Category Archives: Data Analysis

Strong language: what do matric results tell us about language policy in basic education?

27 January 2017

This article first appeared on the Mail&Guardian website.

This article is the second one in the DataEducator project, focusing on the NSC 2015 matric results in the City of Johannesburg metro. It explores the links between the languages spoken by learners at home and the languages that they choose for matric subjects. By overlaying the 2011 Census results in the metro we can see if a language is over- or under-represented in the basic education system.

Language policy has been divisive – even explosive – in South Africa for a long time. There are concerns that indigenous South African languages are not afforded the same prestige as English in education and the workplace. The evidence suggests that these fears have substance.

Home language versus the Queen’s language

Almost 37 000 learners wrote matric in the metro in 2015. The first table looks at their achievements in the official South African languages:

2015 NSC matric results in City of Johannesburg - official languages

2015 NSC matric results in City of Johannesburg – official languages












The metro is famous for being a city of migrants and this is reflected in the range of languages chosen. Nine of the eleven South African languages are taken as home languages (excluding isiNdebele and siSwati) and eight are taken as additional languages (Xitsonga is only offered as a home language).

The most striking observation is that English is taken by almost every learner, either as a home language subject or an additional language subject. English is seen as essential for educational and professional success.

According to Census data, only 20 percent of Johannesburgers speak English as their home language, but 47 percent of learners took English Home Language as a matric subject. Every other language is slightly under-represented as a matric choice: isiZulu is spoken by 23 percent of the metro but written by 22 percent of matrics; Sesotho is spoken by 9 percent and written by 8 percent of matrics; Setswana is spoken by 8% of residents but written by just 6 percent of matrics, and so on.

Afrikaans is the most under-represented language in Johannesburg: 7 percent of people speak it at home but only 2 percent of matriculants wrote Afrikaans Home Language in 2015. It is possible that some of this shortfall is made up by students writing IEB exams, but there is evidence of English Home Language (to the exclusion of English Additional Language) being offered at traditionally Afrikaans high schools.

Afrikaans and isiZulu as additional languages

Afrikaans and isiZulu are the only two languages other than English that are taken as additional languages by a significant number of matrics (40 percent of learners write Afrikaans Additional Language and 6 percent write isiZulu Additional Language).

Many learners appear to take Afrikaans Additional as a default option and outcomes are quite poor, with less than half of all learners achieving over 50 percent for the subject. This compares with almost 70 percent of English Additional Language writers and well over 90 percent of Zulu Additional Language writers.

The outcomes for Afrikaans Home Language are also poor when compared with those of other home language choices: only 76 percent of learners achieve more than 50 percent in the subject, compared with over 90 percent for other home languages. English Home Language is the only other exception to this trend, with just 63 percent of learners who write achieving more than 50 percent as a year mark. This supports the assumption that many learners who take English Home Language do not, in fact, speak English as their mother tongue.

Although far fewer learners take isiZulu as an additional language subject compared with Afrikaans, their outcomes are better. Not only do a far higher percentage of learners achieve at least 50 percent for isiZulu, a whopping 19 percent of learners score at least 80 percent for the subject. No other additional language subject taken by a significant number of learners comes even close to these numbers. The sheer number of learners with distinctions in IsiZulu Additional Language (395) is second only to Afrikaans Additional Language (422).

Assumptions and conclusions

English enjoys a privileged position in basic education. It is overwhelmingly the language of instruction in basic and higher education, and this is true at a national level. There is an argument that English is a more practical and efficient choice in a multi-cultural, multi-lingual city like Johannesburg, but not to the extent seen in the numbers.

Afrikaans is also overrepresented as an additional language choice, although outcomes are poorer than for any other language subject. It’s not clear whether this is due to apathy from learners, a lack of prioritisation from education authorities, or some other factor.

Other official languages remain underrepresented, particularly as additional language choices. Outcomes for IsiZulu Additional Language are very good. It is feasible that the number of learners taking isiZulu will increase in the future. This would be a rational choice by learners: language subjects are all Higher Subject choices and are worth more points for determining entry into higher education. If it is easier to achieve in isiZulu than in Afrikaans, more learners will choose the former.

This assumption doesn’t take into account existing infrastructure (including qualified teachers and good learner materials) for the different languages and the perceptions of learners, schools and parents.

DataEducator, an EDGIS project on public education in collaboration with the Mail & Guardian

My company, EDGIS, is collaborating with the Mail & Guardian newspaper to create DataEducator. an interactive, online tool for basic education in South Africa. You can read more about DataEducator here and look at our first article here.

I’ll repost interesting research and articles to this blog from the Mail & Guardian website. Please support us on social media with the #innovateAFRICA hashtag. Send ideas to this blog and do let me know what you think about everything.

NDP: less easy than 1-2-3

3 February 2016

(Today’s music pun comes to you courtesy of the Jackson 5.)

I bought a print copy of the Business Day today, which I don’t normally do (because free online). There’s a front-page article by Ntsakisi Maswanganyi that talks about the poor state of our economy. The interesting bit for me is the World Bank-cited growth rate that the economy needs if we are to meet our original National Development Plan (NDP) goals – it’s calculated at 7.2 per cent, which I assume is the average rate needed from here on until 2030. Continue reading

If a ward falls in the forest…

25 November 2015

So, about the DA winning Johannesburg and Tshwane. You know, it might just happen.  I don’t think it will, but it could. Vague predictions aside, the DA’s future in the big cities is an important part of the story of municipal government – but it’s not the whole story. In fact, statistically speaking, it’s not even most of the story.

I want to crunch the by-election numbers to look at the DA’s successes and failures over the last four years. Like the ANC, it is more likely to publicise the former and downplay the latter. Continue reading