Developing reading assessments in eleven languages to prepare for a nationwide survey

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

Proposal

Following the establishment of reading benchmarks for all eleven official South African languages, the next step was to conduct a nationwide survey—Funda Uphumelele—to provide baseline estimates of the percentages of children reaching key benchmarks by the end of Grades 1, 2, and 3. Preparing for this survey required the development and equating of new grade-appropriate reading passages and assessment instruments across all languages. Because the survey is intended to be repeated over time, careful attention was paid to building assessment banks that would allow for comparability across rounds without over-reliance on the original benchmark assessments.
To achieve accurate measurement and comparability across languages and years, new assessment banks had to meet two criteria: they needed to be of equivalent rigour to the original benchmark instruments, and they needed to be standardised across languages and grades. Assessment tasks generally fell into three categories. The first were tasks sourced from previous surveys that had already been tested and verified. If originally developed in the target language, these were used as is or with minor edits. If developed in another language, they were versioned or translated into the required language. The second category included tasks that existed in draft or untested form. These were reviewed, revised, and adapted by teams of two linguists and foundation phase specialists per language, ensuring they were “pilot ready.” The third category involved tasks that needed to be newly originated.
All instruments were piloted with learners in schools to ensure appropriateness and technical quality. Each item went through two rounds of piloting, with data analysed after each round and feedback provided to the relevant language teams for revision. This iterative process ensured that final instruments were both valid and reliable.
Developing the Letter Sound Knowledge (LSK) chart illustrates the careful design process required even for seemingly simple tasks. The nine African language teams were provided with the Phonics Progression Programme from the revised curriculum, and where available, materials from previous benchmarking exercises. To ensure consistency and comparability across languages, specific guidelines were applied: all three grades were assessed using the same chart; vowels were introduced early; uppercase sounds that differed markedly from lowercase were prioritised; uncommon sounds were delayed; and sounds that did not exist in a language were included but placed toward the end of the chart. These design principles aimed to balance linguistic specificity with cross-language comparability.
Beyond the technical work, collaboration was itself both a good practice and a necessity. Developing credible assessments across eleven languages required cooperation between researchers, linguists, and educators, but it also revealed systemic challenges: limited national capacity, uneven expertise, and gaps in experience in test development. By pooling expertise and building collective processes, the project demonstrates how multilingual assessment design can be advanced in contexts where resources and capacity remain constrained.

Authors

• Lesang Mbali Sebaeng, Department of Basic Education

• Nompumelelo L Mohohlwane, Department of Basic Education

• Asanda Gontse Lobelo, Department of Basic Education

• Zamangwe Zanele Zwane, Department of Basic Education

• STEPHEN G TAYLOR, Department of Basic Education