Socioeconomic factors may have an impact on risk of HIV infection by affecting HIV awareness, sexual behaviour, sexual networks and co-factors for infection. In sub-Saharan Africa, HIV infection rates were associated with increased education, mobility and income in studies completed before 1996. During the 1990s HIV infection rates fell in some countries, particularly among higher socioeconomic status groups. Consequently, some later studies found no association or a negative socioeconomic position gradient in prevalent HIV
infection. Higher levels of condom use have been reported among more educated and wealthier individuals. The social epidemiology of HIV infection among South African young people remains poorly understood. Baseline data from a cluster randomised trial in Sekhukhuneland, South Africa were used to examine the association between socioeconomic factors and HIV risk characteristics among unmarried young people aged 14-25 years in 2001. Participatory wealth ranking generated a sampling frame of 9653 households in eight villages and ranked households by relative wealth into those that were “very poor”, “poor, but a bit better off” or “doing OK”. Household members from a random sample of 200 households in each village were enumerated, including temporary migrants who had permanent links to the rural home but were not currently sleeping there. There were 2486 eligible individuals from 1482 households, of whom 916 males (73.6% of those eligible) and 1003 females (80.8%) were included in the study sample. Of these, 763 males and 880 females had data available on HIV infection. School attendance was high overall and associated with greater household wealth among women. Most of those not attending school were not in regular employment. Temporary migration was increasingly common with greater age but was not associated with household wealth. Low HIV awareness and risky sexual behaviour characteristics, including early age of sexual debut, lack of condom use, multiple partners for men and partnerships between older men and younger women, were widespread. HIV prevalence was 5.6% among males and 12.2% among females. Household wealth was not associated with HIV risk characteristics except that young women from wealthier households more often reported condom use. Young women and men attending school reported lower numbers of sexual partners than those not attending. Those attending school of both sexes reported recent partnerships involving a lower frequency of sexual intercourse. Young women attending school reported more partnerships with men of a similar age to themselves and higher levels of condom use. Young men attending school were less likely to be HIV infected. Temporary migrants of both sexes reported greater levels of HIV awareness than those staying at home, while female migrants also reported more condom use. In this study, young people in all social groups were at high risk of HIV infection. Those who do not attend school and those remaining in rural areas may be at greatest risk of new infection. Programmes that reach these young people and those that increase opportunities for staying in education might help reduce HIV infection rates.