Background
Violence against women is ‘…any act of gender-based violence that results in, or is likely to result in, physical, sexual or mental harm or suffering to women… whether occurring in public or in private life’ [
1]. The Beijing Platform for Action also defines any form of violence against women as the violation of the human rights that impedes the achievement of the objectives of equality, development and peace in any society [
2]. Violence against women by an intimate partner is also identified as one of the most common and widespread forms of gender based violence (hence forth, GBV) in the world [
3]. Global survey on violence against women conducted by World Health Organization revealed that 30% of women (at least one in three) who had been married to or otherwise partnered with had experienced either physical or sexual violence by intimate partners [
4]. The prevalence is; however, the highest in Africa in which about 37% of women have been victims of either physical or sexual violence by an intimate partners [
3,
4]. Ethiopia is one of the African states in which the highest rate of intimate partner violence prevails. In a 2005 WHO report, about 71% of Ethiopian women who had been married to or otherwise partnered with intimate had experienced either physical or sexual violence by their partners. Out of them, about 35% had experienced at least one form of severe physical mistreatment,
inter alia, being hit with a fist or something else, hard kicked, forcefully dragged, beaten, flogged, chocked, burnt, threatened with weapons [
5,
6].
Domestic violence has brought greater impacts on women’s autonomy and sense of worth. It has also impeded in their ability to care for themselves and for their children [
5]. It has reduced their opportunities and rights for work, mobility and participation in education and training, in community activities and in the wider social network [
7]. Domestic violence against women has caused major traumatic disabilities and heart smiting deaths among women of reproductive age in both developed and developing countries in which the enigma in developing countries is estimated to account for 5% for their healthy year’s life loss [
8,
9]. It also has serious repercussions on children’s life which has resulted in poor performances in schools, increased probability of delinquency, leaving homes early, risks of abuse and joining street life, engagement in substance abuse, attempting to commit suicide and disturbing the family, etc. [
7,
10,
11].
Domestic violence has incurred immense economic costs on developing nations for prices and expenditures to prevent detect and offer health, social and legal services to countless survivors of this particular passion [
9,
12]. The violence has caused many women in being absent from work, and the time and medication required to treat physical and psycho-social damages has also reduced workers’ productivity, earned income and effective utilization of accumulated human and social capital [
11].
Despite taking different legal and policy measures to address violence against women (VAW), domestic violence, specifically wife beating, is still the highest as an acceptable norm in Ethiopia [
6,
13]. The scholars identified that through ratified international documents and adopted national instruments such as Ethiopian Women Policy [
14], the Revised Criminal Law [
15], the Revised Family Law [
16], Strategic Plan for an Integrated and Multi-Sectoral Response to Violence against Women [
17], etc. actions have been taken though non-remedial to the quandary. There are two major reasons for the highest prevalence rate of wife beating in Ethiopia. The first is the nationwide existence and acceptance of hostile gender biased attitudes such as admitting domestic violence, in general, and wife beating, in particular, as normal and private matters. The second is the swift implementation of existing national laws to address GBV in the country [
17].
Studies conducted on GBV revealed that men and women’s attitudes towards wife beating are strongly correlated to their exposure with intimate partners’ violence [
11,
18]. Abrahams and colleagues’ study showed that men who knew and believed that wife beating is an acceptable norm were two times more likely to be engaged in intimate partner violence than their counterparts. It is thus possible to argue that wife beating in Ethiopia is mainly a function of the socio-cultural settings in which the communities are living and ineffective legal and protective actions being taken to overcome the problem.
Although the societal acceptance of wife beating is believed to be widespread in Ethiopia, empirical evidences on the extent of refusal of wife-beating and its explanatory factors are non-existent. No study, to the knowledge of the authors, has ever been undertaken to investigate the determinants of wife beating refusal in urban and rural Ethiopia. No documents that show its prevalence, patterns and determinants or endeavours of a contribution towards the initiatives to be taken or scaled up to eliminate the problem exist. This study, therefore, focuses on assessing the levels and patterns of wife beating refusal and its associated socio-cultural and demographic factors in rural and urban Ethiopia.
Methods
Inclusive measurement of attitude towards wife beating requires considering as many indicators as possible [
29]. The available indicators measure the attitude of women in reproductive age towards accepting wife beating for reasons associated with household management and conjugal relationship [
55]. Besides, in traditional societies like Ethiopia, wife beating is considered either as symbol of reflecting the ‘love’ that a person has for his wife or as a means of showing the man’s supremacy over his wife [
6].
In this paper, the study is based on the analysis of Ethiopian Demographic and Health Survey (EDHS) Data in 2011 which is believed by the researchers as a national representative sample of women in the reproductive age. The survey was conducted based on sampling frames derived from the Population and Housing Census conducted by theCentral Statistical Agency of Ethiopia in 2007.
Kebeles- the smallest administrative unit in the nation were used as primary sampling units; while, households constituted secondary sampling units. A total of 16 515 women in the reproductive age group (i.e. 15–49 years) were interviewed. The response rate for the survey was 95%. The data were collected through face-to-face interviews [
55]. In this study, the questions included in the 2011 EDHS to collect information on wife beating were used to construct an index of refusal to wife beating. For this, refusal of wife beating was measured as a composite index consisting of the following items: a husband is justified in hitting or beating his wife if she (a) burns the food (b) argues with him (c) goes out without telling him (d) neglects the children, and (e) refuses to have sexual intercourse with him. Each item has a response of ‘Yes’ and ‘No’, and a value of ‘1’ is given if the woman accepts but ‘0’ otherwise. The composite index runs between ‘0’ when the woman refuses the hitting or beating at all and ‘1’ if she accepts the hitting or beating for any of the reasons. A woman is considered as ‘refusing wife beating’ if she responded ‘No’ to all of the five questions, and ‘accepting wife beating’ otherwise. The Cronbach alpha value of the index of attitude towards wife beating was 0.72 for rural and 0.78 for urban areas. The alpha values are of acceptable standard as both of them are greater than the minimum required level of 0.7 [
56].
The independent variables were region, respondents’ age group, educational level, religion, work status, marital status, number of living children born to a woman and access to media as well as household wealth status. The percentage distribution of women’s attitude towards wife beating by socio-economic characteristics of the respondents was given using tabular presentations. The Statistical Package for Social Sciences (SPSS version 20) was used to analyse the data. A binary logistic regression model was fitted to identify variables that significantly predict respondents’ refusal of wife beating in Ethiopia having controlled other confounding variables. Separate analysis by a place of residence was undertaken as attitude towards wife beating vary between rural and urban areas. The multivariate analysis model was fitted using the actual (i.e. non-weighted) data.
Results
There were considerable differences between rural and urban areas in the percentage distribution of women who had been refusing wife beating for any of the reasons. In urban areas, as it was found out, more than half (54.1%) of the respondents refused wife beating for any of the reasons; while, only about a fourth (24.5%) of the rural women had to refuse wife beating for any of the reasons (see Table
1, Panels I and III). The level of refusing wife beating in rural areas also varied across regions (Table
1, Panel I). That is, about 30% of rural women in Benishangul Gumuz, Oromia and Gamblla had refused wife beating, but in Somali, SNNP and Afar regions, less than 20% of the rural women had refused wife beating. In rural areas of the remaining regions, the percentage of women with refusing attitudes was between 20 and 30%. Unlike this, more than half (>50%) of urban women residing in Addis Ababa, Dire Dawa, Oromia, Harari, Benishangul Gumuz and Southern Nations, Nationalities and Peoples (SNNP) regions refused wife beating. Only 23% of women living in urban Somali, 37.4% of urban Amhara and 42.9% of urban Gambella had refused wife beating (Table
1 Panel III).
Table 1
Percentage distribution of women of reproductive age by their attitude towards wife beating
Region | Afar | 19.8 | 80.2 | 48.7 | 51.3 |
Amhara | 21.3 | 78.7 | 37.4 | 62.6 |
Oromia | 29.8 | 70.2 | 59.2 | 40.8 |
Somalia | 13.3 | 86.7 | 23.0 | 77.0 |
Ben-Gumuz | 32.6 | 67.4 | 53.8 | 46.2 |
SNNP | 18.1 | 81.9 | 51.0 | 49.0 |
Gambella | 29.2 | 70.8 | 42.9 | 57.1 |
Harari | 21.1 | 78.9 | 53.3 | 46.7 |
Addis Ababa | – | – | 76.1 | 23.9 |
Dire Dawa | 27.8 | 72.2 | 62.0 | 38.0 |
Tigray | 27.5 | 72.5 | 47.1 | 52.9 |
Level of education | No Education | 19.4 | 80.6 | 78.3 | 21.7 |
Primary | 29.4 | 70.6 | 61.9 | 38.1 |
Secondary | 49.5 | 50.5 | 52.0 | 48.0 |
Higher | 79.9 | 20.1 | 35.1 | 64.9 |
Age | 15–24 | 28.4 | 71.6 | 53.8 | 46.2 |
25–34 | 22.2 | 77.8 | 55.3 | 44.7 |
35–49 | 21.5 | 78.5 | 53.1 | 46.9 |
Access to media | Not at all | 21.1 | 78.9 | 23.8 | 76.2 |
Sometimes | 24.0 | 76.0 | 50.5 | 49.5 |
Frequently | 31.4 | 68.6 | 60.1 | 39.9 |
work status | Not working | 24.2 | 75.8 | 56.4 | 43.6 |
Family business | 26.0 | 74.0 | 44.9 | 55.1 |
Employed | 23.6 | 76.4 | 54.0 | 46.0 |
Marital Status | Never married | 35.5 | 64.5 | 57.8 | 42.2 |
Currently in Union | 20.4 | 79.6 | 53.3 | 46.7 |
Ever married | 26.7 | 73.3 | 45.9 | 54.1 |
Religion | Orthodox | 26.0 | 74.0 | 56.4 | 43.6 |
Protestant | 20.4 | 79.6 | 52.3 | 47.7 |
Islam | 26.2 | 73.8 | 48.6 | 51.4 |
Others | 17.7 | 82.3 | 41.7 | 58.3 |
Number of living children | None | 32.6 | 67.4 | 57.0 | 43.0 |
1–2 | 23.0 | 77.0 | 56.2 | 43.8 |
3–4 | 19.5 | 80.5 | 48.2 | 51.8 |
5 and above | 20.3 | 79.7 | 42.1 | 54.2 |
Household wealth Status | Low | 22.8 | 77.2 | 39.6 | 60.4 |
Medium | 23.1 | 76.9 | 50.7 | 49.3 |
High | 27.8 | 72.2 | 70.7 | 29.3 |
Total | 24.5 | 75.5 | 54.2 | 45.8 |
Results of the study revealed that refusing wife beating was differed by educational attainment. Even though a direct relationship between educational level and refusal of wife beating was observed, there existed variation between rural and urban areas. More resistance to wife beating by education level was observed in urban areas though the level is nearly the same among those attaining higher educational levels. Likewise, an inverse relationship between refusal of wife beating and number of living children is observed in urban and rural areas though the proportion of refusal was substantially lower among women residing in rural areas (Table
1, Panels I and III).
Women with no access to media in both rural and urban areas had less resistance to wife beating (21.1 and 23.8%, respectively) (Table
1, Panels I and III). Unlike this finding, there was a wider variation among those who have frequent access to media in urban (60.1%) and rural (31.4%) areas. The proportion with a refusing attitude towards wife beating was also higher among women in frequent access to media in urban (50.5%) than in rural areas (24.0%) (Table
1 Panels I and III). Despite variations in the level of refusal to wife beating in urban and rural settings, the age of respondents had nearly shown similar patterns in both rural and urban settings.
Rural women who have different work and wealth status had nearly similar level of refusal to wife beating (Table
1, Panel I). In urban areas, however, about 70.7% of women residing in rich households refused wife beating which was higher than the proportion among the poor (39.6%). Refusal to wife beating was found to be the highest among urban never married women (57.8%); whereas the proportion of urban women who refused wife beating was lower (45.9%) among those who married previously. Differences in the attitude towards wife beating were also observed among urban women due to work status; that is, 56.4% of non-working urban women refused wife beating vis-à-vis 44.9% of urban women working in family business. Similarly, more than half (56.4%) of urban women following Orthodox religion refused wife beating while the proportion who refused wife beating declined to 41.7% among urban women who have been following traditional and other religions. Closer to half (48.6%) of urban Muslim women were also observed to refuse wife beating (Table
1 Panel III).
The multivariate results have also shown significant variations of refusing wife beating by region. In comparison with women living in Tigray region, rural women in Somlia, Afar, Harari and Amhara regions were less likely to refuse wife beating. Their refusal attitude has decreased by 60, 34, 32 and 29%, respectively, and the effect is statistically significant (Table
2, Panel II). Rural women living only in Benishangul Gumuz region were 41% more likely to refuse wife beating (
p < 0.01). However, rural women living in Oromia, Gambella and Dire Dawa regions had same level of refusal to wife beating as equal level as that of women living in Tigray region.
Table 2
Binary logistic regression results of refusing wife beating among women of reproductive age
Region | Tigray | 1339 | ref | 386 | ref |
Afar | 1027 | 0.66** (.13) | 251 | 1.33 (.18) |
Amhara | 1809 | 0.71***(.09) | 259 | 0.66* (.17) |
Oromia | 1766 | 1.09 (.10) | 354 | 1.73** (.16) |
Somalia | 581 | 0.40***(.16) | 307 | 0.60** (.19) |
Ben-Gumuz | 1083 | 1.41** (.10) | 169 | 1.55* (.20) |
SNNP | 1784 | 0.64***(.11) | 238 | 1.06 (.19) |
Gambella | 898 | 0.84 (.12) | 226 | 0.83 (.19) |
Harari | 442 | 0.68* (.15) | 649 | 1.13 (.14) |
Addis Ababa | – | – | 1724 | 3.05***(.13) |
Dire Dawa | 368 | 1.00 (.15) | 722 | 1.84***(.14) |
Level of education | No Education | 7070 | ref | 1133 | ref |
Primary | 3681 | 1.49***(.06) | 2139 | 1.61***(.09) |
Secondary | 243 | 2.27***(.14) | 1140 | 2.70***(.11) |
Higher | 103 | 7.54***(.23) | 873 | 5.12***(.13) |
Age | 15–24 | 4376 | 0.73***(.08) | 2427 | 0.83***(.09) |
25–34 | 3513 | ref | 1729 | ref |
35–49 | 3208 | 1.12***(.07) | 1129 | 1.01 (.10) |
Access to media | Not at all | 5269 | ref | 451 | ref |
Sometimes | 3791 | 0.97 (.05) | 1335 | 1.29 (.13) |
Frequently | 2037 | 1.06 (.07) | 3499 | 1.49** (.13) |
Workstatus | Not working | 5519 | ref | 2386 | ref |
Family business | 2382 | 0.95 (.05) | 465 | 1.17 (.11) |
Employed | 3196 | 0.96 (.04) | 2434 | 0.90 (.07) |
Religion | Orthodox | 3901 | ref | 3053 | ref |
Protestant | 2317 | 0.79** (.09) | 610 | 0.94 (.11) |
Muslim | 4525 | 1.12 (.07) | 1572 | 0.68***(.08) |
Others | 354 | 0.67** (.16) | 50 | 0.80 (.33) |
Number of living children | None | 3134 | ref | 2590 | ref |
1–2 | 2619 | 0.99 (.09) | 1604 | 1.30* (.11) |
3–4 | 2447 | 0.73** (.11) | 685 | 1.27 (.14) |
5 and above | 2897 | 0.78* (.11) | 406 | 1.21 (.17) |
Household wealth Status | Low | 4565 | ref | 1149 | ref |
Medium | 3355 | 1.02 (.06) | 1194 | 1.17 (.09) |
High | 3177 | 1.07 (.06) | 2942 | 1.47***(.09) |
Marital status | Never married | 2227 | 1.54***(.09) | 2153 | 1.17 (.11) |
Currently in union | 7710 | ref | 2405 | ref |
Ever married | 1160 | 1.27** (.08) | 727 | 1.01 (.10) |
Intercept | | 0.35***(.12) | | 0.29*** (.20) |
Total | 11,097 | | 5285 | |
In urban areas, the likelihood of refusing wife beating was significantly higher in Addis Ababa, Dire Dawa, Oromia and Benishangul Gumuz regions. Numerically speaking, the attitude towards refusing wife beating is two times higher in Addis Ababa (
P < 0.001), 84% higher in urban Dire Dawa (
P < 0.001), 73% higher in urban Oromia (
P < 0.01) and 55% higher in urban Benishangul Gumuz (
P < 0.05) when compared to women living in urban Tigray that was taken as a reference (Table
2, Panel II). Nevertheless, women living in urban areas of Somali and Amhara regions have shown a significantly less likelihood of refusing wife beating. That is, their refusal attitudes had decreased by 40 and 34%, respectively (Table
2, Panel IV). What is more, women living in other urban areas of Afar, SNNP, Gambella and Harari regions had adopted similar attitudes towards wife beating as was observed in those women living in urban Tigray region.
In the vitality education was found to occur in direct relationship with attitudes towards refusing wife beating both in rural and urban Ethiopia. Explicitly, women with higher levels of education were 7.5 and 5.12 times more likely to refuse wife beating than those with no education in rural and urban areas, respectively (p < 0.001) (Table
2, Panels II and IV). Compared to women with no education, the likelihood of refusing wife beating had increased by 127 and 170%, respectively, among rural and urban women with secondary level education. The likelihood of refusing wife beating was also higher by 49 and 61% among women who attained primary level education in rural and urban areas, respectively (
P < 0.001) (Table
2, Panels II and IV).
Younger women (15–24 years) living in rural and urban areas were 27 and 17% less likely to refuse wife beating, respectively, (
P < 0.001) compared to women in peak reproductive ages (25–34 years). Only 12% rural women of older age (35–49 years) were in a position more likely to refuse wife beating (
P < 0.001). Similarly, only urban women with frequent access to media and who were living in better off households were 49 and 47% more likely to refuse wife beating (
P < 0.01) compared to women having no access to media and those living in poor households, respectively (Table
2 Panel IV). Rural women who were following Protestant and other religions such as Catholic and traditional beliefs were found 21 and 33% less likely to refuse wife beating, respectively; whereas, only 32% of Muslim women living in urban areas were less likely to refuse wife beating (
P < 0.01).
Increasing the number of living children decreases rural women’s likelihood of refusing wife beating; whilst having no marriage commitment (i.e. never getting married to and/or dissolving a previous marriage) tended to increase the chances of women’s refusal to wife beating in rural areas (Table
2 Panel II). Contrasting this finding, having one or two living children tends to maximise refusal attitudes towards wife beating in urban areas (Table
2 Panel IV).
Conclusion
Most of the studies in developing countries on wife-beating have focused mainly on the actual prevalence of physical violence and its determinants rather than the underlying attitudes towards wife beating. Unlikely, this study took the advantage of women’s responses about their attitudes towards refusing wife beating for any of the reasons mentioned in the 2011 Ethiopian Demographic and Health Survey. Having knowing the extent of the reasons for refusing wife-beating, the researchers believed that remedial mechanisms of changing the social and cultural practices of Ethiopian societies to abolish gender based violence, in general, and wife beating, in particular, could be devised. Moreover, understanding the underlying attitudes towards wife-beating refusal is supposed to assist in designing effective and efficient intervention programmes that could address this perilous subject. Some of the intervention mechanisms that should be included could be promoting gender equity and equality and IEC campaigns to enhance men’s involvement in the initiatives to abolish GBV including wife beating. Increasing women’s access to educational and employment opportunities are also believed to bring the most significant impacts. These not only can enhance the status of women that shall narrow the gender gap between couples who share household responsibilities and resources but also resist against the violations of their rights. In addition, there is a strong need to revise and strengthen national laws and policies that protect human rights, in general, and women’s right, in particular, with a major emphasis on eradicating gender based violence. Giving due consideration to traditional and customarily rules and regulations that promote gender equity and equality, and strengthening the support of religious institutions in fostering equal opportunities for all irrespective of sex could also help to achieve the targets of ending violence against women.
Limitations of the study
A detailed understanding of attitudes towards wife beating requires utilisation of a broad set of indicators that can capture multiple dimensions of women’s views about a woman’s hitting and beating by a husband. In this regard, indicators of perception towards wife beating used in this study may not catch the underlying causes and multiple dimensions of wife beating. In addition, the dependent variable was constructed based on hypothetical questions concerned with wife hitting and beating over certain particular causes, for capturing as many causes of conflict as possible would be unattainable. Likewise, the findings are believed to show factors associated to attitudes towards wife beating and its correlation with a set of socio-demographic, economic and cultural variables. This may enable interested institutions to design and implement proper intervention strategies to tackle and reduce the socially destructive power of masculinity. The study can also be used as an input for further research on the subject matter of interest in vitality.
Acknowledgements
We would like to thank the Central Statistical Agency of Ethiopia and ICF International for providing the data set used to undertake this study. We are also very grateful to Alelign Aschale for carefully editing the manuscript.