Prevalence and correlates of partner violence among South African adolescents

BACKGROUND: Little is known about the prevalence of partner violence among adolescents, nor of the factors with which it is associated. The objectives of this study were to document prevalence rates for partner violence among high school students in Cape Town, and to explore factors that are associated with such violence. METHOD: The sample consisted of 596 Grade 8 and 11 students attending public high schools in Cape Town, who were selected using a multistage cluster design. They completed an anonymous and confidential questionnaire. The dependent variable was whether they had ever perpetrated partner violence, or intended to do so. The independent variables, which were derived from a modified version of the Theory of Planned Behaviour (TPB), included attitudes, subjective norms (general and peer social influence and outcome expectancy) and self-efficacy. Separate multiple logistic regression models were developed of partner violence intentions or behaviours on the scales derived from the TPB and demographic variables. Regression coefficients from adjusted models were used to examine the potential mediating role of partner violence intentions in the association between each scale and partner violence behaviours using the Sobel test. RESULTS: Among participants who reported being in a relationship, 20.7% reported perpetrating partner violence, and 16.4% reported intending to do so. After adjusting for demographic characteristics and other predictor scales, perpetration of partner violence was significantly associated with attitudes and outcome expectancy, while intention to perpetrate partner violence was significantly associated with attitudes and general social influence. The influence of attitude and general social influence on violent behaviours were each partially mediated by partner violence intentions. CONCLUSIONS: There are high levels of partner violence among high school students in Cape Town. Interventions to address this problem should focus on violence-related attitudes and outcome expectancy.