Sonke CHANGE Trial — Community Mobilisation Against Men's Partner Violence

Diepsloot is an informal, peri-urban settlement north of Johannesburg where baseline research found roughly half of men reported having raped or beaten a partner in the previous 12 months. In response, Sonke Gender Justice, working with the University of the Witwatersrand, designed and tested the CHANGE intervention, combining door-to-door conversations with trained community action teams and six structured workshop themes (gender and power, sexuality and sexual and reproductive health, alcohol and violence, anger management, and the law and healthy relationships) delivered to men over roughly 18 months.

CHANGE was evaluated through a cluster-randomised controlled trial with nine intervention neighbourhoods and nine matched control neighbourhoods, following men aged 18-40 over two years (2016-2018), with results published in 2020 in the peer-reviewed journal Trials.

Physical IPV perpetration declined substantially over the study period in both arms — from roughly 40% to 25% — but the decline was essentially the same in intervention and control communities, meaning the trial found no statistically significant additional benefit attributable to CHANGE itself. The authors reported that the intervention was particularly unable to shift the behaviour of men who used the most severe violence, and concluded the community-mobilisation model had 'limited effect' as implemented.

This is included as an honest cautionary case: a rigorously designed and transparently reported trial that did not confirm its central hypothesis. Its value lies in tempering expectations about door-to-door and workshop-based community mobilisation alone, and in informing the redesign of subsequent men's-engagement programming in South Africa.

Read the full analysis: https://www.whatworks.co.za/resources/item/695-can-community-outreach-reduce-men-s-use-of-partner-violence-findings-from-the-change-trial-in-south-africa