Before detailing the research questions addressed by this thesis, it is critical to clearly define learner mobility. The term ―learner mobility‖ is used to refer to the daily travel of learners to a school that is not the school nearest to their home (Karlsson 2007). It is derived from the phrase ―learner migration‖, which has previously been used in the South African literature on the travel patterns of learners (Sekete, Shilubane et al. 2001; Lombard 2007). This term, however, has not obtained widespread usage outside of this fairly limited literature. In addition, it does not adequately differentiate between daily travel or commuting, and genuine migration in which learners spend nights away from their permanent family residence for the purpose of attending a specific school.
As both of these practices are believed to be widespread in contemporary South Africa, it is essential to distinguish between them, particularly as their implications for individual learners, the educational system and society more broadly, along with appropriate policy responses, are likely to differ substantially. In addition, the learners making use of them are expected to represent different groups, particularly with regard to residential location and family socio-economic status (SES) (Paterson and Kruss 1998). As daily travel is the focus of this dissertation, learner mobility is used as a more appropriate descriptor.
It should be stressed that learner mobility is only one expression of school choice in South Africa. Other, and likely widespread, expressions of choice
4
occur when a family chooses their residence on the basis of proximity to specific schools, when a child leaves the public sector to attend a private school, regardless of location, or when a child lives with people other than his or her family in order to attend a specific school. These types of choice are excluded from the above definition of mobility for various reasons, including that it is not possible to measure them with the data available for this project, and that they can be understood to operate differently from the types of movement described above in terms of the resources required and used, the perceived benefits, the population groups most likely to be taking advantage of them, and their implications for public policy. While all these forms of school choice in South Africa certainly warrant further examination, they do not fall within the scope of this thesis.
This thesis takes several different approaches to the measurement of learner mobility, in order to capture a range of different aspects of the phenomenon.
The first approach involves measuring the straight-line distance between a child‘s home, and his or her school. This is selected because it is theoretically sound, capturing to some extent the level of investment required by mobility, and because it can be accurately measured using the available data.
Additionally, and for these same reasons, it has been widely used in existing work on learner mobility, although in some cases distance is supplemented or replaced by travel time (Sekete, Shilubane et al. 2001; South African Human Rights Commission 2004; South African Human Rights Commission 2006;
Pendlebury and Rudolph 2008).
The second approach involves determining whether the learner is travelling to a school inside or outside of the area in which he or she lives. Unfortunately, identifying in a consistent and useful way whether a school and a home address are in fact located in the same area is very challenging in contemporary South Africa. Educational districts, as used by the Gauteng Department of Education (GDE) for administrative purposes, do not generally align with electoral wards,
5
census districts, or neighbourhoods. Most schools do not have a clearly defined catchment area, unless it has been defined by the school itself. Electoral wards and census districts have also undergone recent and substantial changes, and in many cases bear limited resemblance to historically defined neighbourhoods.
In this context, the most feasible way to measure travel between areas proved to be by locating both school and home within their relevant census districts, using GIS coordinates. Although imperfect, this data does at least provide a preliminary measure of whether a child attends school in the area in which he or she lives.
Finally, the third approach to measuring mobility is to identify whether or not a child is attending the grade-appropriate school nearest to his or her home. This is used primarily as an indicator of whether or not a child and his or her family are engaging in school choice at all. Of course, this measure is imperfect, as a child attending the school closest to his or her home may have chosen this school deliberately, while a child attending a school further from home may do so for completely involuntary reasons. Nonetheless, this measure is used as the best available indicator to provide an approximate measure of the extent of engagement in school choice amongst children in post-Apartheid Johannesburg-Soweto.
It is critical to note that many other definitions of learner mobility are of course possible, and in various contexts may indeed be more appropriate. For example, when parents and learners make school enrolment choices, they certainly take into account a wide range of factors other than distance. Given South Africa‘s long and racially-defined history of learner migration and differential educational opportunities, the historical racial designations of schools is one factor that is likely to play a significant role (Paterson and Kruss 1998; Lombard 2007). Factors such as travel time, travel cost, safety, school reputation, and school fees, among others, are also likely to be of particular importance (Chisholm 2004; Maile 2004; Lemon 2005).
6
The measures of learner mobility used in this thesis, and discussed above, include both a continuous measure (distance from home to school), and binary measures (attending a school in the area in which a child lives, and attending the school nearest to the child‘s home). This is appealing, as mobility can be thought of as both binary – occurring or not occurring – and as a continuum – with the extent of mobility determined by the distance travelled to school. In addition, combining both binary and continuous measures allows for the application of a broader range of different analytical approaches.