Curriculum and language reform in South Africa and its neighbours fail first at teacher preparation

Introduction

South African education is often talked about as if it were a case apart. It has well-known universities and a national curriculum that has been rewritten and tidied more than once. Yet when people line it up against Kenya, Rwanda, Ghana, Botswana or Zimbabwe on the things that actually reach children, such as how a new curriculum lands in classrooms, which language a six-year-old is taught in, and who trained the teacher, the differences are smaller than the reputation suggests. Kenya’s new competency-based curriculum was introduced in what one reviewer calls a haphazard way (M’mboga Akala, 2021). Rwanda moved to English as the medium of instruction and primary teachers trained in other languages struggled (Mahoro et al., 2024). South African student teachers in primary programmes score barely higher in English after four years than they did in their first (Roberts, Moloi and Mort, 2024).

This paper argues that the comparison points less to a gap in curriculum design or in language policy than to a shared weak point, which is how teachers are prepared and supported when a reform or a language policy arrives (M’mboga Akala, 2021; Teane and Hughes, 2024). I look at curriculum reform first, then at language of instruction and early reading, then at teacher education, and then at how far the evidence can bear a comparison at all. The sources are mostly small qualitative studies from the countries named, together with one direct South Africa and Zimbabwe comparison (Machisi, 2023), so the argument is a cautious one. I take the view that South Africa is better read as one member of a regional pattern than as the exception, and that this changes what it should borrow.

Curriculum reform lands on teachers first

The Kenyan case is the best documented of the reforms here. Gichuru et al. (2021) trace the country’s curriculum changes in a scoping review, showing that they were shaped by missionary entry, colonial rule and later national development goals, and that globalisation has pushed policymakers towards new curriculum strategies. That history matters because it seems to make the move to a competency-based curriculum one more step in a long chain and not a sudden break. M’mboga Akala (2021), working from secondary literature and data, is less kind about the present step. Early challenges included inadequate human and material resources, a mismatch between curriculum content and what teachers actually do, and weak public participation, and the reform was rolled out, in her words, haphazardly. Her fix is to resource schools, involve stakeholders properly and retrain teachers.

One of the three early challenges in the Kenyan review gets less attention than the other two, and that is weak public participation. M’mboga Akala (2021) recommends meaningful engagement of all stakeholders, which reads like a standard line until it is set beside Gichuru et al. (2021), who describe a curriculum that is meant to give citizens the knowledge, skills and values for national development. If parents and communities were not brought along, a competency-based curriculum may be hard to explain to them in terms of what children will be able to do. Neither source measures that directly, so this is an inference from two reviews and not a finding. It does suggest that a reform can fail quietly at the community end as well as at the classroom end.

Ghana’s version comes from the other end of the school system. Buabeng and Amo-Darko (2025) interviewed eight senior high school teachers in the Cape Coast Metropolis about the Standards-Based Curriculum, which pushes teaching towards critical thinking, creativity and problem solving. The teachers described too little training, little practical professional development, and thin resources including ICT and STEM materials. Many slid back to traditional methods, which defeats the point of a competency-based reform. Eight teachers in one metropolis cannot speak for Ghana. Still, the story matches Kenya’s closely enough, and the two studies, from different countries and different school levels, land on the same list of missing things (Buabeng and Amo-Darko, 2025; M’mboga Akala, 2021).

South Africa differs in being further along. Its Curriculum and Assessment Policy Statement has been in schools for years, so the problems are those of a settled curriculum and not of a launch. Machisi (2023) is the one source in this paper that compares South Africa directly with another country, using policy documents, syllabi, examination papers and research reports for secondary mathematics in South Africa and Zimbabwe, along with the author’s own experience. The recommendations are telling. For South Africa he suggests, tentatively, longer examination time, a higher pass mark, continuous assessment spread over a longer period and perhaps fewer subjects in Grades 11 and 12. For Zimbabwe he suggests more resources, compulsory mathematics for advanced-level candidates, trial examinations and regular deployment of mathematics curriculum advisors to support teachers in schools (Machisi, 2023).

Put side by side, the three studies show two stages of the same problem. Where a curriculum is new, as in Kenya and Ghana, the complaint is that teachers were not ready and schools were not equipped (Gichuru et al., 2021; Buabeng and Amo-Darko, 2025). Where it is established, as in South Africa and Zimbabwe, the proposals turn to assessment design and, in Zimbabwe’s case, to people who visit schools and help teachers (Machisi, 2023). I would read the curriculum advisors idea as the quietest and perhaps most useful finding. It treats support for teachers as a standing part of the system and not as a launch-week workshop, which is what Kenya’s reviewers wish had happened (M’mboga Akala, 2021).

Language of instruction and the first years of reading

Language policy is where South Africa looks most different on paper. The policy intention is home-language teaching in the Foundation Phase, with English taking over as the main language of learning later on. That makes the evidence from Rwanda, where English became the medium of instruction, a useful contrast (Mahoro et al., 2024). Bowden et al. (2022) followed one mathematics teacher in a rural lower-secondary class, using an ethnographic case study with lesson transcripts and interviews. Policies for learner-centred education and English medium instruction both ran into the realities of the room. The teacher kept control of whole-class interaction, and English restricted students’ talk and their use of the textbook, though students still built activities together and conveyed mathematical meaning. The authors conclude that policy should follow classroom realities and not abstract pedagogical or linguistic ideals.

The Rubavu District study fills in the picture from primary schools. Mahoro et al. (2024) surveyed 52 teachers and interviewed ten head teachers across ten schools, and all of the teachers reported using teaching aids and oral activities. Stories, songs, games, code-switching and role play were common. Poem recitation, impromptu speaking, presentations and group discussion were not used. The authors note that teachers trained in Kinyarwanda or French struggled after English became the medium, partly because of limited proficiency and a shortage of resources, and they recommend in-service training in English. Taken with Bowden et al. (2022), this suggests that in Rwanda the language switch has not been matched by the training and materials that would make it workable, although two studies do not make a national verdict.

Botswana and Namibia give a different angle, since both involve children learning to read in a language that is not their first. Teane and Hughes (2024) spoke to 18 teachers in three primary schools in Botswana about second-language reading and found low reading development in lower primary linked to scarce resources, teachers who lacked expertise, and children arriving without any preschool. In Namibia, Shuukwanyama et al. (2022) asked four mathematics lecturers at a University of Namibia campus about mother-tongue instruction for the early grades. The lecturers saw its benefits but were not all in favour, particularly at university level, and cited students’ backgrounds, a lack of resources and too few lecturers who speak the mother tongue. These are small samples. Even so, both sets of teachers describe the same thing, a language policy that depends on people and materials that are not there.

South Africa’s contribution is a different kind of evidence. Ardington et al. (2021) used longitudinal Early Grade Reading Assessment data on nearly 14,200 children to set benchmarks for oral reading fluency in isiZulu, isiXhosa and Siswati, with a lower threshold of 20 correct words per minute and an upper benchmark of 35. This is not a study of teachers, and the benchmarks are necessary but not sufficient for reading with meaning, as the authors say. What it does show is that African-language reading in the early grades can be measured with a precision the Rwandan, Botswanan and Namibian studies do not attempt. If the comparison tells us anything, it is that South Africa has a stronger yardstick and the neighbours have the same classroom problems (Ardington et al., 2021; Teane and Hughes, 2024).

It is worth noticing how lopsided the evidence is. South Africa can say how many correct words a minute an early-grade child reads in isiXhosa, isiZulu or Siswati, and compare that with a threshold (Ardington et al., 2021). The Rwandan, Botswanan and Namibian studies mostly report what teachers and lecturers said, or what happened in one class, and none of them gives a learner-level measure that could be set beside the South African one (Bowden et al., 2022; Teane and Hughes, 2024; Shuukwanyama et al., 2022). That limits the comparison in a practical way. A reader cannot tell from these studies whether reading in Botswana’s lower primary is better or worse than in South Africa’s Foundation Phase, only that teachers in both seem to name the same obstacles. I would treat a shared yardstick for early reading in more than one African country as the obvious next step for comparative work.

Teacher education is where the threads meet

Every study so far points back to who is in front of the class (Buabeng and Amo-Darko, 2025; Teane and Hughes, 2024). The South African evidence on this is unusually concrete. Roberts, Moloi and Mort (2024) report on an English language and literacy test given from 2020 to 2022 to BEd students preparing to teach in primary schools at six public universities. Fourth-year students averaged 55 per cent (n = 731) and first-year students 52 per cent (n = 1,177). A gap of three percentage points after four years of study is, in the authors’ own view, a cause for concern. It does not tell us how well these students teach, and the test is a first version. It does suggest that programmes are not building English proficiency in a systematic way, which matters in a system where English is the main language of later schooling.

Set that beside Rwanda, where teachers trained in other languages were expected to teach in English (Mahoro et al., 2024), and beside Namibia, where lecturers worried about too few colleagues able to teach in the mother tongue (Shuukwanyama et al., 2022). The direction of the policy is opposite in the two cases, one pushing towards English and the other towards mother tongues, but the failure point is the same. A language policy is a promise about teachers’ competence in a language, and teacher education is where that promise is either kept or broken. South Africa’s own result suggests the promise is shaky even where the policy favours home languages early, because the pipeline that produces teachers is not delivering strong command of English at the end (Roberts, Moloi and Mort, 2024).

The South African authors do say what they want the test to be used for, which is to help lecturers design programmes that build English language and literacy gradually from first year to graduation (Roberts, Moloi and Mort, 2024). Namibia’s lecturers point at a different design question, because their study ends by recommending two parallel language-in-education policies, one for schools and one for the university, with English and the mother tongue given equal weight in tertiary courses (Shuukwanyama et al., 2022). Put together, the two raise something neither addresses on its own. The South African test measures English, so it says nothing about how well a student teacher can teach reading in an African language such as isiZulu or isiXhosa, which is what early-grade reading benchmarks are about (Ardington et al., 2021). That is a gap in what is being measured, as far as these sources show, and it may be worth closing before anyone redesigns the programmes.

Curriculum reform runs into the same wall. Ghana’s teachers wanted practical and continuing professional development and did not get enough (Buabeng and Amo-Darko, 2025). Kenya’s reviewer asks for retraining (M’mboga Akala, 2021). Zimbabwe’s proposed answer is curriculum advisors who come back to schools regularly (Machisi, 2023). Botswana’s teachers needed expertise in second-language reading (Teane and Hughes, 2024). All of these are versions of one request, which is sustained support for teachers after they have qualified. Initial training, whatever its quality, cannot carry a reform on its own, and none of the countries in these studies seems to have found a settled way to provide the follow-up.

How far the comparison can be pushed

A fair reading has to admit how uneven this evidence is. The studies cover different school levels, with secondary classrooms in Rwanda and Ghana, primary classrooms in Botswana and Rubavu, and university lecturers in Namibia. Sample sizes are tiny: one teacher in Bowden et al. (2022), eight in Buabeng and Amo-Darko (2025), four lecturers in Shuukwanyama et al. (2022) and 18 teachers in Teane and Hughes (2024). Machisi (2023) rests on documents and the author’s own experience, and Gichuru et al. (2021) is a review of literature and not of classrooms. Only Ardington et al. (2021) and Roberts, Moloi and Mort (2024) work with large numbers, and both are South African. So the apparent similarity across countries is partly an effect of looking at small qualitative studies, which may well report resource shortages whatever the setting.

There are also real differences hidden by the common list of complaints. Kenya and Ghana are describing reforms in their first years, while South Africa and Zimbabwe are describing mature systems (Machisi, 2023). Rwanda’s switch to English is a language-of-instruction change in the middle of schooling, which is a different problem from teaching early reading in an African language (Mahoro et al., 2024; Ardington et al., 2021). And a South African reader should not assume the Botswana or Namibia findings travel across the border without checking. What I would claim is narrower than a regional law: in each country where teachers were asked, the explanation they gave for weak results pointed to preparation, resources and follow-up support before it pointed to the design of the curriculum or the choice of language.

What South Africa might take from the comparison

If the pattern holds, South Africa’s priorities look a little different. The first is to treat in-service support as part of the curriculum itself, along the lines of Zimbabwe’s curriculum advisors and not through occasional workshops (Machisi, 2023). The second is to use the strong South African reading benchmarks to check whether early-grade teaching is actually moving children over the thresholds, so that the debate about language policy can be carried on with numbers (Ardington et al., 2021). The third is to take the English proficiency of BEd graduates seriously as a design problem for initial teacher education, since the Rwandan and Namibian studies suggest what can happen when language policy runs ahead of teacher competence (Roberts, Moloi and Mort, 2024; Mahoro et al., 2024). Kenya’s lesson on public participation is probably the easiest to forget and the hardest to repair afterwards (M’mboga Akala, 2021).

Conclusion

Comparing South Africa with Kenya, Rwanda, Ghana, Botswana and Zimbabwe suggests that the weak point is shared. New curricula arrive before teachers are ready (Buabeng and Amo-Darko, 2025; M’mboga Akala, 2021), language policies depend on teachers’ command of a language that initial training has not secured (Mahoro et al., 2024; Roberts, Moloi and Mort, 2024), and early reading suffers where resources and expertise are thin (Teane and Hughes, 2024). South Africa’s advantages are measurement and a mature curriculum, not an escape from the problem (Ardington et al., 2021; Machisi, 2023). The evidence is small-scale and the comparison is rough, so none of this settles anything. It does point to teacher preparation and continuing support as the place where reform effort in all these countries is most likely to be won or lost.

Reference list

Ardington, C. et al. (2021) Benchmarking oral reading fluency in the early grades in Nguni languages. International Journal of Educational Development, 84, article 102433. doi:https://doi.org/10.1016/j.ijedudev.2021.102433

Bowden, R. et al. (2022) Learner-centred education and English medium instruction: policies in practice in a lower-secondary mathematics class in rural Rwanda. Compare: A Journal of Comparative and International Education, 54(2), pp. 294-313. doi:https://doi.org/10.1080/03057925.2022.2093163

Buabeng, I. and Amo-Darko, B. (2025) Curriculum reforms without foundation: the effects of inadequate preparation in curriculum reforms on Ghanaian teachers and the education system. Curriculum Perspectives, 45(2), pp. 133-147. doi:https://doi.org/10.1007/s41297-025-00309-7

Gichuru, F.M. et al. (2021) Education curriculum transitions in Kenya: an account and progress to competency-based education policy. Curriculum Perspectives, 41(2), pp. 153-162. doi:https://doi.org/10.1007/s41297-021-00137-5

Machisi, E. (2023) Secondary mathematics education in South Africa and Zimbabwe: learning from one another. Contemporary Mathematics and Science Education, 4(1), article ep23007. doi:https://doi.org/10.30935/conmaths/12812

Mahoro, C. et al. (2024) Techniques used by primary teachers when using English as medium of instruction in Rubavu District, Rwanda. African Journal of Empirical Research, 5(1), pp. 445-455. doi:https://doi.org/10.51867/ajernet.5.1.43

M’mboga Akala, B. (2021) Revisiting education reform in Kenya: a case of Competency Based Curriculum (CBC). Social Sciences & Humanities Open, 3(1), article 100107. doi:https://doi.org/10.1016/j.ssaho.2021.100107

Roberts, N., Moloi, Q.M. and Mort, T. (2024) Assessing student teachers’ knowledge of English to inform curriculum design in initial teacher education. South African Journal of Childhood Education, 14(1), article a1538. doi:https://doi.org/10.4102/sajce.v14i1.1538

Shuukwanyama, T.T. et al. (2022) The language of instruction in mathematics teacher education for the early grades. South African Journal of Childhood Education, 12(1), article a1108. doi:https://doi.org/10.4102/sajce.v12i1.1108

Teane, F.M. and Hughes, V. (2024) The school milieu factors contributing to the second-language reading development of learners: primary school teachers’ experiences in Botswana. Per Linguam, 40(2), pp. 24-46. doi:https://doi.org/10.5785/40-2-1141