Effectiveness of Potential Interventions to Change Gendered Social Norms on Prevalence of Intimate Partner Violence in Uganda: a Causal Inference Approach.

Evidence of the effectiveness of programs to change gendered social norms related to intimate partner violence (IPV) is growing, but their potential to significantly impact actual occurrence of IPV at population level is lacking. We study whether modest changes in gendered social norms related to wife-beating can result in significant changes in the incidence of emotional, physical, and sexual IPV among ever married women in Uganda. We employ an imputation-based causal inference approach, based on nationally representative Demographic Health Survey data. The steps are (1) model the association between adjusted neighborhood norms and experiences of IPV using a random effects logistic regression model, (2) impute unobserved counterfactual probabilities of experiencing IPV for each woman while manipulating her neighborhood norms by setting it to different values, (3) average the probabilities across the population, and (4) bootstrap confidence intervals. Results show that statistically significant inverse associations between more prohibitive neighborhood IPV norms and women's experiences of different forms of IPV at the population level exist. The effect is however small, that even if an entire community disapproves of wife-beating, incidence of IPV falls by about 10 percentage points to 48.5% (95% CI 46.0%-50.9%) from the observed value of 57.6% (95% CI 55.2%-59.9%). Furthermore, changes in neighborhood social norms are found to have no statistical significant effect on the incidence of sexual violence. In conclusion, changing gendered social norms related to wife-beating will not result in significant reductions in different forms for IPV at the population level.