The Incidence of Wife Beating in Malawi: Do Male Attitudes Matter?

Abstract

Understanding male motivation for domestic or intimate partner violence is a vital component within the logic of change for reducing the incidence of abuse. This article analyzes male justifications for domestic violence using data from the Malawi 2015 Demographic and Health Survey. Five justifications for wife beating are recorded in the male survey, and we define agreement with any indicator as indicative of a justification of wife beating. Binary logistic regression is used to link both attitudinal and behavioral factors (alcohol consumption) to women's reports of wife beating. Male justification is only weakly related to the actual incidence of violence within the relationship: there is no association between attitudinal justification of wife beating and reported violence even in the absence of control variables. In contrast, alcohol utilization is strongly associated with the incidence of violence and is robust to controls. This suggests that directly targeting males through interventions designed to change attitude may have only limited success in reducing the incidence of intimate partner violence, and that behavioral change in the form of an antialcohol strategy would be a more productive intervention.

Introduction

Spousal or intimate partner violence is a widespread and persistent problem globally. Roughly one third of women globally have ever experienced violence of one form or another (Stöckl et al. 2014). Experience of wife beating is a strong predictor for multiple negative outcomes such as homicide or suicide, mental health problems such as depression, alcoholism or post-traumatic stress disorder, increased risk of miscarriage, abortion or HIV infection (Andersson et al. 2007; Campbell 2002) as well as difficulties in walking, memory loss, and dizziness (Ellsberg et al. 2008).

There is considerable variation in the incidence between countries (Stöckl et al. 2014), with 42% of women in Malawi having experienced wife beating from their current spouse (Office/Malawi, National Statistical, and ICF 2017). The form of this abuse in Malawi varies: 20% report physical violence such as being slapped, shaken, or punched, 13% some form of emotional violence, 13% sexual violence such as assault or rape, and 3% very severe violence such as being attacked with a weapon, strangled, or burned (Bazargan-Hejazi et al. 2013). These different forms of abuse are not exclusive and will tend to co-occur in different ways: there is evidence of clusters of controlling behavior, emotional abuse as well of more general abusive patterns incorporating both violence and sexual abuse (Chikhungu et al. 2019).

Multiple reasons for the persistent prevalence of domestic violence have been offered. Attitudes regarding domestic violence are not surprisingly a key predictor of the incidence and expression of violent behavior (Archer and Graham-Kevan 2003). This is couched within traditionally gendered expectations: early marriage, especially for teenage brides and those in polygynous unions are more likely to be accepting of wife beating (Rani et al. 2004). Other traditional union formation indicators, such as dowry or bride price, are also strongly predictive of wife beating.

As well as reinforcing economic power dynamics whereby male control over economic resources affects economic autonomy (Anderson et al. 2007), the autonomous status of such provision is predictive of wife beating: female control protecting against the incidence of abuse (Pallikadavath and Bradley 2019). Autonomy in general is associated with lower level of wife beating: greater female control over household decisions tends to reduce the incidence of wife beating (Oyediran 2016), whereas patrilineal inheritance structures are associated with higher rates of wife beating in Malawi (Telalagic 2012). That said, female paid employment tends not to ameliorate the effect of economic disparity or autonomy: indeed females who work for wages tend to experience violence to a greater extent (Lenze and Klasen 2017). This is potentially explained as a departure from “idealized” wife forms.

Indeed, the transgression of traditional female gender roles is often cited as an acceptable reason for wife beating (Oyediran 2016; Rani et al. 2004) or reinforcing masculine status (Jewkes 2002), for instance burning food or neglecting child care (Uthman et al. 2009). Such punishment is seen as a legitimate form of discipline, reinforcing sexist gender roles and sanctioned under these circumstances (Glick et al. 2002; Rani et al. 2004) with sexist attitudes promoting ambivalence to such violence (Glick et al. 2002). The discipline aspect of wife beating is often internalized: leaving the house without permission or talking back to the husband are commonly cited by both males and females as justification for wife beating (Uthman et al. 2009).

Male attitudes on whether violent acts to their partners is justified or not may be attributed to culture, and cultural factors have been shown to play a part in violence against women in Malawi and elsewhere (Alio et al. 2011; Bisika 2008; Chikhungu et al. 2020; Hyde-Nolan and Juliao 2011). The strong influence of permissibility of wife beating has naturally led to policy attempts that affect attitudinal change with a view to reducing wife beating focusing on attitudinal change. These attempts at attitudinal change operate through a number of channels, including leveraging community and social attitudes through peer groups and institutions such as churches, peer networks (Shwachman Kaminaga 2017) as well as more direct interventions such as mass media campaigns (Flood and Pease 2009).

In Malawi, Action Aid is involved in training men and boys to develop their knowledge on human and women rights (ActionAid 2016). Such initiatives may remove the inferiority complex that has been linked to higher levels of violence perpetrated by men toward their income earning partners (Ahinkorah et al. 2018; Chikhungu et al. 2019; Sohini 2016), supported by the resources theory that proposes that men with higher resources command higher power and those with less resources use violence to maintain power (Warner et al. 1986).

The efficacy of these interventions and campaigns is mixed however; although mass media has been associated with some of the largest effects in program outcome, some analyses suggest that in certain contexts campaigns are associated with an increase, not decrease, in wife beating (Okenwa-Emegwa et al. 2016). Furthermore, the direction of the relationship between women's working status and violence against women may be context specific (Bamiwuye and Odimegwu 2014). Behavioral change interventions for tackling violence against women in Malawi have not directly addressed the problem of alcohol abuse although the Malawi National Plan of Action to Combat Gender-based violence of 2014–2020 recognizes the role of alcohol intake in perpetuating gender based violence (Malawi Government 2014).

Alcohol is a significant predictor of wife beating, and interacts with both attitude and social context (Greene et al. 2017; WHO 2018). The intersection of both acceptance of male violence and the use of alcohol is a key part of the normalization process, whereby persistent violence associated by alcohol use is misconstrued to be a special circumstance (Okenwa-Emegwa et al. 2016). This marginalizes the identification of violence as an endemic problem. This article explicitly operationalizes this relationship: we initially model the relationship between attitude and the incidence of expressive male violent behavior and the effect this has on the incidence of wife beating. This offers a test of two major theoretical explanations for the causal relationship between alcohol and the incidence of wife beating (Greene et al. 2017); the proximate effect model that offers the explanation that alcohol directly influences the likelihood of spousal violence by lowering inhibition or behavioral function (Steele and Josephs 1990), or by impeding the ability to identify or process cues such as sexual refusal (Abbey et al. 2004). Indeed, the effect of alcohol consumption on raising the incidence of wife beating is well documented (Greene et al. 2017; Hindin et al. 2008).

An alternative explanation is the indirect effects model, which suggests that alcohol use is causally related to the incidence of wife beating but operates contingent on other factors, such as underlying relationship characteristics or attitudinal beliefs (McKenry et al. 1995). Our analysis examines both of these explanations, testing not only the effect of alcohol and male attitude, but also the interaction between them, operationalizing both proximate and indirect effects of alcohol. Establishing the model of change is vital in terms of designing policy to reduce the incidence of wife beating; programs should be targeted at the causal factor. We, therefore, examine the association between male expression of situations under which wife beating is justified, and the female report of having been beaten by their spouse. This allows us to establish the effect of attitude on behavioral outcomes. By including reported alcohol consumption we are able to establish both the direct influence of alcohol on violence, and quantify the interactive effect establishing where attitudinal expression is altered by alcohol use.

Data

Data for this analysis are drawn from the 2016 Malawi Demographic and Health Survey. The United States Agency for International Development funds the Demographic and Health Survey (DHS) program in many developing countries. In the 2015 Malawi Demographic and Health Survey, ICF International provided technical assistance throughout the study. The sample was selected based on the 2008 Malawi Population and Housing Census frame. Standard enumeration areas were selected based on probability proportional to size. Thirty households were picked from each enumeration area to make a sample size of 27,516 households.

The relevant questionnaires for this analysis are the men's questionnaire and the women's questionnaire. The men's questionnaire collected information from 7478 out of 7903 eligible within the enumerated households, with a response rate of 94.6. The woman's questionnaire collected information from 24,562 women out of 25,146 women aged 15–49 years who were eligible for the interview representing a 98% response rate.

One third of the sampled households received domestic violence questions (6379 households). The domestic violence module collected data on different types of violence: physical, sexual, and emotional, and level of marital control by husband. Men's responses were matched to women's (and domestic violence module) responses through household identifiers, and we, therefore, exclude all cases from the women's sample where there is no (a) domestic violence module and (b) no valid male response. We use female reports of domestic violence since these are considered to reduce the level of potential underreporting of abuse compared with men. Our final analytic sample is, therefore, 3784 male–female pairs.

The survey also provided information on background characteristics, including household wealth status and demographic characteristics of all household members such as age, sex, and relationship to household head. Sample weights were included to adjust for differences in the final probability of selection at household level. Further details of study design and data collection are reported on the National Statistical Office of Malawi website: http://www.nsomalawi.mw/.

Method

The DHS collects data from males on the circumstances under which they believe wife beating or wife beating is acceptable. Five indicator variables are available (wife goes out without telling husband, wife neglects the children, wife argues, wife refuses to have sex, and wife burns food), which take the value 1 if the male respondents believe that beating is justified under these circumstances, and zero if not. These responses are aggregated into an overall indicator, which takes the value 1 if the male agrees with any of the five indicators, and zero only if they disagree with all five. Binary logistic regression is used to link the indicator of male attitudes to whether the female partner reports having ever experienced wife beating. Female reports are derived from the individual questionnaire of the DHS, and are linked to male attitudinal indicators by the indicator of the person number of the spouse within the household. Having ever experienced wife beating is classified as having ever experienced either severe or less severe violence (as defined by the DHS) by the current husband or partner, and takes the form of a binary indicator variable $y_{i} \{\begin{matrix} 0 = H a s n o t e x e p r i e n c e d a b u s e \\ 1 = H a s e x p e r i e n c e d a b u s e \end{matrix}$ . We use a logistic regression model to model the probability of a woman reporting domestic abuse defined as $π_{i} = P r (y_{i = 1})$ , with the model defined as $l o g i t (π_{i}) = x' β$ . The vector x has all variables of interest and control variables, whereas $β$ contains the estimated coefficients.

The following model building strategy is used to build the analytic model. Model I specifies the probability of having experienced domestic abuse as a function of indicator of whether male partner believed wife beating is justified—this allows us to establish whether male attitudes are related to the incidence of domestic abuse. Model II introduces a term for the female report of whether the male partner consumes alcohol, allowing us to establish whether the effect of male attitude is affected by relevant behavioral factors. Model III introduces and interaction term between the male attitude variable and the use of alcohol. This allows us to examine whether attitudinal patterns are moderated by alcohol consumption. Model IV introduces controls for age group, educational attainment, wealth quintile, and region of residence to determine whether the results from models I–III are robust to background characteristics. The authors complied with all guidance for the ethical use of secondary data for DHS and the University of Portsmouth University Ethics Committee.

Findings

The results of the regression analysis are presented in Table 1. Model I assesses the relationship between the male attitude variable and the incidence of domestic abuse. The model finds no association between male attitude, with the coefficient for agreement with any reason for wife beating nonsignificant. Model II adds a term for whether the male respondent usually consumes alcohol. This is strongly associated with the incidence of domestic abuse, with the probability of domestic abuse estimated at 334% higher for women whose partners consumed alcohol regularly compared with those where the partner did not consume alcohol and this effect was significant at the 1% level.

Table 1.

Results of Binary Logistic Regression

	Model I		Model II		Model III		Model IV
	Odds ratio	Confidence interval for odds ratio	Odds ratio	Confidence interval for odds ratio	Odds ratio	Confidence interval for odds ratio	Odds ratio	Confidence interval for odds ratio
Violence justified (ref = no agreement)
Agreement	1.27	0.96–1.69	1.15	0.85–1.55	1.24	0.82–1.87	1.10	0.81–1.49
Alcohol consumption (ref = no)
Yes			3.34^**	2.82–3.95	3.38^**	2.83–4.03	3.30^**	2.77–3.92
Interaction
Agreement × yes					0.86	0.47–1.55
Age group (ref = 15–19)
20–24							1.48^*	1.02–2.15
25–29							1.67^**	1.14–2.44
30–34							1.60^*	1.09–2.35
35–39							1.47	0.97–2.23
40–44							1.53	0.97–2.23
45–50							1.33	0.78–2.27
Region (ref = north)
Central							0.66^**	0.52–0.84
South							0.78^**	0.62–0.97
Education (ref = none)
Primary							1.02	0.78–1.34
Secondary							0.65^*	0.46–0.91
Higher							0.22^**	0.07–0.65
Wealth quintile (ref = poorest)
Poor							0.98	0.75–1.28
Middle							1.06	0.81–1.39
Rich							1.01	0.72–1.34
Richest							1.01	0.75–1.34

p < 0.01, ^*p < 0.01.

Model III adds an interaction term to the model, allowing the effect of alcohol to be affected by male attitude. The main effect of attitude is not significant, nor is the interaction term. Alcohol usage retains its significance, with the odds of having experienced domestic abuse 338% higher for women whose partner consumes alcohol regularly than among those with partners not consuming alcohol. This indicates that the effect of alcohol is significant, but homogenous with respect to male attitude.

Model IV introduces control variables. Owing to the nonsignificance of the interaction term in Model III, this was dropped from this final model. Again, male attitude is not significant, but alcohol is significant at the 1% level, with the probability of domestic abuse 330% higher where the male partner consumes alcohol. A number of background variables are significant as well. Compared with the references category, women in all older age groups were more likely to have experienced domestic abuse, with women aged 20–24 years 48% more likely to have experienced abuse, women 25–29 years 67% more likely and women 30–34 years 60% more. There are also strong regional effects, with women living in the Central region 34% less likely to experience abuse than those in the Northern region, and women in the South 22% less likely to experience abuse. Education plays a protective effect, women with secondary education 35% less likely to experience domestic abuse compared with women with no education. Women with higher education are even less likely to experience abuse, with the probability of abuse 78% lower for women with higher education than women with no education. There is no significant effect of wealth quintile in the presence of other controls.

Predicted probabilities in Figure 1 indicate the stark contrast in the difference between the attitudinal response and the use of alcohol. The predicted probability of experiencing abuse where the male spouse expresses agreement with any reason for wife beating is 27.4%, compared with 22.5% where no agreement is expressed—a difference of 4.9% points. In contrast, the predicted probability depending on whether the male spouse shows a drastic difference between women whose spouses did not consume alcohol (16.0%) and those who's partners regularly consumed alcohol (39.1%)—a 23% point gap.

FIG. 1.

Predicted probability of wife beating by (a) male agreement that violence is justified and (b) male alcohol consumption.

Conclusions

This article aimed to establish association between male attitudinal justifications of wife beating, and to link these to behavioral outcomes. There is no evidence of agreement with any of the reasons for justifying wife beating affecting the incidence of abuse. None of the regression coefficients in the main or interaction effects models are significant, even in the absence of control variables. This is of particular policy relevance, suggesting that efforts to reduce the incidence of wife beating through promoting attitudinal change have only a limited scope in terms of actually achieving their aims of reducing the incidence of wife beating. Indeed, previous literature has noted the limited (or worse, harmful) effect of programs designed to improve attitude as a means of reducing wife beating (Flood and Pease 2009; Okenwa-Emegwa et al. 2016). Alcohol consumption, in contrast, remains a persistent and highly predictive indicator of whether a woman experiences spousal abuse or not. Alcohol consumption alone is the important factor: the lack of an interactive effect indicates that alcohol consumption is not moderated by greater disapproval of wife beating. Instead, there is a consistent and strong relationship: consumption of alcohol greatly increases the risk of experiencing violence to the female partner.

This is highly informative for policy. Although considerable efforts have been directed at reducing the acceptability of wife beating, primarily directed at attitudinal change through programs such as ActionAid (2016), our findings suggest that this is not likely to be a highly effective route to reducing behavioral outcomes in the current Malawian context. Rather, focusing on the behavioral predicates of violence—such as alcohol consumption—may be more effective.

Discussion

The lack of association between attitude and wife beating was key result. Other research has shown an association between attitude and the incidence of wife beating and indeed this forms part of the indirect effect model (Greene et al. 2017; McKenry et al. 1995). The lack of any significant interaction effect in our analysis would be sufficient to falsify or at least indicate no evidence for the indirect effect model; the lack of any association at all is more surprising, indicating that there is no effect of male attitude on wife beating whatsoever. All causation then runs through the behavioral channel: alcohol is strongly associated with abuse, a relationship that does not disappear in the presence of control variables.

There are some potential reasons for the differing nature of the relationship in our research. It should be noted that the prevalence of agreement for justifications of wife beating is small. It is entirely possible that many males are not being honest regarding the circumstances under which they would beat their spouse. Even within the Malawian context where wife beating is common, there is some stigma against expressing justification for wife beating—although reported rates are high, only 7.6% of men report that wife beating is justified under any circumstances (Office/Malawi, National Statistical, and ICF 2017)—and there is a strong possibility of misclassification, with social desirability bias depressing the proportions of men prepared to admit that they believe wife beating was justified.

Although the results of our analysis provide a valuable insight into the role of alcohol and the direct effect on wife beating, it should be noted that there are some limitations on the applicability of this finding, and that wider research is needed before fully moving to policy. In particular, there is considerable variation in both the level of wife beating and the interaction between wife beating and alcohol both at a cross national (Greene et al. 2017; Uthman et al. 2009) and even within Malawi itself linked to patrilineal customs (Berge et al. 2014; Telalagic 2012). Indeed, alcohol usage as a predictor of wife beating is couched within social and community context: there are significant effects of the prevalence of alcohol consumption at an aggregate/community level even controlling for individual level consumption (Greene et al. 2017) indicative of wider social markers of masculinity. The use of community and social institutions (Flood and Pease 2009) as well as peer networks (Shwachman Kaminaga 2017) therefore need to be integrated into tackling wife beating. This is important in cognizant of the fact that alcohol consumption on its own does not lead into aggression, but aggression is an outcome of the combined effect of provocation (instigation) possession of aggressive traits (impellors) and an impediment of self-regulatory capacity (inhibition) as purported by the I³ theory (Eckhardt et al. 2015).

The need to specifically tailor programmatic interventions (Okenwa-Emegwa et al. 2016) applies not only within the geographic and cultural context but within the gender context as well. Although this article has focused on male attitudes, it should be noted that the role of female attitude is a vital component in understanding the acceptability of wife beating. This is particularly important in the Malawian context, since female justification tends to be higher in the sub-Saharan region (Sardinha and Catalán 2018; Tran et al. 2016). Indeed, where gender symmetry has been examined, it has been found that women are often more likely to justify wife beating than their male partner (Uthman et al. 2009). Therefore, although male attitudinal change and education have previously been seen as a route for reducing the incidence of wife beating, the lack of effect in this study as well as the relatively high proportion of female acceptance suggest that involving men alone will have only limited success.

Footnotes

Authors' Contributions

M.A. was responsible for the conceptualization of the article, data analysis, and article preparation. L.C. was responsible for the conceptualization of the article and its preparation.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Abbey

, Zawacki

, Buck

, et al. (2004). Sexual assault and alcohol consumption: What do we know about their relationship and what types of research are still needed?. Aggress Violent Behav. 9, 271–303.

ActionAid (Producer). (2016). Community based approaches to tackling gender-based violence in Malawi: Lessons and challenges of involving women and men. https://actionaid.ie/wp-content/uploads/2016/11/Briefiing-Document-on-Malawi-Research-on-Gender-Based-Violence.pdf (accessed April 15, 2020).

Ahinkorah

, Dickson

, Seidu

A-A

. (2018). Women decision-making capacity and intimate partner violence among women in sub-Saharan Africa. Arch Public Health, 7, 6, 5.

Alio

, Clayton

, Garba

, et al. (2011). Spousal concordance in attitudes toward violence and reported physical abuse in African couples. J Interpers Violence. 26, 2790–2810.

Anderson

, Weisman

, Camacho

, et al. (2007). Women's satisfaction with their on-going primary health care services: A consideration of visit-specific and period assessments. Health Serv Res. 42, 663–681.

Andersson

, Ho-Foster

, Mitchell

, et al. (2007). Risk factors for domestic physical violence: national cross-sectional household surveys in eight southern African countries. BMC Womens Health. 7, 11.

Archer

, Graham-Kevan

. (2003). Do beliefs about aggression predict physical aggression to partners?. Aggressive Behav. 29, 41–54.

Bamiwuye

, Odimegwu

. (2014). Spousal violence in sub-Saharan Africa: Does household poverty-wealth matter?. Reprod Health, 1, 1, 45.

Bazargan-Hejazi

, Medeiros

, Mohammadi

, et al. (2013). Patterns of intimate partner violence: A study of female victims in Malawi. J Inj Violence Res. 5, 38–50.

10.

Berge

, Kambewa

, Munthali

, Wiig

. (2014). Lineage and land reforms in Malawi: Do matrilineal and patrilineal landholding systems represent a problem for land reforms in Malawi?. Land Use Policy. 41, 61–69.

11.

Bisika

. (2008). Do social and cultural factors perpetuate gender based violence in Malawi. Gender Behav. 6, 1884–1896.

12.

Campbell

. (2002). Health consequences of intimate partner violence. Lancet. 359, 1331–1336.

13.

Chikhungu

, Bradley

, Jamali

, Mubaiwa

. (2020). Culture and domestic violence amongst ever-married women in Malawi: An analysis of emotional, sexual, less-severe physical and severe physical violence. J Biosocial Sci. 1–15. DOI: 10.1017/S0021932020000140

14.

Chikhungu

, Amos

, Kandala

, 2nd, Palikadavath

. (2019). Married women's experience of domestic violence in Malawi: New evidence from a cluster and multinomial logistic regression analysis. J Interpers Violence. 886260519851782. DOI: 10.1177/0886260519851782

15.

Eckhardt

, Parrott

, Sprunger

. (2015). Mechanisms of alcohol-facilitated intimate partner violence. Violence Against Women. 21, 939–957.

16.

Ellsberg

, Jansen

HAFM

, Heise

, et al. (2008). Intimate partner violence and women's physical and mental health in the WHO multi-country study on women's health and domestic violence: An observational study. Lancet. 371, 1165–1172.

17.

Flood

, Pease

. (2009). Factors influencing attitudes to violence against women. Trauma Violence Abuse. 10, 125–142.

18.

Glick

, Sakalli-Ugurlu

, Cristina Ferreira

, Aguiar de Souza

. (2002). Ambivalent sexism and attitudes toward wife abuse in Turkey and Brazil. Psychol Women Q. 26, 292–297.

19.

Greene

, Kane

, Tol

. (2017). Alcohol use and intimate partner violence among women and their partners in Sub-Saharan Africa. Glob Ment Health (Camb). 4, e13.

20.

Hindin

, Kishor

, Ansara

. (2008). Intimate partner violence among couples in 10 DHS countries: Predictors and health outcomes. Retrieved from https://dhsprogram.com/publications/publication-as18-analytical-studies.cfm

21.

Hyde-Nolan

, Juliao

. (2011). Theoretical basis for family violence. In Family Violence: What Health Care Providers Need to Know. FS Ross, S Sarina, eds. (Jones and Bartlett Learning, Ontario).

22.

Jewkes

. (2002). Intimate partner violence: Causes and prevention. Lancet. 359, 1423–1429.

23.

Lenze

, Klasen

. (2017). Does women's labor force participation reduce domestic violence?. Evidence from Jordan. Feminist Econ. 23, 1–29.

24.

Malawi Government

MOHNU

. (2014). National Plan of Action to Combat Gender Based Violence in Malawi 2014–2020. (Ministry of Gender, Children, Disability and Social Welfare, Lilongwe, Malawi).

25.

McKenry

, Julian

, Gavazzi

. (1995). Toward a biopsychosocial model of domestic violence. J Marriage Family. 57, 307–320.

26.

Office/Malawi, National

Statistical

, and ICF . (2017). Malawi demographic and health survey 2015–16. Retrieved from https://dhsprogram.com/publications/publication-fr319-dhs-final-reports.cfm

27.

Okenwa-Emegwa

, Lawoko

, Jansson

. (2016). Attitudes toward physical intimate partner violence against women in Nigeria. SAGE Open. 6, 2158244016667993.

28.

Oyediran

. (2016). Explaining trends and patterns in attitudes towards wife-beating among women in Nigeria: Analysis of 2003, 2008, and 2013 demographic and health survey data. Genus. 72, 11.

29.

Pallikadavath

, Bradley

. (2019). Dowry, ‘dowry autonomy’ and domestic violence among young married women in India. J Biosocial Sci. 51, 353–373.

30.

Rani

, Bonu

, Diop-Sidibe

. (2004). An empirical investigation of attitudes towards wife-beating among men and women in seven Sub-Saharan African countries. Afr J Reprod Health. 8, 116–136.

31.

Sardinha

, Nájera Catalán

. (2018). Attitudes towards domestic violence in 49 low- and middle-income countries: A gendered analysis of prevalence and country-level correlates. PLoS One. 13, e0206101.

32.

Shwachman Kaminaga

. (2017). Acceptance of intimate partner violence in rural Malawi: An empirical analysis on the impacts of lineage and AIDS conversation networks. J Interpers Violence. 886260517703375. https://doi.org/10.1177/0886260517703375.

33.

Sohini

. (2016). Women's labour force participation and domestic violence. J S Asian Dev. 11, 224–250.

34.

Steele

, Josephs

. (1990). Alcohol myopia: Its prized and dangerous effects. Am Psychologist. 45, 921–933.

35.

Stöckl

, March

, Pallitto

, Garcia-Moreno

. (2014). Intimate partner violence among adolescents and young women: Prevalence and associated factors in nine countries: A cross-sectional study. BMC Public Health. 14, 751.

36.

Telalagic

. (2012). Domestic production as a source of marital power: Theory and evidence from Malawi. 1243. Cambridge Working Papers in Economics. Faculty of Economics, University of Cambridge. Retrieved from https://ideas.repec.org/p/cam/camdae/1243.html

37.

Tran

, Nguyen

, Fisher

. (2016). Attitudes towards intimate partner violence against women among women and men in 39 low- and middle-income countries. PLoS One. 11, e0167438.

38.

Uthman

, Lawoko

, Moradi

. (2009). Factors associated with attitudes towards intimate partner violence against women: A comparative analysis of 17 sub-Saharan countries. BMC Int Health Hum Rights. 9, 14.

39.

Warner

, Lee

. (1986). Social organization, spousal resources, and marital power: A cross-cultural study. J Marriage Family. 48, 121–128.

40.

WHO. (2018). The ecological framework. The global campaign for violence prevention. Retrieved from http://www.who.int/violenceprevention/approach/ecology/en/ (accessed July 28, 2018).