Objective.

To identify predictors of violence against women in high prevalence departments of Peru.

Methods.

Observational, analytical and cross-sectional study. A secondary analysis of the Demographic and Family Health Survey-ENDES 2019 was carried out. The sample was probabilistic, area-based, stratified, multistage and independent in each department studied. A total of 108 cohabiting or married women who responded to the violence module were included. Descriptive analyses and univariate and multivariate logistic regressions were carried out for complex samples, taking into account sociodemographic characteristics as predictors through the calculation of the odds ratio (OR) with its 95% confidence intervals.

Results.

The probability of suffering some type of violence was 3.56 times higher in cohabiting women than in married women, adjusted for having suffered violence in childhood, not belonging to the soup kitchen and not being affiliated to the Cuna Más National Program (ROC area = 72.27%). The probability of suffering physical violence was 2.91 times higher in women with a history of violence in childhood and 3.91 times higher in the case of economic violence; in the case of psychological violence, the probability was 4.30 times higher in cohabiting women than in married women and 4.18 times higher in the case of sexual violence.

Conclusions.

There are predictors such as the fact of being a cohabitant, having suffered violence in childhood and not belonging to social programs, which are predictors of violence against women.