Analysing the role of language in the context of education, employment and income in South Africa

Although most of the South African population speak an African language as their home language, English remains the lingua franca and continues to dominate economic and political life. This thesis explores the role of language in the context of the labour market, the schooling system and survey data collection primarily using data from the National Income Dynamics Study (NIDS) survey. Firstly, I assess the impact of English proficiency among African home language speakers on employment probabilities and wages. After accounting for endogeneity and measurement error in the employment and wage models, I estimate that being proficient in English is associated with a 23-25 percentage point increase in the probability of employment, and a wage premium of 33 percent. These results contribute to the limited research on this topic in South Africa. Secondly, I explore the extent to which English and Afrikaans dominate as the Languages of Learning and Teaching (LoL T's) in the foundation phase of schools where a share of the learners speak an African language as their home language. Despite overwhelming evidence that mother-tongue instruction in the early years of schooling is superior pedagogically, parental preferences towards it remain low and I find that most schools in urban areas still choose to teach in English or Afrikaans. Mother-tongue instruction tends to occur in rural areas, or poorer urban communities, and where learners within a school come from relatively similar language backgrounds. Thus, when devising implementation plans to assist schools in adopting mother-tongue instruction, there may be a need for different strategies across different areas and schools. Finally, l investigate the role of language in determining the quality of the data that are used to investigate the phenomena in the first two sections. Through a multilevel analysis of the NIDS survey data, I find that when the interviewer and respondent are from the same language group, the respondent is more likely to participate in the survey. This is an important methodological finding as it implies that matching interviewers and respondents according to language may have a positive impact on survey response and hence the representativity of the data.

Grantor

School of Economics

Year dc:date.issued

2020

Author dc:creator

• Kahn,Amy

Advisors dc:contributor.advisor

• Leibbrandt, Murray
• Branson, Nicola

• economics

Handle dc:identifier.uri

http://hdl.handle.net/11427/37453

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/37453