The development of reading benchmarks in South Africa’s eleven languages

• In Event: Measuring reading benchmarks across linguistic divides: The Funda Uphumelele National Reading Survey of South Africa

Proposal

South Africa is a multilingual context with eleven official spoken languages. These languages differ markedly in orthography—spelling and writing conventions—and in morphology—the way words are constructed and related to each other. Such linguistic differences strongly shape the relationship between components of reading, the strategies learners employ, and the pacing that is appropriate for reading development. This paper describes the development of reading fluency benchmarks that are sensitive to this multilingual and linguistically diverse context. While reading norms exist for English (Hasbrouck & Tindal 2006), until recently there was little research to guide the development of benchmarks for African languages.
We argue for the necessity of language-specific benchmarks that recognise not only linguistic features, but also the contextual realities of South African schooling. Establishing distinct accuracy–speed and fluency–comprehension relationships for each language reflects the complexity of reading development in this setting. At the same time, the process requires a balance between technical rigour and educational utility, given the low and unequal levels of learning. The failure to adapt to learning realities has resulted in the setting of unattainable benchmarks in several countries, including Ethiopia, Ghana, Kenya, Liberia, Malawi, Rwanda, Tanzania, and Zambia where only 5% of learner reach the benchmarks (RTI International 2017), severely limiting their usefulness for tracking incremental progress or guiding classroom practice.
The paper presents three approaches used across South African languages, each with different levels of efficiency, cost, and effort. The first approach relied on pre-existing datasets and was relatively inexpensive; the second involved modest additions to ongoing data collections; and the most resource-intensive required dedicated new data collection. Across these approaches, challenges emerged with established methods for setting benchmarks (Jukes et al. 2018a; 2018b). Existing approaches, largely based on Early Grade Reading Assessments (EGRAs), often employ a one-minute reading time limit, which restricts comprehension questions to those parts of a passage read in that short window. In low-literacy contexts, this tends to inflate the observed relationship between fluency and comprehension and sets benchmarks unrealistically high, limiting their diagnostic value.
Our analysis began with an exploration of the relationship between accuracy and speed in reading, and then how fluency interacts with comprehension. In English, accuracy in recognising letters, syllables, and words typically develops first, followed by rapid increases in reading rate as decoding becomes automatic, rapid and effortless (Fuchs et al. 2001; Spear-Swerling 2006). Yet such trajectories are highly dependent on language typology. Accuracy–speed and fluency–comprehension relationships have been insufficiently studied in African languages. To avoid imposing assumptions from English or other languages, we adopted non-parametric techniques to identify empirical regularities across South Africa’s eleven languages.
By grounding reading benchmarks in the linguistic and educational realities of each language, this work aims to improve measurement and support more equitable opportunities for learning in a divided society. In contexts where multilingualism often reflects deep historical and social divisions, developing valid, context-sensitive benchmarks is one step toward ensuring that all children are given a fair chance to read with comprehension.

Authors

• Nompumelelo L Mohohlwane, Department of Basic Education

• Lesang Mbali Sebaeng, Department of Basic Education

• Zamangwe Zanele Zwane, Department of Basic Education

• STEPHEN G TAYLOR, Department of Basic Education

• Asanda Gontse Lobelo, Department of Basic Education