Sage Journals: Discover world-class research

Abstract

This study bridges a knowledge gap in existing literature by investigating correlates of male perpetrated intimate partner violence in Ghana. Based on a review of literature, it was hypothesized that some sociodemographic factors would be associated with the experience of male perpetrated physical, psychological, and sexual intimate violence in Ghana. To examine this hypothesis, data were collected from a convenience but nonreferral sample of 443 women, using a structured instrument. Descriptive statistics showed that 27%, 62%, and 34% of the sample, respectively, had experienced past-year physical, psychological, and sexual violence. Multivariate logistic regression analyses were performed, using some 15 sociodemographic factors as predictors. The analyses showed that women in rural settings had 1.71 and 2.20 times the odds of experiencing psychological and physical violence, respectively; women who were younger than their husbands had 2.67 and 5.71 times the odds of experiencing psychological and sexual violence, respectively; women whose husbands were unemployed had 2.41 and 2.58 times the odds of experiencing psychological and physical violence, respectively; women whose husbands had nonmarital sexual partners had 2.10 and 2.33 times the odds of experiencing psychological and physical violence, respectively; women who rated their health as good had 2.10 and 2.39 times the odds of experiencing psychological and sexual violence, respectively; and women whose husbands did not appreciate them had 2.22 and 2.57 times the odds of experiencing physical and sexual violence, respectively. Low education and polygamous marriage were also related to psychological violence. Policy and practice implications of the findings are discussed.

Keywords

intimate partner violence correlates of partner violence Ghana woman being younger than the partner partner having nonmarital partners

There is now a large body of literature on intimate partner violence (IPV) since the World Health Organization (WHO; 2002) drew attention to the alarming rate of IPV against women across cultural contexts. Recent large-scale multicountry studies have confirmed this by reporting higher prevalence rates of physical, psychological, and sexual violence among women (Abramsky et al., 2011; Garcia-Moreno, Jansen, Ellsberg, Heise, & Watts, 2006; Graham, Bernards, Wilsnack, & Gmel, 2011; Hindin, Kishor, & Ansara, 2008; H. Johnson, Ollus, & Nevala, 2008). The growing consensus is that IPV against women is a public health problem that should be tackled (Loseke, Gelles, & Cavanaugh, 2005; WHO, 2002). Past scholarship has offered various explanations for the occurrence of IPV against women, including the view that IPV is a behavior learned during childhood (Ehrensaft et al., 2003; Gover, Kaukinen, & Fox, 2008; Straus, Gelles, & Steinmetz, 1980), or that it is the result of personality disorders (Dutton, 1994, 2006; Dutton & Starzomski, 1997; Dutton & van Ginkel, 1997; Holtzworth-Munroe, Rehman, & Herron, 2000) or that it is due to the influence of alcohol and other substances (Field, Caetano, & Nelson, 2004; Foran & O’Leary, 2008; Snow, Sullivan, Swan, Tate, & Klein, 2006). However, feminist scholarship argues that male perpetrated IPV is a characteristic of strong patriarchal attitudes/beliefs held among men that leads to subjugation and violence against women (VAW; Dobash & Dobash, 1980; Jewkes, Levin, & Penn-Kekana, 2002; M. P. Johnson, 2006; Loseke et al., 2005; Pence & Paymar, 1993; Stark, 2007). In Ghana, for instance, scholars have noted the cultural predisposition of men and boys to dominate (Mann & Takyi, 2009; Takyi & Mann, 2006), especially, husbands’ quest to maintain control over wives, which has sometimes led to homicide (Adinkrah, 2008a, 2008b, 2014).

IPV is a problem in Ghana, registering a high prevalence (Coker-Appiah & Cusack, 1999; Ghana Statistical Service [GSS], 2009; Issahaku, 2015) with a set of cultural beliefs supporting it (Mann & Takyi, 2009; Takyi & Mann, 2006). For example, in a national survey 33% of women reported partner physical and sexual violence (Coker-Appiah & Cusack, 1999). Again, the 2008 Ghana demographic and health survey (DHS) found that 58% and 42.8% of women have suffered physical and sexual violence, respectively, by a current or former husband (GSS, 2009). Furthermore, Adinkrah (2008a, 2008b, 2014) has analyzed spousal killing cases among Ghanaian couples, indicating that in this society IPV can sometimes lead to fatalities.

This article contributes to existing literature by investigating correlates of male perpetrated IPV in Ghana. Although the seriousness of the issue has necessitated passage of legislation criminalizing domestic violence (Republic of Ghana, 2007), we currently do not know, scientifically, what factors are associated with IPV in Ghana, information that will inform policies and programs to address the issue. The article uses data from the administrative northern region of Ghana to bridge the existing gap.

Literature Review

Unlike earlier literature which explained male perpetrated partner violence as a behavior learned through experience and exposure (Straus et al., 1980), or the result of personality disorder (Dutton, 1994, 2006; Dutton & Starzomski, 1997; Dutton & van Ginkel, 1997), or a psychiatric problem (West, Roy, & Nichols, 1978), or simply the exercise of male power and control—masculinity (Jewkes et al., 2002; M. P. Johnson, 2006; Pence & Paymar, 1993;Stark, 2007), the focus of current scholarship is on identifying factors that correlate with IPV so as to identify pragmatic policy and practice interventions to combat violence. Correlates of male perpetrated IPV are described as factors that are associated with increased risk of perpetration or victimization. Several such factors have been identified in studies conducted across different national and geographic contexts. The following is a brief review of the literature on correlates of male perpetrated IPV.

Correlates of Male Perpetrated IPV

Several sociodemographic characteristics have been identified as correlates that are associated with the risk of IPV victimization among women. Depending on the social context, some factors are more prominent than others as risk markers for IPV perpetration/victimization, although some correlates are common to most studies perhaps because these contexts are fundamentally similar.

Alcohol use

Literature examining the association between alcohol use and male perpetrated IPV is growing. Studies in North America and other parts of the world have found a significant relationship between alcohol consumption and IPV perpetration/victimization. Several studies in the United States of America (U.S.) identify alcohol drinking as a key correlate of IPV (Fals-Stewart, 2003; Fals-Stewart, Leonard, & Birchler, 2005; Field et al., 2004; Snow et al., 2006; Temple, Weston, Stuart, & Marshall, 2008). Field et al. (2004) in a U.S. study investigating aggressive behavior, alcohol use, impulsivity, and attitudes to domestic violence using a probability sample of couples, found that 11% of 1,163 respondents who reported alcohol use also reported perpetrating domestic violence and those alcohol users were 3.2 times more likely to perpetrate IPV. In two samples of U.S. men entering domestic violence or alcoholism treatment, Fals-Stewart (2003) and Fals-Stewart et al. (2005) have found that alcohol use is associated with higher odds of frequency and severity of partner violence. Similarly, in another U.S. study, Snow et al. (2006) found that “problem drinking” was strongly related to perpetration of physical violence (r = .17) and to psychological violence (r = .23), suggesting that more drinking is associated with more perpetration of violence. Outside of the United States, Abramsky et al. (2011) have found a link between alcohol use and IPV against women. In their analysis of data drawn from 10 countries involving a sample of 24,097 women aged 15 to 49, Abramsky et al. identified problem drinking as a key correlate of IPV. According to the researchers, “In all sites odds of IPV were higher in relationships where one or both partners had problems with alcohol . . . ” (p. 13). In another study involving 13 countries, Graham et al. (2011) found that aggression toward a partner was more severe when either the male partner or both had been drinking prior to the incident. According to Graham et al. (2011) although alcohol may not be the cause of IPV, it can worsen the violent incident. However, not all alcohol users perpetrate IPV and men and women who use alcohol do not commit violence in equal measure (Giancola et al., 2002; Graham et al., 2011; Graham, Plant, & Plant, 2004; Thompson & Kingree, 2006). Therefore, it is not clear what in alcohol predisposes users to aggressive behavior.

Age and age disparity between couples

Current literature suggests that there is an association between age and IPV against women. First, some research has shown that there is an increased risk of violence in older couples (Luke, Schuler, Mai, Thien, & Minh, 2007; Vakili, Nadrian, Fathipoor, Boniadi, & Morowatisharifabad, 2010). Luke et al.’s (2007) study of couple attitudes and marital violence involving a sample of 465 women in Viet Nam found that there was an increased likelihood of older men hitting their wives and an increased likelihood of older women being hit by their partners. Vakili et al. (2010) also found a significant positive correlation between age of woman and IPV when they analyzed data from a sample of 702 women drawn from an Iranian county. Similarly, in an African American female subsample, a U.S. study (Lacey, West, Matusko, & Jackson, 2016) found that women aged 35 and above (35-64) were at increased risk of severe partner physical violence. Other studies show that age disparity is related to IPV, especially, that women are at heightened risk of IPV victimization if they are younger (in most cases, less than 30 years; Faramarzi, Esmailzadeh, & Mosavi, 2005; Hadi, 2000; Hindin & Adair, 2002; Obi & Ozumba, 2007). For example, a study conducted with 2,400 women in Iran (Faramarzi et al., 2005) found that younger women (≤20 years) had 2.23 times the odds of experiencing IPV compared with older women. Similarly, in his analysis of data from 500 women in various parts of Bangladesh, Hadi (2000) found that women who were less than 30 years were at increased risk of sexual violence during pregnancy, menstruation, and the neonatal state. However, a 10-country study (Abramsky et al., 2011) produced mixed results on the relationship between age and IPV. Abramsky et al. found a strong association between younger age of women and increased risk of IPV across the sites. But they also report that older age of the woman was also associated with increased risk of IPV. The findings on the effect of age and age disparity on IPV are not conclusive. Although we can say that both younger and older women are faced with increased risk of IPV, the same is not suggested about IPV perpetration among younger and older men. Luke et al. suggest that perhaps because older men and women have been exposed to the potential of violence for longer periods in their relationships, they are more likely to report having hit or been hit.

Current employment status

Employment status of either or both of the couple is an important correlate of IPV, with most studies indicating that unemployment is significantly related to IPV. For example, Vakili et al. (2010) found that women who were not currently employed had a higher mean reported IPV and women whose partners were currently not employed also had a higher mean reported IPV. In addition, Faramarzi and colleagues (2005) found that women who reported not currently employed had 2.75 times the odds of experiencing IPV. Similarly, a Lagos, Nigeria, study involving 400 women (Onigbogi, Odeyemi, & Onigbogi, 2015) found that women whose partners were unemployed had 5.89 times the odds of experiencing sexual violence. Furthermore, in their study of 600 men from south-east Nigeria, Obi and Ozumba (2007) found that among men who reported perpetrating IPV there was a significant association between a spouse’s unemployment status and physical, psychological, and sexual IPV. Findings of a U.S. study using a sample of 848 couples (Cunradi, Todd, Duke, & Ames, 2009) support the fact that unemployment of either or both partners increases the risk of IPV. However, as Naved and Persson (2005) have found in a Bangladeshi sample of 2,702 women, sometimes women who are employed (earning an income) are also at risk of IPV.

Residential setting

Studies on correlates of IPV have found that the setting in which the couple resides, rural versus urban, makes a difference in the experience and/or perpetration of IPV. Some studies find that women in rural settings are at a higher risk of IPV or that men in rural areas have a higher likelihood of perpetrating partner violence. The work of Vakili et al. (2010) is important on this point. Vakili et al. found that women in rural residence showed a higher mean of reported IPV compared with women in urban residence. This finding corroborates what was reported by Haj-Yahia (2000). In an analysis of data from two surveys involving 2,400 and 1, 334 samples of women, respectively, Haj-Yahia found that, across these surveys, residing in a rural setting was consistently significantly related to the risk of experiencing physical, psychological, and sexual abuse. However, perhaps the effect of residence on IPV is context dependent. This is because a study conducted in Chebu, the Philippines (Hindin & Adair, 2002) found that women in rural settings are less likely to experience IPV compared with women in urban areas. This suggests that the effect of residential setting on IPV varies from context to context.

Other correlates

A number of other factors have been identified as correlating with IPV in the literature. Four of these that are of interest to the present study include level of (formal) education, type of marriage (polygamous vs. monogamous marriage), presence of nonmarital sexual partners, and number of children in the marriage. Higher formal educational attainment is generally found to be a protective factor against IPV (Abramsky et al., 2011; Haj-Yahia, 2000; Luke et al., 2007; Onigbogi et al., 2015). These studies suggest that increased formal education of both partners or of the male partner is associated with a decreased risk of IPV against women. Type of marriage correlates with IPV such that women in polygamous marriage are shown to have a higher risk of IPV victimization (Abramsky et al., 2011; Onigbogi et al., 2015). Similar to a polygamous context, research has shown that women whose partners have other nonmarital sexual partners are at increased risk of IPV (Abramsky et al., 2011; Townsend et al., 2011). Studies investigating the relationship between number of children and partner violence indicate that an increase in number of children is associated with increased risk of IPV (Acevedo, Lowe, Griffin, & Botvin, 2013; Obi & Ozumba, 2007; Onigbogi et al., 2015). For example, after controlling for other factors, a U.S. study using a mixed sample (57.9% women; Acevedo et al., 2013) found that participants who had two or more children had 2.21 times the odds of reporting IPV compared with those who had only one child. Although not found in many studies, perhaps because they are not regarded as important factors, polygamous marriage, presence of nonmarital sexual partners, and increased number of children are important considerations for a study in a context like Ghana where, among many cultural groups, these are almost the norm rather than the exception.

Focus of the Present Study

Although some work has been done on domestic violence in Ghana and the issue has received legislative attention, no research has explored correlates of male perpetrated IPV in the country. The present study therefore contributes to existing scholarship by investigating factors associated with IPV in Ghana. Based on what has been found in the literature, it was hypothesized that some sociodemographic factors, including age, employment status, type of marriage, alcohol use, residential setting, and level of formal education, among others, will be significantly associated with physical, psychological, and sexual IPV experience among women. The study used data from a survey with women in Northern Ghana to examine the stated hypothesis, making it possible to see how the Ghanaian situation compares with what has been found in the literature.

Study Context

The research was carried out in the administrative northern region of Ghana. Ghana, formerly the Gold Coast, is a coastal country in West African which shares borders with Togo, Cote d’Ivoire, and Burkina Faso. The population of Ghana is estimated to be over 24 million, of which 51% is female (GSS, 2012). The research was carried out in northern region, one of the administrative regions within the larger geographic area known as northern Ghana. Northern region lies in a semiarid savanna zone, with rain-fed single-season agriculture as the main economic activity. The region, with an estimated population of over 2 million (GSS, 2012), occupies 29.5% of Ghana’s total land mass, is the largest of the ten administrative jurisdictions of the country, and is divided into 20 local government areas—district assemblies. Custom and tradition in northern region place a high premium on marriage and the family institution in which women are subordinated to men. For example, although young men and women look forward to marriage and family life, it is the man who should initiate the process, pay the bride price and bring the woman to his father’s compound, and act as head and breadwinner (Nukunya, 2003). In addition, men have permission to take an additional wife whereas the woman is expected to commit to one husband (Mahama, 2004). Again, in matters of inheritance and ascension to chieftaincy positions—traditional leadership—men (sons and nephews) rank ahead of women. This unequal placement of women in relation to men in this society provides an important lens to look at men’s controlling behavior and violence toward women. (A more detailed description of the study context has been provided in another article by Issahaku [2015].)

Method

Participants

Four hundred and forty-three (443) women participated in the study. These were recruited from outpatient populations at six district hospitals, using targeted sampling approaches. As explained by Singleton and Straits (2005), although targeted sampling is a nonprobability sampling approach it is not necessarily purposive or convenience sampling because it captures a cross-section of the population of interest at strategic locations. Established ethical guidelines suggested that institutional settings that provide services to women are safe grounds to recruit women into research on a sensitive topic such as VAW (WHO, 2001). This guided the choice of district health facilities as recruitment sites. Furthermore, the Research Ethics Board of the University of Toronto advised that to ensure privacy and confidentiality, participants be contacted at health centers with daily outpatient populations large enough to provide anonymity. Information received from the regional directorate of health services indicated that these facilities record an average daily outpatient department (OPD) attendance of more than 300 people. Women, who visit these facilities for their own health care needs or as caregivers to sick children and relatives, constitute a substantial proportion of the OPD populations. (Participants of this study were not at the health centers to receive treatment as victims of IPV. These facilities are not violence support service centers and do not assess for violence as part of the routine care.) At the selected hospitals, women in the OPD population who were within the age range of 18 to 49 years and in intimate relationships were eligible to participate in this study. (Most studies on IPV focus on women of this age range because that is the period during which many women are in active marital unions [see Abramsky et al., 2011; H. Johnson et al., 2008; Kishor & Johnson, 2006].) Posters designed to advertise the study at recruitment sites were not used because it was feared that doing so could attract the attention of men and possibly jeopardize the participation of some women. With this limited publicity, the rule of thumb was that a woman should be approached in such a way that no intruder would know what was being discussed. Any woman willing to participate following a solicitation in face-to-face contact was included in the sample. Recruitment was done concurrently with data collection. Data collectors did a more detailed introduction of this study, emphasizing its purpose and why they were the people being contacted as part of the informed consent process. The consent process highlighted the overall importance of this study, the unique contribution of the information they would provide, as well as the potential risks and benefits to individual participants. It also highlighted the voluntary nature of participation and the right to decline participation or to withdraw from the interview at any time without consequence. At the convenience of participants, instrument administration happened immediately following consent or at a later time. All interviews took place in confidential OPD spaces designated for that purpose by management of the health centers. In all, 600 women were approached to participate in this study. Out of this number, 443 women agreed, resulting in a 74% response rate. (These details have been provided elsewhere [Issahaku, 2015].)

Measures

Participant sociodemographic characteristics alongside measures of physical, psychological, and sexual violence were the important measures for the study. Sociodemographic measures were treated as the independent/predictor variables, whereas the physical, psychological, and sexual violence measures were the dependent variables.

Participant characteristics

The following characteristics of participants and their partners were taken: age relative to husband’s—participants were asked whether they were younger than the husband (yes/no); type of marriage (polygamous/monogamous); number of years of marriage; number of children in the marriage; years of education (participant’s and partner’s); woman and her partner’s current employment status (employed/not employed); whether husband is a chief (yes/no); whether husband drinks alcohol (yes/no); couple’s residence (rural/urban); participant’s assessment of her current health status (good/poor); whether husband has nonmarital sexual partners (yes/no); whether husband appreciates (commends) her for her contributions (yes/no); whether participant belongs to any associations (yes/no); and participant’s religious affiliation (Christian, Muslim, Other).

Measures of partner violence

The WHO’s VAW instrument (see Ellsberg & Heise, 2005), which has been validated for wider use in developing countries and was part of Ghana’s 2008 DHS (GSS, 2009), was adapted for this study. For this study, the section on violence during pregnancy in the original instrument was eliminated since the focus was on violence without reference to any event in the life cycle. Moreover, three questions were added to the psychological violence subscale, making it seven items in all for the present study. The three items included refuse to eat your food, refuse to communicate with you, and refuse to perform your family member’s funeral. These questions were added to the scale because of the author’s understanding of the culture and ways where men use different forms of psychological abuse against women. (Other researchers have made similar modifications to the psychological violence subscale for cultural relevance [see Naved & Persson, 2005; Vakili et al., 2010].) Following adaptation, the violence instrument had 16 items in all.

Past-year physical violence was a six-item Yes or No scale measuring physically assaultive acts such as slapping you or throwing something at you that could hurt, push you or shove you or pull your hair, hit you with fist or something that could hurt you, kick you or drag you or beat you up, choke you or burn you on purpose, and threat to use or actual use of weapons against you. Past-year psychological violence was a seven-item Yes or No scale measuring men’s actions that are potentially psychologically injurious. Examples include the following: insult you or make you feel bad about yourself; belittle or humiliate you in public; scare or intimidate you on purpose; threaten to hurt you or someone dear to you, refuse to communicate with you; refuse to eat your food; and refuse to perform your family member’s funeral. Past-year sexual violence was a three-item Yes or No scale. They included the following: forced you to have sexual intercourse, made you engage in nonconsensual sex due to fear, and forced you to perform something sexual that feels degrading. Reliability analysis showed Cronbach’s alphas of .72 (N = 6), .74 (N = 7), and .61 (N = 3), respectively, for the physical, psychological, and sexual violence subscales.

Data Collection

Data were collected as part of the author’s doctoral dissertation research. The major medium of data collection was face-to-face interviews. Assistants interpreted the English instrument in structured interviews for women who did not have English proficiency. The four main local languages used in these interviews were Dagbanli, Twi, Hausa, and Gonja. On average, interviews lasted for 45 min. The data collection team of 12 female assistants consisted of practicing nurses, high school teachers, and upper-year university students. Data collectors were trained in VAW research based on the WHO’s (2001) guidelines. They were also given an orientation on solution-focused brief counseling to help participants deal with emotions during the interviews and to refer participants who needed further help. Participant safety was taken care of by ensuring, among other things, the privacy and confidentiality of the interviews and ensuring that experienced nurse practitioners were on standby to offer emotional support to participants. Participants who could read and write in English were allowed the choice of independently completing the questionnaire. At each of the health centers, an experienced female nurse practitioner was on standby to provide support. Each participant received GHS 2 (two Ghana Cedis) as appreciation of their time and help with the study. Protocols for this study went through full review by the University of Toronto Research Ethics Board.

Data Analysis and Results

Data were analyzed using the statistical package for the social sciences (SPSS) Version 20. Frequency distribution, bivariate chi-square, and multivariate logistic regression analyses were performed to make sense of the data. Frequency distribution analysis (see Table 1) showed the following: 46% came from rural areas, 67% identified as younger than their partners, 35% identified their marriage as polygamous, and 70% were currently employed. Other characteristics include the following: mean length of marriage was 9.66 years (SD = 8.47), mean number of children was 2.55 (SD = 1.87), mean years of education for participants was 7.89 (SD = 7.67), mean years of partner education was 9.14 (SD = 8.24), 66% of participants were Muslim, 17% rated their general health as poor, 19% reported their partners take alcohol, 20% reported their partners have nonmarital sexual partners, 22% reported their partners were not currently employed, 25% reported they belong to an association, 11% indicated their partner is chief, and 27% reported their partners do not appreciate. Frequency distribution of the violence measures showed that 69% of the sample (N = 443) had experienced past-year IPV, 62% had experienced psychological violence, 27% had experience physical violence, and 34% had experienced sexual violence (Table 1).

Table 1.

Participant Sociodemographic Characteristics and Experience of IPV (N = 443).

Variable	n	%	M	SD
Years of marriage	427		9.66	8.47
Number of children	440		2.55	1.87
Woman’s years of education	398		7.89	7.67
Husband’s years of education	359		9.14	8.24
Woman’s employment status
Employed	310	70
Not employed	133	30
Husband’s employment status
Employed	344	78
Not employed	99	22
Husband is a chief
Yes	46	11
No	397	89
Husband uses alcohol
Yes	82	19
No	361	81
Woman’s self-rated health
Poor	76	17
Good	367	83
Residential status
Rural	204	46
Urban	239	54
Woman younger than partner
Yes	291	67
No	142	33
Religious affiliation
Christian	141	32
Muslim	291	66
Other	8	2
Type of marriage
Monogamy	288	65
Polygamy	155	35
Husband has nonmarital partners
Yes	88	20
No	355	80
Husband appreciates her
Yes	321	73
No	122	27
Belongs to an association
Yes	113	25
No	330	75
Experienced psychological violence
Yes	275	62
No	168	38
Experienced physical violence
Yes	120	27
No	323	73
Experienced sexual violence
Yes	151	34
No	292	66
Experienced any violence
Yes	306	69
No	137	31

Note. IPV = intimate partner violence.

Bivariate relationships

Table 2 presents cross-tabulations for the bivariate analyses. As can be seen, all the demographic variables were significantly related to at least one of the violence measures at the alpha level of .05 or lower as follows: Type of residence was significantly associated with psychological violence (χ² = 7.84, p < .005), physical violence (χ² = 25.53, p < .001), and sexual violence (χ² = 7.91, p < .005). Woman’s age relative to her partner’s was significantly associated with psychological violence (χ² = 12.16, p < .001) and sexual violence (χ² = 14.32, p < .001) but not physical violence (χ² = .023, p < .878). Religious affiliation was significantly associated with physical violence (χ² = 13.65, p < .001) and sexual violence (χ² = 9.38, p < .009) but not psychological violence (χ² = 5.14, p < .077). Type of marriage was significantly associated with physical violence (χ² = 5.29, p < .021) but not sexual violence (χ² = 3.42, p < .064) or psychological violence (χ² =.100, p < .762). Woman employment status was significantly associated with physical violence (χ² = 16.74, p < .001) but not sexual violence (χ² = 3.34, p < .068) or psychological violence (χ² = 1.15, p < .283). Woman association membership was significantly associated with psychological violence (χ² = 4.40, p < .036) but not physical violence (χ² = 3.52, p < .061) or sexual violence (χ² = 1.44, p < .230). Husband employment status was significantly associated with psychological violence (χ² = 17.80, p < .001), physical violence (χ² = 30.22, p < .001), and sexual violence (χ² = 13.04, p < .001). Husband is a chief was significantly associated with physical violence (χ² = 13.23, p < .001) but not sexual violence (χ² = 4.82, p < .090) or psychological violence (χ² = 5.51, p < .064). Husband alcohol use was significantly associated with psychological violence (χ² = 8.20, p < .042) and physical violence (χ² = 17.65, p < .001) but not sexual violence (χ² = 7.07, p < .070). Presence of nonmarital sexual partners was significantly associated with psychological violence (χ² = 22.38, p < .001), physical violence (χ² = 26.49, p < .001), and sexual violence (χ² = 17.81, p < .001). Husband appreciates was significantly associated with physical violence (χ² = 8.12, p < .004) and sexual violence (χ² = 10.04, p < .002) but not psychological violence (χ² = 1.11, p < .292). Finally, woman’s assessed health status was significantly associated with psychological violence (χ² = 8.23, p < .044), physical violence (χ² = 19.98, p < .001), and sexual violence (χ² = 17.88, p < .001).

Table 2.

Chi-Squares of Relationship of Demographic Variables to Risk of Psychological, Physical, and Sexual Violence.

Demographic variable	Psychological violence		χ² p value	Physical violence		χ² p value	Sexual violence		χ² p value
	%	%		%	%		%	%
	Yes	No		Yes	No		Yes	No
Type of residence			.005			.000			.005
Rural (n = 204)	69	31		41	59		41	59
Urban (n = 239)	56	44		19	81		28	72
Woman is younger			.000			.878			.000
Yes (n = 291)	67	33		29	71		40	60
No (n = 142)	50	50		30	70		22	78
Religious affiliation			.077			.001			.009
Christian (n = 141)	61	39		26	74		28	72
Muslim (n = 291)	60	40		30	70		37	63
Other (n = 8)	100	0		88	12		75	25
Marriage type			.762			.021			.064
Monogamy (n = 288)	61	39		26	74		31	69
Polygyny (n = 155)	63	37		36	64		40	60
Woman employment status			.283			.000			.068
Employed (n = 310)	60	40		24	76		32	68
Unemployed (n = 133)	65	35		43	57		41	59
Belong to an Association			.036			.061			.230
Yes (n = 113)	70	30		36	64		39	61
No (n = 330)	59	41		27	73		33	67
Husband employment status			.000			.000			.000
Employed (n = 344)	56	44		23	77		30	70
Not employed (99)	80	20		52	48		49	51
Husband is chief			.064			.001			.090
Yes (n = 45)	78	22		51	49		49	51
No (n = 397)	60	40		26	74		33	67
Husband alcohol use			.042			.001			.070
Yes alcohol (n = 87)	76	24		50	50		46	54
No alcohol (n = 356)	60	40		26	74		31	69
Husband has other sexual partners			.000			.000			.000
Yes (n = 88)	81	19		51	49		43	57
No (n = 355)	51	49		19	81		23	77
Husband appreciates			.292			.004			.002
Yes (n = 321)	60	40		26	74		30	70
No (n = 122)	66	34		39	61		46	54
Woman assessed health			.044			.000			.000
Poor (n = 76)		53	47	17	83		17	83
Good (n = 367)		62	38	30	70		36	64

In subsequent logistic regression analyses, these variables were used to predict violence. This was done in consonance with the specificity assumption which requires that all relevant predictors be included in the model. In addition, because of their importance in the literature, three measures not shown in the chi-square analysis were included in the logistic regression models as follows: years of education for participant and partner were dichotomized into 0 to 7 years = elementary/no education and 8 years and above = some/post-secondary education; and number of children in the marriage was dichotomized into 0 to 2 and 3 or more children. The religious affiliation variable was dichotomized into Christian and Muslim/Other. Other assumptions of the logistic regression analyses were met as follows: the outcome variables were dichotomous, statistically independent, mutually exclusive and collectively exhaustive, and the sample size (N ≥ 400) was large enough to produce meaningful regression coefficients.

Predictive models

Separate logistic regression models were used to assess how the 15 variables predicted IPV. These predictors were type of residence, type of marriage, whether woman is younger than partner, religious affiliation, whether woman belongs to an association, woman employment, partner employment, whether partner is chief, partner alcohol use, whether partner has nonmarital sexual partners, woman’s assessed health, whether partner appreciates, partner education, woman education, and number of children. These are all dichotomous variables. In each case, the reference category (dummy coded 0) is not shown in the logistic regression table. Urban residence, woman is not younger than partner, currently employed, does not belong to an association, no nonmarital partners, health assessed as poor, monogamous marriage, Christian, no partner alcohol use, partner appreciates (yes), some/post-secondary education, 0 to 2 children, respectively, are the reference categories. For each predictive model the nonsignificant Hosmer–Lemeshow goodness-of-fit statistic supported the claim that the model was a good fit for the data.

The first model assessed how well the variables predicted the odds of experiencing psychological violence in the sample. Model coefficients and related statistics are presented in Table 3. An examination of the z (Wald) scores and the probabilities suggests that the population correlation coefficient is probably greater than 0 for at least seven of the predictors. The coefficients for type of residence, woman younger, partner employment, type of marriage, presence of nonmarital sexual partners, woman’s assessed health, and partner education variables were statistically significant, indicating the following: women who resided in rural settings were significantly more likely to experience psychological violence, p < .050, odds ratio (OR) = 1.71, 95% confidence interval (CI) = [1.00, 2.91]; women who are younger than their partners were significantly more likely to experience psychological violence, p < .001, OR = 2.67, 95% CI = [1.58, 4.51]; women in polygamous marriage were significantly more likely to experience psychological violence, p < .041, OR = 1.73, 95% CI = [1.02, 2.93]; women whose partner was not currently employed were significantly more likely to experience psychological violence, p < .016, OR = 2.41, 95% CI = [1.18, 4.95]; women whose partner had nonmarital sexual partners were significantly more likely to experience psychological violence, p < .023, OR = 2.06, 95% CI = [1.10, 3.85]; women who assessed their health as good were significantly more likely to experience psychological violence, p < .030, OR = 2.10, 95% CI = [1.10, 4.10]; and women whose partner had elementary/no education were significantly more likely to experience psychological violence, p < .035, OR = 2.07, 95% CI = [1.05, 4.10].

Table 3.

Odds Ratio of Risk of Experiencing Psychological Violence Using Logistic Regression.

Demographic variable	p	Exp(B)	95% CI for Exp(B)
Resides in rural setting	.050	1.707	[1.001, 2.910]
Woman younger than partner	.000	2.671	[1.582, 4.509]
Polygamy	.041	1.732	[1.024, 2.930]
Muslim/Other	.640	0.871	[0.487, 1.555]
Woman not employed	.638	0.867	[0.479, 1.570]
Partner not employed	.016	2.413	[1.176, 4.951]
Belongs to an association	.645	1.143	[0.648, 2.015]
Partner is a chief	.363	1.461	[0.646, 3.305]
Partner takes alcohol	.294	1.448	[0.726, 2.889]
Partner has nonmarital partners	.023	2.062	[1.103, 3.853]
Partner does not appreciate her	.451	1.246	[0.704, 2.203]
Woman assessed health as good	.029	2.095	[1.080, 4.066]
Woman has elementary/no education	.233	0.664	[0.339, 1.302]
Partner has elementary/no education	.035	2.074	[1.052, 4.090]
Couple has three or more children	.934	1.023	[0.601, 1.740]
Constant	−2.003
−2LL = 411.382
Nagelkerke R² = .196
Hosmer–Lemeshow χ² = 16.860, p = .187

Note. CI = confidence interval; LL = Log Likelihood.

Next, a model predicting the odds of experiencing physical violence in the sample was established. Model coefficients and related statistics are presented in Table 4. An examination of the z (Wald) scores and the probabilities suggests that the population correlation coefficient is probably greater than 0 for at least four of the predictors. The coefficients for type of residence, partner employment, presence of nonmarital partners, and partner appreciation variables were statistically significant, indicating that women residing in rural setting were significantly more likely to experience physical violence, p < .015, OR = 2.20, 95% CI = [1.16, 4.18]; women whose partner was not currently employed were significantly more likely to experience physical violence, p < .010, OR = 2.58, 95% CI = [1.26, 5.29]; women whose partner had nonmarital sexual partners were significantly more likely to experience physical violence, p < .012, OR = 2.33, 95% CI = [1.20, 4.51]; and women whose partner does not appreciate were significantly more likely to experience physical violence, p < .015, OR = 2.22, 95% CI = [1.17, 4.22].

Table 4.

Odds Ratio of Risk of Experiencing Physical Violence Using Logistic Regression.

Demographic variable	p	Exp(B)	95% CI for Exp(B)
Resides in rural setting	.015	2.204	[1.163, 4.175]
Woman younger than partner	.268	1.432	[0.758, 2.706]
Polygamy	.398	0.772	[0.424, 1.407]
Muslim/Other	.815	0.916	[0.437, 1.919]
Woman not employment	.146	1.670	[0.837, 3.332]
Partner not employment	.010	2.581	[1.260, 5.290]
Belongs to an association	.358	0.724	[0.363, 1.441]
Partner is a chief	.335	1.535	[0.642, 3.672]
Partner takes alcohol	.121	1.812	[0.855, 3.840]
Partner has nonmarital partners	.012	2.325	[1.199, 4.508]
Partner does not appreciate her	.015	2.223	[1.170, 4.222]
Woman assessed health as good	.054	2.516	[0.968, 6.423]
Woman has elementary/no education	.086	0.489	[0.216, 1.106]
Partner has elementary/no education	.502	1.310	[0.596, 2.882]
Couple has three or more children	.201	1.506	[0.803, 2.824]
Constant	−3.279
−2LL = 305.310
Nagelkerke R² = .281
Hosmer–Lemeshow χ² = 6.998, p = .537

Note. CI = confidence interval; LL = Log Likelihood.

The final model assessed how the variables predicted the odds of experiencing sexual violence. Model coefficients and related statistics are presented in Table 5. An examination of the z (Wald) scores and the probabilities suggests that the population correlation coefficient is probably greater than 0 for at three of the predictors. The coefficients for woman younger, partner appreciation, and woman’s assessed health were statistically significant, showing that women who were younger than their partner were significantly more likely to experience sexual violence, p < .001, OR = 5.71, 95% CI = [2.88, 11.33]; women whose partner does not appreciate were significantly more likely to experience sexual violence, p < .01, OR = 2.57, 95% CI = [1.40, 4.71]; and women who assessed their health as good were significantly more likely to experience sexual violence, p < .041, OR = 2.39, 95% CI = [1.03, 5.52].

Table 5.

Odds Ratio of Risk of Experiencing Sexual Violence Using Logistic Regression.

Demographic variable	p	Exp(B)	95% CI for Exp(B)
Resides in rural setting	.917	1.031	[0.583, 1.824]
Woman younger than partner	.000	5.709	[2.876, 11.333]
Polygamy	.281	1.350	[0.782, 2.333]
Muslim/Other	.322	1.412	[0.713, 2.795]
Woman not employed	.315	1.382	[0.735, 2.597]
Partner not employed	.459	1.304	[0.646, 2.631]
Belongs to an association	.659	1.146	[0.626, 2.099]
Partner is a chief	.148	1.866	[0.801, 4.347]
Partner takes alcohol	.269	1.495	[0.733, 3.050]
Partner has nonmarital partners	.266	1.426	[0.763, 2.665]
Partner does not appreciate her	.002	2.566	[1.397, 4.711]
Woman assessed health as good	.041	2.388	[1.034, 5.515]
Woman has elementary/no education	.790	1.100	[0.543, 2.230]
Partner has elementary/no education	.709	0.874	[0.432, 1.771]
Couple has three or more children	.091	1.648	[0.924, 2.939]
Constant	−4.148
−2LL = 352.659
Nagelkerke R² = .246
Hosmer–Lemeshow χ² = 11.127, p = .195

Note. CI = confidence interval; LL = Log Likelihood.

The results of the logistic regression analyses indicate that type of residence, woman’s younger age, type of marriage, partner employment status, partner’s nonmarital sexual partners, woman’s assessed health, partner education, and whether or not partner appreciates are the correlates of IPV in this sample. Women residing in rural settings had 1.71 and 2.20 times the odds of experiencing psychological and physical violence, respectively; women who were younger than their husbands had 2.67 and 5.71 times the odds of experiencing psychological and sexual violence, respectively; women whose husbands were currently not employed had 2.41 and 2.58 times the odds of experiencing psychological and physical violence, respectively; women whose husbands had nonmarital sexual partners had 2.10 and 2.33 times the odds of experiencing psychological and physical violence, respectively; and women who assessed their health as good had 2.10 and 2.39 times the odds of experiencing psychological and sexual violence, respectively. Furthermore, women whose husbands did not have the habit of appreciating (commending) them had 2.22 and 2.57 times the odds of experiencing physical and sexual violence, respectively, women in polygamous marriage had 1.73 times the odds of experiencing psychological violence, and women whose husbands had elementary or no education had 2.10 times the odds of experiencing psychological violence. In all three models, the relationship between alcohol use and violence was not significant.

Discussion

Results of this study corroborate previous research on IPV among women. Consistent with findings of studies in contexts similar to the Ghanaian society there were high rates of physical, psychological, and sexual IPV in this study. For example, similar to findings in multicountry studies which show that prevalence rates of IPV among women range from 15% to 71% across countries (Garcia-Moreno et al., 2006) or 10% to 69% across countries (WHO, 2002), the rates found in the present study were 69% overall, 62% psychological violence, 34% sexual violence, and 27% physical violence. These are also comparable with results of Vakili et al.’s (2010) study in Kazeroon, Iran, in which 82.6%, 30.9%, and 43.7% of the sample, reported psychological, sexual, and physical IPV, respectively. It is possible that IPV was underreported in our sample, giving that domestic violence is a sensitive subject and coupled with the social norm in the Ghanaian society which enjoins individuals not to discuss the family’s secret with outsiders.

Consistent with previous research on correlates of IPV, our findings show that some factors do influence the risk of IPV victimization among, with rural residence, woman being younger than partner, partner having nonmarital sexual partners, partner not appreciating the woman, woman assessing her health as good, partner not currently employed, polygamous marriage, and partner having only elementary or no education increasing the risk of IPV. Our finding that women in rural settings are at a higher risk of IPV compared with women in urban residence confirms the work of other researchers (Faramarzi et al., 2005; Haj-Yahia, 2000; Vakili et al., 2010). The finding that women in rural residence were 1.71 times and 2.20 times more likely to experience physical and psychological violence, respectively, corroborates Vakili et al.’s (2010) reported higher mean violence among rural women. It is not known for sure what links rural residence to the risk of IPV. Perhaps isolation and the strength or otherwise of patriarchal traditions may help in understanding the link. It is possible that women in rural communities are more easily isolated by their husbands than those in urban cities. This is because most rural communities do not have social and recreational centers and, perhaps there are only a few friends for the woman to hang out with, making it possible for men to easily confine and have their way with their wives. Moreover, as Nukunya (2003) has noted of Ghana, perhaps patriarchal tradition, which legitimizes a husband’s authority, is much stronger in the rural society. However, the work of Hindin and Adair (2002) which more IPV among women in urban settings contradicts our position. Indeed, Hindin and Addair reported that women in the remotest of rural settings were the least at risk of IPV. This suggests that IPV among women may not be unique to any residential setting.

Our finding that younger women are at increased risk of IPV is consistent with previous research (Faramarzi et al., 2005; Hindin & Adair, 2002; Obi & Ozumba, 2007). In most of these cultures, age comes with power and privilege and, especially in a patriarchal society where men occupy a position of privilege, the younger age of the woman relative to partner aggravates her vulnerability. Therefore, consonant with Faramarzi et al. (2005) who report that younger women were 2.23 times more likely to experience IPV, we found that younger women had 2.67 and 5.71 times the odds of experiencing psychological and sexual violence, respectively. However, other studies have shown that both younger and older women are at risk of IPV (Abramsky et al., 2011; Lacey, West, Matusko, & Jackson, 2016; Luke et al., 2007). Although women who are younger than their partners may be especially vulnerable to IPV, what the literature suggests, generally, is that IPV risk among women does not respect a woman’s age relative to her partner’s. In societies such as Ghana, age is a tool in the system of social-gender stratification based on judgments of maturity, responsibility, and cultural role significance (Mahama, 2004; Nukunya, 2003; Ortner, 1972; Rosaldo, 1974), in which case a man is always “older” than a woman because of the privilege attached to his sense of judgment and social roles. As a result, perhaps women in general and married women in particular are always under the disciplinary authority of the man, including his use of violence.

Our findings that women whose partners were currently not employed and women whose partners had nonmarital sexual partners had higher odds of experiencing violence confirm previous research. The finding that women whose partners were unemployed had 2.41 and 2.58 times the odds of experiencing psychological and physical violence, respectively, adds to the Nigerian study (Onigbogi et al., 2015) in which women whose husbands were unemployed had almost six times the odds of experiencing sexual violence. Perhaps the association between male unemployment and IPV is explained by the concept of “male identity crisis” (Bradley, 1999; Jewkes et al., 2002; Moore, 1994), which is characterized by feelings of vulnerability, loss of control, loss of meaning, and loss of status arising from an imminent inability to fulfill the role of breadwinner (Ehrenreich, 1995; Mohammed, 2004). Under such circumstances “of threatened egotism” (Baumeister, Smart, & Boden, 1996, p. 5), men may use violence at the least provocation. Luke et al. (2007) discussed this issue in terms of men’s possession of little resources or low status in their Vietnamese study. Although not significant in our study, other studies show that unemployed women are at increased risk of IPV (Cunradi et al., 2009; Obi & Ozumba, 2007; Vakili et al., 2010). Again, consistent with what has been reported by others (Abramsky et al., 2011; Onigbogi et al., 2015; Townsend et al., 2011) are our findings that (a) women whose partners had nonmarital sexual partners were 2.10 and 2.33 times more likely to experience psychological and physical violence, respectively, and (b) women in polygamous marriage had 1.73 times the odds of experiencing psychological violence. A context of nonmarital sexual partnerships is similar to polygamous marriage perhaps only that nonmarital partners do not reside in the conjugal home. Women in these contexts usually view each other rivals (Benneh, 1994; Kuenyehia & Aboagye, 2003; Lamphere, 1974; Mahama, 2004). Therefore, women whose partners have other sexual partners may be challenging their husbands over the issue and get physically beaten. Women in this context may also face psychological violence in the form of insults, negative comparisons, belittlement, and cessation of communication.

Other correlates found in the study, including the effect of education, woman’s assessed health, and whether or not partner appreciates the woman are an important contribution to the literature. Our finding that women whose partners had only elementary or no education were at increased risk of IPV is consistent with research which reports that increased educational achievement is associated with decreased risk of violence (Abramsky et al., 2011; Ishida, Stupp, Melian, Serbanescu, & Goodwin, 2010; Onigbogi et al., 2015). The relation between education and IPV suggests that men with higher educational attainment are perhaps more accepting of egalitarian values and respect women’s rights to personal safety from harm. In addition, perhaps men who have attained a level of formal education understand the negative implications of IPV and are also aware of laws which protect women against domestic violence. Unique to our study, and puzzling, are the findings that women who assessed their health as good and women whose partner did not appreciate them were at increased risk of IPV. The effect of woman’s health on IPV in our study can be likened to the observation in other studies that when women control resources and decision making they are at increased risk of violence (Hindin & Adair, 2002). Because health is wealth and wealth is strength perhaps women who assess their health as good act tough, challenge conservative male perceptions of women as the weaker sex, and appear threatening to men who use violence to feel in control. It is also possible that in the social context of our study men selectively beat or sexually abuse healthy women to avoid responsibility for any serious repercussions. Our finding that if a partner does not appreciate the woman she is at increased risk of IPV is also consistent with the finding that even when women are working and contributing income to support the family they are still at risk of violence (Abramsky et al., 2011; Naved & Persson, 2005). Indeed, dual-earner families are now the norm rather than the exception (Bradley, 1999). Hence, in addition to their child care and homecare work, most women are out there working to contribute income to their families. As a result, nonappreciation of the woman is in itself violence and to follow it up with physical and sexual violence is double jeopardy which we can hardly explain. However, if men in this society have the opportunity to establish sexual liaison with other women then, perhaps it is understandable why they can afford not to appreciate their wives.

Limitations of the study should be taken into account when interpreting the results. The sample employed in this study was regional, nonrandomly selected, and characterized by a volunteer bias. In addition, the measures used in the study were self-reports which are characterized by a response bias and where underreporting is highly likely. Another limitation is that we did not introduce any control variables in the analyses. The noted limitations notwithstanding, however, 443 is a reasonably large sample and recruiting from public places which attract a cross-section of the general population of women ensured that some diversity was reflected in the data. Therefore, given that there has been no prior empirical investigation of correlates of IPV in Ghana, the study makes an important contribution to the literature and has implications for policy and practice in the country.

Conclusion and Implications

IPV victimization is a serious issue among women in Ghana, with nearly seven in 10 women experiencing violence at some point in their marriage. This corroborates and adds to findings of research in other contexts (Garcia-Moreno et al., 2006; Vakili et al., 2010; WHO, 2002). The association between some sociodemographic factors and IPV found in this study is also consistent with and adds to existing literature. Although not exclusive to them, women in rural areas, women who are younger than their partner, who assess their health as good, whose partners have other sexual relationships, and whose partners do not appreciate them are particularly vulnerable to IPV victimization. In a patriarchal society like Ghana where men “construct women as legitimate vehicles for the reconfirmation of male powerfulness through beatings . . .” (Jewkes et al., 2002, p. 1615), younger women, less exposed women (in rural areas), strong/healthy women, and women challenging men over their extramarital escapades are at increased risk of violence. For similar reasons, the risk of violence is higher for women whose husbands have only elementary or no education or whose husbands are unemployed. If postsecondary education and a white-collar job enhance a man’s social standing, then with neither of these he will have to assert his place in relation to the woman by some other means which, unfortunately in this case, is violence.

Findings of this study have implications for policy and practice in Ghana. A number of such implications are outlined here for serious consideration by government, civil society organizations (CSOs), and nongovernmental organizations (NGOs) who are stakeholders and partners in promoting family and community well-being. To begin, the age at which young people establish marriage relationships needs a critical review. Maturity in terms of self-discipline, personal responsibility, and accountability for behavior should be factored into the definition of legal age of marriage. If maturity comes with education and young people need up to 18 years to finish high school and enter college, then it makes sense to raise the age of marriage to 21/22 years. However, although Ghana’s law puts the age of maturity at 18 years, it is not uncommon to find 16- to 17-year-old girls in marriage. Indeed, in areas where polygamy is practiced it is possible to find a girl married to a polygamous man who is old enough to be her father. What is seriously required, therefore, is concerted education and awareness about the harmful implications of marrying teenage girls or giving teenage daughters into marriage. Second, the finding that women in rural Ghana are at increased risk of IPV suggests that within the broader framework of tackling violence, rural areas be given special attention. In this regard, it is recommended that (a) disparity in access to quality education and educational attainment between rural and urban Ghana be bridged, (b) education and awareness about the problem and about laws protecting women against violence be targeted to rural communities, and (c) perhaps the provision of woman-friendly recreational facilities in rural communities will give women safe spaces to break from the routine interactions of violence with their husbands. The CSOs and NGOs who already have a strong presence and influence in rural Ghana can partner government and village elders in the education and awareness drives and in providing safe spaces for women. A third policy intervention suggested by this study is education and awareness about the economic and social costs of polygamy and extramarital partnerships. Polygamy and extramarital partnerships are not lawful in Ghana, but they are not also outlawed. Indeed, as Islam permits men to have multiple marital partners, Ghana cannot outlaw these practices. However, we think that perhaps when men become acutely aware of the economic and social costs of multiple partnerships they will engage less in them and that will reduce the associated risk of violence. As we have suggested elsewhere (Issahaku, 2016), traditional and religious leaders who occupy positions of influence, including chiefs and their elders-in-council, Imams, Priests, and Pastors would, in partnership with government, CSOs, and NGOs, play an important role in the education and awareness on the harm of multiple and extramarital partnerships. Finally, if women who assess their health as good are at increased risk of IPV then, as we have suggested elsewhere (Issahaku, 2015), we reiterate the need to encourage routine visits to health clinics among women. At these clinics, in addition to routine health checks, women would be assessed and given support for violence and emotional difficulties. A universal health insurance coverage would significantly remove barriers to routine visits to health centers among women.

Findings of this study also have implications for future research in Ghana, two of which are the following. First, there is need for a larger scale study employing random sampling techniques to further investigate factors associated with IPV in the country. This would fill one of the limitations of the present study. Second, and perhaps more important, future research should try to explain/theorize the relationship between IPV and the identified correlates through qualitative designs. This would enhance understanding of the mechanism between these factors so that appropriate strategies would be developed to disrupt the relationship.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) received no financial support for the research,authorship,and/or publication of this article.

Author Biography

Paul Alhassan Issahaku PhD,is an assistant professor in the School of Social Work at Memorial University of Newfoundland. He has been a lecturer in the Department of Social Work at University of Ghana,Legon,and a research associate and course instructor in the Factor-Inwentash Faculty of Social Work at University of Toronto,Canada. His research interests comprise issues of family welfare and social policy,with a focus on Ghana.

References

Abramsky

Watts

C. H.

Garcia-Moreno

Devries

Kiss

Ellsberg

. . . Heise

(2011). What factors are associated with recent intimate partner violence? Findings from the WHO multi-country study on women’s health and domestic violence. BMC Public Health, 11, Article 109. doi:10.1186/1471-2458-11-109

Acevedo

B. P.

Lowe

S. R.

Griffin

K. W.

Botvin

G. J.

(2013). Predictors of intimate partner violence in a sample of multiethnic urban young adults. Journal of Interpersonal Violence, 28, 3004-3022. doi:10.1177/0886260513488684

Adinkrah

(2008a). Husbands who kill their wives: An analysis of uxoricides in contemporary Ghana. International Journal of Offender Therapy and Comparative Criminology, 52, 296-310. doi:10.1177/0306624X07307119

Adinkrah

(2008b). Spousal homicides in contemporary Ghana. Journal of Criminal Justice, 36, 209-216. doi:10.1016/j.jcrimjus.2008.04.002

Adinkrah

(2014). Homicide-suicide in Ghana: Perpetrators, victims, and incidence characteristics. International Journal of Offender Therapy and Comparative Criminology, 58(3), 364-387. doi:10.1177/0306624X12470530

Baumeister

R. F.

Smart

Boden

J. M.

(1996). Relation of threatened egotism to violence and aggression: The dark side of high self-esteem. Psychological Review, 103, 5-33. doi:10.1037/0033-295X.103.1.5

Benneh

(1994). Family development in Ghana: An overview. In Adayfio-Schandorf

(Ed.), Family development in Ghana (pp. 5-12). Accra, Ghana: Ghana Universities Press.

Bradley

(1999). Gender & power in the workplace: Analyzing the impact of economic change. Hamshire: Macmillan Press.

Coker-Appiah

D. C.

Cusack

(1999). Breaking the silence & challenging the myths of violence against women & girls children in Ghana (Report of a national study on violence). Accra, Ghana: Gender Studies & Human Rights Documentation Center.

10.

Cunradi

C. B.

Todd

Duke

Ames

(2009). Problem drinking, unemployment, and intimate partner violence among a sample of construction industry workers and their partners. Journal of Family Violence, 24, 63-74. doi:10.1007/s10896-008-9209-0

11.

Dobash

R. E.

Dobash

(1980). Violence against wives: A case against patriarchy. London, England: Open Books.

12.

Dutton

D. G.

(1994). Behavioral and affective correlates of borderline personality organization in wife assaulters. International Journal of Law and Psychiatry, 17, 265-279. doi:10.1016/0160-2527(94)90030-2

13.

Dutton

D. G.

(2006). Rethinking domestic violence. Vancouver, Canada: University of British Columbia Press.

14.

Dutton

D. G.

Starzomski

A. J.

(1997). Personality predictors of the Minnesota Power and Control Wheel. Journal of Interpersonal Violence, 12, 70-82. doi:10.1177/088626097012001005

15.

Dutton

D. G.

van Ginkel

(1997). The interaction of cultural and personality factors in the etiology of wife assault. In Sev’er

(Ed.), A cross-cultural exploration of wife abuse (pp. 101-121). Lewiston: The Edwin Mellen Press.

16.

Ehrenreich

(1995). The decline of patriarchy. In Berger

Wallis

Watson

(Eds.), Constructing masculinity (pp. 284-290). New York, NY: Routledge.

17.

Ehrensaft

M. K.

Cohen

Brown

Smailes

Chen

Johnson

J. G.

(2003). Intergenerational transmission of partner violence: A 20-year prospective study. Journal of Consulting and Clinical Psychology, 71, 741-753. doi:10.1037/0022-006X.71.4.741

18.

Ellsberg

Heise

(2005). Researching violence against women: A practical guide for researchers and activities. Geneva, Switzerland: World Health Organization and Program for Appropriate Technology in Health.

19.

Fals-Stewart

(2003). The occurrence of partner physical aggression on days of alcohol consumption: A longitudinal diary study. Journal of Consulting and Clinical Psychology, 71, 41-52. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/12602424

20.

Fals-Stewart

Leonard

K. E.

Birchler

G. R.

(2005). The occurrence of male-to-female intimate partner violence on days of men’s drinking: The moderating effects of antisocial personality disorder. Journal of Consulting and Clinical Psychology, 73, 239-248. doi:10.1037/0022-006X.73.2.239

21.

Faramarzi

Esmailzadeh

Mosavi

(2005). Prevalence and determinants of intimate partner violence in Babol city, Islamic Republic of Iran. Eastern Mediterranean Journal, 11, 870-879. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16761656

22.

Field

C. A.

Caetano

Nelson

(2004). Alcohol and violence related cognitive risk factors associated with the perpetration of intimate partner violence. Journal of Family Violence, 19, 249-253. doi:10.1023/B:JOFV.0000032635.42145.66

23.

Foran

H. M.

O’Leary

K. D.

(2008). Problem drinking, jealousy, and anger control: Variables predicting aggression against a partner. Journal of Family Violence, 23, 141-148. doi:10.1007/s10896-007-9136-5

24.

Garcia-Moreno

Jansen

H. A. F. M.

Ellsberg

Heise

Watts

C. H.

(2006). Prevalence of intimate partner violence: Findings from the WHO multi-country study on women’s health and domestic violence. The Lancet, 368, 1260-1269. doi:10.1016/S0140-6736(06)69523-8

25.

Ghana Statistical Service. (2009). Ghana demographic and health survey 2008. Accra, Ghana: ICF Macro.

26.

Ghana Statistical Service. (2012). 2010 population and housing census: Summary report of final results. Accra: Author.

27.

Giancola

P. R.

Helton

E. L.

Osborne

A. B.

Terry

M. K.

Fuss

A. M.

Westerfield

J. A.

(2002). The effects of alcohol and provocation on aggressive behavior in men and women. Journal of Studies on Alcohol, 63, 64-73.

28.

Gover

A. R.

Kaukinen

Fox

K. A.

(2008). The relationship between violence in the family of origin and dating violence among college students. Journal of Interpersonal Violence, 23, 1667-1693. doi:10.1177/0886260508314330

29.

Graham

Bernards

Wilsnack

Gmel

(2011). Alcohol may not cause partner violence but it seems to make it worse: A cross national comparison of the relationship between alcohol and severity of partner violence. Journal of Interpersonal Violence, 26, 1503-1523. doi:10.1177/0886260510370596

30.

Graham

Plant

(2004). Alcohol, gender and partner aggression: A general population study of British adults. Addiction Research Theory, 12, 373-384.

31.

Hadi

(2000). Prevalence and correlates of marital sexual violence in Bangladesh. Journal of Interpersonal Violence, 15, 787-805. doi:10.1177/088626000015008001

32.

Haj-Yahia

M. M.

(2000). The incidence of wife abuse and battering and some sociodemographic correlates as revealed by two national surveys in Palestinian society. Journal of Family Violence, 15, 347-374. doi:10.1023/A:1007554229592

33.

Hindin

M. J.

Adair

L. S.

(2002). Who’s at risk? Factors associated with intimate partner violence in the Philippines. Social Science & Medicine, 55, 1385-1399. doi:10.1016/S0277-9536(01)00273-8

34.

Hindin

M. J.

Kishor

Ansara

D. L.

(2008). Intimate partner violence among couples in 10 DHS countries: Predictors and health outcomes (DHS Analytical Studies No. 18). Calverton, MD: Macro International.

35.

Holtzworth-Munroe

Rehman

Herron

(2000). General and spouse-specific anger and hostility in subtypes of martially violent men and nonviolent men. Behavior Therapy, 31, 603-630. doi:10.1016/S0005-7894(00)80034-9

36.

Ishida

Stupp

Melian

Serbanescu

Goodwin

(2010). Exploring the associations between intimate partner violence and women’s mental health: Evidence from a population-based study in Paraguay. Social Science & Medicine, 71, 1653-1166. doi:10.1016/j.socscimed.2010.08.007

37.

Issahaku

P. A.

(2015). Health implications of partner violence against women in Ghana. Violence and Victims, 30, 250-264. doi:10.1891/0886-6708.VV-D-13-00075

38.

Issahaku

P. A.

(2016). Policy suggestions for combating domestic violence in West Africa. International Journal of Sociology and Social Policy, 36, 66-85. doi:10.1108/IJSSP-03-2015-0033

39.

Jewkes

Levin

Penn-Kekana

(2002). Risk factors for domestic violence: Findings from a South African cross-sectional study. Social Science & Medicine, 55, 1603-1617. doi:10.1016/S0277-9536(01)00294-5

40.

Johnson

Ollus

Nevala

(2008). Violence against women: An international perspective. New York: Springer.

41.

Johnson

M. P.

(2006). Conflict and control: Gender symmetry and asymmetry in domestic violence. Violence Against Women, 12, 1003-1018. doi:10.1177/1077801206293328

42.

Kishor

Johnson

(2006). Reproductive health and domestic violence: Are the poorest women uniquely disadvantaged? Demography, 43, 293-307. doi:10.1353/dem.2006.0014

43.

Kuenyehia

Aboagye

E. O.

(2003). Parenting, fostering, adoption and domestic employment. In Kuenyehia

(Eds.), Gender relations in the family: A West African perspective (pp. 38-69). Accra, Ghana: WalWA.

44.

Lacey

K. K.

West

C. M.

Matusko

Jackson

J. S.

(2016). Prevalence and factors associated with severe physical intimate partner violence among U.S. Black women: A comparison of African American and Caribbean Blacks. Violence Against Women, 22, 651-670. doi:10.1177/1077801215610014

45.

Lamphere

(1974). Strategies, cooperation, and conflict among women in domestic groups. In Rosaldo

M. Z.

Lamphere

(Eds.), Woman, culture, and society (pp. 97-112). Stanford, CA: Stanford University Press.

46.

Loseke

D. R.

Gelles

R. J.

Cavanaugh

M. M.

(2005). Introduction: Understanding controversies on family violence. In Loseke

D. R.

Gelles

R. J.

Cavanaugh

M. M.

(Eds.), Current controversies on family violence (2nd ed., pp. ix-xix). Thousand Oaks, CA: Sage.

47.

Luke

Schuler

S. R.

Mai

B. T. T.

Thien

P. V.

Minh

T. H.

(2007). Exploring couple attributes and attitudes and marital violence in Vietnam. Violence Against Women, 13, 5-27. doi:10.1177/1077801206295112

48.

Mahama

(2004). History and traditions of Dagbon. Tamale, Ghana: GILLBT Printing Press.

49.

Mann

J. R.

Takyi

B. K.

(2009). Autonomy, dependence or culture: Examining the impact of resources and socio-cultural processes on attitudes towards intimate partner violence in Ghana, Africa. Journal of Family Violence, 24, 323-335. doi:10.1007/s10896-009-9232-9

50.

Mohammed

(2004). Unmasking masculinity and deconstructing patriarchy: Problems and possibilities within feminist epistemology. In Reddock

R. E.

(Eds.), Interrogating Caribbean masculinities: Theoretical and empirical analyses (pp. 38-67). Jamaica: University of the West Indies Press.

51.

Moore

H. L.

(1994). A passion for difference: Essays in anthropology and gender. Bloomington: Indiana University Press.

52.

Naved

R. T.

Persson

L. A.

(2005). Factors associated with spousal physical violence against women in Bangladesh. Studies in Family Planning, 36, 289-300. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16395946

53.

Nukunya

G. K.

(2003). Tradition and change in Ghana: Introduction to sociology (2nd ed.). Accra: Ghana Universities Press.

54.

Obi

S. N.

Ozumba

B. C.

(2007). Factors associated with domestic violence in south-east Nigeria. Journal of Obstetrics and Gynaecology, 27, 75-78. doi:10.1080/01443610601056509

55.

Onigbogi

M. O.

Odeyemi

K. A.

Onigbogi

O. O.

(2015). Prevalence and factors associated with intimate partner violence among married women in an urban community in Lagos State, Nigeria. African Journal of Reproductive Health, 19, 91-100. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/26103699

56.

Ortner

S. B.

(1972). Is female to male as nature is to culture? Feminist Studies, 1(2), 5-31.

57.

Pence

Paymar

(1993). Education groups for men who batter: The Duluth model. New York: Springer Publishing Co.

58.

Republic of Ghana. (2007). Domestic Violence Act, 2007 (Act 732). Accra: Author.

59.

Rosaldo

M. Z.

(1974). Woman, culture, and society: A theoretical overview. In Rosaldo

M. Z.

Lamphere

(Eds.), Woman, culture, and society (pp. 17-42). Stanford, CA: Stanford University Press.

60.

Singleton

R. A.

Jr. Straits

B. C.

(2005). Approaches to social research. New York, NY: Oxford University Press.

61.

Snow

D. L.

Sullivan

T. P.

Swan

S. C.

Tate

D. C.

Klein

(2006). The role of coping and problem drinking in men’s abuse of female partners: Test of a path model. Violence and Victims, 21, 267-285. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16761854

62.

Stark

(2007). Coercive control: The entrapment of women in personal life. Oxford: Oxford University Press.

63.

Straus

M. A.

Gelles

R. J.

Steinmetz

S. K.

(1980). Behind closed doors—Violence in the American family. Garden City, NY: Anchor Books.

64.

Takyi

B. K.

Mann

(2006). Intimate partner violence in Ghana, Africa: The perspectives of men regarding wife beating. International Journal of Sociology of the Family, 32, 61-78. Retrieved from http://www.jstor.org/stable/23028786

65.

Temple

J. R.

Weston

Stuart

G. L.

Marshall

L. L.

(2008). The longitudinal association between alcohol use and intimate partner violence among ethnically diverse community women. Addictive Behaviors, 33, 1244-1248. doi:10.1016/j.addbeh.2008.05.005

66.

Thompson

M. P.

Kingree

J. B.

(2006). The roles of victim and perpetrator alcohol use in intimate partner violence outcomes. Journal of Interpersonal Violence, 21, 163-177.

67.

Townsend

Jewkes

Mathews

Johnston

L. G.

Flisher

A. J.

Zembe

Chopra

(2011). HIV risk behaviors and their relationship to intimate partner violence (IPV) among men who have multiple female sexual partners in Cape Town, South Africa. AIDS Behavior, 15, 132-141. doi:10.1007/s10461-010-9680-5

68.

Vakili

Nadrian

Fathipoor

Boniadi

Morowatisharifabad

M. A.

(2010). Prevalence and determinants of intimate partner violence against women in Kazeroon, Republic of Iran. Violence and Victims, 25, 116-127. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/20229697

69.

West

D. J.

Roy

Nichols

F. L.

(1978). Understanding sexual attacks. London, England: Heinemann.

70.

World Health Organization. (2001). Putting women first: Ethical and safety recommendations for research on domestic violence against women. Geneva, Switzerland: Author.

71.

World Health Organization. (2002). World report on violence and health. Geneva, Switzerland: Author.