Power and Control in Kathmandu

Abstract

We argue that the concept of power has been inadvertently sidelined in recent theory and research on husband violence. Three types of relationship power may matter with respect to husband violence: attempted power, actual power, and achieved power. Analyses of a randomly selected representative sample of 270 married or partnered women in Kathmandu showed that actual power was related to husband violence prevalence, severity, and injury. Achieved power was related to husband violence prevalence and severity, and attempted power was related to husband violence injury. Implications are discussed.

Keywords

domestic violence power coercive control intimate partner violence Nepal

Introduction

It is well documented that domestic violence occurs in a wider context of inequality between men and women (Dhakal, 2008). Classic feminist theory of domestic violence roots its explanation in patriarchy (Dobash & Dobash, 1979). Hence, high rates of domestic violence (Paudel, 2007 reports 35% of Nepali women suffer from gender-based violence in the home) in the context of social inequality suggest circumstances ripe for an analysis of the relationship between relationship power and domestic violence in Nepal. Domestic violence is a serious problem in Nepal, where violence by a family member is responsible for about 80% of violence against women (Dhakal, 2008). A growing literature documents domestic violence in Nepali families in the immigrant community (cf. Yoshihama, Ramakrishnan, Hammock, & Khaliq, 2012). However, despite widespread reports of beatings, dowry-related femicide, and a sense that the 1,100 cases of domestic violence recorded by the police every year constitute the “tip of the iceberg” (Dhakal, 2008), domestic violence remains understudied in Nepal itself.

Despite classic studies suggesting the importance of power as an important predictor of domestic violence (cf. Coleman & Straus, 1986; Dobash & Dobash, 1979), accounts of power differentials as explanations of domestic violence have lost traction in recent years. This is partly the result of direct critiques (cf. Dutton & Nicholls, 2005) and partly the unintended result of a focus on the importance of control attempts in the theoretical literature (cf. Johnson, 2008). Control is arguably a narrower construct than power, with more immediate consequences of being manipulated against one’s will for the disadvantaged party. It is also, however, arguably a subset or specific kind of power. Thus, the common phrase “power and control” may be somewhat redundant. However, use of the term control in domestic violence research is overwhelmingly associated with control tactics. The power and control wheel of Pence and Paymar (1993) is a list of strategies that perpetrators use to try to get power over victims. This stress on control attempts is continued in work on Johnson’s popular typology at both the theoretical and empirical levels (Johnson, 1995, 2008; Johnson & Leone, 2005). Thus, it is that much of the recent literature on domestic violence focuses not on whether and how perpetrators have power over victims, but how and whether they try to get power over victims. A perpetrator may work very hard to try to get power without actually obtaining it. Ironically, much of the research about power and domestic violence is not about power as it is usually understood.

A recent theoretical account restores power to a central place in its understanding of domestic violence (Emery, 2011). Drawing on Emery (2011), we argue that power, with respect to husband violence, may matter in three ways: (a) whether the perpetrator is motivated to try to get power over the victim, (b) the actual power the perpetrator has over the victim, and (c) whether the victim accepts or legitimates the perpetrator’s power over her. We refer to these as (a) attempted power, (b) actual power, and (c) achieved power, respectively. Attempted power captures the control tactics extensively theorized and studied empirically in the research literature. Actual power helps recenter the discussion on the subjective lived experience of women. Achieved power puts stress on the legitimacy of power differentials which, according to Emery (2011), has implications for domestic violence and the long-term stability of power differentials. This article adds empirical nuance to the theoretical discussion by examining the roles of men’s attempted power, actual power, and achieved power in predicting any violence by husbands, the severity of violence by husbands, and injury from violence by husbands in Kathmandu.

Power in the Relationship: How and Why It May Matter

Although recent research on power argues that there are six bases of power (French & Raven, 1959), we argue, in agreement with Emery (2011), that this is substantially beyond the remit of the original subject. For example, we would argue that the so-called expert power of a doctor is not properly power, but rather influence.¹ Weber (1922/1978) initially defined power as an individual or group’s ability to “realize their own will in a social action even against the resistance of others who are participating in the action” (p. 926). Weber argues that this is sociologically amorphous, and hence the concept requires a more precise definition. For this reason, he limited the definition to “the probability that a command will be obeyed” (p. 53).² In terms of intimate relationships, this maps onto the probability that an ultimatum made by one partner will be obeyed. In some intimate relationships power, so defined, may be very fluid, varying from domain to domain as well as over the course of time. However, insofar as systematic sociostructural differences in walk-away costs for men and women persist over time, power in most relationships is likely to have a stable component. Stability in the distribution of power over time and domain is even more likely when relationships are comparatively ordered rather than chaotic (see Emery, 2011).

For the purposes of this article, domestic violence is defined as “an act or threat of physical force in a relationship between sexually intimate partners” (Emery, 2011, p. 528). In classic feminist accounts of domestic violence, violence is used in a systematic and more or less instrumental fashion to achieve control over women (Dobash & Dobash, 1979; Martin, 1981; Pence & Paymar, 1993). These accounts conceptualize domestic violence as a form of patriarchal order backed by force (Emery, 2011). In these accounts, the power of a man in a particular relationship rests on the use of violence, but that power also projects into the larger community, affecting cost–benefit calculations of women in non-violent relationships. Likewise, in feminist accounts, inequalities in the larger society exacerbate the power dynamics and scope of violent control (Stark, 2007) experienced by battered women (Yllo & Bograd, 1988; Yllo & Straus, 1990). In this interpenetration of patriarchy and violence, violence supports patriarchal control in the relationship and the larger society, and patriarchal control in the larger society either openly supports or minimizes and obfuscates violent control of women in intimate relationships. The feminist account, in turn, is arguably rooted in a larger sociological tradition that ties power and the use of force to the existence of social order (Dahrendorf, 1970; Durkheim, 1893/1984; Emery, 2011).

Emery (2011) argues that power has been unintentionally sidelined in recent theoretical considerations of domestic violence because of theoretical stress on a control motive (cf. Johnson, 2008) rather than achieved control (power). Empirical studies have documented a relationship between power in relationships and domestic violence (Babcock, Waltz, Jacobson, & Gottman, 1993; Coleman & Straus, 1986; Frieze & McHugh, 1992; Kim & Emery, 2003; Sagrestano, Heavey, & Christensen, 1999). However, some studies (cf. Babcock et al., 1993) find that it is when husbands have less power that they are more violent. Empirical findings that continue to link power and patriarchal beliefs to domestic violence (e.g., Emery, Kim, Song, & Song, 2013) suggest that the perpetrator’s perception of having less power, rather than the perpetrator’s actual power, may be driving that relationship (Sagrestano et al., 1999).

Emery (2011) does not make concrete predictions about the prevalence of domestic violence and unequal power distribution in marriages or marriage-like relationships. Instead, he argues that the etiology of domestic violence must be understood as fundamentally tied to the power context. That is, domestic violence in relationships with large power differentials must be understood as having a different constellation of causes than domestic violence in relationships with low power differentials. Such an argument suggests that power is an important moderator of other variables but not necessarily a main effect. Emery’s (2011) work does suggest, indirectly, that husband violence severity and injury may be higher in relationships characterized by large power differentials. We suggest that a more nuanced understanding of the role of power in the relationship is needed to appropriately understand domestic violence. Power may matter as a motive, it may matter in terms of actual inequality in the relationship, and it may matter in terms of achieved or normative control.

Control Motive: Attempted Power

Johnson’s (2008) conceptualization of the role of power is rooted in the empirical work of Holtzworth-Munroe and Stuart (1994). In this conception, it is the perpetrator’s motive to get power over the victim that is most important. When husband violence is used by perpetrators as part of a larger campaign to get power over their victims, it qualifies as Johnson’s (2008) most severe type of violence: intimate terrorism. Johnson uses the motive for control to distinguish between common couple violence and what was first described as patriarchal terrorism (Johnson, 1995) and later intimate terrorism (Johnson, 2008). Johnson follows Holtzworth-Munroe and Stuart (1994) in arguing that perpetrators with strong control motives are more likely to have underlying Axis II psychopathology that will be associated with more severe and lethal violence. Recent research has found that victims of intimate terrorism were more likely to have posttraumatic stress disorder (PTSD), miss work, be injured, use pain killers, and suffer from an escalating pattern of violence than others (Graham-Kevan & Archer, 2003; Johnson & Leone, 2005). Johnson’s (2008) typology is probably the dominant one in the field (Emery, 2011). Given the theoretical clarity and the empirical support for Johnson’s (2008) account, this accepted position is probably merited. Johnson’s (1995) argument that differences in qualitative and quantitative work on husband violence boiled down to a missing data problem represented an enormous step forward in the field of domestic violence.

Emery (2011), however, argues that Johnson’s (2008) conceptualization somewhat misses the point, and that power, rather than control motive, is really the variable of interest. This is because attempted power does not include a rather essential item: whether the attempt to get power over the victim was successful. Unsuccessful control attempts might render some perpetrators pathetic, rather than terrifying, in the victim’s eyes. Thus, it could be the case that control motive is correlated with husband violence and its sequelae because of its high correlation with actual power in the relationship. That is, it could be that attempts to get power only matter insofar as they actually succeed.

Actual Power

It is actual power in the relationship on which the feminist tradition focuses (Dobash & Dobash, 1979). Apart from the control needs of perpetrators posited by Holtzworth-Munroe and Stuart (1994), disempowerment of women facilitates the perpetration of violence because women with less power cannot as easily make the use of violence costly to the perpetrator. Emery (2011) argues that power in non-violent relationships rests on the principle of least interest (Waller, 1937), but that violence undermines power that rests on this principle.

Violent control, when it is effective, systematically deprives women of resources that would allow them to impose the relationship walk-away costs on the perpetrator (leave the relationship) and undermines their ability to acquire resources and invoke social control. This is why the image of the perpetrator who attempts to grab power using violence but cannot because he is wheelchair bound (Emery, 2011) poses an intuitive obstacle to Johnson’s (2008) concept of intimate terrorism. When the victim is not, in fact, disempowered and the perpetrator’s attempts to grab power fail, the image ceases to terrify. However, when power disparities in the intimate relationship are large, a simple threat, gesture, or facial expression may be sufficient to invoke terror. Actual power disparities in the relationship facilitate the perpetrator’s use of violence by increasing the likelihood that he will continue to get away with it. Although many instruments have been developed to measure women’s power in intimate relationships [the most famous example being Blood and Wolfe’s (1960) decision power index], such approaches can be problematic because men may delegate decision-making power about routine matters while still retaining most of the power in the relationship. Sometimes, the simplest approach is to ask women, straightforwardly and subjectively, about how much power they feel they have in the relationship. Indeed, Frieze and McHugh (1992) found that battered women subjectively reported having less power in their relationships than other women.

Achieved (Legitimated) Power: Patriarchal Beliefs of Wives

If a man’s actual power in the relationship decreases the costs for the use of violence, power, including the right to use violence, that is legitimated by women further decreases the costs for the use of violence because women with patriarchal beliefs may accept “correction” (Martin, 1981) of an errant partner as a man’s right. Women attacked in such a circumstance are unlikely to seek help, to leave, or seek to impose costs unless the violence ranges outside whatever bounds have been established as acceptable “correction.” It may be that women with such beliefs are more likely to respond with shame and attempts to conceal the violence from those who might intervene. Such patriarchal beliefs may or may not represent a Gramscian (Gramsci, 1992) form of hegemonic control, false consciousness, or brain washing. Certainly, many perpetrators seek to establish acceptance of domestic violence by the victim, but whether the patriarchal beliefs emanate from the perpetrator, from the larger society, or have been actively developed by women as part of identity construction has little bearing on their importance vis à vis perpetration of violence. Women’s patriarchal beliefs decrease costs of violence directly by making remedy seeking less likely. Patriarchal beliefs also decrease costs of violence indirectly by leaving the power structure in the intimate relationship uncontested. Although Emery, Kim, et al. (2013) found a significant interaction between men’s patriarchal beliefs and their abuse of the children in predicting abuse of wives, they did not find an interaction between women’s patriarchal beliefs and women’s abuse of the children in predicting women’s violence against their husbands. However, Emery, Kim et al. (2013) did not examine women’s patriarchal beliefs as a predictor of men’s violence.

Based on the foregoing argument, we postulate the following:

Attempted power will be associated with any violence by the male partner, the severity of violence by the male partner, and injury from violence by the male partner.

Actual power in the relationship will be associated with any violence by the male partner, the severity of violence by the male partner, and injury from violence by the male partner.

Achieved power (patriarchal beliefs of female partners) will be associated with any violence by the male partner, the severity of violence by the male partner, and injury from violence by the male partner.

Neutralizing Beliefs and Other Controls

A strong theoretical and empirical literature suggests that the neutralizing beliefs of the perpetrator are a key variable in understanding crime generally and domestic violence specifically (Matza, 1964/1990; Mooney, 2007; Ptacek, 1988). The general idea is that most men possess, to varying degrees, socialized norms against wife abuse. To temporarily suspend these norms and excuse the perpetration of violence, perpetrators of domestic violence must develop neutralizing beliefs. Because such beliefs may be correlated with the patriarchal beliefs of female partners, we control for the male partner’s neutralizing beliefs about husband violence as reported by the female partner.

We also control for some of the antecedents of relationship power suggested by Emery (2011) and Emery and Wu (2016). These include the female to male income ratio, education ratio, and age ratio. They also include female to male height ratio and female to male weight ratio. These latter two ratios are also of interest because Straus and Gelles (1990) suggest that sex differences in height and weight are partially responsible for differences in domestic violence injury between men and women. All models also control for a social desirability of reporting husband violence measure, whether the family owns a car, household income, and the female partner’s age, education, caste, and number of years in the current relationship. Analyses of husband violence injury also control for the severity of the violence.

Method

Data

The Kathmandu Families and Neighborhoods Study (KFNS) is a representative random probability cluster sample of married or partnered women in 20 wards in Kathmandu.³ Wards were selected using probability proportional to size (PPS) sampling; households were selected using simple random sampling of ward lists. Face-to-face interviews were conducted orally in the women’s homes. The KFNS was designed to collect data on domestic violence and community characteristics; power analysis suggested that a sample size of 300 was adequate. The first author developed a two-day domestic violence interview training which the second author delivered to the interview team in Kathmandu. Interviewers were strictly trained to protect participant privacy both during and after the interviews, making sure that participants could not be monitored by partners or others and that the completed surveys were not accessible to anyone outside the study team. Participants were advised that participation was voluntary, and that answers were confidential. Participants were provided with a small gift upon interview completion. Female partners could not be interviewed in every case. When husband respondents were removed, 270 women remained. A total of 11 people refused to be interviewed or were unavailable due to relocation because of construction. This yields a study response rate of 96%.

Among the women in the sample, 26% had no education, 22% had completed primary education, 35% had completed middle or high school, and 16% had any post-high-school education. The average years of education for women in the sample was 7.7. This is quite close to the national average for females estimated by the Central Intelligence Agency (CIA n.d.). Among those interviewed, 30% had a monthly household income between 15,000-20,000 Nepalese rupees per month (US$172-US$229/month). This was also the median household income. This amount seems like a reasonable estimate, as the CIA estimates annual per capita income for Nepal at US$1,300 per annum (2012), and household incomes should be higher both because data are from the capital and most households consisted of more than one wage earner (86% of women interviewed were employed full- or part-time; 92% of husbands were employed full- or part-time).⁴ The women had been married 11 years on average, and were, on average, 33 years of age. Thirty-nine percent of the women’s self-reported caste were coded as being “high,” 25% as “middle,” and 35% as “low.” Ninety-one percent of the sample had at least one child. Thirty-eight percent lived in households that owned at least one car.

Measures

Husband violence prevalence, severity, and injury

The measure of husband violence is a modified version of the Straus and Douglas (2004) Conflict Tactics Scale Short Form (CTS2S), which has been found to have good concurrent validity with the longer form, the CTS2 (Straus & Douglas, 2004). It includes six physical violence items: (a) slapped; (b) pushed, grabbed, or shoved; (c) hit with object; (d) punched, kicked, or bit; (e) beat-up; and (f) used or threatened with a knife or gun, and two injuries from physical violence items (had a sprain, bruise, small cut, or felt pain the next day because of a fight with the partner and had to see a doctor [MD] because of a fight with partner). Possible responses were once, twice, 3-5 times, 6-10 times, 11-20 times, or more than 20 times in the past year.

Logistic regression coefficients (run separately for violent acts by men and women) from husband violence injury regressed on number of each type of violent act were used to weight the violent acts prior to combination in a scale. Thus, a one-unit increase in the Husband Violence Severity scale is associated with the same increase in the log-odds of injury across the entire CTS scale. A one-unit increase in husband violence severity is associated with 1.4 times higher odds of any report of injury from husband violence in the past year (χ² = 40.9, df = 1, p < .001). This measure of husband violence severity, when used as a control in analyses for husband violence injury, maximizes the explained variance between violence severity and injury by design. This allows for a conservative test of other independent variables that are truly not confounded by their relationship with violence severity. Reliability for the Husband Violence Severity scale was acceptable (Cronbach’s α = .81). Husband violence injury items were combined into a simple scale indicating the number of reported injuries per year (Cronbach’s α = .75).

The CTS as a measure of husband violence is controversial and has been critiqued for ignoring the context and consequences of the violence (Straus & Gelles, 1990). It was also the subject of fierce debate because in Western contexts it tends to find approximately equal numbers of violent acts by husbands and wives (Johnson, 1995). Nonetheless, the CTS shows evidence of reliability and construct validity, and tends to reveal higher rates of domestic violence than other measures (Straus & Gelles, 1990). In non-Western contexts, the CTS often shows more violence by husbands than by wives (cf. Kim & Emery, 2003).

Attempted power/control motive

To measure attempts to control wives, we used the Johnson and Leone (2005) measure of intimate terrorism implemented on the National Violence Against Women Survey. These items were as follows: Your husband: (a) tries to limit your contact with family and friends, (b) is jealous or possessive, (c) insists on knowing who you are with at all times, (d) puts you down in front of others, (e) makes you feel inadequate, (f) shouts or swears at you, and (g) prevents you from knowing about or having access to the family income. Possible responses were never, rarely, sometimes, often, and always. Reliability was excellent (Cronbach’s α = .89).

Actual power in the relationship

Actual power in the relationship was measured by asking the women, subjectively,

on a scale of 0 to 100 where 0 means you have no power, 50 means you share power equally with your spouse, and 100 means you have total power: how much power do you have in your relationship with your spouse?

Although subject to response bias, it may be less subject to other forms of bias. For example, individual women in the sample can be presumed to know whether they really have absolute control over some decisions, or whether that decision power is, for example, simply delegated for convenience. Unlike the measures of attempted power and achieved power, which capture the husband’s attempts to get power over his wife or, arguably, achieved power over his wife, actual power captures women’s power in the relationship as subjectively experienced.

Achieved (legitimated) power: Patriarchal beliefs of wives

Patriarchal beliefs of wives were adapted from Emery, Kim et al.’s (2013) measure. These items were as follows: (a) men should be the leaders in society; (b) men should take the initiative in romantic relationships; (c) wives should do most household chores; (d) the family’s economic decisions should be made by the husband; (e) wives should follow their husbands’ opinions about the wife’s job; (f) the husband’s opinion is more important than the wife’s in making important decisions about the children, school, work, and so on.; and (g) from time to time, it is ok for husbands to use violence against their wives to preserve the husband’s authority. Possible responses were strongly agree, agree, disagree, and strongly disagree. Reliability for this scale was acceptable (Cronbach’s α = .70).

Husband’s neutralizing beliefs about domestic violence

Husband’s neutralizing beliefs were measured by wife report. Neutralizing beliefs about domestic violence were measured using a scale created by the first author designed to adapt Matza’s (1964/1990) theory to domestic violence. The four items were as follows: Your spouse thinks (a) fights between spouses are not a big deal as long as nobody gets seriously hurt; (b) if a husband hits his wife, or vice versa, the person who got hit probably deserved it; (c) if a husband hits his wife, he probably did it for her own good; and (d) there are crimes a lot worse than hitting one’s spouse. Possible responses for the items were strongly agree, agree, disagree, and strongly disagree. Reliability for the scale was good (Cronbach’s α = .89).

Social desirability of reporting husband violence

Social desirability bias was controlled by measuring the respondent’s willingness to disclose husband violence to others. This was measured with endorsement of the following six items: If my spouse ever hit me, I would try to keep it secret from (a) my friends, (b) my family, (c) my neighbors, (d) my coworkers, (e) my boss, (f) everyone. Possible responses were very likely, likely, unlikely, and very unlikely. Reliability was excellent (Cronbach’s α = .92).

Analytic Issues

Analyses were conducted in Stata version 11. The interclass correlation coefficient (ICC) suggested that variation in wards was responsible for less than 3% of the variation in both husband violence severity and injury, obviating any need for multilevel modeling. Both logistic and ordinary least squares (OLS) regression models were corrected for clustering of standard errors within wards. Inferences for logistic regression odds ratios derive from likelihood ratio rather than WALD tests. In regression models for husband violence severity and injury, the log was taken to reduce skew in the data. This allows the interpretation of coefficients as the estimated percent change in severity or injuries for a one-unit increase in the independent variable. The distribution of self-reported power (actual power) in the relationship also appears to be somewhat skewed (χ² = 59, df = 2, p < .001). However, this scale was not transformed via log-normal to retain interpretability. The three measures of power were all standardized to enable easy comparison. Models were also run using the measure of attempts to grab power (coercive control) as an indicator of intimate terrorism, as used in Johnson and Leone (2005). Our approach followed theirs, which was to create a variable that was 1 (intimate terrorism) if the continuous measure was 2 standard deviations above its mean and 0 otherwise. Running the models with the indicator of Johnson’s (2008) intimate terrorism concept instead of the continuous attempted power measure did not substantially change the direction or significance of relationships in the models (analyses not shown).

Model diagnostics were run using OLS regression models for husband violence severity and injury, and a linear probability model (OLS) for any physical violence by the husband to make use of the full range of diagnostics available in Stata. When the dependent variables were suitably transformed, none of the models showed evidence of linearity problems when Pregibon’s link tests were run. Variance inflation factors (VIFs) were all less than 5 in all models, suggesting that multicollinearity is not a serious problem. Studentized residuals were calculated to test for outliers. Three significant outliers were found in the model for logged husband violence injury, and four in the model for husband violence severity (all outliers were positive). Removal of outliers did not change the direction and increased the significance of the independent variables of interest in the model for logged husband violence injury or severity. Wife to husband income ratio became a significant predictor of logged husband violence severity when the outliers were removed. All models are shown with the outliers included. The linear probability model for any physical violence by the husband showed no evidence of linearity or multicollinearity problems and had no significant outliers.

Results

Descriptive statistics are reported in Table 1. In addition to the sample characteristics discussed in the “Method” section, Table 1 suggests that wives in the sample are indeed smaller and lighter than their husbands on average (both ratios are less than 1). Table 1 also suggests that wives in the sample have, on average, nearly 1 year less education than their husbands and are nearly 4 years younger. The table also shows some anomalies; namely that wives, on average, appear to have more education and income than their husbands as indicated by the income and education ratios (both greater than 1). However, these ratios are highly skewed. The median wife to husband income ratio is 0.014. The median wife to husband education ratio is 1. Thus, the high averages appear to be driven by a few women in the sample who had much more education and income than their husbands. This is exacerbated by the use of ratios. As noted in the “Method” section, the distribution of self-reported power (actual power) in the relationship is skewed. Hence, although the mean appears to indicate that wives in the sample report they have more power than husbands (69%), the median indicated that the central tendency was to report equality (50%). Because we were concerned that results might be driven by the 20 women who had much more education (more than 4 times as much) than their husbands, we ran one set of models including those women and another set excluding them. We present the results of the inclusive models. We note the two instances when results were not the same in the text below. The two differences in results appear to stem from one case, for which the dfbeta statistic was greater than the $2 / \sqrt{n} criterion$ .

Table 1.

Sample Descriptive Statistics.

Variable	n	M	SD
Any violence by husband	270	54.07%	0.50
Any injury from husband violence	270	9.26%	0.29
Injuries	270	1.38	6.32
Logged injuries	270	−1.89	1.34
Husband violence severity	269	1.58	3.24
Logged husband violence severity	269	−0.82	1.62
Husband attempted power	270	13.95	5.59
Wife’s power, self-report	235	69.14	26.64
Achieved power (wife’s patriarchal beliefs)	270	19.71	2.93
Husband’s neutralizing beliefs	270	9.35	2.77
Wife to husband height ratio	269	0.94	0.07
Wife to husband weight ratio	268	0.94	0.53
Wife to husband income ratio	270	4.00	16.16
Wife to husband education ratio	268	9.46	35.58
Wife to husband age ratio	269	0.89	0.11
Standardized social desirability of reporting domestic violence	270	11.21	4.56
Does family own a car?	266	37.97%	0.49
Wife’s age	269	32.85	7.48
Husband’s age	269	36.79	8.71
Wife’s years of education	269	7.69	5.68
Husband’s years of education	269	8.41	5.36
Household income (rupees)	269	17,625	2,220
Wife lower caste	251	35.46%	0.48
Wife middle caste	251	25.10%	0.43
Wife high caste	251	39.44%	0.49
Years married	270	11.13	7.28

Rates of husband violence were very high: 54% of the sample reported some physical violence by the husband in the last year. Given the high rate, it seems important to unpack this figure insofar as possible. Among the 270 wives interviewed, (a) 37% reported being slapped in the past year; (b) 13% reported being pushed, grabbed, or shoved; (c) 28% reported being hit with an object; (d) 21% reported being punched, kicked, or bit; (e) 41% reported being beaten up; and (f) 4.5% reported that they were victimized or threatened with a knife or gun. The first two items are sometimes coded as “minor,” with the remaining items coded as “severe.” No violence is ever minor. Likewise, some of the “severe” items may or may not be as severe as some of the “minor” items, depending, for example, on whether the object was a pillow or a brick (Item 3), and whether the slap was on the arm or on the face (Item 1). For comparative purposes however, the combined prevalence rate for the first two (minor) items was 40%, and the combined prevalence rate for the last four (severe) items was 50%. The domestic violence injury prevalence rate was 9.2% for the last year. Among respondents, 4.8% reported between 1-10 husband violence injuries, 2.2% between 11-20 husband violence injuries, and 2.2% more than 20 husband violence injuries in the last year.

In the current instance, we were concerned because our sample shows a substantially higher rate of husband violence than other studies of Nepal (54% in our Kathmandu sample vs. 35% reported by Paudel, 2007). Despite this disparity, we are inclined to trust our data. First, as noted above, the CTS tends to produce higher endorsement rates of husband violence than other approaches (Straus & Gelles, 1990). Second, though high, the rate is consistent with the five interviews conducted by the second author (three of the five—60%—reported some form of physical violence by the husband in the past year). Third, the second author felt that the other two women she met were concealing physical violence by their husbands. There are additional grounds to consider 54% as a potential underestimate. Among respondents, 43% said that it was very likely that they would “try to keep it secret from everyone” if ever hit by their husbands. Among those 43%, 78 women (29% of the sample) reported no physical violence by their husbands. Hence, we cannot rule out the possibility that 54% is still an underestimate. Reports of husband violence were strongly related to the “keep it secret from everyone” item, with those reporting that they were more likely to keep it secret reporting less husband violence (χ² = 42.5, df = 3, p < .001).

Actual Power and Achieved Power, but Not Attempted Power, Associated With Any Violence by the Husband

The more power women had, the less they reported husband violence. Table 2 shows the results of the logistic regression model for any physical violence by their husband. It is noteworthy that although a 1 standard deviation increase in self-reported power is associated with 2 times lower odds of reporting any violence by the husband (χ² = 16.4, df = 1, p < .001) and a 1 standard deviation increase in achieved power (wives’ patriarchal beliefs) is associated with 1.5 times higher odds of any husband violence (χ² = 5.3, df = 1, p < .05), attempted power, the variable used by Johnson and Leone (2005) to measure intimate terrorism, is not significantly associated with any husband violence (χ² = 1.1, df = 1, p = .30). Neither is the non-significant attempted power odds ratio in the predicted direction.

Table 2.

Logistic Regression of Any Violence by the Husband (N = 223).

Variable	Any violence by the husband
Variable	B	SE B	OR
Standardized husband attempted power	−0.20	0.19	0.82
Standardized wife’s power, self-report	−0.72	0.20	0.49***
Standardized achieved power (wife’s patriarchal beliefs)	0.42	0.12	1.52*
Standardized husband’s neutralizing beliefs	0.29	0.13	1.33^†
Wife to husband height ratio	−2.96	2.53	0.05
Wife to husband weight ratio	0.06	0.17	1.06
Wife to husband income ratio	0.02	0.01	1.02
Wife to husband education ratio	0.004	0.004	1.00
Wife to husband age ratio	2.04	1.65	7.69
Standardized social desirability of reporting domestic violence	0.67	0.20	1.95***
Does family own a car?	−0.08	0.41	0.92
Wife’s age	−0.06	0.02	0.94
Wife’s education in years	−0.05	0.03	0.96
Wife’s lower caste	0.67	0.41	1.96^†
Wife’s middle caste	0.87	0.54	2.38*
Household income	−0.08	0.31	0.92
No. of years wife has been married	−0.03	0.03	0.97

Note. Standard errors corrected for clustering by ward. Significance tests for each coefficient are based on likelihood ratio, not WALD tests. OR = odds ratio.

†

p < .10. *p < .05. **p < .01. ***p < .001.

A 1 standard deviation increase in wife-reported husband’s neutralizing beliefs was not associated with a significant increase in odds of husband violence at the .05 level (χ² = 2.9, df = 1, p = .09). Among the control variables in the model, only the social desirability of reporting husband violence measure (χ² = 12.5, df = 1, p < .001) and being in middle rather than high caste (χ² = 4.1 df = 1, p < .05) were significantly associated with reporting husband violence. A 1 standard deviation increase in reported likelihood of keeping husband violence secret was associated with nearly 2 times lower odds of reporting husband violence. Being in middle versus high caste was associated with 2.4 times higher odds of reporting husband violence.

Actual Power and Achieved Power, but Not Attempted Power Predict Husband Violence Severity

Table 3 shows the results of the regression model for logged husband violence severity. A 1 standard deviation increase in self-reported power was significantly associated with a 37.5% decrease in husband violence severity (p < .01); a 1 standard deviation increase in wives’ patriarchal beliefs was significantly associated with an 18.5% increase in husband violence severity (p < .01). However, attempted power was not significantly associated with logged husband violence severity (B = 0.13, p = .24). The coefficient, however, is in the expected direction.

Table 3.

Ordinary Least Squares Regression of Logged Husband Violence Injury and Logged Severity (N = 223).

Variable	Logged husband violence injury		Logged husband violence severity
Variable	B	SE B	B	SE B
Severity of husband violence	0.25***	0.06
Standardized husband attempted power	0.17*	0.08	0.13	0.11
Standardized wife’s power, self-report	−0.12*	0.06	−0.38**	0.11
Standardized achieved power (wife’s patriarchal beliefs)	−0.06	0.05	0.18**	0.06
Standardized husband’s neutralizing beliefs	0.15	0.09	0.14*	0.06
Wife to husband height ratio	0.49	0.76	−2.14	1.66
Wife to husband weight ratio	0.02	0.05	0.09	0.06
Wife to husband income ratio	0.01	0.01	0.01	0.005
Wife to husband education ratio	0.001	0.001	0.001	0.002
Wife to husband age ratio	0.70	0.61	1.27	0.84
Standardized social desirability of reporting domestic violence	0.07	0.14	0.51***	0.12
Does family own a car?	0.01	0.13	0.15	0.21
Wife’s age	−0.01	0.02	−0.02	0.02
Wife’s education in years	−0.01	0.02	−0.02	0.02
Wife is lower caste	−0.14	0.14	0.39^†	0.21
Wife is middle caste	−0.02	0.20	0.42	0.28
Household income	−0.11	0.11	−0.02	0.20
No. of years wife has been married	0.01	0.01	−0.01	0.02

Note. Regression standard errors corrected for clustering by ward.

†

p < .10. *p < .05. **p < .01. ***p < .001.

A 1 standard deviation increase in wives’ reports of husbands’ neutralizing beliefs is associated with a 14% increase in the severity of husband violence (p < .05). This was only marginally significant when women who were much more educated than their husbands were removed (B = 0.133, p < .10). A 1 standard deviation increase in reported likelihood of keeping husband violence secret was associated with a 51% decrease in the severity of husband violence (p < .001). None of the other control variables were significantly associated with husband violence severity. The model explained 32% of the variance in logged husband violence severity.

Actual Power and Attempted Power, but Not Achieved Power Predict Husband Violence Injury

Table 3 also shows the results of the regression model for husband violence injury. As for husband violence prevalence and severity, wives’ self-reported power was a significant predictor of injury. A 1 standard deviation increase in self-reported power was significantly associated with a 12.3% decrease in injury from husband violence (p < .05). However, this finding became marginally significant when women with much more education than their husbands were removed from the model (B = −0.111, p < .10). Unlike the pattern for prevalence and severity, attempted power (coercive control) was significantly associated with husband violence injury. A 1 standard deviation increase in attempted power was associated with 17% more injuries from husband violence (p < .05). On the other hand, achieved power does not appear to be as important as it is for husband violence prevalence and severity. Wives’ patriarchal beliefs were not associated with injury from husband violence (B = 0.06, p = .25). However, controlling for the severity of husband violence in the injury model may make the results somewhat conservative.

None of the other variables were significantly associated with injuries from husband violence, including the wife to husband height and weight ratios. However, the coefficients for husband’s neutralizing beliefs, height ratio, and wife to husband age ratio are relatively large. We cannot rule out the possibility that statistical power played a role in the null findings in these cases. The model explained 42% of the variance in logged husband violence injury.

Discussion

This is a preliminary study that takes its point of departure from a relatively new theoretical perspective (Emery, 2011), with results that also depart from the previous literature. We therefore wish to emphasize caution in the interpretation of our results. Our findings for the prevalence of husband violence were substantially higher than those for other studies; this is particularly unusual because other studies examined Nepal as a whole, but our work focused on the capital, where one might suppose that husband violence would be somewhat less prevalent.⁵ We found that more than half of our sample reported some physical violence by the husband in the past year, and half of all women in the sample reported what is conventionally termed severe violence.

However, in a larger developmental context, a 54% husband violence rate may not be so unusual. Nepal suffers from high rates of poverty and recently emerged from 10 years of civil war. Rates of husband violence have been reported to be as high as 70% in neighboring India; 56% of Indian women felt that wife beating could be justified in some circumstances (“India Tackles Domestic Violence,” 2006). Our approach may have revealed more violence than other measures, as the CTS is known to find higher rates of endorsement of violence than other measures (Dutton & Nicholls, 2005). The sensitivity training provided to interviewers may also have helped increase the disclosure rate, as may have the introduction to the scale that is designed to reduce the stigma of discussing victimization.⁶

Subjective Reports of Actual Power in the Relationship Consistently Linked to Husband Violence Prevalence, Severity, and Injury

Controlling for two other forms of power, including the measure Johnson and Leone (2005) used to capture intimate terrorism, wives’ subjective reports of actual power in the relationship were significantly linked to husband violence prevalence, severity, and injury. Although Emery (2011) did not make arguments about power inequality with respect to prevalence, the fact that power in the relationship matters when attempts to grab power are held constant is consistent with Emery’s (2011) position that power is of fundamental importance in understanding husband violence. Although it appears to be important, the motivation of the perpetrator to attempt to grab power may be limited as an explanatory concept because attempts to grab power may or may not succeed.

We again wish to stress both that our findings are preliminary, and that we do not believe that attempted power is unimportant. Our findings do suggest that attempts to grab power (coercive control) are important in understanding injury from husband violence in Kathmandu. However, the findings appear to leave room for Emery’s (2011) claim that power itself, rather than attempts to grab it, is more fundamental to an understanding of husband violence. Much more extensive research is needed to evaluate whether Emery’s (2011) five-part typology (Anarchic, Violent Conflict, Tolerant Dictatorship, Despotic Dictatorship, Totalitarian Dictatorship) can deliver as much or more empirical clarity as Johnson’s (2008) four-part typology (situational couple violence, violent resistance, intimate terrorism, and mutual violent control). Specifically, Emery’s (2011) typology was based on differing levels of order and on norms (the relationship of the violence to the norms shared by the couple) in addition to power. Future research needs to examine these three variables simultaneously. Moreover, Emery’s (2011) work suggests that power should moderate the relationships of other predictors with husband violence. Hence, future studies should measure and examine interaction effects between the three criteria used by Emery (2011): power, order, and legitimacy.

Usefulness of Patriarchy as a Concept Does Not Appear to Be Dead

Despite claims that the gender paradigm has been shown by empirical research to be irrelevant (cf. Dutton & Nicholls, 2005), our article has shown that wives’ patriarchal beliefs, achieved power, are associated with husband violence prevalence and severity, as Emery, Kim et al. (2013) showed that the relationship between perpetrator’s patriarchal beliefs and husband violence severity appears to be switched on by child abuse. The findings for actual power and achieved power suggest that women with less power in intimate relationships may be at higher risk of victimization and victimization by more severe abuse. We argued that women who believe male privilege, or even violence, is acceptable may be at greater risk because they are less likely to report or seek other remedies when husband violence occurs. This may indeed be the case. Although our findings are preliminary, when combined with the perpetrator findings of Emery, Kim et al. (2013), they suggest that ruling out the relevance of patriarchy with respect to husband violence would be premature.

Attempted Power, or Johnson’s Concept of Control Attempts, Remains Important

Although not significant in predicting husband violence prevalence or severity, attempts to grab power were significantly related to reported injury from husband violence. Intuitively, it makes sense that whether a husband is attempting to control his wife as part of a violent context should matter, because it may indicate underlying perpetrator pathology that points to higher risk for the victim. For example, Holtzworth-Munroe and Stuart (1994) argue that the desire to control is related to borderline personality type symptoms, which are in turn, among other things, linked to murder suicides when women attempt to leave. Other research has indeed linked such attempts to gain power over the victim with more serious violence and more negative outcomes for the victim (Graham-Kevan & Archer, 2003; Johnson & Leone, 2005). Our point is not that such attempts are unimportant, but simply that whether the attempts succeed or fail may be equally or even more important.

Husbands’ neutralizing beliefs about domestic violence, as reported by wives, were found to be significantly associated with husband violence prevalence and severity. The construct of neutralizing beliefs is ripe for additional domestic violence research in its own right. Our research supports Mooney’s (2007) findings in that respect. In our findings, neutralizing beliefs do not appear to explain the relationship between power and husband violence. Somewhat surprisingly, husband–wife differences in height and weight do not appear to be associated with increased husband violence injury, as suggested by Straus and Gelles (1990). This suggests that although Straus and Gelles (1990) were right about the fact that women suffer more injuries from domestic violence than men do, they may have been wrong about the reason. Given the effect sizes, however, we cannot rule out that a larger sample size might have found a relationship for height differences.

More broadly, the results suggest that male power in relationships, as it relates to husband violence, deserves a more nuanced and less one-dimensional approach than those traditionally used. Each of the three forms of male power we examined was uniquely significant (controlling for other forms) for at least one of the dependent variables. Past research (cf. Coleman & Straus, 1986) has limited itself to consideration of one measurement of power, like final say, and then made assessments of the relationship between power and husband violence. Our findings suggest that such an approach may be limited at best and at worst erroneous, and that further work on husband violence and power with a more nuanced theoretical approach is much needed.

Limitations

There are a number of limitations that affect the strength of the conclusions that can be drawn from these analyses. The study is non-experimental, cross-sectional, and subject to reporting biases. Findings are thus associations and not necessarily causal relationships. Findings are only generalizable to the city of Kathmandu, and although we have controlled for a measure of social desirability bias about husband violence reporting, we cannot rule out the possibility that this bias is still present in our data. Reliability is also limited by the fact that we only interviewed one partner from each couple. Likewise, it is a serious concern that the prevalence rate for husband violence and employment rates in our study are higher than that found in other studies of Nepal; all of our findings require replication. When women with much more education than their husbands were removed, the relationship between actual power and injury from husband violence became marginally significant. This change appears to have been driven by a single outlier.

The erudite reader will also point out that a serious threat to the validity of conclusions is reverse causality bias. That is, it may be reasonable to argue that power disparities decrease costs of husband violence and hence play a causal role, but it is also reasonable to argue that husband violence disempowers wives. This throws into question the direction of the relationships found in the data, particularly for actual and achieved power. Longitudinal studies are needed to sort out this directionality problem. Furthermore, although power analyses were used to determine sample size, those analyses were focused on husband violence and neighborhood variables rather than power. Hence, we cannot rule out that some of the non-significant findings might have occurred because of a lack of statistical power.

Conclusion

This article took a first step in adding to the research literature on power and domestic violence by measuring three types of power and husband violence prevalence, severity, and injury in Kathmandu. It also provided an initial test of Emery’s (2011) theoretical proposition that power is an important variable in understanding husband violence and is not adequately captured by attempted power. The findings that actual power reported by wives is associated with husband violence prevalence, severity, and injury require replication in a longitudinal context. Achieved and attempted power both require additional research. However, our findings do suggest that power is an important, if neglected variable in understanding husband violence. Furthermore, regardless of directionality, it is evident that from the perspective of Nepali women in Kathmandu, husband violence is still a crime of power. In this, the findings of early feminist researchers such as Dobash and Dobash (1979) and Brownmiller (1975) appear to continue to merit consideration.

Footnotes

Acknowledgements

We thank Hon. Eknath Dakhal, the Minister of Cooperatives and Poverty Alleviation, and Mr. Santosh Paudel who made this study possible. We also wish to thank Keshav Shrestha, the Kathmandu research team, and the participants.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Measure development for this study was facilitated by an earlier grant from the H.F. Guggenheim Foundation and National Science Foundation of China Grants 71002014 and 71110107027. Data collection was facilitated by support from Namseoul University.

Notes

Author Biographies

Clifton R. Emery, PhD, is an associate professor of social welfare at Yonsei University in Seoul, South Korea. He is also the associate director of the Yonsei Institute for North Korea Studies and holds honorary appointments at the University of Bath and the University of Hong Kong.

Sirjana Thapa, MA, is a PhD student at the School of Social Welfare at Yonsei University. She is an assistant professor of child welfare at Namseoul University in South Korea.

Shali Wu, PhD, is an associate professor of marketing at the School of Management at Kyunghee University in Seoul, South Korea. Her research interests include perspective taking and decision making in cross-cultural contexts.

References

Babcock

Waltz

Jacobson

Gottman

(1993). Power and violence: The relation between communication patterns, power discrepancies and domestic violence. Journal of Consulting and Clinical Psychology, 61, 40-50.

Balogun

Owoaje

Fawole

(2012). Intimate partner violence in southwestern Nigeria: Are there rural-urban differences? Women & Health, 52, 627-645.

Blood

Wolfe

(1960). Husbands and wives. Glencoe, IL: Free Press.

Brownmiller

(1975). Against our will: Men, women and rape. New York: Fawcett Books.

Central Intelligence Agency (CIA). (n.d.). World factbook: Nepal. Retrieved from https://www.cia.gov/library/publications/the-world-factbook/geos/np.html

Coleman

Straus

(1986). Marital power, conflict, and violence in a nationally representative sample of American couples. Violence and Victims, 1, 141-157.

Dahrendorf

(1970). On the origin of inequality among men. In Laumann

E. O.

Siegel

P. M.

Hodge

R. W.

(Eds.), The logic of social hierarchies (pp. 16-44). Chicago, IL: Markham.

Dhakal

(2008). Nepalese women under the shadow of domestic violence. Lancet, 371, 547-548.

Dobash

(1979). Violence against wives: A case against the patriarchy. New York: Free Press.

10.

Durkheim

(1984). The division of labor in society. New York: Free Press. (Original work published 1893)

11.

Dutton

Nicholls

(2005). The gender paradigm in domestic violence research and theory: Part 1-the conflict of theory and data. Aggression and Violent Behavior, 10, 680-714.

12.

Emery

(2011). Disorder or deviant order? Re-theorizing domestic violence in terms of order, power, and legitimacy: A typology. Aggression and Violent Behavior, 16, 525-540.

13.

Emery

Kim

Song

(2013). Child abuse as a catalyst for wife abuse? Journal of Family Violence, 28, 141-152.

14.

Emery

(2016). Who calls the shots in Beijing? Resources, culture and equality in family decision power. Under review.

15.

French

Raven

(1959). The bases of social power. In Cartwright

(Ed.), Studies in social power (pp. 150-167). Ann Arbor, MI: Institute for Social Research.

16.

Frieze

McHugh

(1992). Power and influence strategies in violent and nonviolent marriages. Psychology of Women Quarterly, 16, 449-465.

17.

Graham-Kevan

Archer

(2003). Intimate terrorism and common couple violence: A test of Johnson’s predictions in four British samples. Journal of Interpersonal Violence, 18, 1247-1270.

18.

Gramsci

(1992). Selections from the prison notebooks. New York: International Publishers.

19.

Holtzworth-Munroe

Stuart

(1994). Typologies of male batterers: Three subtypes and the difference among them. Psychological Bulletin, 116, 476-497.

20.

India tackles domestic violence. (2006, October 26). BBC News. Retrieved from http://news.bbc.co.uk/2/hi/south_asia/6086334.stm

21.

Johnson

(1995). Patriarchal terrorism and common couple violence: Two forms of violence against women. Journal of Marriage and Family, 57, 283-294.

22.

Johnson

(2008). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Boston, MA: Northeastern University Press.

23.

Johnson

Leone

(2005). The differential effects of intimate terrorism and situational couple violence: Findings from the National Violence Against Women Survey. Journal of Family Issues, 26, 322-349.

24.

Kim

Emery

(2003). Marital power, conflict, norm consensus, and marital violence in a nationally representative sample of Korean couples. Journal of Interpersonal Violence, 18, 197-218.

25.

Martin

(1981). Battered wives. San Francisco, CA: Volcano Press.

26.

Matza

(1990). Delinquency and drift. New Brunswick, NJ: Transaction Publishers. (Original work published 1964)

27.

Ministry of Health and Population. (2012). Nepal demographic and Health Survey 2011. Kathmandu, Nepal. Retrieved from http://www.measuredhs.com/pubs/pdf/FR257/FR257%5B13April2012%5D.pdf

28.

Mooney

(2007). Shadow values, shadow figures: Real violence. Critical Criminology, 15, 159-170.

29.

Paudel

(2007). Domestic violence against women in Nepal. Gender Technology and Development, 11, 199-233.

30.

Pence

Paymar

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

31.

Ptacek

(1988). Why do men batter their wives? In Yllo

Bograd

(Eds.), Feminist perspectives on wife abuse (pp. 133-157). Newbury Park, CA: Sage.

32.

Sagrestano

Heavey

Christensen

(1999). Perceived power and physical violence in marital conflict. Journal of Social Issues, 55, 65-79.

33.

Stark

(2007). Coercive control: How men entrap women in personal life. New York: Oxford University Press.

34.

Straus

Douglas

(2004). A short form of the revised conflict tactics scales, and typologies for severity and mutuality. Violence and Victims, 19, 507-520.

35.

Straus

Gelles

(1990). Physical violence in American families. New Brunswick, NJ: Transaction Publishers.

36.

Waller

(1937). The rating and dating complex. American Sociological Review, 2, 727-734.

37.

Weber

(1978). Economy and society. Los Angeles: University of California Press. (Original work published 1922).

38.

Yllo

Bograd

(1988). Feminist perspectives on wife abuse. Newbury Park, CA: Sage.

39.

Yllo

Straus

(1990). Patriarchy and violence against wives: The impact of structural and normative factors. In Straus

Gelles

(Eds.), Physical violence in American families (pp. 383-399). New Brunswick, NJ: Transaction Publishers.

40.

Yoshihama

Ramakrishnan

Hammock

Khaliq

(2012). Intimate partner violence prevention program in an Asian immigrant community: Integrating theories data, and community. Violence Against Women, 18, 763-783.