Spousal violence is any behaviour within a relationship that causes physical, sexual and psychological harm including any act of physical aggression, sexual coercion and psychological abuse. Spousal violence can be from man to woman, and it can also be from woman to man. The patriarchal nature of the study population-Yoruba of Southwest Nigeria warranted this study to focus only on the violence perpetrated from man to woman. And for this reason, spousal violence is defined here as violence experienced by women at the hands of their current or earlier spouse(s). The term “spouse” includes legally married husband and those male partners who are not legally the husbands of the respondents but are living together, or have lived together with the woman as if married. The measure of spousal violence in this study does not include emotional violence; it is restricted to physical and sexual violence because emotional violence may not be easily measured objectively. Intimate partners violence occurs in all countries, irrespective of social, economic, religious or cultural groups [1].

Spousal violence is a violation of human rights and a public health problem which is associated with detrimental emotional, psychological, social and physical outcomes. It can lead to serious injury, disability or death. It can also lead to varieties of health problems; such as stress induced physiological changes, substance use, lack of fertility control and loss of personal autonomy as it is often seen in abusive relationship [2]. Women organization around the world has long drawn attention to violence against women and to intimate partners’ violence in particular. Through their effort violence against women has now become an issue of international concern [1]. Sexual violence is a major public health and human right problems in Botswana and Swaziland as reported by [3]. Abused women have higher rates of unintended pregnancies and abortion, sexually transmitted infections including HIV and mental disorders such as depression, anxiety, sleep and eating disorders. When this violence occurred during pregnancy, it is associated with adverse pregnancy outcome such as miscarriage, pre-term and still births [4,5].

Spousal violence may not only affect the women involved but may also damage the health and well-being of the children in the family. Studies have shown that the children of abused mothers have lower rates of immunization and higher rates of diarrhea disease and are more likely to die before the age of five years [6]. In the same vein, findings from a number of other studies shows that witnessing spousal violence can also negatively affect the normal development of children in the family [7]. A child who grew up in families where there is violence suffers a range of behavioral and emotional disturbances that can be associated with perpetration of violence later in life.

However, violence against women is increasingly recognized as a health issue in nearly all the countries of the world and attention is turning to the measurement of its health and other consequences for women and their families [8]. Although a large and growing literature exists on the correlates of spousal violence, a complete picture of the risk factor for violence is yet to emerge. Several variables have been linked with spousal violence, among which are women’s education and husband’s/partner’s education [9-12], wealth [13-15], use of alcohol [16,17], attitudes to violence [18], and women having a say regarding decision about their health care by the woman [19].

A South African survey of over 280,000 pupils shows that up till the age of 15 years, about 9% of both boys and girls reported forced sex in the past year. This rose to 13% for males and 16% for female by age 19 years [20], while a Nigerian study found that 45% female and 32% of male reported having had forced sexual intercourse at the age 15 years [21]. Patriarchal norms continue to relegate many women in Yoruba speaking areas of sub-Saharan Africa to a sub-ordinate position relative to men, thus providing fertile ground for men’s perpetration of violence against women. One study which covers 17 sub-Saharan Africa countries show that spousal violence was widely accepted under certain circumstances; “neglecting the children” was the most common reason reported, this is followed by “going out without informing the husband” and “suspecting a wife of being unfaithful” [22]. Heise [23] in an ecological approach to abuse conceptualizes violence as a multifaceted phenomenon grounded in interplay among personal, situational and socio-cultural factors. Although drawing on the conceptual advances of earlier theorist; Heise goes beyond their work in three significant ways. First, it uses the ecological framework as a heuristic device to organize the existing research base into an intelligible whole, whereas other theorists present the framework as a way to think about violence. However, the present study is interested in the influence of social and economic factors on spousal violence.

In the light of the above, the major goal of the present effort is to report the prevalence rate (ever experience and experience in the 12 months preceding the survey) of spousal violence among currently married or cohabiting women (15-49 years) in the study population. And to identify the key characteristics associated with experiencing spousal violence. This will further enhance the understanding of spousal violence and its correlates, and it will also guide the formulation of policy against this ugly menace especially in Africa south of Sahara. Attempts will also be made to compare the present estimates to the estimates from Demographic and Health Survey conducted in Nigeria in 2008 [24].

In clear terms the research questions for this study are as follows:

i. What is the prevalence rate of spousal violence in southwest Nigeria?

ii. What are the key characteristics associated with spousal violence in Southwest Nigeria?

iii. Has there been any significant departure in spousal violence estimates of the present study from the 2008 National Demographic and Health survey conducted in the country?

iv. What is the position of estimates in the present study to the estimates from the 2002 World Report on violence and health?

Data for this study was collected in Southwest Nigeria. Nigeria is located west of Africa (Latitude 5°N and 14°N, and Latitude 3°Eand 15°E), and it occupies an area of approximately 913, 768 km2 with a population of about 150 million people [25]. The country has, within the past five decades experienced tremendous changes as a result of the juxtaposition of the traditional and modern socio-economic mechanisms; such as a massive and rapid urbanization, industrialization and commercialization. The country is located north of the Atlantic Ocean and South of Niger Republic and Chad. It borders Cameroon in the east and Republic of Benin in the West. Southwest is one of the six geo-political zones in the country and it has six states, these are: Ondo, Oyo, Ogun, Osun, Ekiti and Lagos.

The sample size for the study consists of 300 currently married or cohabiting women within the age range of 15-49 years. The selection of the sample was on the basis of fifty respondents per state. The selection of fifty respondents in each state was done using multistage sampling technique based on the 2006 census enumeration area demarcation in the country. In preparation for the 2006 census in Nigeria, the country was demarcated into Supervisory Areas (SAs). There were over a thousand SAs in each of the states, each SAs was further demarcated into 10 Enumeration Areas (EAs). In the first stage, two SAs were randomly selected from each of the states, in the second stage; five EAs were systematically selected from each of the SAs. In the last stage five households from each of the EAs were systematically selected for participation in the survey. In other words, five households were selected from each of the EAs and this gave a total of twenty five households from each of the two EAs. This resulted in a total sample of fifty from each of the states. One currently married/cohabiting woman age 15-49 years in the selected household was eligible for the interview. Eligible respondents were identified through the household questionnaire and in situations where there were more than one eligible respondents in a household the lottery method was used to select one respondent out of the eligible respondents.

The interview took place in a confidential atmosphere, where complete privacy was maintained between the interviewer and the respondent. In situation where privacy cannot be obtained, the interview was postponed to a later agreed date.

Demographic and Health Survey (DHS) began collecting information on spousal violence in 1990 with the Columbia survey. The following year, DHS programme had developed a standard module and methodology for the collection of data on spousal violence. There are three core questionnaires in the DHS survey: the household questionnaire, women questionnaire and the men questionnaire. This study however limits itself to the adoption of the household questionnaire and women questionnaire. The men’s questionnaire was excluded because spousal violence as operationally defined in this study is violence experience by the women at the hands of men. This is the most common type of violence in the study population. Besides, men may not give the correct position of violence perpetrated by them. In addition to that, the patriarchal nature of the study population may not allow the men to perceive some of their actions as an act of violence against the women.

The household questionnaire was used to identify all the usual household members and visitors in the selected household, and to determine the eligibility of all household members for the individual women’s survey. The household survey also collected some basic information on the characteristics of each person in the household and on the household assets and amenities.

Measurement of spousal violence was done through the use of modified version of Conflict Tactics Scale (CTS) [26], which included questions that asked respondents whether their current or most recent husband/partner ever perpetrated any of a series of specific acts of spousal violence. Respondents who answered “Yes” to a particular item were then asked about the perpetration in the 12 months preceding the survey. Women who reported at least one of these acts were classified as having experienced spousal violence, while those who reported none of this act were classified in the “no violence” category. The series of specific acts included are: pushing, shaking, slapping, throwing things at the respondents, arm twisting, punching with the fist or something else that can hurt, kicking, dragging, burning, threatening or attacking with a knife, gun or other types of weapon. Also included are: being forced to have sexual intercourse even when the respondents did not want it and being forced to perform sexual act which the respondents did not want.

Other information collected from the respondents apart from the socioeconomic and the biological variables of age, age at marriage, and number of living children includes union status, women’s attitude to wife beating, childhood experience to violence and union status. Women’s attitude to wife beating was measured through description of five scenarios. The scenarios are: if the respondents go out without telling their husband, if she neglected the children, if she argues with the husband, if she refuse to have sex with the husband and if she burns food. Women who agreed with one or more of the above are regarded to justify wife beating. Exposure to childhood violence was measured by asking the respondents if they can recall whether their father ever beat their mothers. Respondents with such memory are regarded as being exposed to violence during childhood. They were also asked whether they are married or cohabiting to measure union status. And finally, respondents were asked if they make decisions regarding their health alone or they make it jointly with their husband. Data analysis was done using Statistical Packages for the Social Scientist (SPSS). This involves largely the generation of simple percentages and the estimation of logistic regression for the multivariate analysis. Logistic regression was employed in order to get the “net effect” of each of the variable on the relationship of interest.

Prevalence of spousal violence

The spousal violence indicator used in this study as earlier stated is derived from responses to a set of questions on violence based on Straus [26] Modified Conflict Tactics Scale (CTS). The respondents were asked whether their current or last husband/partner ever did any of the acts listed on Table 1. In an event where respondents listed more than one act, they were asked to indicate the most frequent act and this was recorded for the respondents. Table 1 shows the percentage distribution of respondents who ever experienced each specific act of violence by their husbands/partners and those who experience spousal violence in the 12 months preceding the survey. Out of the total sample of 300 women, 142 of them had ever experienced spousal violence, while 96 of them experienced spousal violence within the 12 months preceding the survey. This yielded a prevalence rate of 47.3% and 32% for ever and current experience of spousal violence respectively. As expected, the overall prevalence of spousal violence in the 12 months preceding the survey (32%) was substantially lower than women’s lifetime experience (47.3%).These rates were higher than what Sambisa et al. [27] reported for Bangladesh. They reported a physical spousal violence rate of 31% in the past year to the survey. However, the rates were higher than the estimates of 15% for ever experience, and 4% for current experience of spousal violence in Japan and it is lower than the estimates of 54% for Ethiopia as reported by Garcia et al. [28] in findings from WHO multi-country study on women’s health and domestic violence against women. The most reported acts of violence include pushing or kicking (31%), slapping and punching (15.5%), and arm twisting and throwing things at (14.1%). This also is in alignment with the report of Sambisa et al. [27] earlier cited.

Table 1: Percentage distribution of respondents by their ever experience and current experience of specific type of violent act.

The more severe acts of violence such as burning and threatening or attack with a knife or a gun is less common, or were less reported because in most cases victims of this forms of violence were likely to have been separated from such husbands/partners. About one in every twenty (4.9%) of the respondents reported that they were forced to perform sexual acts they did not want, while 12.7% reported that their husbands/partners forced them to have sexual inter course when they did not want. It is obvious from the table 1 that respondents reported more of physical violence (82.4%) than sexual violence (17.6%). The pattern is similar when the lifetime experience of the respondents is compared with their current experience. This is not surprising because Yoruba men are assumed to have unlimited sexual license over their wives. As such, it takes a lot of courage and self assertion for any wife/ woman to consider whatever assault she received from her husband/ partner as sexual violence.

The current and lifetime prevalence rate of spousal violence recorded in this study is indeed very high. This is higher than most recorded in history. In 2005, Kishor reported a prevalence rate of 17% for Cambodia, 29% for Haiti and 22% for Dominican Republic [5]. Nigerian Demographic and Health Survey (NDHS) (2003) did not collect any information on spousal violence but NDHS (2008) reported a prevalence rate of 28% and 7% for physical and sexual violence respectively for Nigeria. The disparity between physical and sexual violence reported in this study follows the pattern of the reported disparity in the 2008 NDHS.

Correlates of spousal violence

There is a large and growing literature on the correlates of spousal violence; a complete picture of the risk factor for violence is yet to emerge. Understanding the phenomena of spousal violence requires an analysis that goes beyond only the examination of the characteristics of the victim but also required, is the understanding of the characteristics of the perpetrators, the life experiences and behaviours of both the victim and perpetrator, the nature of the relationship of the couple, and the household and community context in which the violence occurs. Some of these variables are examined in Table 2.

Table 2: Percentage distribution of respondent by ever experience of spousal violence and socioeconomic variables (N = 142).

Table 2 shows the percentage distribution of respondent’s lifetime experience of spousal violence by their current husband/partner. From the Table 2, spousal violence is associated with the age of the women. The youngest women age group (15-19yrs) was least likely to experience spousal violence. The largest proportion (56.4%) of reported spousal violence was within age group 25-34 years. Experiences of spousal violence increases progressively with age; from age group 15-19 years to 30-34 years, after which age it began to decline. The pattern is similar to findings from WHO multi-country study which reported a rate of 48% for urban Bangladesh for women aged 15-19 years old within the past 12months; while the rate was only 10% for women aged 45- 49 years old [28]. This is probably due to the fact that as women age increases beyond age 34 years, their social status also increases and they become less vulnerable to spousal violence. On an equal note, it can be reasonably assumed that the longer couple stays/lives together, the more they understand themselves and the lower the risk of spousal violence. In the same vein, if it is reasonably assumed that women and their husbands are in the same age cohort, it can be hypothesized that younger men tend to be more violent than older men, and that violence tend to start early in many relationships.

Age at first marriage/cohabiting was also revealed to be associated with women’s experience of spousal violence in panel ii of Table 2. Respondents who were less than 20 years old when they first got married or started cohabiting were more likely to report spousal violence than those who were 20 years or older when they first got married. More than two in every five (43.7%) in age group 15-19 years had ever experienced spousal violence as against only 2.8% who experienced the same in age group 30 years and above. It is probable that a women who married at a later age has likely had the opportunity to pursue higher education, to be employed, either of which may enable the women to cultivate a sense of autonomy and can make her less vulnerable to spousal violence.

Women’s educational status was also found to be associated with spousal violence as revealed in panel iii of Table 2. Most respondents (72.4%) who reported ever experience of spousal violence are either in the “no formal education” or “primary education” categories. Women who had post secondary education are least likely to experience spousal violence. Education does not only liberate women from the claws of custom and tradition, it also has a protective value on the women, enhances the degree of control a woman has over her body and also increases her socio-economic status. Garcia et al. [28] reported that education have a protective effect on women even when controlling for income and age.

Employment status of the respondents was also related to the experience of spousal violence. Women who were unemployed were significantly more likely to experience spousal violence than women in the self employed category. About half of the respondents in the unemployed category have been victims of spousal violence sometime in the past, while only 16.9% of self employed respondents have ever experience spousal violence. This may not be unconnected with the fact that women who work are often assumed to be more empowered economically and by extension socially, and they may be less likely to experience spousal violence.

The relationship between the socio- economic status of the respondent’s household and the respondents experience of spousal violence was explored in panel v of Table 2. The socio-economic status of the household was measured using on index which was constructed using household asset data. These include: ownership of a number of consumer items ranging from television to a bicycle or car, as well as dwelling characteristics such as source of drinking water, sanitation facilities and types of flooring materials. Each asset was assigned a weight generated through principal component analysis. The resulting asset scores were standardized in relation to a normal distribution with a mean of zero [29].

Each household was then assigned a score for each asset, and the score were summed for each household. Individuals were ranked according to the score of the household in which they were interviewed. The sample was divided into quintiles from one (lowest) to five (highest). The index is consistent with expenditure and income measure [30] and it has been validated in a number of countries [31,32]. Women from the lowest 40% were classified as low, while those in the highest 40% were classified as high, and those in the 20% in between were classified as being in the middle socio-economic status. The result indicated that household socio-economic status was associated with respondent’s experience of spousal violence. Women from the low socio-economic household were most likely to experience the highest rate of spousal violence, while women who live in households with the highest socioeconomic status experienced the least (9.9%) violence. This is about six times (5.7) lower than the experience of 56.3% by women who lives in low socio-economic status households.

The number of living children that respondents had was also related to the risk of spousal violence. As revealed in panel vi of Table 2, women who had more than two living children were more likely to experience violence from their spouse/partner than women who had fewer children. Also women who had no children were more likely to experience spousal violence than women who had one child. The life time experience of spousal violence by women who had more than two children (45.1%) is about five times (4.6) higher than the experience of women with only one child, and about four times (3.9) higher than the experience of women who had no child. This can be explained by the fact that not having children is a potential source of conflict in African marriage, and also when the children are more than what the resources of the husbands can adequately take care, this can also generate tension and conflict.

Panel vii of Table 2 shows the relationship between union status and spousal violence. The panel revealed that cohabitation rather than being married is related to higher experience of spousal violence. The Table 2 indicates that 69% of women who are cohabiting have experienced violence from their partner. This is against 31% of their counterparts who are married. This is not surprising because Yoruba men may not accord a cohabiting partner the normal and expected type of respect due to a properly married wife. Marriage among the Yorubas of Southwest Nigeria is a unification of two extended families where mutual respect not only for the man and woman involved in marriage, but also in between the two families may make the couple less vulnerable to violence. This same trend was reported by WHO, more violence was reported among women who were cohabiting than those who were married. Comparable estimates of women who justify wife beating are 8% in Ethiopia, 70% in Peru, 55% in Tanzania and 30% in Thailand [1]

Panel viii of Table 2 revealed the relationship between women’s attitude to wife beating and spousal violence. Following the patterns of Demographic and Health Survey, five scenarios were described, and respondents were asked to indicate whether they agree or disagree that wife beating is justified. The scenarios are: (a) if she goes out without telling her husband, (b) if she neglected the children, (c) If she argues with her husband, (d) If she refuses to have sex with her husband and (e) If she burns the food. Respondents who agree with one or more of the above are regarded to justify wife beating. From panel viii of Table 2, respondents who agreed that wife beating was justified in at least one of these situations were most likely to experience spousal violence (62.7%) than women who did not agree with a single reason (27.3%). This is still a reflection of the patriarchal nature of Yoruba society where women more often than none are seen as object to be acquired and possessed by men. They are to be seen and never to be heard.

Husbands/partners consumption of alcohol was also strongly related to respondent’s experience of spousal violence. Respondents were significantly more likely to experience spousal violence, if their husbands/partners often got drunk (71.1%) than if they did not drink alcohol at all or never got drunk (28.9%). Alcoholic intoxication easily triggers aggression and this may partly explain the differences in the spousal violence experience of respondents whose husbands/partners drink and those whose husbands/partners do not. The World Report on violence and health complimented this position by stating that women who live with heavy drinker run a far greater risk of physical violence and that men who have been drinking inflict more serious violence on their partners. The report went further that women in Canada who live with heavy drinkers were 5times more likely to be assaulted by their partner than those who live with non-drinker [1].

Table 2 also evaluated the relationship between respondent’s experience of spousal violence and exposure to violence during childhood. This was explored in panel x. The women were asked if they recall whether their fathers ever beat their mothers. As evidenced from the Table 2, there is a relationship between respondent’s experience of spousal violence and exposure to parental violence. Respondents who were exposed to parental violence were significantly more likely to experience violence from their husbands/partners (69.7%), than respondents who were not exposed to parental violence (30.3%). Ordinarily, women who have childhood exposure to spousal violence may perceive it as a normal part of marriage relationship, and may therefore not see anything bad in it if they are having the same experience in their own marriage.

Women who make decision about their own health care jointly with their husbands/partners were significantly less likely to experience spousal violence compared with women who makes these decision on their own. From panel xi of Table 2, about 60% of women who make decision about their own health care alone had experienced spousal violence. This is against 12.7% for respondents who make decision with their husbands/partners. Making a joint health decision is an indication of mutual respect for each other, and to the husband/partner; it is a pointer to the fact that he cares about the health of his spouse. This may partly account for the disparity in the spousal violence experience of respondents in the three categories.

Multivariate analysis

Logistic regression analysis was employed to find out the net effect of each of the examined variables in Table 2 and to also establish the genuineness or spuriousness of some of the relationships on this same table. The result of the analysis is presented in Table 3.

Table 3: Logistic Regression Estimates of respondents ever experience of spousal violence (Presented as odds).

In the Table 3, a set of modeling blocks approach was used where each set of covariates were entered in stages. There were three stages in all. Stage 1 was a straight forward relationship between the respondent’s biological characteristics of age, age at first marriage and the number of living children on one hand, and ever experience of spousal violence on the other as the independent variable. In stage 2, the social variables of educational status of the respondents, employment status and the socio economic status of the household were introduced into the relationship. And in stage 3, the exposure variables of husbands/partners alcoholic consumption, women who make decisions about their own heath, union status, exposure to violence during childhood and women attitude reflecting the acceptance of men’s right to beat their wives were introduced. This approach allows the examination of the extent of confounding between the blocks of factors by observing the manner in which each subsequent factor affects the relationship between the variables entered in earlier stage and women risk of violence. The data presented for this analysis are odds ratios and their respective p-values. The total sample for the survey was included in the analysis.

The estimates of the regression model in stage 1 show that the biological variables of age, age at first marriage/cohabitation and number of living children are all individually strong determinants of spousal violence. It is sufficient to say that the three variables were measured as continuous variables and the result is consistent with the findings on Table 2 regarding the three variables. The odd ratio for age indicates an inverse effect, the higher the age of the women, the less likely the women experience spousal violence. The odd ratio revealed that with each additional year, the likelihood of spousal violence experience is 1.21 times lower than the previous year. The result for age at marriage shows a similar pattern though with a lower odd ratio. Number of living children indicated the expected positive relationship; the odd ratio of 1.51 implies that with each additional child, the risk of women’s experience of spousal violence increases by 1.51 times. The result for number of living children is significant at 5%, while the result for age and age at marriage are of no statistical significance.

Stage 2 shows the effect of adding the social variables of women’s education, women’s employment status and socio-economic status of the household into the relationship. The inclusion of these variables does not alter the significant effects of the biological variables earlier included, although there were slight changes in their odd ratio. Of particular interest is the change observed in the odd ratio for number of living children, where the level of significance increases from 5% to 1%. The expected inverse relationship between women’s educational status and spousal violence was noticed in stage 2. Women who had post secondary education are 2.91 times less likely to experience spousal violence than women who had no education. The same pattern was displayed in the odd ratio for women employment status though with reduced odds. Women who are self-employed are 1.18 times less likely to experience spousal violence than women who are unemployed. The result for socio-economic status of the respondent’s household was not surprising. Women in the high socio-economic status are 2.73 times less likely to experience spousal violence than women who are in the low socio-economic status household. All the odds are significant at 1% level and they are in conformity with the findings on Table 2.

The introduction of the exposure variables of husbands/partners alcoholic consumption, women who makes decision about their own health, union status, exposure to violence during childhood and women’s attitude reflecting the acceptance of men’s right to beat their wives were introduced into the relationship in stage 3. The addition of these variables does not alter in any significant way the effects of the biological and social variables earlier introduced in stage 1 and 2 respectively. Nevertheless, there were slight changes in their odd ratios but the level of significance remains largely unchanged. As expected and in conformity with the results on Table 2, women whose husbands/ partners consume alcohol have a higher odd of exposure to spousal violence than women whose husbands/partners do not drink alcohol. The odd ratio of 1.72 implies that women whose husbands/partners drink are 1.72 times more likely to experience spousal violence than women whose husband/partners do not drink, although the result is not statistically significant. In the same vein, women who make decision about their own health care are also 1.72 times more likely to experience spousal violence than women who make decision about their own health jointly with their husbands/partners.

The relationship between union status of respondent and spousal violence was also evaluated in stage 3. Women who are married are 1.11 times less likely to experience spousal violence than women who are cohabiting with their partner. Exposure to violence during childhood and women attitude reflecting the acceptance of men to beat their wives showed expected pattern, and at a very high significant level. Women who are exposed to parental violence as a child are 2.98 times more likely to experience violence from their husbands/partners than women who do not have such exposure. Women whose attitude is favorable to wife beating is 3.01 times more likely to experience spousal violence than women who are not favorably disposed to such attitude. The two results are significant at 1% level.

The results on Table 3 is largely a confirmation of our findings on Table 2, and it goes a long way to suggest the important position of number of living children, women’s education and socio-economic status of women’s households in the determination of spousal violence in the study population. Other variables that are also significant are: union status, women’s attitude to wife beating and women’s exposure to violence during childhood. However, it can therefore be hypothesized that women who have ever experience spousal violence differ significantly from women who have not experience spousal violence because they have fewer number of children, they are educated, they belongs to higher socio-economic status households, they have been properly married, they are not favourably disposed to wife beating and they have not been exposed to violence during childhood.

Violence in the 12 months preceding the survey was noticeably lower than violence ever experienced in the current relationship. The analysis was limited to the current relationship to forestall under reporting often associated with memory lapses. Physical violence was also far below sexual violence. Kicking/pushing is the most frequent act of physical violence reported (31%); this is closely followed by slapping (15.5%) and arm twisting/throwing things at (14.1%). Sexual violence is less reported not because it is less common but probably because the issue of sex in Southwest Nigeria is still regarded as a taboo and should not be openly discussed.

However, on the key characteristics associated with spousal violence in southwest Nigeria as contained in the second research question in this study. Young age at marriage appears to be a risky factor for being a victim of spousal violence in this study. This is in agreement with the submission of Romans et al. [33] in a study they conducted in Canada and also of the study of Sambisa et al. [27] in Bangladesh. This is also in agreement with the findings from WHO multi-country study on women’s health and domestic violence. The study reported that younger women especially those 15-19 years were at a higher risk of ‘current’ (within the last 12months) physical or sexual violence or both. Age is a biological variable that has to do with maturity. The responsibilities associated with marriage for a woman is so enormous in Yoruba speaking area of Southwest Nigeria that it would require a lot of maturity which can only come with age to cope. Another correlate identified in the study is the women’s levels of education. Ackerson et al. [11] in a study he conducted in India reported that low level of education is however, the most consistent factor associated with spousal violence. William also reported that the risk of spousal violence is lower among women with post secondary education. A higher level of education may act as a protective factor against women exposure to spousal violence. This is justified both in Tables 2 and 3 of the analysis. Education is an eye opener, it guarantees the women some level of self autonomy, and it also empowers them to be economically independent. Hence educated women commands a little respect especially when the husband/partner is less educated.

Heise et al. [34] in a study conducted for the World Health Organisation shows that spousal violence cut across all socioeconomic groups but that women living in poverty are disproportionately affected. This position is also supported in the present analysis. The inverse relationship between socio-economic status of the women’s household and exposure to spousal violence is indicated in the analysis. However, it is not clear why poverty increases the risk of this form of violence; whether it is because of the low income itself, or because of the other factors that accompany poverty, such as overcrowding and hopelessness. In some cases living in poverty may likely generate stress, frustration and a feeling of inadequacy for having failed to live up to their culturally prescribed role of providers. Poverty may also provide ready material for marital disagreement.

Moreover, women’s favorable attitude to spousal violence is also indicated in this study as a determinant of spousal violence. One of the prominent theories that explain perpetration of spousal violence is the maintenance of patriarchy and male dominance within a society [35], hence women acceptance of the justification of wife beating. Patriarchal male dominance norm reflects gender inequality, and legitimizes spousal violence against women. This is especially true of Yoruba speaking society of southwest Nigeria, this is also true of many societies of Africa south of Sahara. This inequality creates power hierarchies where men are perceived in the society as economically and religiously superior, and of higher status compared to women [36]. As such, men are socialized to believe they are superior to women, should dominate their partners and endorse traditional gender roles. Women subordination and submission is then considered to be normal, expected and accepted. This gender inequality and male dominance reduces the opportunity for women to be involved in decision making at every level, decreases the resources available to them and increases the acceptance of the use of violence against women.

However, women who reported a history of spousal violence were significantly more likely to experience spousal violence than women who reported no such history. This indeed is still a reflection of the patriarchal nature of Yoruba speaking society of southwest Nigeria where “father hitting mother” is more of a normal behaviour than an exception. Childhood exposure to parental violence, especially female child reinforces the belief in the child the right of the man to exhibit violent acts towards his wife/partner. Hence the situation is assumed to be normal when the child eventually becomes a victim later in life when she is married.

The prevalence rate of spousal violence in the study population is very high and having established in the literature that spousal violence can lead to serious injury, disability or death. It can also lead to a variety of health problem, such as miscarriage, abortion, mental disorder etc. In the like manner, spousal violence has been associated with detrimental emotional, psychological and physical outcome. It is therefore expedient that efforts should be directed towards its amelioration if not complete eradication. In the light of this, the following policy is proposed by this study.

There should be social and economic changes which empower women to take their rights. Child marriage should be protected by law and the education of girl child should be made compulsory. This will not only increase the age at which women gets married, but will also empower them economically to protect themselves against violence. Policy makers must examine the social and cultural practices which undermine gender equality. The patriarchal nature of Yoruba society should be demystified. Boy and girl child should be socialized to see themselves as not only equals but that they also have equal opportunities in life. The idea of typifying different sets of toys for the boy and girl child should be discouraged.

Policy makers must identify effective ways of raising awareness about spousal violence and develop supportive social structure for victims. There should be a general awakening of women to take control of their bodies and lives. The women non-governmental organizations should take up this responsibility. And good parenting skill should be integrated into the existing maternal health programme. Social, political, community and religious leaders should be enlisted in speaking out against spousal violence, while micro finance programmes for women financial empowerment should be introduced.