List experiment and misreporting of sexual violence in Brazil

Abstract

This paper aims to analyze the potential omission and misreporting from women victims of sexual violence committed by perpetrators other than the intimate partner or ex-partner. For this, the List Experiment method has been applied to a sample of 4,746 women who participated in the PCSVDF-Mulher in 2019 along seven Brazilian capitals (Fortaleza, Recife, Salvador, Porto Alegre, São Paulo, Goiânia and Belém). The results of the experiment indicated an estimated prevalence of sexual violence of about 4.1%, however, no sensitivity bias or misreporting was found. Furthermore, results based on econometric models estimated by Maximum Likelihood indicated that older women and more educated women were more likely to omit the sexual violence suffered when asked directly, in contrast to questioning via the list experiment, whose reports were higher. Likewise, women from São Paulo were more likely to omit, whereas women from Recife, Salvador and Goiânia tended to omit less, indicating possible heterogeneities among the cities. These results, therefore, suggest initial policy prescriptions by pointing out which groups of sexual violence victims public authorities and services should give more attention to at the time of the complaint in order to improve their disclosure of victimization.