Creating a Frame: Experimenting with Random Sampling within Non-Homogeneously Distributed Urban Migrant Populations

Gayatri Singh, Brown University
Benjamin D Clark, University of the Witwatersrand
Kennedy N Otwombe, University of the Witwatersrand

This paper will present key elements of a random sampling methodology utilized in a recently concluded survey in Johannesburg (South Africa) with four multinational migrant communities namely, Zimbabwean, Congolese (DRC), Somali and rural-urban South African migrants. The study made use of a mixed methodology approach involving qualitative and spatial methods to ‘create’ a sampling frame for the final survey phase. The need for using these methods to generate a creative sampling frame arose from the lack of baseline household distribution information for these population groups and unique terrain of the survey site, inner city Johannesburg, where both foreign and South African migrants often inhabit dilapidated buildings that are difficult to access. A full enumeration of the method, the data collection strategies employed and challenges faced in the data collection will be provided in this paper that argues the feasibility of random sampling approaches within a non-homogeneously distributed population context.