Lost in Space: Residential Sampling and Johannesburg's Forced Migrants

The study of displaced peoples regularly compels analysts to reconsider the relationship between people and space. After several decades of critical scholarship, most research on forced migration begins, at the very least, with an awareness of the difficulties involved in representing mobile populations through the myopic and deeply territorialized lens of the nation-state. However, the silences and dissonances encountered in our attempts to capture the nature of migrant lives and experiences often compel us to go further, to acknowledge the statist and sedentary character of our own ways of knowing. This paper takes us through an example of this brand of epistemological tension, narrating and reflecting on residential sampling problems in the African Cities Project (ACP): a survey of refugees in inner city Johannesburg, Maputo, Nairobi and Lubumbashi. Surveys of marginal and dispersed populations utilise various forms of spatial classification to generate representative samples. While the problems associated with tracing people ‘on the move’ are widely accepted, this survey began with the assumption that once displaced populations had settled, they could be a) adequately categorized according to residential sampling ‘frames’ (addresses and area codes) and; b) readily accessed in their homes. This paper explores the theoretical and practical limits of this premise. It shows how, under the combined strain of an unknown urban landscape, the unique blends of public/private space in African cities, and respondents’ feelings of insecurity in their new environment, a variety of tensions emerged between spatial boundaries and sampling techniques.