Introduction
South Africa’s ‘learning crisis’ can be traced back to the early grades, where far too few learners master the fundamental literacy and numeracy skills necessary for successful progress through school. There are myriad reasons behind this failure but the lack of resources and teacher expertise are among the primary drivers. Foundation Phase teachers have not been provided with opportunities to acquire specialized knowledge in teaching learners to read, particularly in African languages, nor do their classrooms have sufficient and appropriate African language reading materials. Evidence shows that teacher content knowledge in mathematics is particularly poor, with resulting poor content-specific pedagogical knowledge. This lack of knowledge and resources, combined with large and heterogenous classes, makes it challenging for teachers to effectively integrate resources and implement effective pedagogical practices. This leads to limited curriculum coverage and classrooms dominated by teacher-led instruction and communalized whole-class activities with limited targeted individual practice and feedback[i].
In order to address the challenges posed by large and diverse classrooms, policymakers in several countries have explored utilizing teacher assistants or youth volunteers to provide assistance with classroom management and additional support to students in small groups or one-on-one. The extremely high rate of youth unemployment in South Africa makes harnessing the potential of unemployed youth a particularly promising approach. In 2020, the Department of Basic Education (DBE) and the Presidency introduced the Presidential Youth Employment Initiative-Basic Education Employment Initiative (BEEI), which has now entered its fourth phase. The initiative has provided training and placement for around 840,000 young people on contracts lasting from five to eight months as education or general assistants in public schools, at a cost of R25.5 billion[ii]. However, there has been no evaluation of the programme’s impact thus far.
With the goal of enhancing teachers’ content knowledge and pedagogical skills, Funda Wande created a comprehensive set of high-quality materials for Reading for Meaning (Funda Wande) and Calculating with Confidence (Bala Wande) in various African languages. These materials are fully aligned with the CAPS curriculum and integrated in a manner that enables teachers to move effortlessly between different resources. Along with these materials, Funda Wande established a Teacher Assistant (TA) programme in Limpopo. This programme aims to develop a model for selecting, training, and supporting matriculants to aid teachers within a structured program to enhance early literacy and numeracy outcomes. If successful, this model could shape the design of ongoing BEEI and other youth initiatives.
SALDRU was tasked with conducting an independent evaluation of the Funda Wande intervention in Limpopo with the aim of assessing whether the Funda Wande materials, training, and support in the form of teacher assistants are effective in making progress towards the programme’s stated goal of all learners reading for meaning and calculating with confidence by the end of grade 3. The impact evaluation uses a randomized control trial with 120 no-fee schools in the Limpopo districts of Capricorn North and Capricorn South, randomized into one of three arms – Learner and Teacher Support Materials (LTSM), LTSM together with a TA (LTSM+TA) assigned to each classroom, and control – over three years (2021-2023). The schools are all no-fee (quintile 1-3) Sepedi language of learning and teaching in the Foundation Phase and were selected in collaboration with the Limpopo Department of Education[iii].
During the third term of 2021, the evaluation team conducted literacy and numeracy assessments with around 2400 grade 1 learners across the 120 schools. These same learners were re-assessed in the third term of 2022 when most of them were in grade 2[iv]. This article shares the main findings from the second wave of the evaluation.
Key findings
The provision of Funda Wande selected, trained and supported teacher assistants to classrooms over six school terms has proved highly effective in improving foundational reading and mathematics skills. By the third term of grade 2, learners in classrooms with Funda Wande teacher assistants significantly outperformed their peers in comparable control schools by 0.5 standard deviations in both literacy and numeracy. These gains are substantial; translating to around 1.25 years of learning in these skills in ‘business as usual’ environments. An examination of the sub-tasks shows learners in classrooms with Funda Wande teacher assistants scoring significantly higher across the full range of sub-tasks in both the Early Grade Mathematics Assessment (EGMA) and the Early Grade Reading Assessment (EGRA).
Figure 1. Treatment effects – Early Grade Mathematics Assessment
Figure 2. Treatment effects – Early Grade Reading Assessment
Learning gains in the materials only arm are more modest but still significant. Learners in the LTSM schools outperformed their peers in control schools by 0.26 standard deviations on the EGMA and 0.2 standard deviations on the EGRA. This demonstrates that a comprehensive set of high quality, integrated resources that are aligned to the curriculum and provided to learners in a one-to-one ratio can result in substantive improvements in early grade literacy and numeracy.
Learning gains are also educationally meaningful with a significantly higher proportion of learners in schools with Funda Wande teacher assistants achieving learning outcomes in line with curriculum expectations. EGRA results can be compared against the DBE’s newly established grade-specific benchmarks for Sepedi[v]. By the third term of grade 2, the percentage of learners reaching the grade 2 benchmark was only 22 percent in control schools compared to 32 percent and 39 percent in materials only and teacher assistant schools respectively (Figure 3). This signals that the Funda Wande program has been effective in moving a significant proportion of learners to the expected performance level for their grade. There is clearly a substantial way to go to get every learner to meet the benchmarks, but these results are very encouraging, particularly considering the learning losses associated with COVID-19 related school disruptions[vi]. Looking to the benchmark for the end of grade 1, we find substantially more learners in teacher assistant schools were meeting this earlier benchmark (60 percent) compared to learners in control schools (44 percent).
Figure 3. Percent of grade 2 learners reaching grade-specific Sepedi benchmarks
Teacher assistants are positively received by teachers and perform a range of roles in the classroom that are well aligned to their training and the programme’s expectations. Unprompted, teachers report assisting learners during learner activities (91 percent), assisting the teacher with marking (81 percent), monitoring learners (77 percent) and working with learners in small groups (70 percent) as the most common roles of the teacher assistants in their classrooms.
Figure 4. Roles of teacher assistants as reported by teachers in LTSM+TA schools
Reported use of the Funda Wande and Bala Wande materials is very high in both treatment arms. Around 90 percent of teachers report using the teacher guides and learner activity booklets daily. Due to concerns around desirability bias associated with teacher self-reports around the use of intervention LTSM, we conducted an audit of the workbooks of two randomly selected learners in each school to provide an objective assessment of levels of use.
Learners in Funda Wande schools complete many more pages in their learner workbooks than their peers in control schools. Not only do they complete more pages, but these pages are more likely to be marked, particularly in the teacher assistant schools. This suggests that the Funda Wande program results in better curriculum coverage and more individualized feedback.
Figure 5. Number of partially or fully completed pages in mathematics workbooks
The workbook audit reveals considerable variation in usage across treatment schools suggesting that the Funda Wande programme is not uniformly adopted and coverage for some teachers is very low. Understanding the drivers of this variation is key for Funda Wande as they continue to develop sustainable and scalable models of support.
Conclusion
The implementation and accompanying evaluation of the Funda Wande programme in Limpopo provides the first rigorous evidence on the potential of unemployed youth to contribute to efforts to address the learning crisis in South African schools. The evaluation found that teacher assistants can be highly effective in improving foundational literacy and numeracy outcomes when rigorous recruitment and selection is in place, high-quality learner and teacher materials are available in classrooms, comprehensive training is provided, and continuous support is offered. These results are particularly encouraging considering the limited exposure that learners had to the programme in the first year due to ongoing COVID-19 related closures and rotational timetabling.
The BEEI has established the organisational, political and financial feasibility of operating a teacher assistant programme at scale. The Funda Wande programme demonstrates that it also feasible to meaningfully impact learning outcomes and provides a model for recruitment, training, and mentoring to achieve that. However, the intervention and accompanying evaluation have some limitations. The impacts are measured over a fairly short period that includes months of disrupted schooling due to COVID-19. It will be important to ascertain the extent to which initial gains are maintained or possibly compounded in the medium and long term. The schools are based in two education districts in one province. However, these low resourced and under-performing no-fee schools share characteristics with the vast majority of schools in our education system. One key concern with scaling up this programme is the ability to continue to selectively recruit youth. The very high youth unemployment rate, together with the large number of applicants for the BEEI positions suggest that this should not be a major constraint to scale. Funda Wande will need to give careful consideration to the adaptation of each component of the intervention to operate at scale.
While the results for the materials only arm of the intervention are more modest, learner gains in these schools are significant and educationally meaningful. This is an important piece of evidence around the impact of well-designed and contextually appropriate resources that are mindful of barriers to use. Earlier studies in Africa have led to a common wisdom that resources without ongoing support are seldom effective. This is unfortunate and overly simplistic as each of these studies is careful to point out reasons as to why the inputs were effective (e.g. wrong language, wrong level, materials locked away)[vii]. The evaluation shows that well-designed and integrated resources that are aligned with the curriculum and provided to learners in a one-to-one ratio, can lead to significant improvements in foundational literacy and numeracy skills.
References
Ardington, C. and Henry, J. 2021. Funda Wande Limpopo Evaluation: Midline Report. SALDRU. University of Cape Town. Cape Town.
Ardington, C., Wills, G. and Kotze, J. 2021. COVID learning losses: early grade reading in South Africa. International Journal of Educational Development 86
Ardington, C., Mohohlwane, N., Sebaeng, L., Jodar, P., Beggs, C., Makgabo, C., Maledu, A. & Zwane, Z. 2022. Sepedi Early Grade Reading Skills Benchmarks. Technical Report. Pretoria: DBE.
Department of Basic Education and Presidential Employment Stimulus. 2022. Implementation Framework for the Presidential Youth Employment Initiative (PYEI) in the Basic Education Sector: PHASE IV 2023 – 2024.
Glewwe, P., Kremer, M., and Moulin, S. 2009. Many Children Left Behind? Textbooks and Test Scores in Kenya. American Economic Journal: Applied Economics, 1(1), 112–135
Hoadley, U. 2018. Pedagogy in poverty: Lessons from twenty years of curriculum reform in South Africa. Abingdon/New York: Routledge.
Hoadley, U. 2012. What do we know about teaching and learning in South African primary schools? Education as Change, 16 (2):187–202.
Kotze, J., Wills, G., Ardington, C., Taylor, S., Mohohlwane, N. & Nuga-Deliwe, C., 2022. Background Advisory Note to the 2030 Reading Panel: Learning losses due to the COVID-pandemic, 2022. Pretoria: Department of Basic Education.
NEEDU. 2013. NEEDU National Report 2012: The State of Literacy Teaching and Learning in the Foundation Phase. National Education and Evaluation Development Unit. Technical report. Pretoria: Department of Basic Education.
Sabarwal, S., Evans, D., and Marshak, A. 2014. The permanent input hypothesis: The case of textbooks and (no) student learning in Sierra Leone (No. WPS7021; pp. 1–37). The World Bank.
Taylor, N. 2019. Inequalities in Teacher Knowledge in South Africa. In Spaull, N. and Jansen, J.D. (eds.) South African Schooling: The Enigma of Inequality. Springer. Chapter 14.
Van der Berg, S., Hoadley, U., Galant, J., van Wyk, C. & Bohmer, B., 2022. Learning Losses from COVID-19 in the Western Cape. Evidence from Systemic Tests. Stellenbosch: Research on Socio-Economic Policy, Stellenbosch University
[i] See Taylor 2019, Hoadley 2012; Hoadley 2018 and NEEDU 2013, for detailed discussion on how these various factors contribute to the learning crisis in South Africa.
[ii] Department of Basic Education and Presidential Employment Stimulus 2022: page 19.
[iii] See the first midline report (Ardington and Henry 2021) for details around recruitment and randomization of schools and balance across treatment arms.
[iv] The team were able to re-assess 91%, 90% and 91% of learners in the control, LSTM and TA arms.
[v] The Sepedi benchmarks (Ardington et al. 2022) are as follows:
End of Grade 1: all learners should correctly sound at least 40 letters per minute
End of Grade 2: all learners should correctly read at least 40 words per minute
End of Grade 3: all learners should correctly read at least 60 worder per minute
[vi] Various South African studies suggest that learners lost between 40% to 118% of a year of literacy learning due to school closures and rotational time-tabling (Ardington, Wills and Kotze 2021, Kotze et al. 2022, van der Berg et al. 2022).
[vii] See Glewwe, Kremer and Moulin 2019 and Sabarwal, Evans and Marshak 2014 as two key examples.