Diffusion in social norms change about violence against women: A longitudinal analysis of intervention data from a cluster randomised trial.

This study examines the diffusion effects of a Social and Behaviour Change Communication intervention in Nepal targeting gender equity and violence against women. The Change trial involves weekly radio programming, listening and discussion groups (LDGs), and community engagement. This longitudinal study analyses a repeated cross-sectional two-armed, pair-matched, single blinded cluster trial of a 9-month intervention. We used probability proportionate to size methodology to identify 72 wards in the Terai region, half of which were randomly assigned to the intervention. For the community-based survey, 20 women per ward were chosen using simple random sampling (N = 1440). Ten women from each intervention ward (N = 360) were also selected to participate in radio LDGs. Injunctive norms were measured with the Partner Violence Norms Scale-PVNS. Each one person increase in diffusion was associated with a 0.04 (SE = 0.01, p-value < 0.01) higher endline norms score, adjusting for confounders. There was evidence of effect modification with a significant baseline norm by diffusion interaction term (Estimate = -0.12, p-value = 0.04). Findings demonstrated that diffusion was related to endline norms only in communities with lower baseline levels of gender equitable norms. Study findings support the importance of diffusion as a pathway to intervention scale-up and norms change.