A presentation of the minor dissertation* (full text available here) submitted in partial fulfillment of the requirements for the award of the degree of Master of Education in Educational Administration, Planning & Social Policy.

Faculty of Humanities, University of Cape Town, 2017

*Passed with distinction as recommended by Prof Azeem Badroodien (UCT) & Prof Peter Plüddemann (UWC)

Keywords

LoLT | medium of instruction | language-in-education policy | language-in-education legislation | language barrier | disadvantaged learners | disadvantaged schools | learner performance | language policy | English as medium of instruction | learning in South Africa | isiXhosa as medium of instruction | language policy implementation | code switching | code mixing | LoLT transition at Grade 4 | bilingualism | multilingualism | straight-for-English approach | early-exit models | transitional models | late-exit models | parental involvement | cultural hegemony | literacy in South Africa | numeracy in South Africa | language of learning and teaching

Abstract

The research investigated the differential effect of language of teaching and learning (LoLT) on learner performance in relation to learners’ first language (L1) (isiXhosa) compared to their second language (L2) (English) by researching a sample of comparable primary schools that followed one of two LoLT models; either:

· An early-exit transitional LoLT model, which employed isiXhosa as LoLT until Grade 4, where a transition to English LoLT occurred; or

· A straight-for-English LoLT model, which made use of English as LoLT from Grade 1 and throughout all grades.

The research was conducted in two parts, A and B. Part A indicated which LoLT model was associated with higher learner performance, and Part B determined the extent to which Part A’s findings reflected LoLT model implementation, as well as the ways in which this implementation affected performance.

Part A quantitatively compared the performance of 14 714 (average dataset size) learners in 2012 and 2015 Western Cape Education Department (WCED) Systemic Testing of Language and Mathematics from 135 (average dataset size) National Quintile (NQ) 1–3 schools which followed different LoLT models cross-sectionally and longitudinally. Comparisons drawn by the research were tested for statistical significance. Part A found that:

· Grade 3 learners in L1 LoLT contexts performed better than Grade 3 learners in L2 LoLT contexts.

· The performance of learners in the early-exit transitional model of LoLT declined after the point of transition to L2.

· Learners who had had more years of exposure to L2 LoLT performed better in Grade 6 when taught and assessed in the language than learners who had had fewer years’ exposure to the language as LoLT.

Part B qualitatively studied a sample of five schools from Part A identified from each of the LoLT models, visiting them to investigate how LoLT model implementation affected learner performance. 11 teachers were surveyed using comparative rating scale response item questionnaires and 14 teachers as well as members of school management were interviewed using structured interview schedules. Part B found that:

· Particularly where English was LoLT, the greatest challenge was teachers’ and learners’ inability to use the LoLT in order to teach and learn.

· Teachers and learners typically employed mixed-language-use practices to communicate during lessons that deviated from the LoLT, constituting LoLT non-implementation to the extent that LoLT descriptors were inaccurate.

· Schools designated as following a straight-for-English model were unintentionally implementing a rough dual-medium (two-way immersion) LoLT model which outperformed schools following an early-exit transitional LoLT model by Grade 6.

· The transition of LoLT from L1 to L2 at Grade 4 in early-exit transitional models was non-conducive to educational outcomes from the point of transition onwards.

The research concluded that a dual-medium late-exit transitional LoLT model with gradually increasing L2 LoLT/gradually decreasing L1 LoLT use from Grade 1 up to the point of transition should be advocated for within disadvantaged South African contexts of schooling. From the point of complete transition, the L2 should be used as LoLT for the rest of schooling, while the L1 should be maintained, being taught as a subject.