Abstract #165  -  Social network characteristics, demographic homophily and implications for peer influence among young South Africans
Authors:   Presenting Author:   DR Jeffrey Grierson - La Trobe University
Additional Authors:  Prof Anthony Smith, Prof Marian Pitts, Mr Vernon Solomon, Prof Graham  Lindegger, Prof Kevin  Durrheim,
Aim: Effective HIV prevention activities that rely on peer influence require an understanding of the structure and constitution of extant peer networks. We report on patterns of homophily in peer networks from a survey of tertiary students at the University of Kwa Zulu Natal Pietermaritzburg campus.
Method / Issue: Students completed a 50 item anonymous intranet based survey including: demographics (age, gender, racial identification, area of residence); social network characteristics (size, age, gender and racial composition); health status and health seeking.
Results / Comments: 589 students completed the survey. The sample was 57% female, aged 18-54 (mean = 23, median =22), and 36% lived on campus. Racially, 61% reported they were Black, 21% White, 12% Indian and 6% mixed race or other. Students had between zero and more than 50 close friends with a median of 3. Peer networks had a mean density of 0.69. Peer networks were strongly characterised by demographic homophily. Young men reported that 72% of their close friends were male and young women that 59% of their close friends were female. Black respondents reported that 93% of their close friends were also Black; White that 89% of close friends were White; and Indian that 84% of close friends were Indian. There was also marked homophily in behaviour. While only 21% of participants reported being drinkers, students who were drinkers reported that 77% of their close friends were also drinkers, whereas non-drinkers reported that only 37% of their close friends were drinkers. Similarly 53% of smokers’ close friends were also smokers compared to 12% of non-smokers’ friends being smokers.
Discussion: Peer networks are not consistent across this population, but cluster according to key socio-demographic characteristics. Prevention work incorporating peer influence should not assume diffusion across young people, but can utilise the specific forms of peer influence in each of these clusters.
Go Back