The objective of the research is two-fold. Firstly, the research aims to arrive at a meaningful estimate of child labour in South Africa, and secondly, to establish a link between child labour and scholastic retardation. To establish an understanding of the turf, I take the reader through a detailed analysis of why children work, where they work and whom they work for. The study looks at the problems that have defined child labour for many decades and the steps taken both internationally and locally to enhance the efforts for its elimination. It looks at how, internationally, the campaign against child labour has shifted from children engaged in economic activity, to children engaged in hazardous work and the Worst Forms of child labour, which involves the economic exploitation of children by adults, through child prostitution, pornography, elicit trade, armed conflict etc. The definitional problems that have plagued the estimation of child labour in South Africa are reviewed, and I suggest specific approaches to measurement and estimation of child labour in future. I discuss the pertinent issues that need to be addressed to define child labour in South Africa, and I use the 1999 Survey of Activities of Young People (SAYP) to develop a conceptual framework of estimating child labour in South Africa. This is against the backdrop of the apparent disagreement between the main role-players, on the estimated levels of child labour in the country. I apply my model to the SAYP data set, and I estimate child labour by isolating all children in hazardous work, either by working conditions or environment, effect to child’s health and child’s schooling or by the number of hours for which they worked. I am very aware and mindful of the overwhelming need for children to work, among many South African households, simply for household sustenance. I therefore use the concept of long-hour cut offs, for different age groups of children to clearly establish the difference between ‘unacceptable’ child labour and ‘acceptable’ child work. To obtain the second objective of the study - establishing a link between scholastic progression and child labour, I focus on children who were attending school at the time of the survey, in the households under investigation; and, I choose to use the ‘grade-specific scholastic retardation rate’ as the appropriate measure of scholastic progression. By introducing gender as one of the determinants, I construct nine, different but not necessarily mutually exclusive groups of children with apparent variation in the intensity of the child labour characteristic. Then, among the children in each of the nine groups, I calculate grade-specific scholastic retardation rates (SRR) for children who were enrolled in grades 1 to 6 at the time of the survey. I am then able to graphically compare the SRR for the nine different groups, and graphically demonstrate that there is a link between child labour and scholastic retardation. The results of the research show that children in child labour tend to be more scholastically retarded than those who are not engaged in child labour, and that child labour seems to have more adverse effects on boys than girls enrolled in the same grades.