Language is what makes us uniquely human. It “is the house of the being” (Heidegger, 2000, p. 83) and it “is the dress of thought” (Johnson, 1905, p. 67). It is also the vehicle that carries the content of life-giving education for disadvantaged learners in South Africa (SA).
The research was inspired by this, and led to investigating the manner in which the use of language during lessons has affected the extent to which learners within disadvantaged bilingual contexts of primary schooling have been able to learn.
The key issue investigated was the differential impact of the language of learning and teaching (LoLT) on learner performance in relation to learners’ first language (L1) compared to their second language (L2) as measured by the Western Cape Education Department (WCED) Systemic Testing results. The context within which this was investigated is discussed below.
In SA, despite the fact that government allocated a greater percentage (19,3%) of its total expenditure to education than the global average for middle-income countries (15,7%) between 1997–2014 (World Bank, 2016), the learner performance of the country’s education system during this period was of the poorest in the world (Spaull, 2013, p. 3).
While the definite cause of this is subject to debate, the choice of LoLT policy in South African (SA) schools was a significant causal factor (Pinnock & Vijayakumar, 2009, p. 16). This highlights the importance of evaluating the available LoLT policy options for that which is most likely to facilitate optimal performance, raising the question: How does LoLT affect learner performance? This was the research problem.
Responding to this problem was the reason for conducting the research. By answering this key question, the research aimed to contribute to filling gaps in the knowledge commonly held by LoLT policy role players on the topic.
For informed decisions to be made on which LoLT policy is most likely to facilitate learner performance within the context of interest required further understanding of how LoLT affected learning within that context.
The purpose of the research was to find out how LoLT policy, in the form of LoLT models, affected learner performance. In so doing, the research aimed to contribute to a LoLT discussion in SA by producing knowledge that would shed light on how LoLT policy was typically formulated, implemented, and performed within schools. It was hoped that such knowledge could then be used to guide the formulation as well as implementation of LoLT policy in the manner that served the educational interests of disadvantaged learners in SA, optimally. The value of responding to the research problem was significant because of the potential for the research to achieve this.
As is the case in SA, “in multilingual societies, the choice of language of instruction and language policy in schools is critical for effective learning” (Education For All (EFA), 2005, p. 160). It is of crucial importance for a LoLT policy which is appropriate for use in the context of implementation to be determined if effective learning and quality education are to be achieved. The research problem – how does LoLT affect learner performance – was significant, because informed and appropriate LoLT policy choices could be made using knowledge gathered by responding to the problem.
By responding to the problem within the context of interest, the research provided an indication of the model of LoLT that was evidently optimal for facilitating teaching and learning of maximal quality within a system of schooling that required learners to write school-leaving examinations in a L2 – as was the case for the vast majority of disadvantaged SA learners. This indication is critical if schools within such contexts are to make informed decisions when formulating LoLT policy that will facilitate the education of learners to their best long-term advantage.
Additionally, by responding to the problem, the research also generated knowledge of factors surrounding LoLT policy implementation that affected the extent to which it supported quality teaching and learning .This knowledge holds significant value in that it indicates how such factors can be best managed for optimal teaching and learning as an outcome of using the LoLT.
It was necessary for several problem-related questions to be asked in order for the research to provide a comprehensive response to the problem.
These questions are presented below. Together, they provided the structure within which the research was conducted. As the research engaged with each question, a questioning process that operated in a systematic self-reflexive cycle was employed in order to allow the development of new questions as the research progressed.
The purpose of the first set of questions was to find out what options disadvantaged primary schools in SA had when adopting LoLT policy models, as well as to determine the extent to which each option was educationally sound. The answers to these questions were found by reviewing the relevant literature and by visiting schools:
· To what extent did relevant evidence-based research on best practice manifest in LoLT policy at schools?
o What were the LoLT policy options available to schools?
o How did language in education experts view each particular LoLT policy option?
The second set of questions was answered by exploring whether a pattern existed between LoLT policy and learner performance:
· What effect did LoLT policy have on learner performance?
o More specifically, what was the relationship between LoLT policy and learner performance during Grade 3 and 6 WCED Systemic Tests in disadvantaged schools?
Lastly, the third set of questions investigated the nature of how LoLT policy materialised in schools according to role players involved in its implementation:
· How did role players in schools say LoLT policy was implemented?
o What reasons did role players give for having decided upon the LoLT policy at their schools?
o What theories of language had role players drawn upon when they made LoLT decisions?
o What language in education legislation had role players consulted when they made LoLT policy decisions, if any?
o How did role players regard the way that LoLT policy was implemented at their schools?
§ Did role players regard the LoLT policy at their schools as having been successfully implemented or not?
§ What problems or benefits had been experienced by role players in relation to the LoLT policy at their schools?
§ How was the designated LoLT policy reflected by classroom practices, according to role players?
These questions were answered by the research in order to provide a comprehensive response to the research problem by employing the following research design.
The research used a two-part mixed-methods design.
The first part (Part A) provided a quantitative indication of the learner performance associated with the implementation of two different LoLT models. These were the early-exit transitional LoLT model, which employed L1 as LoLT until Grade 4, where a transition to L2 LoLT occurs; and the straight-for-English LoLT model, which made use of a L2 LoLT from the start of primary schooling and throughout all grades thereafter. The performance of learners in schools following each LoLT model within a homogeneous sample was compared cross-sectionally as well as longitudinally using data from WCED Systemic Testing of Language and Mathematics. The findings from these comparisons were described statistically and tested for statistical significance.
The second part of the research (Part B) provided a qualitative assessment of how the two different LoLT models were typically formulated and implemented. To do this, a comparable sample comprised of several schools following each LoLT model was selected from those that were included within Part A. Teachers and members of school management from these schools were surveyed using questionnaires made up of comparative rating scale response items as well as interviewed using a standardised open-ended interview schedule. The data collected was captured, analysed, and presented qualitatively.
Several assumptions were inherent to the nature of the process described above.
The research assumed that the learner performance data used was accurate, and resulted from the use of reliable and valid instrumentation administered and captured under controlled circumstances. It was further assumed that participants engaged with surveys and interviews in a manner that was truthful and reliable.
The research was potentially limited by the extent to which it could be generalised owing to the small size of some data sets, as a result of scarcity. Factors such as perceptual misrepresentations and researcher bias add further potential limitations to the validity of the research.
In terms of scope, Part A used data sets reflecting the learner performance of an average of 128 schools/14 714 learners within disadvantaged contexts of the Western Cape, SA; while within the same context, Part B visited 5 schools and surveyed 11 teachers, as well as interviewed 14 teachers, members of school management.
Chapter 2 sets out the legislative, empirical and theoretical framework within which the investigation was located. A detailed account of the design and methodology employed follows in Chapter 3. Findings from the research are then presented in Chapter 4. Chapter 5 concludes the report. This chapter contains conclusions from the research as well as recommendations for the formulation and implementation of LoLT policy most likely to result in disadvantaged SA learners’ quality of education and brighter future.