Effective scaling up in education requires evidence, planning and collective action

Joy Olivier is a research consultant to The Learning Trust

“We can navigate the tough challenges in the education sector through strong cross-sector partnerships, competent leadership, innovative interventions and smart resource allocation…”

 

Scaling up education initiatives requires a collaborative network of partners

The social sector’s obsession with scale makes sense, given that we want to see big, positive change. However, the reality is that only the public sector can fund at the scale required for real social change. With the relatively tiny ticket sizes of philanthropic and corporate giving, the proposed solution currently touted in grant-making and research circles is to use private funding to test models, measure impact, produce evidence and then hand over to the Government to implement at scale.

This argument holds water when it comes to health; for example, when the problem is HIV/AIDS, and the solution is anti-retrovirals. Or when the problem is Malaria, and the solution is insecticide-treated bed nets. Unfortunately, when it comes to education, there are not any silver-bullet solutions. There is no one simple thing that we can provide which results in, for example, learners being able to read for meaning.

For example, it stands to reason that the continuous training of teachers may lead to improved learning gains for more children. Yet, there is very little evidence that this produces gains, unless teachers receive coaching – which produces small gains, or are assigned carefully recruited and trained teaching assistants – which produces larger gains. While randomised control trials (RCTs) are the gold standard of evidence of impact, they can only tell us whether or not something has worked, and not definitively why. For this, we need qualitative research, and to learn from what is happening on the ground.

Understanding what shifts the needle

South Africa’s education system as a whole was making small but steady gains up until 2019 (when the Trends in International Mathematics and Science Study / TIMSS shows a flatlining on improvement). Then the disruption to schooling caused by Covid-19 wrecked devastating learning loss; the average 10-year-old in 2022 knows less than the average 9-year-old in 2021.

In the last few years, there have been some rigorous studies that help to build our understanding of what potentially does and does not shift the needle on improved learning gains. Much of what has been learned about literacy and numeracy in recent years is discussed in three volumes soon to be published by Oxford University Press. The third volume focuses on interventions, and a typography for these was presented at the 2022  RESEP conference. One of the conclusions from this presentation was that the South African education system needs fewer interventions that are neither at scale nor generating evidence.

Scaling up in a fragmented, under-resourced and under-researched system

Given the sector fragmentation, duplication of effort, limited resources, and lack of research, I can understand where this argument is coming from. However, I would like to table two alternative recommendations:

i. Begin with less-expensive research strategies to ascertain what is worth testing with gold standard research methods

It is important to innovate, implement, iterate, and then implement some more at ever-increasing sizes and in different locations to find out what works and what does not, before moving to the phase of rigorous testing to generate evidence. Randomised control trials (RCTs) are expensive; not only in monetary terms, but also in how they affect implementation, school operations and children on the ground.

Given how much is happening in education at the moment, in order to achieve a proper control group for an RCT, children would need to be “quarantined” from any other interventions. This has worrying ethical implications, especially in the wake of learning loss.

Before adopting gold standard methodologies, let us learn through good quality programme monitoring, thoughtful action research methods, and quasi-control groups. And then, when we think we are really onto something that is showing impact and have political support for wide-scale implementation – and the only way to unlock the requisite public funding is with rigorous evidence – then by all means let us go for gold.

ii. Once we have found the models that show statistically significant impact at a low-enough cost for scale, we are going to need many organisations to implement them.

Rather than eradicating community-based organisations not operating at scale, let us capacitate them to implement evidence-based models.

Given the widely accepted strain under which the schooling system is operating, the long-standing dysfunctionality in many schools, and the fast-building tide of teacher retirements, it is unlikely that interventions can be implemented by the state alone.

And so, we look to civil society and low-cost education interventions. Perennially underfunded, somewhat fragmented, and lacking in expensive evidence, this sector is nonetheless the best bet we have got to partner with the Government for wide-scale implementation in no-fee schools.

Whether it is coaching teachers, training teaching assistants, providing tutors for catch-up programmes, or leveraging ed tech to improve literacy and numeracy, we need to identify the models that work, and partner with Government and civil society to get them implemented.

There are very few education NGOs in South Africa with the bandwidth to scale exponentially wider than their current ambit. Those that have this muscle need to establish the impact of their interventions with independent evaluations and should also look at implementing other models with proven impact. However, the pressure on organisations to quickly demonstrate impact at scale can put them at risk; it takes time to consolidate, attract talent and build the systems and processes to handle large-scale implementation.

We need a coalition of community-based organisations with strong relationships with children, families, and schools, and passionate and committed education and community development practitioners. Rather than writing them off because they are not operating at scale or conducting research, we should:

  • map and coordinate the sector;
  • reduce duplication of effort and streamline resource allocation;
  • build organisational capacity;
  • share approaches that have proven impact, and avail materials, tools, and training to implement them;
  • support organisations to grow their programme delivery teams by tapping the public employment stimulus; and
  • develop tools and systems for sector-wide monitoring, evaluation, and learning (MEL).

Interventions can make an important contribution towards system strengthening

 Some funders are reticent to support interventions that help particular learners but do not address the route systemic issues that manifest the need for ex-state intervention; they would rather invest in strengthening the system, than on interventions in particular grades and subjects. This was described by a presenter at the RESEP conference as a false dichotomy. If we implement at scale and sustain these efforts over time, and if they are aligned with curriculum and accountability goals, then interventions can make a valuable contribution towards system strengthening.

When it comes to implementing at scale, I propose that rather than looking to a small group of well-resourced organisations to tackle learning loss and poor learning outcomes for the entire country, we need to be realistic about what it is going to take to implement at scale. An asset-based approach (map, coordinate, leverage and improve on what already exists), as well as context-responsive, flexible, adaptive, and community-based interventions are needed. We can strengthen the system by strengthening these important players in the education ecosystem.

The learning losses and psychosocial impacts of the pandemic will be with us for many years, and the system will be further impacted by teacher retirements and climate crises. The coming years are going to be tough. However, with strong cross-sector partnerships, competent leadership, innovative interventions and smart resource allocation, the education sector has an opportunity to turn things around. We have no choice but to really go for it.


The research underpinning this article includes the following resources:

Ardington, C. and Henry, J. (2021), Impact Evaluation of Funda Wande Teacher Assistant Intervention in Limpopo Province. SALDRU, University of Cape Town, Cape Town. Read it here.

Fleisch, B. & Dixon, K. (2019). Identifying mechanisms of change in the Early Grade Reading Study in South Africa. South African Journal of Education, 39 (3) p.1-12. Read it here.

Reddy. V. Winnaar, L., Juan, A., Arends, F., Harvey, J., Hannan, S., Namome, C. & Zulu, N. (2020). TIMSS 2019: Highlights of South African Grade 5 Results in Mathematics and Science: Achievement and Achievement Gaps. Read it here.

Spaull. N. & Taylor, S. (forthcoming). A Typology of Reading and Mathematics Interventions in South Africa. In Spaull & Taylor (Eds) Large scale interventions to improve early reading and mathematics in South Africa. Oxford University Press.

Leave a Reply