The study revealed that determinants of IPV among women aged 15 to 24 years operate at different levels. Having at least one living child, witnessing interparental violence, having unfavorable attitudes towards IPV, and having a partner who drinks alcohol were associated with increased odds of IPV at the individual level. Adjusting for all other factors, age was also a significant predictor for IPV. At the community level, geographical zones and living in communities with lower poverty and higher IPV acceptance as a norm increased the odds of IPV.
We found a relatively high prevalence of lifetime IPV of 47.8%. This prevalence is comparable with that reported in the WHO multi-country study on IPV among adolescent girls and young women in which the overall prevalence ranged from 19% in Serbia to 66% in Peru [27]. It is higher than what was found in Nigeria where a 12-month prevalence of IPV of 21% was reported [28]. Similarly, this national prevalence is higher than from a study in Shinyanga district that reported a prevalence of 35.9% [29].
The variation in the prevalence of IPV seen between and within communities and countries could be due to the presence of factors at different levels of the socio ecology that are very different geographically as suggested by the socio-ecological model of violence (30). The variation could also be explained by the difference in the measurement and definition of IPV between studies. While some studies studied the 12 months prevalence of IPV of only physical and sexual violence, our study looked at the lifetime prevalence of physical, sexual, and/or emotional violence. The high prevalence observed, coupled with evidence of the physical, mental and sexual health consequences of abuse, calls for the need for GBV prevention, intervention, and support services to support this important age group. The variation in the prevalence of violence seen within and between countries highlights that violence is not inevitable, and that it can be prevented.
In this study, we found that women with attitudinal acceptance of partner violence were more likely to experience IPV. Similar findings have been reported by other studies in low and middle-income countries [31, 32] and in Vietnam [20]. It is believed that women’s acceptance of IPV is shaped by childhood exposure to family violence [34]. This is further supported by the social learning theory which suggests the importance of observational learning, including the observation of behaviors and the consequences of those behaviors. In intimate relationships, witnessing parental IPV could lead individuals to view violence as acceptable [35].
Moreover, we found a similar relationship of IPV to witnessing interparental violence, as women who witnessed their father abuse their mother were more likely to be victims of IPV. Several other studies have documented similar findings [27, 32, 36]. This calls for the behavioral interventions to utilize attitude change approaches which can lead to change in behavioral practices such as the practice of violence and also having parenting interventions that can reduce negative parenting practices.
Although this study did not assess the severity, partner’s alcohol intake was found to increase the odds of IPV to the woman. This finding is in line with findings of studies from Uganda [37] and other LMICs [38, 39] that have shown a strong and consistent association between IPV and alcohol use among abusive men. Alcohol abuse can be explained to cause disinhibition which may result in diminished ability to avoid violence [36, 39]. Therefore, it is important for interventions effort to address alcohol abuse as IPV victimization is more likely to occur in such environments. However, inconsistent findings between IPV and alcohol use have been reported in a systematic review including studies from the United States, Canada, the United Kingdom, and Australia suggesting the presence of other mediating factors [40].
We found higher odds of partner violence among youth with at least one living child compared to those with no living children. This is similar to findings from a study in Shinyanga [29]. One possible explanation to why women with more children are likely to experience IPV victimization could be women with children could be more tolerant of violent acts so that they could continue caring for their children jointly with their partners [41]. Another possible explanation could be the stress children can bring to the family as larger families have the potential to generate increased frustration due to the need to provide and violence may be the outlet to the frustration generated [42].
The study showed that there is zonal variation in the experience of IPV. The Western, Southern highlands, Southern, and Lake zones have significantly higher odds of IPV. This variation may be due to differences in culture and traditions across regions. Moreover, this has policy implications, as interventions aimed to prevent violence will have to be more holistic in their approach, going beyond the descriptive zonal variables.
Women living in communities with high IPV accepting norms were found to have increased odds of IPV. Similar findings were reported in Kenya, Malawi, Zimbabwe [31], India [43], and Ethiopia [24]. Tolerant community norms regarding acts of violence not only underlie the occurrence of violence but also allow it to persist in the area and may consequently challenge intervention efforts. These norms also affect women by stigmatizing those who report abuse to preserve a moral order [24].
Contrary to our expectations, we found that it is less likely for female youth from a community with a high level of poverty to experience IPV compared to those from a low level of poverty. This is contrary to findings from a study in Tanzania (22). It could be that other mediating factors have mitigated the effects of the community poverty level, although this should be taken with caution because in the current study aggregates from the wealth index were used as a proxy measure for community poverty level. But poverty is a cumulative effect of many factors and not just ownership of assets.
Strength and weakness
This study utilized data from a nationally representative sample, which makes the findings generalizable. Also, the study had enough power to conclude on the factors associated with IPV among women aged 15 to 24 years in Tanzania.
The hierarchical nature of the data allowed us to explore factors operating beyond the individual level, i.e., to look at variables at the community level. Employing the multilevel analysis enabled us to look at factors operating at the community level. The use of this analytical method has a policy implication of making programs examine the contextual factors, which may explain the high prevalence of IPV among women aged 15 to 24 years in Tanzania.
The study has some limitations which need to be considered when interpreting the results. First, the fact the data used was collected using a cross-sectional survey inhibited us from making causal inferences between the exposure variables and our outcome variable, experience of IPV. Second, all variables used in this study, including variables on partner characteristics, were self-reported using a face-to-face interview rather than a self-administered tool. By using this technique respondents might be subjected to give socially desirable responses. Furthermore, we cannot rule out recall bias since respondents were asked to recall past experiences. Third, even though the survey followed the WHO strategies for domestic violence research that helps to minimize under-reporting bias, under-reporting of IPV experiences might still have occurred due to fear of stigma and shame.
Lastly, using secondary data, the study was limited to evaluating the variables as found in the questionnaire of the parent study. Hence, we failed to examine important factors such as community-level crime rates which have been documented elsewhere (43–45).