Sur ce site

Accueil du site > Equipes de Recherche > Mortalité liée au sida et caractéristiques des ménages en Afrique du Sud rurale : implications pour l’utilisation des ressources naturelles et le développement > Methodology

Household characteristics in rural South Africa : implications for natural resources and development


The Agincourt sub-district of the Bushbuckridge is located in the region of the Limpopo Province, South Africa. The Agincourt sub-district, named after one of the local villages, consists of 21 villages, comprising over 12 000 households and 70 000 people. This is the demographic surveillance site of the Agincourt Health and Population Unit (AHPU) of the University of the Witwatersrand. The area is typical of rural communities across South Africa, and is characterised by poverty, high human densities, and a high reliance on both natural resources and on remittances from a large migrant population. Due to poor employment opportunities in the region, a large proportion of adults are migrant labourers, working on commercial farms and in the towns and cities across the country. A significant proportion of households depend on the state pension of an elderly resident as the only reliable source of household income. Sero-prevalence of AIDS is around 18% in the region.

To answer the above mentioned questions, we made use of three data sources :

1) The Ongoing Demographic Surveillance System : the AHPU has collected since 1992 census data at 12-18 month intervals from all 12,000 households in the Agincourt sub-district. The resulting data of particular interest in the present project is household economic status.

2) Natural Resource and Development Survey : the resource use patterns of Agincourt residents were approached through a survey of a random, stratified sample of 248 households.

3) Qualitative Interviews : we undertook in-depth qualitative interviews with 30 households drawn randomly from the sample of 124 households.

Survey fieldwork started in early May 2004. The survey questionnaire was administered by a team of four experienced fieldworkers from the AHPU, supervised be a senior fieldworker and the investigators. The fieldworkers and field supervisor were from the study area and spoke the local language fluently. Households were thus interviewed in their mother tongue. Data collection took approximately six weeks. The 30 qualitative interviews were conducted towards the end of the field campaign, after most households had been surveyed. The researchers were assisted by an interpreter who translated questions and responses. Interviews were recorded using a dictaphone. This data collection took approximately two weeks. Analysis of the data was done by the principal investigators and assisted by graduate students. The survey data were analysed using descriptive statistics and multivariate logistic and ordinary least squares regression models.