The Association between Spousal Gender Based Violence and Women’s Empowerment among Currently Married Women aged 15-49 in Zimbabwe: Evidence from the 2010-11 Zimbabwe Demographic and Health Survey

In Zimbabwe, levels of spousal Gender-Based Violence (GBV) remain a health, human rights and development concern. The main objective of this study was to investigate the association between spousal GBV and women’s empowerment among women aged 15-49 who were currently in union or living with a man. The analysis utilised the 2010-11 Zimbabwe Demographic and Health Survey (ZDHS) data. Spousal violence was measured by different forms of GBV i.e. physical, emotional, sexual violence and overall GBV. Overall, 48.2% of women of reproductive age who are in a union experience some type of GBV. More specifically, 26.9% report physical violence, 25.7% sexual violence and 25.7% emotional violence. The analysis showed that women who did not participate in decision-making at household level were more likely to experience GBV than those who do. Women who have control over their spouses’ earnings were less likely to suffer from GBV. However, ownership of property (land and/or house) was not associated with spousal GBV. Women who were in polygamous unions, whose spouses drank alcohol, earned less than their spouses and had a history of non-spousal physical violence were more likely to experience GBV than all other women. Younger women (aged 15-19 years) were more likely to experience spousal emotional, physical and sexual violence than the older age groups. Recommendations are to improve women’s participation in decision making, mainstreaming GBV in development, and improve the economic and social emancipation of women and girls.


Résumé
Au Zimbabwe, les niveaux de la violence conjugale à caractère sexiste restera toujours l'une des préoccu-pations de la santé, des droits de l'homme et du développement. L'objectif principal de cette étude était d'examiner l'association entre la violence conjugale à caractère sexiste et l'autonomisation des femmes chez les femmes âgés de 15 à 49 ans actuellement en union conjugale ou vivant en couple avec un homme. L'analyse a utilisé des données recueillies dans l'Enquête démographique et de santé au Zimba-bwe (ZDHS) effectuée entre 2010 et 11 [The 2010-11 Zimbabwe Demographic and Health Survey (ZDHS)]. La violence conjugale a été mesurée par différentes formes de violence à caractère sexiste c'est-à-dire physique, émotionnel, la violence sexuelle et la la violence à caractère sexiste globale. 48.2% des femmes en âge de procréer en union conjugale en subit la violence à caractère sex-iste sous une forme ou une autre. Plus spécifiquement, 26.9% d'entre-elles dit avoir été victimes de vio-lence physique, 25.7% d'entre-elles signale avoir souffert des violences sexuelles et 25.7% fait part de la violence émotionnelle. Cette analyse a démontré que les femmes qui ne participent pas à la prise de déci-sion au niveau des ménages étaient plus susceptibles que celles qui y participent de d'être touchées par la violence à caractère sexiste. Les femmes qui disposent à leur gré du revenu de leurs conjoints étaient moins susceptibles de souffrir de la violence à caractère sexiste. Toutefois, la propriété des biens (terres et/ou maison) n'est pas associée à la violence conjugale à caractère sexiste. Les femmes en unions polyg-ames, dont les conjoints consomment de l'alcool, gagnent moins que leurs conjoints et qui ont un lourd passé en matière de la violence physique non-conjugale étaient les plus susceptibles que toutes les autres femmes de subir la violence à caractère Introduction Gender-Based Violence (GBV) has been acknowledged as a serious global health, human rights and development issue (USAID and UNICEF 2012). The United Nations General Assembly 1993, Article 1 defines violence against women as "any act of genderbased violence that results in, or is likely to result in, physical, sexual or mental harm or suffering to women, including threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or in private life". The UN Declaration on the Elimination of Violence against Women (CEDAW) describes violence as "manifestation of historically unequal power relations between men and women, which have led to domination over and discrimination against women by men and to the prevention of the full advancement of women", and as "one of the crucial social mechanisms by which women are forced into subordinate position compared to men" (Articles 1 and 2 of the UN General Assembly Declaration on the Elimination of Violence against Women (1993) and Recommendation 19, paragraph 6 of the 11th Session 1992 of the CEDAW Committee). In 2007, Zimbabwe adopted the Domestic Violence Act (Chapter 5:16, Act 14/2006) which recognises a series of harmful practices or injury, including physical, sexual, emotional and psychological abuse.
Violence against women is a universal problem affecting millions of women everyday worldwide (Oyediran and Isuego-Abanihe 2005). Several studies indicate that women and girls are at higher risk of experiencing violence than men and boys. According to the United Nations (UN 2013) estimates, one in three women worldwide has experienced physical or sexual violence. The most common type of violence against women is domestic, a form of violence perpetrated by a current or ex-husband/ex-partner (WHO 2005;Heise et al. 1994, cited in Naved andPersson 2005). Several studies have reported that spousal GBV is widespread (Abramsky et al. 2011;Saffitz 2010;Koenig et al. 2003;Jewkes et al. 2002). However, prevalence of spousal violence is higher in African and Latin American countries (Abramsky et al. 2011). Studies conducted in Sub-Saharan Africa indicate high prevalence of physical, sexual and emotional violence against women (Nyamayemombe et al. 2010;Saffitz 2010;Koenig et al. 2003;Jewkes et al. 2002;Watts et al. 1998). According to the 2010-11 Zimbabwe Demographic and Health Survey (ZDHS), 48% of women had experienced some form of spousal violence. Also, 30% of women reported that they had experienced physical violence since the age of 15, while 27% of women have experienced sexual violence since the age of 15, of which nine out of ten cases were spousal violence. Nyamayemombe et al. (2010) Similarly, population-based surveys conducted in Eastern and Southern Africa also indicate high prevalence rates of spousal physical violence ranging from 13% in South Africa to 49% in Ethiopia and spousal sexual violence ranging from 7% in South Africa to 59% in Ethiopia" (USAID and UNICEF 2012). Justino's (2008) study shows nearly 50% of women in Mozambique experience domestic violence. Similarly, a study conducted in Botswana by the Botswana Women's Affairs Department (WAD) reported that three out of every five women have been victims of violence largely perpetrated by intimate partners and acquaintances (WAD 1999).
Zimbabwe, despite having adopted the Domestic Violence Act, the challenge is the full implementation of the policy as level of domestic violence remains high as 48% the women have experienced some form of spousal violence (ZIMSTAT and ICF International 2012). Violence against women is one of the most rampant human rights violations (Oxford International 2012). An increasing amount of research highlights the health burdens, and demographic consequences of spousal violence (United Nations 2006). Domestic violence can lead to death and disability (UN Millennium Project 2005cited in USAID 2009). Not only is GBV a major obstacle to women's and girls' development and to the welfare and development of their communities and societies as a whole, but also negatively impacts on the socio-economic development of the country (UN 2013;Oxford International 2012).Unequal power relations between men and women significantly contribute to spousal violence (USAID 2009).
Several studies conducted on GBV, focused on physical, sexual and emotional or psychological violence. Most of these studies have tended to look at individual-level or family factors, sidelining the community or socio-cultural factors, such as gender inequality. Information on the interface between spousal violence and women empowerment remains relatively limited in Zimbabwe, especially on women's empowerment as measured by ownership of assets such as land and house. The 2010-11 ZDHS does not articulate on the linkages between spousal violence and women empowerment, hence the need to explore this relationship in detail to help inform policy makers and GBV programmes in particular. This study was conducted to provide evidence on the scale of spousal GBV by exploring the association between spousal GBV and women's empowerment by incorporating other aspects of women's empowerment i.e. women's ownership of property (i.e. land and house) and control over cash earnings, as well as the two indicators of economic empowerment identified in the ZDHS 2010-11. The specific objectives of this extended analysis were to: (a) determine the prevalence of spousal GBV, including the different forms of GBV (physical, emotional and sexual); (b) assess the indicators of women's empowerment and other factors as determinants of GBV. The results are expected to, not only help inform policies and programmes, but also help to identify targets for GBV interventions.

Previous Studies and Framework for the Analysis
According to Castro (2004), sociological models link violence to gender inequality. The causes of gender violence lie in the way society is organized: for instance, unequal economic opportunities available to women and men, the availability of institutional resources for women who are victims of spousal violence, and the degree of protection offered by the legal system. Psychological models also incorporate individual characteristics as determinants of violence. According to Walker (1984), there is also qualitative evidence that abuse is often accompanied by a curtailment of the victim's economic and social independence.
Several studies have been conducted to ascertain factors associated with spousal GBV, varying from individual to societal and structural factors (e.g. USAID and UNICEF 2012; Abramsky et al. 2011;Nyamayemombe et al. 2010;Jewkes et al. 2002;Oyediran and Isuego-Abanihe 2005). Jewkes et al. (2002) found that domestic violence is strongly associated with the "status of women in a society and the normative use of violence as part of the exercise of power". Nyamayemombe et al. (2010) assessed the relationship between GBV and women's status in relation to; age gap with husband/partner; education gap with husband/partner; person who decides woman's cash earnings; and participation in household decisions. They found that these indicators were predictors of spousal violence.
Previous studies reported that women who experienced domestic violence in childhood are more likely to encounter domestic violence from their husbands or partners (Abramsky et al. 2011;Devries et al. 2010;Jewkes et al. 2002). Tuladhar et al. (2013) found that women's empowerment was inversely associated with greater odds of having experienced spousal violence, but insignificant when age, wealth, caste/ethnicity, and ecological zone were controlled. They concluded that violence is a multi-faceted problem affected by a wide variety of contextual and situational factors.
Previous studies show a positive association between polygamous unions and GBV (Abramsky et al. 2011;Nyamayemombe et al. 2010). Nyamayemombe et al. (2010) reported that women who were in polygamous unions were more likely to experience spousal violence (46.1%) compared to women in monogamous unions (35.2%). Also, domestic violence is positively associated with husband's or partner's alcohol consumption (Abramsky et al. 2011;Saffitz 2010;Devries et al. 2010;WAD 1999). Often women who have daughters only are more likely to be subjected to domestic violence than women with sons. The number of male children could be used also as a proxy for having daughters only. A negative relationship was reported between the number of male children and violence (Saffitz 2010). This could be due to a "man's increased satisfaction with his wife for bearing male children, to male children having more opportunities to contribute to the family income, or to a male child's ability to protect his mother" (Saffitz 2010: 89). The opposite reasons could explain why women with daughters only are at high odds of being abused.
Several studies have reported a significant degree of social acceptability and tolerance for GBV among women (Abramsky et al. 2011;Devries et al. 2010;Saffitz 2010;Oyediran and Isiugo-Abanihe 2005). The reason given for the justification of wife beating by husband was that "victims of GBV had internalised the abuses committed against them and were now part of the status quo who, perhaps unknowingly, propagates such beliefs" (Saffitz 2010:96). However, Jewkes et al. (2002) found no association between GBV and a woman's attitudes towards violence.
Women whose husbands or partners were younger or were of the same age were more likely to experience spousal violence than women with older partners, especially with partners 10 and above years older (Nyamayemombe et al. 2010). Women who had partners with lower educational level were more likely to experience spousal violence than women with husband/partners with same or higher education (Nyamayemombe et al. 2010). Wealth index was also reported to be inversely related to GBV (Oyediran and Isiugo-Abanihe 2005). That is, women in the wealthiest quintile were making significant contributions to family income, thereby raising their socioeconomic status. Younger women are at higher risk of spousal GBV than older women (Abramsky et al. 2011). Previous researches have shown an inverse relationship between level of education attained and experience of spousal violence among women currently in union (Abramsky et al. 2011;Saffitz 2010;Devries et al. 2010;Oyediran and Isiugo-Abanihe 2005). Several studies have shown associations between spousal violence and rural/urban residence and education, for example in Zimbabwe, women from rural areas and less educated were more likely to have experienced spousal violence. In contrast, other studies in South Africa have reported a U-shaped relationship between intimate partner violence (IPV) and education, whereby reduced IPV is seen at the lowest and highest educational levels (Abramsky et al. 2011).
Given the empirical context of spousal violence, there seems to be virtually few studies done to investigate the association between spousal GBV and women's ownership of property, i.e. land and house, which are considered as important indicators of women empowerment, especially land. The study aims to build upon literature on the association between GBV and such indicators as women's ownership of property, i.e. land and house, as these have not been included in previous studies.
The analysis was guided by Heise (1998) model, an ecological framework that is useful in understanding the complex interactions of gender and power. The model recognises the roles played by factors at four different levels of violence causality, i.e. individual, interpersonal, institutional and structural. The first level corresponds to an individual's personal and biological history, encompassing factors of an individual's development that shape responses to inter-personal and institutional stressors. The personal level includes factors such as witnessing spousal violence as a child and being abused as a child. The second level, which is interpersonal, is representative of one's immediate context in which violence takes place which is the family, household, intimate or acquaintance relationships and covers issues such as alcohol consumption and male control of wealth. The third, the institutional level, accounts for institutions and social structures that affect the individual including socioeconomic status and isolation of women. Lastly, the structural level represents one's economic, social structure and cultural environment including factors like acceptance of violence, men's entitlement/ownership of women and masculinity linked to aggression and dominance (Heise 1998;Abramsky et al. 2011;Saffitz 2010).The framework seeks to demonstrate that no single level or factor determines or explains violence but rather a combination of factors that yield violence (Moser and Shrader 1999).
However, to move out of the problem of spousal violence, the empowerment theory must be mentioned. The empowerment theory (Itzhaky and Porat 2005 cited in Payne and Wemeling 2009: 1) asserts that "victimization is not something that happens to an individual because of personal characteristics, family or origin, but rather family violence can happen to anyone who has the misfortune of becoming involved with someone who seeks to maintain power and control over intimate partners or family members". The feminist approach to domestic violence "holds that almost all male-on-female abuse is based on the patriarchal values of our society and that these values are sanctioned by a culture in which male domination of women is both covertly and overtly reinforced " (Lawson 2000; 20 cited in Payne and Wemeling 2009). With the patriarchal system, many men still hold traditional beliefs that they control their wives and therefore justify spousal violence.
This paper therefore seeks to account for variables that are crucial in all the levels of one's social environment by assessing factors that put women at risk of spousal GBV, taking into consideration issues of women's empowerment. Hence, Heise (1998) model was adopted and modified to suit the local context incorporating some aspects of women's empowerment. See Figure 1.

Study Design
Our analysis utilizes data from the 2010-11 ZDHS, which collected data from a nationally representative sample of women age 15-49 and men age 15-54. For this paper, data from the women's questionnaire was used for analysis. The unit of analysis is on 4,094 women aged 15-49 currently in union or living with a man who participated in the Household Relations Module. The analysis is on women who are currently in union because the study is on spousal violence. The 2010-11 ZDHS has for the second time collected data on gender-based violence with the first collected in the 2005-06 ZDHS. The 2010-11 ZDHS had included for the first time a module on women's empowerment which covers women's control over earnings, women's participation in decision making, and women's access to property i.e. land and/or house.

Variables
Our outcome or dependent variables include the different forms of GBV, i.e. spousal physical, emotional and sexual violence, and any GBV. See Figure 1. Independent variables include: women's empowerment was our key variable, which is measured by: women's control over their cash earning; women's control over husband's/partner's cash earnings; women's ownership of assets (house, land); women's participation in making household decisions; women's attitudes towards wife beating; and women's earnings relative to husband's/partner's. Other factors associated with GBV include: history of abuse; intergenerational exposure to violence; consumption; age difference with husband/partner; education gap with husband/partner; whether husband drinks alcohol, and daughters only. Socio-demographic factors include: residence (urban, rural); wealth quintile; religion; education of woman, and daughters only.

Measures
The definition of domestic violence, a form of gender-based violence, used in the analysis was the one defined in the 2010-11 ZDHS, which is "any act of violence resulting in physical, sexual or psychological harm or suffering to women, girls, and also men, including threats of such acts, coercion, or arbitrary deprivation of liberty" (ZIMSTAT and ICF International 2012). Spousal violence was measured by different forms of GBV i.e. physical, emotional and sexual violence, as well as the overall GBV. The following questions were asked on each: Spousal emotional violence: ever been humiliated or threatened with harm or insulted or made to feel bad by husband/partner. Spousal sexual violence: ever been physically forced into unwanted sex or forced into other unwanted sexual act or physically forced to perform sexual acts by husband/partner. Spousal physical violence: ever been pushed, shook or had something thrown or slapped or punched with fist or hit by something or kicked or dragged or strangled or burnt or threatened with knife/gun or other weapon or arm twisted or hair pulled by husband/partner. Spousal GBV: ever experienced spousal physical, sexual or emotional violence.

Data Analysis
The analysis was done using STATA Version SE Version 12. The analysis used univariate, bivariate and multivariate methods. The univariate analysis was used to determine the prevalence of spousal physical, emotional and sexual violence and the overall GBV. In the bivariate analysis, logistic regression was conducted to determine the relationship between each key independent variable (women's empowerment and other factors) and outcomes (spousal physical, emotional and sexual violence and the overall GBV). We used Pearson's chi-square to determine a relationship and the p-value to determine whether the difference between the variables is statistically significant at 95% confidence level. Lastly, we used multivariate logistic regression to examine the association between the outcome variables and women's empowerment variables plus other compounding variables, controlling for socio-demographic factors (age group, residence, region, and wealth quintile). The variables put in the model were guided by our logical framework. The domestic violence weight (d005) was used to restore the representativeness of our sample.
In reporting the results, we considered odds ratios (ORs): OR <1 indicating a negative relationship, OR >1 indicating a positive relationship, and OR=1 indicating no association. Statistical significance was considered at the 5% level.
Limitations of the Study The analysis was based on the 2010-11 ZDHS data. The results cannot be generalised to ever/never married women as the analysis is based on women who are currently in union, excluding those women who had been in union, i.e. widows, divorcees and those in separation, and have experienced spousal violence. The ZDHS data cannot be disaggregated to provincial and district level. This disaggregation is critical for programming as it would enable to identify the areas, besides the rural urban differences, that are most affected where intervention focus is required. Another limitation is that the gender-based questions only targeted women and excluded the perceptions of the men.

Sample Description
A total of 4,094 women who were currently married, aged 15-49 years old and who consented to be interviewed for the household module from the ZDHS 2010-11 were analysed. Table 1 shows the socio-demographic characteristics of the respondents. Almost a quarter of the respondents (24.2%) were aged 25-29 years old while the 20-24, 30-34 and 35-39, cohorts constituted 21.0%, 17.6% and 14.4%, respectively. The remaining cohorts aged 15-19, 40-44 and 45-49 constituted 8%, 8.5% and 6.2%, respectively. The median age was 27 years (IQR = 15) whereas the mean age was 28.1 years (SD = 9.3). A third of the women (33%) were from urban areas, 61.4% had attained secondary education, 43.1% of the respondents were of Apostolic Faith religion and 11.5% of the women were in polygamous unions. Almost half of the women (45.4%) were married to a man who drank alcoholic beverages. Less than a quarter (22.3%) of the women had daughters only. Very few women, approximately 3%, were older than their current husband/partner and less than 1% was not sure of their husband/partner's age. The majority of the women (62.4%) reported that they did not have any cash earnings whilst 24.5% earned less than their husband/partner, 7.2% earned about the same and 5.9% earned more than their husband/partner. Less than a tenth of the women (6.7%) had a history of non-spousal physical violence while 37.3% of the women had their mother beaten up by their father.
Respondents were asked questions to determine their empowerment status. The majority of the women (68.8%) reported control of the husband/partner's earnings was done jointly and 12.9% reported that these earnings were solely controlled by the husband/partner. The analysis also showed that 48.7% and 48.4% of the women did not own any land or house, respectively. Three-quarters of the women (75%) reported that they participated in all household decisions while 4.6% did not participate in any decision. Two in five women (40.7%) believed that a man is justified in beating his wife.  Table 2 shows the results of bivariate analysis performed to examine the association between each form of spousal violence and women's empowerment. Overall, 48.2% of women studied had experienced some type of GBV -specifically, 24.9% reported emotional violence, 26.9% reported physical violence, and 25.7% reported sexual violence.
As Table 2 shows, women who had no control over their husband's/partner's cash earnings were more likely to suffer from spousal emotional violence (32.7%) compared with women who jointly controlled their husband's/partners earnings, or women who alone decided how their man's earnings were used, 24% and 23%, respectively (p<0.001). Women who believed that a man is justified in beating his wife were more likely to suffer from spousal emotional violence (30.1%) than women who believed that a wife should not be beaten (21.4%; p<0.001). Women's participation in household decision making was also associated with spousal emotional violence. Women who did not participate in household decision making were more likely to experience spousal emotional violence (33.2%) than women who participated in some or all decisions, 30.9% and 22.8%, respectively (p<0.001). No significant associations were found between spousal violence and other measures of women's empowerment, i.e. women's control over her cash earnings, and ownership of house or land.
Physical violence was associated with almost all the measures of women's empowerment, with the exception of women's control over her cash earnings and ownership of a house. Women who had sole control over their cash earnings were more likely to suffer from spousal physical violence (32.5%) than women with no control or had joint control, 25.6% and 27.4%, respectively (p<0.05). Women who jointly owned land with their husband/partner were more likely to experience spousal physical violence (29.7%) than women who did not own land or solely owned land, 25.2% and 26.3%, respectively. As with emotional violence, women who believed that a man is justified in beating wife were more likely to experience physical violence (33.5%) than those who did not believe in wife beating (22.4%, p<0.001). Unlike emotional violence where women who did not participate in household decisions were more likely to suffer from emotional violence, women who participated in some decisions were more likely to experience physical violence (34%) than those who did not participate or participated in all decisions, 28% and 24.9%, respectively (p<0.05).
Women who believed that men were justified in beating their wives were more likely to experience sexual violence than women who did not believe in wife beating, 29.9% and 22.9%, respectively (p<0.001). Also, women who participated in some household decisions were more likely to experience sexual violence (32.7%) than women who made all decisions (23.9%) or did not take part in any decisions (25%) (p<0.001). No significant associations were found between sexual violence and other measures of women's empowerment i.e. women's control over her and husband's/partner's cash earnings, and ownership of house or land.
On the overall GBV, bivariate associations were observed on only three measures of women's empowerment. First, women solely controlled their cash earnings were more likely to suffer from GBV (54.2%) than women who had joint control or no control of their cash earnings, 49.4% and 46.4%, re-spectively (p<0.05). Second, women whose husband/partner solely controlled his cash earnings or had joint control were more likely to experience some GBV (51.4% and 52.5%, respectively) compared with women who solely controlled their husband's/partner's earnings (46.5%; p<0.05). . Lastly, women were also likely to experience GBV when they believed a man is justified in beating his wife (56.6%) than those who did not believe in wife beating (42.4%; p<0.001). No significant associations were found between GBV and ownership of house or land, and women's participation in household decisions.

Factors Associated with Spousal Emotional, Physical and Sexual Violence and overall GBV
Multivariate logistic regression was employed to examine the association between spousal violence and socio-demographic variables, as well as women's empowerment indicators. All predictors were entered in the model simultaneously, but the results are presented in Tables 3 and 4.

Emotional Violence
Results in Table 3 Table 4 show that women who reported that the spouse/husband/partner has full control of his cash earnings were about twice more likely to suffer from emotional violence compared to women who reported that they control their husband/partner's earning (AOR=1.78; 95% CI=(1.27, 2.47)). Women who believed men were justified in beating their spouses were 1.35 times more likely to suffer from spousal emotional violence than those who did not feel wife beating was justified (AOR = 1.35; 95% CI = (1.14, 1.60)). Women exposed to non-spousal violence ever since the age of 15 years were 1.70 times more likely to suffer from spousal emotional violence than those who were not exposed (AOR=1.70; 95% CI = (1.43, 2.03)). Spousal emotional violence was significantly higher in women who were married to men who drank alcohol than those with husbands/partners who did not drink (AOR = 1.34; 95% CI = (1.11, 1.62)). Women whose fathers ever beat up their mother were 2.32 times more likely to have experienced spousal emotional violence than women whose fathers did not beat up their mothers (AOR = 2.32; 95% CI = (1.75, 3.08)).

Physical Violence
Much like emotional violence, age was also significantly associated with spousal physical violence, as older women aged 35 years and above were less likely to suffer spousal physical violence compared to 15-19 year olds after controlling for all other variables (See Table 3). Women who had attained more than secondary (tertiary) education were less likely to have experienced spousal physical violence than women who had no education (AOR = 0.34; 95% CI = (0.13, 0.88)). Women who were in polygamous marriages were 1.77 times more likely to suffer spousal physical violence than those in monogamous marriages (AOR = 1.77; 95% CI = (1.35, 2.31)). Table 4 shows that women whose cash earnings are controlled by others were significantly less likely to experience spousal physical violence compared to women who control their own cash earning. Joint ownership of land increased physical violence on women (AOR=1.36; 95% CI = (1.07, 1.73)). Women with history of non-spousal violence were about twice as likely to experience physical violence compared to women who were not (AOR = 2.11; 95% CI = (1.78, 2.50)). Table 4 also shows that women with husbands/partners who drank alcohol were 1.60 times more likely to suffer spousal physical violence than women with husbands/partners who did not drink alcohol (AOR = 1.60; 95% CI = (1.33, 1.92)). Women whose fathers ever beat up their mothers were more likely to experience physical violence than women who were not exposed (AOR = 1.65; 95% CI = (1.18, 2.30)).

Sexual Violence
Age was significantly associated with spousal sexual violence, results in Table 3 show that as age increased the odds of suffering from spousal sexual violence decreased. Women with secondary education were 1.95 times more likely to experience spousal sexual violence than women with no education (AOR = 1.95; 95% CI = (1.02, 3.73)). Table 3 also shows that women from the fourth wealth quintile were 1.36 times more likely to experience spousal sexual violence than women from the lowest quintile (AOR = 1.36; 95% CI = (1.03, 1.78)). Table 4 shows that women who reported that wife beating was justified were 1.26 times more likely to experience spousal sexual violence than women who reported that wife should not be beaten (AOR = 1.26; 95% CI = (1.07, 1.49)). The odds of spousal sexual violence amongst women who had history of physical violence from other people besides their husband/partner were 1.5 times higher than those without any history of physical violence (AOR = 1.46; 95% CI = (1.24, 1.73)). Women who were married/cohabitated with men that drank alcohol were 1.41 times more likely to suffer spousal sexual violence than those with men who did not drink (AOR = 1.41; 95% CI = (1.19, 167)). In addition, women who reported that their fathers beat up their mothers were twice as likely to experience sexual violence as those who did not (AOR = 1.97; 95% CI = (1.45, 2.67)). The different measures of women's empowerment were not significantly associated with sexual violence.

Gender Based Violence (Emotional, Physical and Sexual)
General trends in the odds of women being abused by their spouses show that the older the woman the less likely she experienced gender based violence when compared to 15-19 year old women. Women who were in polygamous marriages (AOR = 1.45; 95% CI = (1.12, 1.88)), married to men who drank alcohol (1.47; 95% CI = (1.26, 1.71)), had at least a son (AOR = 1.43; 95% CI = (1.02, 2.01)), had a history of non-spousal physical violence (AOR = 1.96; 95% CI = (1.72, 2.24)), those who reported that their fathers beat up their mother (AOR = 2.42; 95% CI = (1.84, 3.17)) were more likely to experience gender based violence than all other women. The rest of the indicators of women's empowerment were not statistically significantly associated with GBV, except for women who believed that men were justified in beating their wives compared to women who did not believe in wife beating (AOR = 1.48; 95% CI = (1.26, 1.73)) and women who owned land jointly with their spouse (AOR=1.32, 95% CI = 1.07, 1.63). Significance level: * p < .05; ** p < .01

Discussion
Gender based violence has become a major development challenge across the world in general and in Zimbabwe in particular. The extent of GBV among currently married women in Zimbabwe is very high and skewed among the young women in union. The study showed that 25.7% had experienced intimate sexual violence. About 27% of currently married women had experienced spousal physical violence, while 24.9% reportedly experienced emotional violence. The overall prevalence of spousal gender based violence (GBV) was 48.2%. Several studies in Africa have also reported high levels of spousal violence (Justino 2008;WHO 2005;Oyediran and Isuego-Abanihe 2005;WAD 1999;Watts et al. 1998).
The main objective was to investigate the association between women's empowerment and spousal GBV. The analysis shows that most of the indicators of women's empowerment were not associated with spousal GBV, except for women's control over husband's cash earnings and women's ownership of land. Women who have no control over their husband's cash earnings are more likely to suffer from emotional violence. Given that the majority of women do not have cash earnings, the plight of women is compromised by their inability to control cash earnings. Women are the custodians of children, as they ensure that children have access to food, health care, education among other essentials. Failure to control earnings therefore exposes them to violence.
Women activists have always advocated for ownership of land by women as a form of empowerment. However, no significant relationship was found between ownership of land by the wife alone and GBV. This could be due to the fact that very few women own land on their own (9.4%). Women who owned land jointly with their spouses were more likely to suffer from GBV and in particular, physical and sexual violence than women who do not own land. What is needed is to lobby for increase single ownership of land by women. The analysis has shown that ownership of house was not associated with spousal GBV. This finding contributes to the body of literature on women's empowerment and GBV.
The findings in this study show that younger women aged 15-19 are more likely to experience spousal emotional, physical and sexual violence than the older age groups. Previous studies have also reported similar findings (Nyamayemombe et al. 2010). This could be explained by the following factors: the prevalence of intergenerational sexual and marital relationships, marital communication and formation challenges owing to the age gap, and mismatch in marital expectations, among other factors. With age, most relationships stabilize and communication improves, hence the need for GBV behaviour change communication to be targeted at adolescents. Given that link between GBV and HIV, addressing this misnomer becomes critical. There is therefore need for targeted interventions that address male norms and behaviours, but also emancipate young and adolescent girls. There is need to create not only young women friendly GBV services and facilities, but also include boys and young men in addressing GBV.
Currently married women in polygamous relationships were more likely to experience GBV, and in particular emotional and physical violence, compared to women in monogamous relationships. Evidence shows that polygamous relationships are generally characterized by competition for resources, particularly the husband and income, which in the end culminates in emotional and physical violence.
The results in this study, like other preceding studies such as Nyamayemombe et al. (2010), point to the effect of alcohol consumption on GBV. Alcohol consumption was significantly associated with GBV and in all forms of GBV. Currently married women whose husband/partner consumes alcohol were likely to experience spousal GBV. Other studies, such as Abramsky et al. (2011), Saffitz (2010, and Devries et al. (2010) have also found that alcohol consumption, coupled with income challenges at household level, as a cause for spousal emotional, physical and sexual abuse. Evidence shows that men abuse alcohol to deal with stress.In addition, alcohol hinders judgement.
Exposure to violence is also critical in understanding GBV. The analysis shows that currently married women who witnessed or experienced GBV prior to marriage were more likely to suffer spousal emotional or physical abuse. Having a history of abuse puts women at higher odds of being sexually abused as reported in other studies (Tuladhar et al. 2013;Abramsky et al. 2011;Devries et al. 2010;Jewkes et al. 2002). In the same realm, the same women felt this was justifiable. This shows the extent to which exposure to a certain negative behaviour can be perpetuated and become detrimental to society in general and women in particular.
The analysis also shows that women with at least one son were more likely to suffer emotional violence than women with no children. There was no significant association between women with daughters only and GBV. This is contrary to other studies on spousal violence; Saffitz (2010) who reported a negative rela-tionship between number of sons and spousal violence. As indicated earlier, the number of male children could be used also as a proxy for having daughters only.

Conclusion
The study found that GBV is widespread in Zimbabwe, with nearly half of the women reported having experienced some form of violence; 26% of women had experienced spousal sexual violence, 27% spousal physical violence, and 25% spousal emotional violence. Spousal GBV is a major problem in Zimbabwe and there is need to take a comprehensive and multi-sectoral approach to addressing this major impediment to development. Young married women are at the risk of spousal GBV. There is need to emancipate and empower young women and adolescents. At the same time, age-sensitive development interventions need to be designed. It is also important that GBV is incorporated in the secondary and tertiary schools.
The link between polygamy and GBV is strong. Compounded by the risks of HIV and AIDS, the culture of polygamy needs to be addressed. Behaviour communication has failed in areas where polygamy is accepted. There is need to look for other alternative strategies that should be targeted at both men and young women in such unions, while the efforts to address intergenerational sexual relationships should be strengthened. Emancipation and empowerment of women remains critical and targeted to include religious sects.
Alcohol abuse is a major factor in GBV. Given the documented link between alcohol abuse and HIV, it also becomes critical that the country adopts an alcohol policy that protects both the consumers and the families.
As part of the empowerment process, there is need to engage men, women, youth and girls in processes that ensure women and girls access income earning opportunities. In the same vein, there is need for targeted, gender sensitive interventions that destigmatize women and girls' participation in the mainstream economy. The same interventions should also ensure that women are inculcated to appreciate that they can make a difference in their own lives and not to be viewed as objects and mere recipients in the economic sphere. This should start at primary school level. Institutions of higher learning, including colleges and vocational training centres, should continue to create a platform for women to participate more.
Given that the decision making process at household level remains skewed, there is need for the engagement of men, community leaders and women in an intensive and broad based process that identifies challenges and constrains to women participation in decision making. Once done, there is need for a comprehensive and focused behaviour change communication package that addresses the challenges and constrains. More likely, the outcome of this process will also point to the need to economically empower women, while educationally emancipating them.
Overall, GBV remains a major challenge in Zimbabwe, amid calls to mainstream GBV in all development work, create youth friendly GBV services (including screening), economic and social emancipation of women and girls, offer opportunities for rehabilitation of survivors of GBV (including the youth), and address some mal-behaviours as polygamy and alcohol abuse through targeted behaviour change communication. While the Domestic Violence Act (Chapter 5.16): 2007 and its Regulations (2008)makes it criminal to perpetrate acts of GBV, there is need to do more by creating an environment of non-tolerance, while addressing the socio-cultural barriers which stigmatize victims from reporting. Engaging the communities, especially men, community leaders and the youth becomes critical.
To conclude, the analysis has contributed to the body of literature by analysing the association between women's ownership of property, as an indicator of women's empowerment, and GBV. Joint ownership was found to be associated with spousal GBV. The rest of the findings on the association between GBV and factors, including partner's alcohol consumption, age and polygamy, confirm findings from previous studies conducted in Zimbabwe and other African countries, and thus can be used for programme development. The study was limited in that it was cross sectional and focused on women who were in union. Some of the divorces could have been a result of the GBV. Therefore, there is need for further research on women who are not in union, particularly the divorced. The sole ownership of either land or a house by women was not associated with GBV and this could be a result of few women who actually own land or a house. There is need to lobby for increase in sole ownership on property such as land. Despite the limitations, the findings of the study help to identify targets for GBV intervention.