping educators with effective classroom management techniques. These have been tailored for the unique challenges of MGML classrooms under the Ennum Ezhuthum initiative. During the sessions, hands-on strategies for crucial elements of classroom management were actively employed and demonstrated by teachers.

There is still a long way to go to institutionalize the need-based approach. For instance, the current training modules are based on trends observed at the statelevel. But the true spirit of need-based training lies in being able to differentiate at the cluster level (a group of 20-30 teachers). Furthermore, it is not only the targeting that needs to be improved, it is also the thinking and the agency. How do we equip districts and blocks to decide on the training needs over time?

On the other hand, is the reliability of data. As with many self-reported datasets, there is strong evidence of upward bias of data.

Which means that having access to reliable data is a critical prerequisite for empowering educators and administrators to conduct need-based training. This requires sustained effort on training the observers and setting up verification mechanisms.

For example, we conducted an experiential training to improve the inter-rater reliability of observations. In these sessions, the observers were asked to observe the same classroom situation and complete the observation tool. They were then asked for their rationale and thinking. They discussed this in small groups. Lastly, the ideal response was provided as feedback to the observers.

After this training, we observed that the reliability of data improved by 46%. This further validated the need for training administrators on effective observation and the strength of the data-led approach to need-based teacher training.

To further democratize the training structure, we developed asynchronous training nanomodules. These are specialized learning units that focus on delivering specific skills, knowledge, or competencies in a concise format. These were created using the trends identified from the data. Through this modular approach to training, each module is designed to address a particular aspect or topic that will support in reinforcing concepts pre/post training.

These modules are made available through the EE Champions playlist on the state-run YouTube channel for easy reference. The design is inclusive to meet the needs of access, for example, availability of subtitles, setting the pace of the video, flexibility to pause/repeat the content anytime from anywhere. Artefacts (images and videos) shared by teachers on the Telegram group are leveraged to be presented as exemplar or best practices along with highlighting the theory behind the practice.

Embedding evaluation and learning as part of training design

Kirkpatrick’s framework is anchored into our training pathway through four levels of evaluation. These serve as a continuum to each other. These four levels include – reflective spaces for cross-pollination of ideas (teacher demo and feedback), acquisition of knowledge through case lets (translating theory to practise), behavioural shifts identified through classroom observation data, and tracking overall impact in teacher-student outcomes.

We measure reaction through observation and teacher feedback forms. Quizzes and demonstration observations provide data on learning. The classroom observation tool gives feedback on change in behaviour. And the culmination of these efforts shows us the improvement in teacher and student outcomes.

This was an intentional aspect of the design. Most policies and reforms are good on paper. However, these are plagued by implementation inconsistencies. Hence, getting consistent and constructive feedback cyclically on both the outcomes and processes was important. It was a critical necessity to ensure the success of the training efforts undertaken by the department.

One of the long-standing use cases for the data collected via training is for improving teacher motivation and excellence in practice. We use the data to recognize the teachers for best teacher’s award, incentives for grants and other recognition avenues. This continuous use of data for recognition of teachers has allowed us to strengthen classroom practices over time.

We continue to navigate in establishing a strong system of self-directed educators. In this process, a learning we ground ourselves to is that the training model must be iterative and responsive to the stakeholders we serve. Hence, we must always reflect critically, diagnose challenges and course-correct, as we strengthen implementation of a statewide policy. However, our anchor for these decisions must be the outcomes – both teacher and student outcomes.