Explaining trends and patterns in attitudes towards wife-beating among women in Nigeria: analysis of 2003, 2008, and 2013 Demographic and Health Survey data

Theoretical framework and study hypotheses

Most studies show that collectively, patriarchal societies expect women to play certain conventional roles, such as birthing and caring for children, taking care of the house, being subservient to their spouses, and submitting to their husbands’ sexual needs (Hindin 2003; Oyediran and Isiugo-Abanihe 2005; Rani et al. 2004; Lawoko 2006; Lawoko 2008). In male-controlled cultures, changes in what women demand of their sexual partners by the women are cited as a key reason for marital violence (Watts and Mayhew 2004). This study builds on the work of Rani et al. (2004), in assuming in patriarchal societies, transgression on gender roles is seen by both men and women as a justifying condition for violence.

The social learning theory and the socio-ecological model provide the conceptual framework for explaining women’s perceived attitude of justifying wife-beating. This theory posits that people observe and model others’ behaviors, attitudes, and emotional reactions, particularly those of significant others at home, school, and community levels. Social values, norms, and gender roles in a patriarchal society are conveyed within social groups and diffused from generation to generation. Acceptance of wife-beating is premised on the patriarchal norm of male domination. The effects of changes and adoption of foreign values can alter the social learning process by encouraging individuals (including women) to challenge established social values and question those that cause harm/are no longer acceptable. Rani and colleagues posit that these enabling factors will work through three mechanisms by (1) producing a conflict between reality and the myth of male dominance, (2) exposing to more egalitarian social networks and social structures other than kinship, and (3) exposing communities to modern norms and values through the media (Rani et al. 2004).

Many Nigerian societies maintain rigid gender norms. Many social practices, such as uneven distribution of power in marital relations, polygamy, acceptance of male promiscuity, power of extended family over married couples, restriction of women’s mobility, and the almost universal prevalence of bride wealth, explain the persistence of male superiority and widespread violence against women in Nigerian culture.

Other factors may moderate or influence cultural support for marital violence in Nigeria. Enabling factors—increased access to education, employment in the formal sector, participation in public roles, including the political sphere, and improved fiscal responsibility within the household—may lead women to challenge or question social norms about gender roles. Both education and urbanization may increase the likelihood of women having non-supportive attitude towards wife-beating over time.

However, multi-level factors (at the individual, micro, exo, and macro levels) reinforce cultural norms that condone violent punishment of wives. At the individual level, limitations in education and access to employment represent disabling factors for women. At the micro (household) level, women’s lack of autonomy to make decisions about household purchases can put them at risk. Also, exposure to violence within the family and acceptance of wife-beating to resolve conflicts may perpetuate acceptance of such violence—even if enabling factors are present. For instance, if a woman increases her financial contributions to household expenses, the chance of conflict may increase, because these contributions could diminish men’s authority. Yet men who contribute the bulk of household income would be uninhibited about violence, because they hold fiscal dominance. Similarly, high female autonomy as measured by women’s participation in household decision-making should be associated with reduced acceptance of spouse-beating.

Interlinked factors at the exo level also underpin marital violence. The broader social, political, and economic environment (such as the limited availability of legal redress for battered women and women’s related failure to seek such support) reinforce violence at home. Exo factors also influence the structure and availability of microsystems that can influence the spousal relationship. Finally, normative social values (macro factors) are subject to social, political, and economic conditions (Bukatko and Daehler 2001) that can influence women’s perceptions of wife-beating. Spousal age and differences between spouses’ respective ages can also affect the gender power balance, in that an individual performs unique social roles at specific ages.

The multiple influences mentioned above suggest two study hypotheses about the effects of social learning and socio-ecological models on women’s attitudes towards wife-beating: (1) women who are more educated and who are urban dwellers will have lower odds of reporting that wife-beating is justified and (2) women who belong to ethnic groups or religions with more conservative values, or who live in regions dominated by such groups, or that are populated by groups tending to condone domestic violence, will have higher odds of reporting that wife-beating is justified.

Study design and data

The present study used data from the Nigeria Demographic and Health Surveys or NDHS: nationally representative sample surveys conducted in 2003, 2008, and 2013. The National Population Commission of Nigeria collected these data with technical assistance from Macro International as part of activities funded by the US Agency for International Development through the MEASURE DHS+ project. The surveys are designed to collect, analyze, and disseminate data about maternal health, sexual and reproductive health, and malaria among women aged 15 to 49, along with information on the demographic, social, and economic characteristics of these women and members of their households. The NDHS also examines attitudes towards gender roles, exposure to messages about health care in the media, and knowledge of and attitudes towards the use of reproductive health services.

The NHDS took nationally representative samples of women, covering both rural and urban households using stratified and multistage cluster sample designs. To examine trends in, prevalence of, and factors associated with acceptance of wife-beating, the analytic samples considered both ever-married and single women. The three survey datasets primarily contain individual sampling weights for women, which were obtained by further correction for women’s individual nonresponse based on household sampling weights (National Population Commission (Nigeria) and ICF International, 2014). The datasets were pooled and thus comprised 7620 respondents from the 2003 NDHS, 33,385 respondents from the 2008 NDHS, and 38,948 respondents for the 2013 NDHS.

Dependent variable

The NDHS collects data on women’s attitudes towards wife-beating through hypothetical questions about violence that could take place when a husband becomes angered by things that his wife does. The respondents were asked whether a husband would be justified for hitting or beating his wife in the following situations: (a) she goes out without telling him, (b) she neglects the children, (c) she argues with him, (d) she refuses to have sex with him, and (e) she burns food during cooking. A binary variable (yes/no) was created for each question. All “don’t know” responses, comprising between 2 and 5%, were treated as missing values.

From these dichotomous (yes/no) variables, an index representing respondents’ attitudes towards wife-beating was created, with a score range of 0–5. The index places individual women on a continuous scale of relative acceptance of the husbands’ right to beat his wife in at least one of the five scenarios. The index was then categorized into a binary variable as women who does not support husband beating their spouse under any circumstances and those who support wife-beating for at least one reason. A binary outcome variable was created for acceptance of wife-beating, coded as “0” if the respondent did not agree that a husband is justified in beating his wife in any of the situations and coded as “1” if the respondent agreed in at least one situation. A logistic regression model was used to examine the effects of selected demographic and socioeconomic variables on women’s attitudes towards wife-beating (as the outcome variable).

Independent variables

The independent variables, then, were the following: (a) age, (b) marital status, (c) educational status, (d) religion, (e) women’s autonomy in household decision-making, (f) wealth index, (g) number of living children, (h) place of residence, (i) region of residence, (j) exposure to media, (k) ethnic affiliation, and (l) spousal age difference. The inclusion of the covariates was informed by previous studies on factors associated with acceptance of wife-beating, especially in the context of sub-Saharan Africa (Hindin 2003; Rani et al. 2004; Ogunjuyigbe et al. 2005; Oyediran and Isiugo-Abanihe 2005; Lawoko 2006; Oyedokun 2007; Uthman et al. 2009).

Statistical analysis

Profile of the study population

About two thirds of the study population were currently in a monogamous union. Over half of women reported being involved in household decision-making in 2003 and 2008, though this proportion decreased dramatically to 37.8% in 2013: a surprising finding, given expectations of improvements across the continuum of women’s empowerment, but possibly due to continuing adherence to conservative values. About two fifths of the women resided in households categorized in the lower two quintiles of the wealth index; nearly two fifths had three or more living children. Nearly two thirds of respondents lived in rural areas, and more than half resided in the northern regions of Nigeria. There was no significant variation in the distribution of respondents by background characteristic across the three surveys.

Patterns and trends

In 2003, 62.4% of women reported believing that wife-beating would be justified for at least one of the scenarios presented in the DHS. In 2008 and 2013, 45.7 and 37.1% of women, respectively, reported that wife-beating would be justified for at least one of the five scenarios presented during the interview. Thus, the proportion of women supporting wife-beating declined by 25.3 percentage points over the period 2003–2013. Figure 1 reveals that the proportion of women believing that wife-beating is justified for each of the five reasons declined over the 10-year period. The trends for all five scenarios are similar in that they show a sharp drop in both 2008 and 2013, except for the “burning of food” category, which was stable between 2008 and 2013. Overall, the proportion of women reporting that wife-beating is justified for at least one of these reasons declined over time. For instance, about 24% of the women reported that wife-beating is justified for all five of the reasons in 2003, as compared with 14.1 and 11.1% in 2008 and 2013, respectively. Multivariate analysis of the trend showed that women’s support for patterns of wife-beating decreased significantly in Nigeria between 2003 and 2013.

Figures 2, 3, and 4 show regional variation in support for acceptance of wife-beating among Nigerian women in 2003, 2008, and 2013, respectively, when the wife goes out without telling her husband. Higher support was observed in the Northeast region in 2003, while a lower level of support were observed in the Southeast. In 2013, the lowest levels of support for the husband’s right to beat his wife if she goes out without telling him were observed in the Southwest.

In 2008, the proportion of women justifying violence against wives who argue with their husbands was similar across regions, although support was somewhat lower in the Southwest. However, the proportion of women who reported that a man is justified beating or hitting his wife if she goes out without telling him increased in 2008, compared with 2003. The patterns of women supporting wife-beating for any reason varied between 2003, 2008, and 2013 across the regions, as shown in Figs. 5, 6, and 7, respectively. Agreement that wife-beating is justified in at least one scenario is relatively low in the Southwest as compared with other regions. The percentage of women supporting a husband’s right to beat or hit his wife varies by region within the 10 years under review.

In Table 2, a chi-square test identified the following factors as having a significant association (p < 0.05) with women reporting that wife-beating is justified for at least one reason (consistent across each of the three survey years): (a) respondents’ age, (b) type of union, (c) educational status, (d) wealth index, (e) exposure to media (listening to the radio, watching television, and reading newspapers/magazines), (f) place of residence (urban versus rural), (g) region, (h) ethnicity, (i) religious affiliation, (j) spousal age difference, (k) autonomy in household decision-making, and (l) number of living children. The analysis reveals that the percentage of women reporting that wife-beating is justified also decreased significantly (p < 0.05) across the years for all subgroups of the study population. The patterns and trends are similar for ever-married, never-married, and all women in the three survey waves.

The multivariate models in Table 3 show the odds ratios estimating the likelihood that female respondents would accept that husbands are justified in hitting or beating their spouse under at least one circumstance. Models were estimated for ever-married, never-married, and all women, using pooled data from the 2003, 2008, to 2013 NDHS. These odds ratios from the logistic regression show the net effect of personal attributes on women’s perceptions of domestic violence, specifically wife-beating. The results of the multivariate model for all women clearly reveal that the estimated odds of women agreeing that husbands are justified in beating their wives were lower in 2008 and 2013 than in 2003. The patterns of women accepting wife-beating are similar in all three models in Table 3 for the survey years. After controlling for the influence of year, the model shows that the likelihood of supporting wife-beating decreased with age among ever-married women (p < 0.001). The results also show that the likelihood of women accepting men’s right to beat their wives increased with number of living children among ever-married women (p < 0.003).

The analysis further reveals that (a) education, (b) religious affiliation, (c) ethnicity, (d) place of residence, (e) region, (f) exposure to mass media, (g) household wealth quintile, and (h) female empowerment, as measured by participation in household decision-making, are significant predictors of women agreeing that husbands have the right to beat their spouses in the multivariate model for all women.

Diminished support for domestic violence among Nigerian women

The analysis revealed that in 2003, 62.4% of women surveyed perceived that husbands are justified in beating their wives for at least one of the reasons mentioned in the survey; this percentage declined in the subsequent surveys to 45.7% in 2008 and to 37.1% in 2013. Although the study focused on the acceptance of wife-beating among women, the consistency of the results with those of studies examining men’s attitudes towards wife-beating in 18 countries in Africa, Asia, Latin America, and the Caribbean (MacQuarrie et al. 2015) suggests that the current findings are relevant to broader issues of gender dynamics and normative values.

Patriarchal beliefs still prevail, remaining an important factor for the perpetuation of gender-based and especially domestic violence in sub-Saharan Africa, including in Nigeria. However, this analysis reveals evidence of a cultural shift, with a declining proportion of women believing that wife-beating is justified. Anecdotal evidence of this cultural shift also abounds. For instance, wives of political officeholders, including the spouses of the President and State Governors, have launched projects to address gender-based violence and support gender equality in Nigeria. These interventions seek to create and advocate for female empowerment at several levels: helping women play more significant roles in management and the professions, gain political influence within elected and appointed positions, and participate in household decision-making.

The findings of this study offer further evidence that in any society, normative values, including norms surrounding violence against women, are dynamic. In this regard, Nigeria is no exception. The study analysis shows that ever-married women support wife-beating more than do their never-married counterparts. This may be explained in the context of the institution of marriage in Nigeria, where the man is the head of the household and women who are married are strongly committed to maintaining the relationship regardless of the circumstances. The percentage of women justifying husbands’ right to beat their spouses has been declining in the past 10 years across segments of society.

The observed relationship between a woman justifying wife-beating and marital status is expected. It may be that ever-married women, having experienced domestic violence firsthand, might provide a response to the DHS that fits in with community attitudes. By contrast, never-married women may have better knowledge about rights, gender equality, and female empowerment, which could encourage this group of women to challenge traditional norms that require women to be submissive to their husbands. Never-married women are more likely to be educated than their ever-married counterparts.

The study results show a negative relationship between education and household wealth quintile and women’s acceptance of wife-beating. The results indicate that patterns of support for wife-beating decline with increasing level of education and increasing household wealth quintile. The findings correlate with MacQuarrie et al. (2015), who showed that men’s acceptance of wife-beating declines with increasing levels of education and household wealth quintile. A plausible explanation for these findings is that educated women and those living in households in a higher wealth quintile are likely to be in (or potentially be in) a more modern and egalitarian relationship than uneducated women and women at the lowest wealth quintile.

The analysis reveals that even given the general decline in support for wife-battery, older Nigerian women are less likely to support the practice than are younger women. Women’s acceptance of wife-beating declined with increasing age for both ever-married and never-married categories. This decline may be due to women’s better understanding of dyadic relationships and power play within the marital union as they grow older. Older women may also be more likely than younger women to have a good understanding of the norms and values that protect women within marriage or prevent men from being aggressive. For instance, most Nigerian cultures have mechanisms for rebuking men who are aggressive in formal or informal marital relationships. Younger married or never-married women may have been exposed to more recent messages that promote women’s agency; they may thus hinge their relationships on equality, which is a core value of women’s advocacy.

Geographic trends

Regional variations in attitudes towards wife-beating suggest that though violence against women in Nigeria is universal, it occurs in different forms and degrees from region to region. These results corroborate those of a study done in Zimbabwe (Hindin 2003), and these heterogeneous tendencies may be attributed to variability in ethnic or cultural norms and socioeconomic differences. For instance, the North-Central and South-South regions tend to be multicultural and exhibit variability in educational levels among different ethnic groups, while the Northeast and Northwest regions are more rural, dominated by specific cultural and traditional norms that are seen as vital. Exploring these disparities may require a model that captures multilevel, random, or contextual effects, such as ethnicity, level of community development, or other regional variables. Such models would be an interesting and worthwhile extension to the regression model used in this analysis.

Overall, this study reveals a declining trend in women’s acceptance of wife-beating, except in the Southeast region, an Igbo-dominated region where acceptance increased. This may be due to the Igbo tradition of payment of expensive bride’s wealth by the husband’s family to the wife’s, which supports the husband’s dominance and reflects a traditional gender hierarchy in which men are considered superior to women. The bride’s wealth payment among Igbo speakers may promote violence against women, since it symbolizes “loss of rights” by the bride and “transfer of rights” to the groom. Past study has revealed that Igbo women experienced more domestic violence as compared with their Yoruba and Hausa/Fulani counterparts (Tenkorang et al. 2013). Experience of wife-battery in this region may explain the increased support for wife-beating among Igbo women; expressing support for this practice during the survey may be a way of providing a response that aligns with community norms.

Religion and support for domestic violence

Religion, as an indicator of social control, appears to exert a different effect on married versus never-married women. Its effect on the dynamics of normative values, including gender-related values such as support for wife-beating, is changing over time (Norris and Inglehart 2002). The study found that in 2013, Protestant women were less likely than Catholic women to support wife-battery—a finding that has been observed elsewhere (for example in Burkina Faso: see Pambe et al. 2014). The liberal gospel and doctrines professed by Protestants may explain this association.

It should be noted that women who are Christian, whether Catholics or Protestants, are less likely than Muslim women to hold positive attitudes towards wife-beating (Obeid et al. 2010). The position of Islamic doctrine regarding violent acts against wives has mixed interpretations, partly due to the multiple factors that determine a women’s status and how these factors relate to patriarchal ideologies. However, the study’s findings on Muslim women’s views about wife-beating may be explained by Islam’s patriarchal position on women’s submission to their husbands. The effect of religion on violence against women could be interpreted as having the same effect as does culture. Beliefs and norms do change over time, and religion has an influence on both, which in turn have an impact on gender-based violence (Norris and Inglehart 2002; Rashid et al. 2014).

These results suggest a disparity in the attitudes towards wife-beating between Christians and Muslims in Nigeria. The reason for the North/South disparity cannot be fully determined from this analysis. Potential explanations might include the influence of religious traditions, teachings, doctrines, religious communities, and institutions that convey values and belief systems to their members. More research is necessary to determine the underlying causes of the difference.

Regardless, understanding how religious groups in Nigeria relate to gender-based violence will help guide policies and programs to reduce the burden and consequences of domestic violence against women. In addition, adherents to particular religions often receive direct support or counseling through relationships with religious leaders, whose guidance could influence support for wife-beating. Religious texts and teachings can serve as resources to shape perceived attitudes towards abuse and to manage episodes of abuse. Yet religion can also be misused to excuse or condone abusive behavior. In the context of violence against women, religious teachings and communities will always play a role and may never be neutral.

Limitations and further research

This study has limitations, some of which are inherent in the DHS datasets used and the quantitative approach, which enables only certain types of exploration of attitudes. Some of the variables used may raise selection bias; endogenous and appropriate techniques are required to model the relationships that may exist between the response variable and the explanatory variables. For instance, a mixed regression model that incorporates random effects may be appropriate. Random effects may capture some of the unobserved and unmeasured population effects that influence normative values. The use of a multilevel modelling technique would account for the hierarchical structure of the data, adjusting for estimated coefficients and other cluster-level sources of unobserved heterogeneity (variance) that may be associated with attitudes or perceptions towards justification of wife-beating or of domestic violence.

In addition to the quantitative approaches mentioned above, there is a need to collect qualitative data to explain the underlying context of the continuing acceptance of wife-beating among Nigeria women.

Traditional attitudes and behavior relative to gender appear to be changing in Nigeria, and this study suggests a steady decline in women’s acceptance of gender-based violence. Study findings reveal that demographic and socioeconomic characteristics have important and similar effects on cultural values, especially gender norms, that affect the attitudes of Nigerian women towards wife-beating—confirming the findings of other researchers working within the same theoretical framework (Hindin 2003; Oyediran and Isiugo-Abanihe 2005). These findings should inform the strategies of both policymakers and advocates for women’s empowerment.

The traditional gender attitudes and behavior may be changing in Nigeria. This is the first study to document trends in Nigerian women’s acceptance of gender-based violence at home. Acceptance has declined overtime.