Measuring the effects of community polygyny on intimate partner violence: a multilevel modeling using nationally representative cross-sectional data.

backgroundIntimate partner violence (IPV) has received increasing attention the last three decades and it has been recognized as health, social, and human right issues across the world. Worldwide, sub-Saharan Africa is the most affected region. In the search of putative factors associated with IPV, women in polygamous marriages were found to be more exposed to IPV compared with those in monogamous marriages. However, previous research focused on polygyny at individual level; therefore, ignoring possible heterogeneity of the "acceptance of polygyny" across communities. This paper developed the concept of "community polygyny" in Central Africa and tested its associations with IPV. Furthermore, the paper tested interactions between the community polygyny and (i) polygyny at women's level; (ii) women's education; and (iii) urban residence.

methodsThe paper used recent Demographic and Health Surveys of four countries in Central Africa (Democratic Republic of the Congo, Cameroon, Gabon, and Chad). Multilevel binary logistic regression analyses (additive and multiplicative models) were carried out. Findings were reported as adjusted odds ratios (aOR) at 95% Confidence Interval (95%CI). MAIN

findingsFindings indicated an inverse-relationship between the percentage of women living in polygamous marriages and IPV. Indeed, the percentage of polygamous marriages was higher in Chad (34.3%) compared with the Democratic Republic of the Congo (18.6%), Cameroon (14.6%), and Gabon (13.9%). However, the percentage of women who experienced intimate partner violence in the last 12 months was lower in Chad (18.9%) compared with the Democratic Republic of the Congo (43.5%), Gabon (40.3%), and Cameroon (30.8%). In multivariate results, findings showed that a sizeable percentage of intraclass correlation (ICC) of IPV in the selected countries was explained at cluster level: ICC = 18.2% [95%CI: 15.0%-21.9%]; ICC = 14.3% [95%CI: ICC = 10.9%-18.5%]; ICC = 7.8% [95%CI: 5.3%-11.3%]; and ICC = 29.5% [95%CI: 23.7%-35.9%] in the Democratic Republic of the Congo, Cameroon, Gabon and Chad, respectively. Community polygyny (from Model 2) showed different patterns. In the DRC, it was positively and significantly associated with the likelihood of IPV (aOR = 2.890; 95%CI: 1.461-5.719). In contrast, it showed a negative association with IPV in Cameroon (aOR = 0.278; 95%CI = 0.143-0.539); in Gabon (aOR = 0.504; 95%CI = 0.237-1.074); and Chad (aOR = 0.749; 95%CI = 0.319-1.763).

conclusionFindings from this study substantiates the importance of moving beyond individuals' characteristics and incorporating the collective mindset of polygyny to fully capture the effects of polygyny on intimate partner violence in Central Africa. Previous research pointed out the negative effects of polygyny at individual level. This study showed that community polygyny, since communities might differ on the acceptance levels towards polygyny, is equally important to understand how polygyny could affect the prevalence of intimate partner violence in Central Africa. Therefore, interventions aimed at eliminating intimate partner violence should integrate communities' influences on intimate partner violence in Central Africa and worldwide.