Predictors of Discordance and Concordance in Reporting of Intimate Partner Violence: Evidence From a Large Sample of Rural Ethiopian Couples

Abstract

Intimate partner violence (IPV) is a major worldwide health challenge, and addressing this challenge requires high-quality data. This analysis uses a large-scale survey of 5033 households in rural Ethiopia in which both men and women were surveyed about past-year IPV in order to quantify the degree of discordance, including both husband only reporting and wife only reporting, for multiple forms of IPV (emotional, physical, and sexual). In addition, logistic regression is employed to analyze the effects of demographic characteristics and individual norms and behaviors on the probability of discordant reporting. The results suggest that almost half of households (44%) are characterized by discordant reporting in at least one dimension of IPV. Given the high level of discordance, 61.4% of households report any physical and/or sexual IPV using the household-level measure, compared to a rate of 41.9% from the women’s data only. In addition, men who report more gender-equitable attitudes and behaviors (failing to concur with justifications for IPV, reporting higher support for gender equitable norms, and reporting a higher level of female engagement in decision-making and intrahousehold task-sharing) are more likely to be members of wife only reporting households: that is, they are less likely to report perpetration of IPV. Women who report more gender-equitable attitudes and behaviors, by contrast, are more likely to be members of husband only reporting households.

Bidirectional intimate partner violence: Evidence from a list experiment in Kenya

“…Using this method, the prevalence of IPV is demonstrated by the constant (Column 7).10 We useMcKenzie and Woodruff (2014) to estimate whether our sample is large enough to detect differences, and find that we have 99.99% to detect a 10% difference in prevalence between household groups, 93% power to detect a 5% difference, and we can detect an effect as low as a 4.1% difference with 80% power.11 There are, however, a greater number of studies that analyze discordance in spouse reporting of IPV perpetrated against women, in which the wife is asked to report her experience of IPV and the husband is asked about perpetrating violence. See, for example,Khawaja and Tewtel-Salem (2004),Yount and Li (2012),Yoshikawa et al (2021),Leight et al (2022).12 Indeed, in Ethiopia,Gilligan et al (2021) find that list experiments lead to lower reporting of IPV among women compared to direct elicitation. This fleeing behavior is not homogeneous; it is more common among women under 30 and those who have higher self-efficacy.…”