Keywords

1 Introduction: Religion and Intimate Partner Violence

There is widespread consensus that intimate partner violence (IPV) is a serious public health problem that affects a sizable number of Americans. Although estimates of IPV rates vary depending on the sample studied and how questions are worded, commonly cited estimates are that from 7.6 to 11.5 % of men and 12 to 25 % of women are physically and/or sexually assaulted by an intimate partner each year (Tjaden and Thoennes 2006; Black et al. 2011). Although some researchers have argued that women are as likely as men to assault their intimate partners (see, for example, Archer 2000; Dutton 2006; Straus 2007; Dutton et al. 2009), the vast majority of studies show that there are important differences in IPV perpetrated by women and men. For example, studies show that women’s and men’s motivations for using violence against an intimate partner differ. Men are more likely to use violence when they perceive themselves losing control of the relationship or when they interpret their partners’ words or behavior as challenges to their authority. In contrast, women are more likely to use violence, especially severe physical violence, in self-defense, when they believe they are in imminent danger of being attacked, or in retaliation for being attacked (Barnett et al. 1997; Dobash et al. 1998; Miller 2001; Rajan and McCloskey 2007). Moreover, men’s violence is typically more severe; women are more likely to be injured when they are assaulted by their intimate partners, and their injuries are more serious and more likely to require medical treatment or hospitalization (Archer 2000; Tjaden and Thoennes 2000; Menard et al. 2009). And men are more likely than women to kill their intimate partners (Fox and Zawitz 2007).

In light of the data on the incidence and seriousness of male-perpetrated intimate partner violence, it is not surprising that a great deal of research has been undertaken to identify factors that may increase or decrease men’s likelihood of abusing their female partners. Researchers report, for instance, that social class and employment status are related to IPV perpetration, with lower socioeconomic status and poverty, as well as unemployment, being associated with an increased risk of IPV perpetration by men (Renzetti 2009; Renzetti and Larkin 2009; Brush 2011). Researchers have also found that younger men (aged 18–30), men with substance abuse problems, and those who endorse male dominance in intimate relationships are also at greater risk for IPV perpetration (Luthra and Gidycz 2006; Bennett and Bland 2008; DeKeseredy and Dragiewicz 2009; Basile and Black 2011).

Interestingly, however, few researchers until relatively recently have examined how religion may be related to men’s IPV perpetration. This is surprising because there are several good reasons to hypothesize that religion may influence the likelihood of IPV perpetration. On one hand, many religious traditions have supported and reinforced patriarchal ideology, such as the notion of male dominance and the importance of men as authority figures and instrumental leaders of their households. One might speculate, therefore, that men who are strongly religious and adhere to these patriarchal ideals might be more inclined to use violence against an intimate partner whom they judge as having committed a “sin” or who usurped male authority in some way (Nason-Clark 2000). On the other hand, given that religious values and teachings also encourage pro-social behavior, including empathy and a concern for the welfare of others, one might hypothesize that adherence to these kinds of religious tenets would reduce men’s likelihood of using violence against an intimate partner (see, for example, Johnson 2011). In this case, religion might motivate intimate partners to be more accommodating to one another and to compromise with one another when there is disagreement, so that conflicts are more likely to be defused rather than to escalate to violence.

But the study of religion and intimate partner violence is complicated by a number of factors. For one thing, religion and religious expression are highly diverse. Dollahite et al. (2004), for instance, note that the United States is one of the most religious and religiously diverse countries in the world today in terms of the population’s voluntary participation in religious activities and institutions. They cite research that identifies more than 2,600 distinct faith communities in the U.S. and Canada, with 216 major Christian denominations alone. As Dollahite et al. (2004: 412) also point out, “there is as much diversity within major faith groups as between them.” Social scientists often assign these diverse faith communities to rather broad categories, such as “Liberal, Moderate, Conservative,” but critics of this approach maintain that, in practice, these labels may have little to do with specific markers of religiosity or with individuals’ identification with religious movements, and they tend to “conflate religious, economic, social and political ideas into one monolithic measure” (Woodberry et al. 2012: 65). One’s affiliation with a particular faith group may be less important with regard to behavioral outcomes than how one is religious.

In this chapter, we discuss research studies, including our own recent work, that have examined the effects of religiosity on male perpetration of intimate partner violence. And in doing so, we will identify some of the major gaps in this body of research and suggest areas and topics for future study that would help us better understand the potential of religiosity as a protective factor or a risk factor for IPV perpetration.

2 Dimensions of Religiosity and Their Relationship to IPV Perpetration

Dollahite et al. (2004: 413) remark that religiosity is not only complex, it is also multifaceted. They identify three major dimensions of the construct: (1) religious practices (for example, outward, observable expressions of faith), (2) religious beliefs (for example, personal beliefs, framings, meanings), and (3) religious communities (support, involvement and relationships grounded in a congregation or other religious group). Although each of these dimensions of religiosity may be distinguished from one another, they are undoubtedly interrelated. Yet, researchers examining the effects of religiosity on IPV perpetration have typically measured religiosity narrowly in terms of just one of these dimensions or even with a single indicator of one dimension.

2.1 Religious Practices and IPV Perpetration

Most research on the relationship between religiosity and IPV perpetration has used rather narrow measures of religious practice. Ellison and colleagues (Ellison and Anderson 2001; Ellison et al. 2007), for example, measured religiosity in terms of frequency and consistency of attendance at religious services. Their research consistently supports the hypothesis that religiosity reduces likelihood of IPV perpetration. Ellison and Anderson (2001) found that among both men and women, those who attended religious services more often and more consistently were less likely to perpetrate IPV, even after controlling for religiosity’s effects on increasing social integration and social support, and decreasing substance abuse and psychological problems. Moreover, the protective effects of religiosity with regard to IPV perpetration remained significant regardless of whether data measuring IPV perpetration were from self-reports or from partner reports – an important finding given the tendency of IPV perpetrators to underestimate the frequency and severity of their violence. In a more recent analysis, Ellison et al. (2007) found that the protective effect of religiosity (that is, church attendance) varies by race/ethnicity: It strongest for African American men and women and also for Hispanic men, groups that are at especially high risk for IPV perpetration.

Cunradi et al. (2002) also found that attendance at religious services is associated with reduced IPV perpetration, but they report that alcohol consumption appears to be significantly more important than religiosity in predicting men’s likelihood of perpetrating IPV. In a study of a large, nationally representative sample of married adults, DeWall (2010) measured religiosity in terms of frequency of engaging in religious services as well as participants’ self-assessed importance of religious beliefs in their daily lives. He found that although alcohol abuse was associated with greater IPV perpetration, this was the case only among participants low in religiosity. In his study, high religiosity served as a protective factor against IPV perpetration and also appeared to break the commonly observed link between alcohol use and IPV.

2.2 Multidimensional Measures of Religiosity

Researchers critical of using attendance at religious services as the sole measure of religiosity have instead used scales composed of multiple items that may subsequently be summed to yield a religiosity “score.” For instance, in our recent study of religiosity and IPV perpetration (Renzetti et al. 2012), we used the Religious Commitment Inventory (RCI-10; Worthington et al. 2003) to capture types of religious practice (apart from attendance at religious services) as well as the importance of religious beliefs. The RCI-10 is composed of 10 items with a 5-point Likert response scale ranging from “not at all true of me” (1) to “totally true of me” (5). Items cover such aspects of religious practice as reading books or magazines about one’s faith, making financial contributions to one’s religious organization, and spending time in private religious thought or meditation. Items measuring the importance of religious beliefs include, “My religious beliefs lie behind my whole approach to life,” and “Religious beliefs influence all my dealings in life.”

A national sample of 251 adult men, ranging in age from 18 to 85 (mean = 43.3 years) who had been in a committed intimate relationship for at least 5 years completed a web-based survey that included the RCI-10 (to measure the independent variable, religiosity) as well as four subscales from the revised Conflict Tactics Scales (CTS2; Straus et al. 1996) – minor physical assault, major physical assault, minor psychological aggression, major psychological aggression (to measure the dependent variable, IPV) – using an 8-point Likert scale (0 = never happened, 8 = happened more than 20 times). Perpetration of these acts was measured for the past year.Footnote 1

Scores on the RCI-10 could range from 10 to 50; the mean religiosity score for this sample was 31.43 (SD = 3.76). The sample was then divided into two groups – low religiosity (scores ranging from 10 to 29) and high religiosity (scores ranging from 30 to 50) – for bivariate analysis.

Scores on the scale of minor physical violence ranged from 0 to 30 (Mean = 3.68, SD = 8.08), and 26.1 % of the men reported perpetrating at least one act of minor physical violence in the past year. In addition, 19.9 % reported perpetrating as least one act of major physical violence in the past year (range = 0–24, Mean = 4.56, SD = 10.64). Not surprisingly, perpetration of psychological aggression was more common: 66.7 % of the men reported perpetrating at least one act of minor psychological aggression during the past year (range = 0–24, Mean = 4.44, SD = 5.04); 24.1 % reported perpetrating at least one act of major psychological aggression in the past year (range = 0–24, Mean = 3.49, SD = 7.49). For bivariate analysis, results for each type of violence were collapsed into two groups: men who had perpetrated at least one act of that type of violence in the past year, and men who had not perpetrated that type of violence in the past year.

Chi-square statistics were calculated to test the relationship between religiosity and IPV perpetration. In all analyses, we controlled for race/ethnicity and income, since previous studies have shown that both race/ethnicity and income have independent effects on the likelihood of IPV perpetration. None of the results of these tests were statistically significant, although they all approached significance. Men who scored high in religiosity on the MCI-10 perpetrated fewer acts of both minor (p = .07) and major psychological aggression (p = .07) than those men who scored low in religiosity. But to our surprise, men who scored high in religiosity on the MCI-10 also perpetrated more acts of both minor (p = .10) and major (p = .06) physical violence than men who scored low in religiosity.

Although these findings are at first glance puzzling – why would religiosity be associated with a reduction in psychological IPV, but an increase in physical IPV? – they remind us, as noted previously, that religiosity in and of itself may not be meaningfully related to likelihood of perpetrating IPV without considering how a person is religious.

2.3 Kinds of Religiousness and IPV Perpetration

Pargament (2002) points out that the influence of religiosity on specific behavioral outcomes depends to some extent on the degree to which one’s religion is integrated into one’s everyday life. As social psychological research on religiosity has shown, in order for religious values to influence attitudes and behavior, they must be internalized (Blais et al. 1990). When religious values are internalized, the individual “takes on” the values as his or her own. Drawing on self-determination theory, Richard Ryan and his colleagues (Ryan et al. 1993) identified two styles of religious internalization that affect how people regulate their own behavior. They label these introjected religious self-regulation and identified religious self-regulation.

According to self-determination theory, introjection is a self-regulating style in which an individual is motivated to adopt a value or rule (or a religious teaching or practice) because he or she wishes to gain or wishes not to lose the approval of others; they want to avoid guilt and shame. In contrast, identification is a self-regulating style in which an individual is motivated to adopt a value or rule (or a religious teaching or practice) because it gives him or her enjoyment, it is intrinsically satisfying to him or her, and it corresponds to his or her personal values and beliefs. Ryan et al.’s (1993) research and that of others (for example, Blais et al. 1990) show very different emotional and behavioral outcomes for each of these types of religious self-regulation. Those high in introjected regulation have higher rates of depression, lower self-esteem, lower perceived self-efficacy, and tend to use less adaptive coping strategies. In contrast, those high in identified regulation are more likely to express greater life satisfaction, higher self-esteem, a greater sense of self-efficacy, and tend to use more positive and adaptive coping strategies (Ryan et al. 1993). Consequently, in our recent study of the relationship between religiosity and IPV perpetration, we also examined whether how men are religious, in terms of their style of religious self-regulation, affects their likelihood of perpetrating IPV.

The 251 men in our sample, completed the Religious Self-Regulation Questionnaire (RSQ; Ryan et al. 1993), which is composed of 12 items – 6 that measure introjected religious self-regulation, and 6 that measure identified religious self-regulation – using a 7-point Likert scale (1 = not at all true, 7 = very true). Items that measure introjected religious self-regulation include: “One reason I think it’s important to actively share my faith with others is because I would feel bad about myself if I didn’t,” and “A reason I think praying by myself is important is because if I don’t, God will disapprove of me.” Items that measure identified religious self-regulation include: “When I turn to God, I most often do it because I enjoy spending time with Him,” and “An important reason why I attend church is by going to church I learn new things.”

We found that men who scored high in introjected religious self-regulation perpetrated significantly more acts of minor physical violence (r = .312, p < .001), major physical violence (r = .317, p < .001), minor psychological aggression (r = .221, p < .001), and major psychological aggression (r = .315, p < .001). With regard to the relationship between identified religious self-regulation and IPV perpetration, however, all of the correlations were weak and none was statistically significant, although they were in the expected direction. That is, men who scored high on identified religious self-regulation perpetrated fewer acts of minor physical violence (r = -.062, ns), major physical violence (r = −.062, ns), minor psychological aggression (r = −.062, ns), and major psychological aggression (r = −.064, ns).

It is worth reiterating that introjected religious self-regulation has been found to be associated with several negative psychological outcomes, including depression, low self-esteem, low perceived self-efficacy, and use of less adaptive coping strategies. Importantly, research with IPV perpetrators has found that they typically have low self-esteem and low perceived self-efficacy, and they tend to use negative coping strategies (Edleson and Tolman 1992; Holtzworth-Munroe and Mehan 2004). In contrast, men (and women) who are high in identified religious self-regulation have been found to express greater life satisfaction, high self-esteem, high perceived self-efficacy, and tend to use positive, adaptive coping strategies. Although the correlations between identified religious self-regulation and IPV perpetration in our study were weak and not statistically significant, they were in the expected direction. Thus, religiosity per se, whether high or low, may be less functionally important than how adherents to a religion are religious. This point is further underlined when one examines the intersection of beliefs and behavior.

2.4 Religious Beliefs, Gender Attitudes, and IPV Perpetration

Researchers have found no relationship between religious denomination (for example, Cunradi et al. 2002) or denominational homogamy (for example, Ellison et al. 1999) and likelihood of perpetrating intimate partner violence. Nevertheless, some denominations are associated with more conservative values and beliefs – with regard to gender roles and intimate relationships as well as other social issues – than other denominations, and within a specific denomination, congregations often vary in their level of traditionalism. Given that researchers report that traditional, as opposed to egalitarian, attitudes about gender roles – that is, beliefs that support stereotypical traits and behavior for women and men – are one of the most consistent predictors of attitudes that support violence against women, one might expect that adherents to religious beliefs, values, and teachings that promote traditional gender roles would be more likely to perpetrate, or at least to legitimate, violence against an intimate partner, particularly if that partner violates strongly held norms and beliefs (Berkel et al. 2004). The research that has tested this hypothesis has produced equivocal findings at best. For example, while Berkel et al. (2004) report that the college students in their sample who held traditional gender role attitudes were more likely than those who held egalitarian gender role attitudes to express more support for the use of violence against women, Cunradi et al. (2002) found that rates of IPV perpetration were higher among men in religious groups classified as liberal than those in religious groups classified as fundamentalist. Recall that in our study (Renzetti et al. 2012) men who scored high in religiosity as measured by the RCI-10 were significantly less likely to perpetrate both minor and major psychological IPV. Interestingly, we (Renzetti et al. 2011) also found that this relationship was even stronger for men who endorsed items on the Old Fashioned Sexism Scale (p < .001).Footnote 2

Cunradi et al. (2002: 149) speculate that such findings reflect the centrality of the patriarchal family unit among fundamentalists, which may lower their IPV risk, whereas couples in more liberal religious groups “may be less socially constrained by religiously ordained norms and rules governing discourse within family life.” As Pargament (2002) points out, while psychologists have been critical of religious fundamentalism for its association with rigid thinking and authoritarianism, it holds some clear advantages for adherents, including greater marital happiness, satisfaction and commitment, reduced marital conflict, and more effective marital problem-solving strategies. Additional research shows that highly religious couples report that their religious beliefs help them cultivate a shared vision and purpose, enhance their relational virtues, reduce stress levels in their marriage, facilitate relational reconciliation following a conflict, and enhance partners’ willingness to forgive one another (Lambert and Dollahite 2006; see also Wilcox and Wolfinger 2007).

Pargament (2002: 176) cautions that the helpfulness of religiousness may vary across context and situation and, moreover, that “the efficacy of religion may have less to do with specific religious beliefs and practices and more to do with the degree to which religion is well integrated into individuals’ lives.” Both of these hypotheses are certainly worthy of further empirical investigation with regard to how religiosity influences men’s likelihood of perpetrating intimate partner violence.

3 Directions for Future Research

As we noted in the introduction, research on the relationship between religiosity and intimate partner violence perpetration is relatively undeveloped. Much of the research to date has been limited by narrow measures of religiosity that are unidimensional and fail to capture the complexity of this construct. We sought to overcome this limitation by using the RCI-10 and the RSQ. Another promising measure is the Brief Multidimensional Measure of Religiousness/Spirituality (BMMRS; Fetzer Institute 1999), which is designed to tap 12 different dimensions of religiosity, including daily spiritual experiences, religious and spiritual coping, and religious support. This last element is noteworthy, since most studies have neglected the “religious communities” dimension of religiosity identified by Dollahite et al. (2004). Pargament (2002: 174) also stresses the need for religious measures “linked to particular contexts and situations” because these are “stronger predictors of well-being than global distal religious measures.” In addition, he reminds us that some of the influence of religion and religiosity on behavior may be direct, but some may also be mediated “by more finely delineated religious dimensions” (178). Future studies of IPV perpetration, therefore, should develop explanatory models that include both direct and mediated effects of religion/religiosity.

A related concern stems from the measures of intimate partner violence that are typically used in these studies. In secondary analyses of large, national survey datasets (for example, Ellison et al. 2007), IPV perpetration is usually measured with one or two items, such as “During the past year, how many fights with your [partner] resulted in you hitting, shoving or throwing things at her?” Methodological research has shown that such narrow measures result in gross underestimations of IPV frequency and severity (Jaquier et al. 2011). The most commonly used instruments, though, are the Conflict Tactics Scales (CTS) and Revised Conflict Tactics Scales (CTS-2). Both of these have been extensively criticized, largely for their inability to capture the context, meaning, and motivations underlying the violent event (see DeKeseredy and Schwartz 1998).

In addition to more nuanced measures of both religiosity and IPV, future studies should include more diverse samples. Although there are some notable exceptions (see, for example, Landau et al. 2002; Nojomi et al. 2007; Vakili et al. 2010), much of the research to date has sampled largely Christian groups, many of which reside in the U.S. or other Western industrialized nations. Future research should strive to include members of non-Christian religions and spiritual groups, including New Age, “congregation free” spirituality, and other religious minorities (Dollahite et al. 2004). Comparative research conducted in societies in which non-Christian groups dominate and in which rates of IPV have been found to vary significantly from rates in the U.S and Western industrialized countries would also enrich our understanding of the direct and mediated effects of religion/religiosity on intimate partner violence perpetration. And given that levels and types of religiosity as well as rates of IPV perpetration vary by place of residence (e.g., rural, suburban, urban communities; southern United States relative to other regions of the country), researchers need to attend to regional variations in their studies of the relationship between religion/religiosity on intimate partner violence perpetration.Footnote 3

Nearly all of the research on religiosity and IPV has been correlational and cross-sectional. Such studies do not allow us to draw conclusions about causality. Although we assume that specific religious beliefs and practices promote or inhibit certain behaviors, such as IPV perpetration, it may be the case that aggressive, violent, or hostile people, particularly those who are hostile to women, are attracted to or seek out religions that support their attitudes. Longitudinal studies would likely give us greater confidence in our causal inferences about the effects of religiosity on IPV perpetration, especially given that religious beliefs and commitment typically change over time for many people (Pargament 2002; Dollahite et al. 2004).

Finally, we join several other scholars who call on researchers to be more open to and less negatively biased toward individuals, couples, and families “who take their religion very seriously” (Stark and Finke 2000: 14; see also Dollahite et al. 2004; Johnson 2011). The empirical literature indicates that most religions, even in their fundamentalist forms, may be both beneficial and detrimental to adherents (Pargament 2002). However, we are unlikely to identify the benefits as well as the detriments of religion in relation to intimate partner violence perpetration if our research is colored by what Johnson (2011: xii) calls “the last acceptable prejudice” in an age of political correctness. As Pargament (2002: 168) point out, “Questions about the general efficacy of religion should give way to the more difficult but appropriate question: how helpful or harmful are particular forms of religious expression for particular people dealing with particular situations in particular social contexts according to particular criteria of helpfulness or harmfulness?” To discover the complex but accurate answers to this question, though, requires us, as social scientists, to remove the blinders of our prejudice against highly religious people.