New data points to solutions for SA’s child literacy woes

We’ve all seen the headlines about how our 10-year-olds can’t read for meaning and how poorly South Africa’s children do compared with those in other countries. But we need to move forward and set reasonable targets so we can clearly see why we have this problem and what can be done to change it, said Dr Nompumelelo Nyathi-Mohohlwane, the director: reading in the Department of Basic Education (DBE).

She was speaking at the first Funda Uphumelele National Survey (Funs), which measures the reading ability of children in grades 1 to 3.

We need to understand what’s below the tip of the iceberg, is the analogy she uses.

That’s what the Department of Basic eEducation has been doing for the past seven years. Teams of linguists and data analysts have painstakingly researched the skills and markers that show how children learn to read in each of the 11 official languages. This helped to establish reading benchmarks for each of the foundation phase grades.

Then the department conducted a survey, which assessed 27,838 learners in 710 public schools across all the provinces and in all 11 languages, and measured the percentage of children who reach the critical reading benchmarks by the end of grades 1, 2 and 3.

The results were released in November, and they are far from fun.

Only about one in three children – an alarmingly low statistic – reached minimum benchmarks for reading set for each of the grades in their home language. This means that “a staggering two million children are not at the minimum required reading level in the foundation phase”, a 2030 Reading Panel summary states.

Worse than that, 15% of children could not read a single word by the end of Grade 3. That means that in a classroom of 40 children, there could be six who cannot read.

These results are shocking, but they shouldn’t really surprise us given previous surveys like the Progress in International Reading and Literacy Study (Pirls), which assesses Grade 4 reading comprehension, and the South African Systemic Evaluation (Sase), which assesses Grade 3s.

What’s different this time, though, is that it is possible to identify which parts of reading are not successful. There are now clear measurable benchmarks for reading in 11 languages and each grade that include skills that can be taught using proven methods.

The context

In grades 1, 2 and 3, which is when children are taught how to read, they can attend a school where they are taught in their home language, but many of them don’t.

Children learn to read most effectively in the language they understand best, which is their home language, said Minister of Basic Education Siviwe Gwarube at the release of the results.

“When you learn to join the letters, i-n-d-l-u together to form a word – indlu – the lights will come on if you know what that word means (a house),” said Gwarube.

Having a solid foundation in their home language made it easier for children to build bilingual proficiency, which for most children in this country meant learning English as a first additional language usually from Grade 4, she said.

Yet English is the most common language of teaching and learning in South Africa. About 30% of the Grade 1, 2 and 3 learners who took part in the survey were assessed in English because of this, even though it’s the home language of only about 10% of them.

IsiZulu is the most commonly spoken home language, followed by IsiXhosa.

What are the targets?

Learning to read is a complex process that culminates in written reading comprehension, but there are a range of subskills that have to be mastered before a child can understand.

For example, a child must be able to produce the correct sound represented by a letter in a particular language. The Grade 1 benchmark is that a child should correctly sound out 40 letters in a minute.

Another example is the ability to accurately read a narrative text – one that is grade and context appropriate – in a particular language. A minimum number of correct words a child should be able to read in a passage per minute (CWPM) was established for grades 2 and 3. These minimums are not the same for all languages.

Learners who read below these fluency rates are unlikely to access higher order reading skills and are unlikely to comprehend what they are reading.

Dr Nyathi-Mohohlwane made it clear why language-specific benchmarks have to be language specific. The text: “There was a stranger who was very hungry. He came to a village and asked for food. Nobody had any food,” is a 21-word text in English, but translates into 11 words in IsiZulu and 33 words in Sepedi.

Learners assessed in English do better

Beyond the disappointing low overall levels of reading achievement, the survey found that learners who were assessed in English were the most likely to reach the reading benchmarks.

But some interesting details surface in the data. In Grade 1, a higher proportion of the children who were assessed in Afrikaans, IsiXhosa and Sesotho reached the benchmark of reading 40 correct letter sounds per minute, than in the other seven languages.

But by Grade 3 the percentage reaching the benchmarks in those three languages had all dropped. In grades 2 and 3 the benchmarks changed from letter sounding to reading a minimum number of words in a minute, so this drop could indicate something about the way these skills are taught in those languages.

On the other hand, a smaller percentage of children assessed in IsiZulu achieved the Grade 1 benchmarks. But by Grade 3 this had turned around and the proportion who reached the benchmarks was higher than all but those assessed in English and Tshivenḓa.

Even though there are advantages in learning to read in one’s home language, many of the children who took part in the survey were not assessed in theirs. Only 33% of those who were assessed in English spoke English at home, the survey found.

Where you live matters

In Gauteng and the Western Cape a higher percentage of learners reached the benchmarks than in any of the other provinces. In Gauteng 64% of schools teach in English. In the Western Cape 40% of the schools teach in English and 38% teach in Afrikaans.

Yet, interestingly, a higher percentage of the Tshivenḓa children who took part in the survey met the reading benchmarks in grades 2 and 3 than any of the other languages apart from English. Tshivenḓa is predominantly spoken in the deep rural areas of Limpopo.

Tshivenḓa is the language of instruction in grades 1, 2 and 3 in only two provinces: Limpopo and Gauteng. Sixteen percent of Limpopo’s schools teach in Tshivenḓa in the foundation phase, according to the Funs technical report. In Gauteng, a tiny 0.4% of schools teach in Tshivenḓa.

Gauteng, Limpopo and Mpumalanga are the most multilingual provinces for grades 1, 2 and 3. However, in Limpopo nearly 60% of schools teach in Sepedi. In Mpumalanga, 27% of schools teach in Siswati and 6% in isiNdebele, while 25% teach in English.

In all the other provinces except the Western Cape one language dominates.

The advantages in Gauteng and the Western Cape are largely attributable to the higher average socioeconomic status of the people living in those provinces, notes the Department of Basic Education in the Funs technical report.

Children in better-off schools do better

Children who attend schools in areas that are economically better off (quintile 5 schools) are roughly twice as likely to reach the reading benchmarks than children in no-fee schools, or quintile 1 to 3 schools, the survey found.

Only 18% of the children who took part in the survey and were assessed in English attended schools in quintile 1-3 (in poorer areas), often called no-fee schools. In the other languages, except Afrikaans, more than 80% of the children were at no-fee schools. Fifty-six percent of Afrikaans children were in no-fee schools.

Gauteng (28%) and the Western Cape (45%) also had by far the lowest proportion of learners in no-fee schools. In the other provinces 70% or more of the learners attended no-fee schools.

Children in better off schools are less likely to be taught in their home language

Less than half of the children in quintile 5 schools (44%) who took part in the reading survey were assessed in their home language compared with 90% of the children from the no-fee schools.

This is because many children from urban and middle-class families attend schools where they are taught in English, according to the Funs technical report.

English-speaking children were the most likely to be assessed in their home language. Ninety-five percent were assessed in their home language, while 72% of IsiZulu-speakers were assessed in their home language and just over half of Sesotho-speaking children were assessed in theirs.

You can see the pattern most clearly in Gauteng, where the lowest percentage of children were in no-fee schools, and only about half of children in the survey were assessed in their home language.

Classrooms in Gauteng often have learners from multiple home-language backgrounds, according to the Funs report. It’s the economic heartland of the country and people migrate there to find work.

Socioeconomic status is clearly reflected in language. But, to a large extent, differences in reading achievement across languages “reflect underlying socioeconomic disparities rather than purely linguistic differences or the quality of provincial administration”, notes the Department of Basic Education.

Many of the children who do not learn in their home language are from affluent families who are able to send them to schools with better resources, and this is why they do better in reading assessments, it says.

Why the new reading benchmarks matter

If the skill of reading is not acquired in the first three grades of schooling the consequences are far reaching. In Grade 4 learners have to move from “learning how to read” to “reading in order to learn” other subjects, like maths.

Learners who do not make this transition swiftly are at risk of performing poorly in later grades and even dropping out of school, says the department. So it’s important to identify weaknesses in early reading skills so they can be addressed as soon as they are discovered.

The new benchmarks establish a definition of reading proficiency that helps teachers identify children who are struggling and identify the remedial support that’s required.

The survey points to a “critical concern”: fewer than half of children reach the reading benchmark by the time they reach Grade 4, which is when they transition from learning in their home language to English.

This shows there is a need to strengthen and extend mother-tongue-based bilingual education beyond the foundation phase, the technical report states. This is what the department is doing, rolling out mother-tongue-based bilingual education in Grade 4 and beyond, according to Gwarube.

The reading benchmarks and the Funda Uphumelele work have set benchmarks that can be used in classrooms, schools and at the provincial and national levels to track progress in foundation phase reading in each language. They can be the guiding lights that can be followed to reach the goal of having every 10 year old able to read for meaning.

As the minister said: “This is the value of good data: it gives us the power to act intelligently, not blindly.” DM

This report and data visualisations were made possible by the support of the Henry Nxumalo Foundation.

First published by The Outlier. For more like this subscribe to the newsletter.