Measuring pathways to youth violence and the possible effects of a sports-based development intervention on youth in South Africa

Doctoral Thesis

2016

Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher

University of Cape Town

License
Series
Abstract
Drawing inspiration from longitudinal-experimental studies of youth violence and intervention in the global North, this study was designed as an ambitious two-site, large-n (n=700), longitudinal (3 waves over 24 months) quasi-experimental panel study to explore pathways to violent behaviour and violence-potential along with possible effects of a sport-based life skills programme among young men and boys. Fieldwork challenges led to a revision of the panel sampling and programming setbacks forced the closure of the second site. Lessons drawn from field research and youth-violence intervention in the urban South African context thus form an important component of the 'lived experience' of this study. These adjustments led to a revised quasi-experimental comparison group design which followed 3181 male subjects in Khayelitsha, 12-23 years-old, over a 12 month period. A unique violence-potential 'scorecard' (comprised of four sub-scales: attitudes towards gang associations, attitudes towards the use of instrumental violence, deviant peer associations, and self-reported fighting) was developed and tested through confirmatory factor analysis and correlation with self-reported violent behaviours and an external assessment (from the primary maternal caregiver). The resulting 'Violence Propensity Score' serves as the primary dependent variable in quantitative analyses. Findings indicate that the theorised risk factors of household deprivation and violent home environment influence parental involvement and, alongside harsh/inconsistent parenting, are significantly associated with a higher Violence Propensity Score, and lower school attitude and attachment, in cross-sectional analysis. In Structural Equation Modelling with longitudinal data, a pathway emerged through which an unstable home environment, influenced by deprivation and violence, affects the quality and consistency of parenting perceived by young male subjects. In turn, early deviant associations and attitudes toward violence and gangs are cultivated and these may have a deleterious effect on schooling and, with this, a subject's orientation toward the future and the present value of investment (of schooling efforts) for delayed gratification. This violence potential and weak school attachment manifests in greater future substance abuse and, in turn, much greater exposure to and acceptance of instrumental violence and criminal associations. Intervention participation in wave 2 did not significantly predict future violence potential, yet rate of participation in wave 3 was cross-sectionally associated with less violence potential. An unanticipated lack of early intervention attendance data and high rates of attrition from the intervention limited the robustness of intervention-related findings. However, area-based police crime data covering the Khayelitsha intervention catchment area showed a potential area effect within an 800m sq. area. Detailed spatial analysis of all crimes in Khayelitsha confirmed the significance of this reduction effect, though it was not unique. Crime incident mapping also revealed implausible concentrations and patterns of crime increase that cast doubt on the veracity and consistency of these data. This study is the first of its kind to quantitatively assess predictors of youth violence and intervention effects in South Africa. It is also the first study internationally to provide confirmatory factor analysis results for a youth violence risk measure for a general population and employ this Violence Propensity Score in longitudinal study. The Violence Propensity Score presented here is designed to be easily implemented and assessed by youth development practitioners (with non-statistical backgrounds), both to target interventions toward higher-risk youth and to easily measure changes in violence-risk over time.
Description
Keywords

Reference:

Collections