Sie sind auf Seite 1von 13

Journal of Family Violence, Vol. 21, No. 1, January 2006 ( C 2006) DOI: 10.

1007/s10896-005-9002-2

Fighting Families: Family Characteristics Associated with Domestic Violence in Five Latin American Countries
Dallan F. Flake1 and Renata Forste1,2
Published online: 8 April 2006

This study uses data from Demographic and Health Surveys (DHSs) to examine the relationship between familial characteristics and the likelihood of experiencing domestic violence in Colombia, the Dominican Republic, Haiti, Nicaragua, and Peru. Logistic regression techniques are used to measure relationships between marital status, family size, partner alcohol use, socioeconomic status (SES), decision-making power, and education homogamy and the likelihood of experiencing partner violence. Cohabitation, female-dominant decision making, and partner alcohol are positively associated with domestic violence across datasets. Family size, SES, and education homogamy emerged as statistically signicant in some, but not all of the datasets. This study helps clarify the prole of the abused Latina and also tests the applicability of current abuse research to a non-Western setting.
KEY WORDS: Latin America; spouse abuse; family violence; marital violence.

INTRODUCTION Domestic violence3 is one of Latin Americas most pressing social problems, as each year between 10 and 35% of Latina women are physically abused by their partners (Buvinic et al., 1999). Whereas the region is notorious for its high rates of political and social violence, much less understood is the violence that occurs behind closed doorsbetween husbands and wives. With so much attention centered on Latin Americas corruption, crime, and political instability, it is easily overlooked that the family is perhaps this regions most violent social institution. Although domestic violence research has reached unprecedented heights, relatively little is known about how spouse abuse functions outside traditional Western regions of study such as North America and Europe. Culture is known to affect the magnitude and characteristics of inti1 Department 2 To

of Sociology, Brigham Young University, Provo, UT. whom correspondence should be addressed at Department of Sociology, Brigham Young University, 2032 JFSB, Provo, UT 84602; e-mail: renata forste@byu.edu. 3 While numerous forms of aggression are incorporated into the term domestic violence, the present study focuses exclusively on domestic violence involving physical abuse between heterosexual partners.

mate violence in different societies (Holtzworth-Munroe et al., 1997), but because few studies compare these issues in different cultural contexts, it remains unclear if presentday abuse research can be applied to non-Western settings. A few foundational studies have been conducted in Latin America (Ellsberg et al., 2000; Gonzales de Olarte & Gavilano Llosa, 1999); however, they tend to focus on women in a single city or country rather than examining broader patterns of domestic violence across Latin America. This study extends the domestic violence literature by examining family characteristics associated with spouse abuse across ve Latin American countries. Colombia, the Dominican Republic, Haiti, Nicaragua, and Peru were selected for analysis because they reect the rich diversity of Latin America. The purpose of this study is twofold. First, it aims to create a more comprehensive prole of the abused Latina to inform researchers, policymakers, and women themselves of potential risk markers for abuse. Second, this study investigates the applicability of Western abuse research to less-developed countries. This may potentially be its most important contribution, as it could shed light on the relevance of current theories, models, programs, and policies to Latin America. 19
0885-7482/06/0100-0019/0
C

2006 Springer Science+Business Media, Inc

20 The Latin American Context Latin America is one of the most culturally heterogeneous regions in the world. The myriad races, ethnicities, languages, and lifestyles preclude the lumping together of Colombians, Dominicans, Haitians, Nicaraguans, and Peruvians as part of a uniform Latin American culture. In examining domestic violence in Latin America, however, most of the countries share two characteristicsa legacy of social violence and rigid gender scriptsthat are integral to understanding spouse abuse in the Latin American context. Latin America has long been one of the worlds most violent regions. Beginning with the Spanish Conquest and extending to the present day, political conict has become a near permanent xture in the lives of millions of Latinos. In Colombia, the constant war between the government and drug cartels has made the country one of the most violent places in the world. Colombia is currently engaged in one of its bloodiest conicts to date, against the powerful guerilla organization FARC. The Dominican Republic, Haiti, and Nicaragua have been equally unstable. Corrupt regimes and bloody civil conicts led to international military intervention in all three countries during the twentieth century. Peru has a similar history of political violence and is presently battling the Shining Path, a Maoist terrorist group whose stated goal is to destroy existing Peruvian institutions and replace them with a communist peasant regime. Since the Shining Path took up arms in 1980, approximately 30,000 persons have died in Peru. Buvinic and colleagues (1999) argue that societies with long histories of wars are vulnerable to outbreaks of social violence. The widespread availability of weapons, coupled with the attenuation of inhibitions against violence that war brings, tend to exacerbate already powerful contributing factors including inequality and high levels of poverty. The effects of political violence on Latin American society have been devastating: The regions homicide rate of almost 30 murders per 100,000 people is more than double the world average (Murray & Lopez, 1996). Excessive political and social violence is associated with higher rates of domestic violence (Messing, 1999). In Latin America, violence is the subject of casual conversation and newspaper headlines. Children who grow up witnessing or experiencing violence can become desensitized to the deleterious effects of aggression and see it as a suitable way of obtaining what they want. The second characteristic of Latin American societygender-based normsreinforces male authority and superiority over females throughout much of Latin America. According to cross-cultural literature, two of the most enduring factors that promote violence against

Flake and Forste women are rigidly dened gender roles and a cultural definition of manhood that is linked to dominance (Counts et al., 1992). The term machismo is often used to describe Latino masculinity, and refers to the cultural expectation that males must show they are masculine, strong, and sexually aggressive, and even able to consume large amounts of alcohol without getting drunk (Giraldo, 1972). Machismo is largely viewed as an expression of an inferiority complex stemming from the Spanish Conquest (Riding, 1985). Hypermasculinity is a culturally accepted response to male dependency, powerlessness, feelings of inferiority, and low self-esteem (McCord & McCord, 1960). Machismo, then, is the combination of feeling inferior and acting superior (Ingoldsby, 1991). Male dominance is reinforced by womens role in Latin American society. Marianismo refers to the expectation that women embrace the veneration of the Virgin Mary in that they are capable of enduring any suffering inicted upon them by males (Stevens, 1973). Latin American women are to be submissive, dependent, sexually faithful to their husbands, and are expected to take care of household needs and dedicate themselves entirely to their husbands and children. Because Latinas identities are dened by their roles as mothers and wives, Latino patriarchy denies women individuality on the basis of gender (Rivera, 1998). Nobel Prize laureate Octavio Paz (1961) observed that a woman who does not conform to the traditional female ideal is viewed as a mala mujer (bad woman) in Latin America.

The Family as an Enabling Context for Abuse In describing the Latino family, social scientists generally focus on two concepts: familism and machismo. Familism refers to the Latino ideal of placing ones family ahead of individual interests, and includes responsibilities and obligations to ones immediate family members and other kin (Ingoldsby, 1991). As previously noted, machismo is the term used to describe Latino masculinity and is characterized by aggressiveness and hypersexuality (Giraldo, 1972). Although familism and machismo may appear at odds with one another, both concepts are evident in Latino families. The combination of familism and machismo may make Latino families more susceptible to domestic violence, since women are expected to fulll familial obligations unconditionally within an overarching patriarchal family system. How can the family be a haven for love, support, and comfort, and yet be the place where one is most likely to be spanked, slapped, beat up, assaulted, or killed (Gelles, 1997)? Family violence researchers have sought

Domestic Violence in Latin America to resolve this paradox by examining how certain family characteristics inuence the likelihood of domestic violence. Family-level explanations of violence tend to focus on issues of stress and power dynamics. Other aggravating factors, including marital status and drinking, are also commonly linked to domestic violence. The present study examines whether family characteristics associated with abuse in Western contexts are similarly related to domestic violence in Latin America. The following factors have been linked to wife abuse in Western literature; each is included in the present analyses. Marital Status Higher rates of domestic violence are consistently found among cohabitors compared to married couples. In an analysis of 14 marital violence studies, Brownridge and Halli (2000) conclude that on average, cohabitors are between 2 and 4 times more likely to engage in physical violence than married couples. Western theoretical explanations often point to the temporary and impermanent nature of cohabitation as a primary reason cohabitors are more abusive than married couples (Nock, 1995). Whereas cohabitation in Western countries usually serves as a trial period preceding marriage, the relationship is much more permanent in Latin America and might best be described as surrogate marriage (Castro Martin, 2002). Cohabitation has been an integral component of the Latin American family system since the colonial period, when the Catholic Church sanctioned informal sexual unions between Spanish colonizers and indigenous women. Because of the unique nature of cohabitation in Latin America, it is reasonable to expect that marital status would inuence domestic violence differently in this region than in Western settings. Understanding the relationship between marital status and partner violence in Latin America is critical, given that cohabitation rates there are increasing. In some countries, more than half of couples opt to cohabit rather than marry (Castro Martin, 2002). If marital status does not inuence violence in Latin America, the growing popularity of cohabitation need not be a concern for antiabuse coalitions. If cohabitation has a similar effect in Latin America as in Western countries, however, the growing prevalence of this relationship type would be reason for great alarm. Hypothesis 1: Cohabiting women are more likely to experience domestic violence than married women. Family Size Numerous studies have found a positive linear relationship between family size and domestic violence

21 (Brinkerhoff & Lupri, 1988; Ellsberg et al., 2000; Farrington, 1977). The general perception among family violence researchers is that large families are more prone to violence because they experience greater stress associated with the necessity to provide for several children (Hoffman et al., 1994). Family size has a high potential for generating frustration because of its low probability of resolution. Violence not only becomes a possible response to this frustration, but also an acceptable one. Family size might be a particularly important characteristic of abuse in Latin America because of the high fertility rate. Although the regions total fertility rate is gradually declining (2.7 children per woman), it remains much higher than rates in North America (2.1) and Europe (1.4) (Population Reference Bureau, 2002). High fertility rates, coupled with widespread poverty, can be a major source of stress for families. Hypothesis 2: Women with larger families are more likely to experience domestic violence than women with smaller families.

Partner Alcohol Use The relationship between alcohol use and domestic violence is complex (Roizen, 1997). While most research conrms that alcohol and violence go hand-in-hand (Hotaling & Sugarman, 1986; Pan et al., 1994), there is little agreement over the exact role alcohol plays in partner violence. Martin (1993) argues that the relationship between alcohol and violence differs depending on factors such as who has been drinking, the drinking context, and the relationship between perpetrator and victim. Theory building is difcult because so many factors combine to determine the link between alcohol and violence (Stith & Farley, 1993). Selective disinhibition theory (Parker & Rebhun, 1995) might be the most promising explanation, positing that alcohols negative effects on peoples perceptions and judgment interact with a complex set of social and psychological factors to result in violence in certain cases. Understanding the relationship between alcohol and violence is particularly important in Latin America because gender scripts encourage heavy alcohol consumption among males (Giraldo, 1972). Although per capita alcohol consumption rates are comparable in Latin America and the United States, the prevalence of problem drinking is relatively high among Latinos (Madrigal, 1998). A polarity has been established in Latin America, where low perception of problem drinking leads to social pressure to drink. The combination of problem drinking and social pressure to drink may make women in Latin America particularly susceptible to violence.

22 Hypothesis 3: Women whose partners sometimes or frequently get drunk are more likely to experience domestic violence than women whose partners never get drunk. Socioeconomic Status It is commonly assumed that women who are poor are more likely to experience violence than women who are not poor (Ellsberg et al., 1999; Heise, 1998; Jewkes, 2002). Poverty is not necessarily viewed as a causal factor, but it is generally assumed to increase the risk of spouse abuse. In 9 of 11 case-comparison studies from the United States, Hotaling and Sugarman (1986) identied family income as a consistent marker of wife assault. The relationship between socioeconomic status and domestic violence is also well established internationally, in Cambodia (Nelson & Zimmerman, 1996), Nicaragua (Ellsberg et al., 2000), Chile (Larrain, 1993), and Thailand (Hoffman et al., 1994). A variety of domestic violence perspectives espouse the idea that domestic violence is more widespread among the poor because families living in impoverished conditions are subject to higher levels of stress than families not living in poverty (Martin et al., 1999). Carlsons (1984) structural theory of intrafamilial violence contends that the inequitable distribution of societal resources causes stress and tension among people with insufcient material resources. When combined with other aggravating factors such as living conditions, overcrowding, a sense of hopelessness, and lack of employment opportunities, poverty can signicantly increase the risk of domestic violence (Gonzales de Olarte & Gavilano Llosa, 1999; Heise, 1998). Poverty may be an especially salient risk marker for abuse in Latin America, where 44% of people live in poverty and 19% live in extreme poverty (Economic Commission for Latin America and the Caribbean, 2002). The prevalence of poverty suggests that millions of Latino families experience high levels of stress and tension associated with economic frustration. The stress associated with poverty may have a pronounced impact on domestic violence in Latin America. Hypothesis 4: Families with lower socioeconomic status are more likely to experience marital violence than families with higher socioeconomic status. Decision-Making Power One of the fundamental differences in the roles enacted by men and women in relationships involves power. A major part of how gender roles are identied in families is through decision-making power. Rettig (1993) argues that decision-making processes are key to understanding

Flake and Forste the dynamics of couple relationships because they reveal interaction and agency within relationships, and can indicate where individuals are acting out or resisting social norms. Decision-making power is an important dimension of marital power, as it represents how much say an individual has in the couple relationship. Coleman and Straus (1990) examined how four types of decision-making relationships inuence spousal violence: egalitarian (couple makes decisions together), divided power (male makes some decisions, female makes others), female-dominant (female makes most decisions), and male-dominant (male makes most decisions). They found violence to be most prevalent among nonegalitarian couples, regardless of whether the man or woman dominated the decision making. Studies by Yllo (1993) and Kim and Sung (2000) reveal similar patterns. Given the rigidity of gender scripts in Latin America, decision making may have a particularly powerful effect on the likelihood of experiencing domestic violence. Female-dominant decision making may heighten the risk of domestic violence. Because of the cultural expectation that men should govern their families by making critical decisions, men whose partners dominate decision making might resort to violence to reassert dominance over their families. Male-dominant decision making may also increase the risk of domestic violence for women in Latin America, even though the man does not feel threatened by his partner. It is likely that dominance in decision making is indicative of a mans dominion over other aspects of the couple relationship. Hypothesis 5: Women in nonegalitarian relationships, regardless of who dominates the decision making, are more likely to experience domestic violence than women in egalitarian relationships.

Education Homogamy Status inconsistencies in relationships, specically with regard to educational attainment, lead to higher levels of spouse abuse in Western contexts (Anderson, 1997). Violence is more likely to occur in nonhomogamous relationships, regardless of whether the male or female has more education. In patriarchal societies, women who have more education than their partners have a high risk of abuse because gender roles entail that husbands have more education than their wives (Okun, 1986; Walker, 1984). OBrien (1971) and Gelles (1974) contend that if a husband does not possess more skills and resources than his wife to legitimate his superior status, he may feel threatened by an educational disadvantage to his wife and may use physical force as a last resort. Men with higher levels of education than their wives are also more likely to

Domestic Violence in Latin America become violent. Goode (1971) explains this phenomenon in terms of access to resources: Men with higher levels of education possess more resources, which means they have the ability to use force. Hypothesis 6: Women with more or less education than their partners are more likely to experience domestic violence than women whose education levels are the same as their partners. Based on this literature review, relationships between family characteristics and domestic violence in Latin America are examined. In so doing, this study aims to shed light on why some Latinas are more likely than others to experience abuse. METHODS Sample This analysis uses Demographic and Health Surveys (DHSs) conducted in Colombia (1995), the Dominican Republic (1999), Haiti (2000), Nicaragua (1998), and Peru (2000). DHSs are nationally representative household surveys with large sample sizes of women ages 1549, which provide data for a wide range of monitoring and impact evaluation indicators in the areas of population, health, and nutrition. The datasets do not include all factors empirically linked to differential rates of domestic violence; thus the results should not be interpreted as denitive. Nonetheless, DHSs do contain measures of several signicant family characteristics of violence, including marital status, family size, partner alcohol use, socioeconomic status, decision-making power, and education homogamy. Restricting the sample to women currently in a union, the sample sizes are 6,082 in Colombia, 588 in the Dominican Republic, 2,275 in Haiti, 6,728 in Nicaragua, and 15,174 in Peru. Sample cases are weighted to adjust for oversampling of particular regions and to compensate for differences in response rates. Measures This study focuses exclusively on physical abuse between heterosexual partners and employs Gelles (1997) denition of violence: An act carried out with the intention or perceived intention of causing physical pain or injury to another person (p. 14). Physical aggression is operationalized differently in each of the datasets. The Colombia DHS asks respondents to list reasons they have been physically hit by their partners. A dichotomous variable was created to measure violence and is coded 1 if the respondent listed a reason she had been abused, and 0 if

23 she had not been abused. In the Dominican Republic and Haiti, a series of questions was asked based on Straus (1990) Conict Tactics Scale. If the respondent answered yes to any of the physical aggression questions, her response was coded 1; if she answered no to all of the questions, her response was coded 0. The Nicaragua DHS asks if respondents have ever experienced any physical violence at the hands of a partner. Yes responses were coded 1; no responses were coded 0. The Peru DHS asks respondents if they have ever been hit, pushed, or shoved by a partner. Again, yes answers were coded 1 and no answers were coded 0. Cases with missing dependent variable data were excluded from the analyses. The percentage of cases dropped is no greater than 10% in each of the datasets. Marital status is measured as a dichotomous variable coded 1 if the respondent is married and 0 if she is cohabiting. Divorced, widowed, and separated women are not included in this study. In Haiti, 97% of the sample are married. Thus, too few women are cohabiting to create a meaningful measure of marital status in Haiti. Based on US statistics, we anticipate cohabiting women to experience more violence than married women (Brownridge & Halli, 2000). Family size is measured by how many living children the woman has. We expect that women with larger families are more likely to experience violence than women with smaller families, since large families have higher stress levels associated with having to provide for several children (Hoffman et al., 1994). Partner alcohol use is measured by how often respondents partners come home drunk. Dummy variables were created for each response option: never gets drunk, sometimes gets drunk, and frequently gets drunk. Because response options in the Nicaragua DHS are slightly different, we adjusted the categories for uniformity. If the male never comes home drunk, the response is categorized as never gets drunk; if he comes home drunk once in a while or once a month, the response is categorized as sometimes gets drunk; and if he comes home drunk twice a month, once a week, or almost daily, the response is categorized as frequently gets drunk. Because alcohol weakens brain mechanisms that normally restrain aggression (Parker & Rebhun, 1995), we expect a positive, linear relationship between partner alcohol use and domestic violence. Decision-making power is determined by a series of questions that ask if the woman, her partner, or somebody else has the nal say in certain household decisions (such as her own health care, making large household purchases, daily purchases, visits to family or relatives, and food to be prepared each day). Although the questions vary slightly across datasets, they are conceptually uniform in that they

24 measure decision-making power with respect to household decisions. Response options include respondent alone, respondent and partner, respondent and other person, partner alone, someone else, and other. Respondents are classied into one of four relationship power types based on their answers to the nal say questions: egalitarian (both partners have an equal say in most issues), divided power (man or woman is dominant in making decisions in different areas), female-dominant (woman makes most decisions), and male-dominant (man makes most decisions) (Straus, 1990). We expect the likelihood of violence to be greater in nonegalitarian relationships because there is a higher probability of conict when couples do not make decisions together (Rettig, 1993). Education homogamy is measured with three dummy variables constructed by subtracting a womans total years of education from her partners total years of education. If the female and male have the same amount of schooling, the response is categorized as homogamous; if the male has more schooling, the response is categorized as male has more than female; if the female has more schooling, the response is categorized as female has more than male. Based on Western data (Okun, 1986), we expect women in nonhomogamous relationships to be more likely to be abused than women in homogamous relationships. Estimation Basic descriptive statistics are initially employed to provide a demographic prole of the samples. As the dependent variable is binary, each dataset is examined separately using logistic regression techniques. The equations express the log odds of being abused (versus not) as a linear function of a set of explanatory variables. The models coefcients represent the increase or decrease in the likelihood of physical abuse, associated with a unit (or category) change in an independent variable. RESULTS Table I presents descriptive statistics on family factors inuencing partner violence. The prevalence of spousal violence in all ve countries is high, ranging from 16% in Haiti to 39% in Peru. The percentages could be much higher, as some women deny, minimize, and underreport abuse (Ellsberg et al., 2001). These data suggest that while domestic violence occurs throughout Latin America, the proportion of women who have experienced violence varies dramatically between countries. Cross-

Flake and Forste national variation in violence rates may be attributable to a wide array of sociodemographic and cultural factors. Differences in survey methodologies and variable operationalization may also help account for differential rates of violence. Couple relationships in Latin America are characterized by a strong afnity to cohabit rather than marry. In Colombia, Nicaragua, and Peru, approximately half of all women currently in relationships are cohabiting; in the Dominican Republic, two-thirds of women cohabit. Other characteristics of Latino families include relatively high fertility, alcohol use, and low socioeconomic status. While poverty affects families throughout Latin America, some countries are much poorer than others. In Colombia, 67% of couples own at least six (of a possible seven) household amenities included in the SES index. In comparison, 5% of Haitians own the same number of amenities. The power dynamics of couple relationships vary across countries, suggesting that patriarchal norms might not be uniform across Latin America. Decision-making power, for example, varies dramatically between countries. In the Dominican Republic and Nicaragua, most couples make decisions together (egalitarian). In Haiti, most couples split household decision-making responsibilities (divided power) rather than make decisions together. In Peru, it is most common for females to control the decision making (female-dominant). There is also variance in education homogamy. The cultural expectation for males to have more education than their female partners persists in all ve countries. In some countries, however, this norm is challenged: In Colombia, the Dominican Republic, and Nicaragua, one-third of women in relationships have more education than their partners.

Logistic Regression Model Odds ratios presented in Table II provide at least partial support for the hypothesized relationships between family characteristics and the likelihood of experiencing partner violence. Marital status and partner alcohol use are the strongest predictors of abuse, having emerged as statistically signicant in all of the datasets. Married women are far less likely to be physically abused than cohabiting women. The effect is strongest in the Dominican Republic, where married women are half as likely to be abused as cohabitors (p < .01). That the Dominican Republic has the lowest percentage of married women (35%) is important, as it indicates that marital status has a pronounced effect on domestic violence in that country. As hypothesized, partner alcohol use increases a womans

Domestic Violence in Latin America


Table I. Demographic and Background Factors Inuencing Partner Violence (Percentages) Colombia (1995) Ever physically abused by partner Marital status Married Cohabiting Family size (living children) 01 24 5+ Partner alcohol use Never gets drunk Sometimes gets drunk Frequently gets drunk Missing SES (07) 02 35 67 Decision-making relationship Egalitarian Divided power Male-dominant Female-dominant Other Education homogamy Homogamous Male has more than female Female has more than male Missing [ N] 19.0 54.1 45.9 28.8 57.7 13.5 8.9 24.0 67.1 26.0 38.0 34.6 1.4 6082 Dom. Rep. (1999) 22.6 35.0 65.0 25.2 62.2 12.6 32.9 25.5 8.0 33.7 7.8 47.5 44.7 44.5 17.3 8.8 26.7 9.5 16.6 42.1 32.0 9.3 588 Haiti (2000) 15.7 97.3 2.7 29.6 45.9 24.5 8.1 9.6 2.6 79.7 65.6 29.4 5.0 19.2 51.4 4.3 22.2 2.9 27.6 46.1 15.9 10.4 2275 Nicaragua (1998) 26.1 46.0 54.0 23.9 50.3 25.8 45.8 40.9 12.0 1.2 23.2 50.4 26.4 60.1 9.6 11.1 13.6 5.6 25.3 37.6 35.2 1.9 6728 Peru (2000) 38.9 55.8 44.2 24.7 55.7 19.6 26.6 65.7 6.9 .8 28.0 31.5 40.5 32.5 11.7 8.5 43.6 3.7 30.1 50.8 18.7 .3 15174

25

Note. Statistics are weighted to represent population parameters. The reported sample sizes are weighted.

likelihood of being assaulted. Women whose partners sometimes get drunk are between 1.3 (p < .001) and 2.5 times (p < .001) more likely to experience violence than women whose partners never get drunk. Frequent drunkenness is associated with an even higher likelihood of violence: Women whose partners frequently get drunk are between 2.6 (p < .001) and 9.8 (p < .001) times more likely to be abused than women whose partners do not get drunk. As hypothesized, women who do not make decisions together with their partners are at a greater risk of being abused than women who share in the decision-making process (egalitarian). In Haiti, Nicaragua, and Peru, women in divided power relationships (she makes some decisions, he makes others) are between 1.2 (p < .01) and 2 times (p < .001) more likely to experience violence than women in egalitarian relationships. Women whose partners control decision making (male-dominant) are between 1.3 (p < .01) and 2.7 times (p < .001) more likely to be abused

than women in egalitarian relationships. Male-dominant decision making is not statistically signicant in the Dominican Republic or Peru. Female-dominant relationships have the strongest and most consistent effect on domestic violence. In each country, women who control the decision-making are much more likely to experience violence than women who share decision making with their partners. The hypothesized relationship between education homogamy and domestic violence is partially supported by these data. In Colombia and Haiti, women with less education than their partners are more likely to experience violence than women with the same level of education as their partners. In Nicaragua and Peru, women with more education than their partners have a higher likelihood of abuse than women who have the same amount of education as their partners. Education homogamy was not found to be associated with spouse abuse in the Dominican Republic.

26
Table II. Family Characteristics and the Likelihood of Experiencing Partner Violence (Odds Ratios) Colombia (1995) Married Family size Partner alcohol use Never gets drunk Sometimes gets drunk Frequently gets drunk Socioeconomic status (07) Decision-making relationship Egalitarian Divided power Male-dominant Female-dominant Education homogamy Homogamous Male has more than female Female has more than male 2 LL Chi-square Df [ N] .747 1.276 1.027 1.000 1.190 1.140 5784.291 132.820 5 6082 Dom. Rep. (1999) .519 .742 1.000 2.441 9.844 1.140 1.000 1.642 1.526 2.057 1.000 1.135 1.185 552.836 75.660 10 588 Haiti (2000) 1.010 1.000 2.484 4.800 1.172 1.000 1.925 2.750 2.318 1.000 1.805 1.067 1866.065 114.232 9 2275 Nicaragua (1998) .593 1.174 1.000 1.298 2.631 1.009 1.000 2.026 1.314 2.082 1.000 1.013 1.158 7245.761 478.436 10 6728

Flake and Forste

Peru (2000) .704 1.210 1.000 1.890 8.233 1.004 1.000 1.197 1.084 1.378 1.000 1.076 1.373 18897.311 1382.728 10 15174

Note. Statistics are weighted to represent population parameters. The reported sample sizes are weighted. p < .05. p < .01. p < .001.

DISCUSSION Although domestic violence is a serious and widespread problem in Latin America, few researchers have sought to explain partner violence in the Latino context. This study examines the magnitude and characteristics of partner abuse in Colombia, the Dominican Republic, Haiti, Nicaragua, and Peru, and is one of the only violence studies to test a model across multiple datasets. This research offers several important contributions to the family violence literature and serves as a foundation for future research in Latin America. A major contribution of this study is that it tests the applicability of Western theoretical and empirical violence models to a non-Western setting. Previous research has focused primarily on wife abuse in North America and Europe. Thus, little is known about the nature of partner violence in cross-cultural settings. To address this gap in the violence literature, the current study examines how Western risk markers for abuse inuence violence in Latin America. To at least some extent, each of the variables tested emerged as an important predictor of domestic violence, indicating that several risk markers for abuse are shared between Latin American and Western countries. This nding has signicant implications for family violence researchers and policymakers, as it suggests that current research and policies might have some relevance

in parts of Latin America. Much more research is needed to fully understand how the national context inuences spouse abuse. This study helps clarify the prole of the abused Latina. If a woman cohabits in Latin America, she is more likely to experience violence than if she is married. While marriage is critical to reducing abuse among Latinas, there is a tendency for women to cohabit rather than marry. That cohabitation rates are increasing in every Latin American country (Castro Martin, 2002) is a major concern, as it signies that more and more women are inadvertently placing themselves at risk of partner violence. Socioeconomic conditions are likely to be part of the explanation for the high prevalence of cohabitation in Latin America. Castro Martin (2002) explains that unlike in developed countries, consensual unions in Latin America are most prevalent among the poor, suggesting that nancial costs may deter couples from formal marriage. Although marriage is generally regarded as more desirable than cohabitation, consensual unions are easier to initiate and are less costly (Greene, 1991). Modifying existing marriage requirements to accommodate the poor could help reduce the incidence of domestic violence by encouraging couples to more fully commit to one another by marrying rather than cohabiting. Partner alcohol use also plays a critical role in partner violence. Of all the family factors included in the present

Domestic Violence in Latin America study, alcohol has the strongest and most consistent effect on the likelihood of experiencing domestic violence. Although alcohol consumption rates in Latin America are not extraordinarily high, problem drinking is more prevalent among Latinos than other groups (Madrigal, 1998). Given the rigid social expectation that macho men should be able to consume large quantities of alcohol (Giraldo, 1972), it is unlikely that alcohol consumption rates can be lowered. A more plausible recommendation would be to educate men and women about drinking responsibly. Latino families should be educated about the risk of domestic violence that accompanies drunkenness. If men can learn to drink without getting drunk, they may be less likely to become violent, so long as casual drinking does not drastically alter their perceptions and judgment (Parker & Rebhun, 1995). In addition to marital status and partner alcohol use, power dynamics inuence domestic violence in Latin America. Decision-making power, in particular, has a pronounced effect on the likelihood of abuse. Our ndings suggest that if couples do not make decisions together, there is a greater likelihood of domestic violence than if they share in decision making. Power dynamics in Latin America are such that when one partner has more decision-making power than the other, there is a greater risk of marital conict and violence. In particular, when females wield more decision-making power than their partners, they are more likely to be abused than when they share decision-making power equally. This nding lends support to theories of patriarchy, which suggest that men who have less power than their partners may turn to violence to reestablish culturally prescribed dominance over women (Straus et al., 1980). If a woman cohabits, has a large family, has a partner who gets drunk, does not share decision-making responsibilities with her partner, or does not have the same level of education as her partner, she is more likely to experience domestic violence than a woman who marries, has a small family, has a partner who never gets drunk, shares decision-making power with her partner, or has the same amount of education as her partner. The prole of the abused Latina appears quite similar to the prole of abuse victims in the United States, Great Britain, Switzerland, and other Western countries. The full picture remains blurred, however, as numerous factors were not tested in the present study. According to the ecological perspective, domestic violence is a multifaceted phenomenon grounded in an interplay of individual, family, community, and national characteristics (Heise, 1998). To understand differences in abuse victims, one must consider the entire ecology of the individual:

27 their home, workplace, church, family and community roles, and the overarching institutional patterns of culture (Bronfenbrenner, 1979). While this study makes important contributions to the understanding of domestic violence in Latin America, it is not without limitations. Two problems arise from the narrow focus of Demographic and Health Surveys, which are not designed primarily for the study of abuse. First, the operationalization of domestic violence questions is not always uniform, making cross-national comparisons somewhat difcult. Second, several family characteristics of abuse in Western cultures were not included in the DHS questionnaires. Religiosity, resource control, and attitudes toward violence may be important risk markers for abuse. These factors were unable to be included in the model, however, because the DHSs do not include measures of these variables. Findings from this study should not be interpreted as denitive, but rather as foundational. Much more research is needed to fully understand the characteristics of domestic violence in Latin America. To compare crossnational results more effectively, future research should employ standardized questionnaires and methodologies. Other types of abuse, including psychological and sexual abuse, and child maltreatment should be included to understand the entire scope of domestic violence. Power dynamics must also be emphasized. Machismo in particular may play a prominent role in explaining spouse abuse in Latin America, and can be measured with a variety of well-established scales, including the Bem Sex Role Inventory (Bem, 1974), Villemez and Tougheys (1977) 28-point Macho Scale, the Hyper-Masculinity Index (Mosher, 1991), and Cuellar and colleagues (1995) 17-item Machismo Scale. A nal recommendation is to extend research to men. The vast majority of violence studies target women because they are generally more willing to participate and share their experiences with abuse. While understanding the risk markers of abuse for women is critical, it is equally essentialif not more so that we uncover the reasons why men hit their partners. To effectively lower rates of intimate violence, we must create a thorough and comprehensive prole of the abusernot just the abuse victim. REFERENCES
Anderson, K. L. (1997). Gender, status, and domestic violence: An integration of feminist and family violence approaches. J. Marriage Fam. 59: 655669. Bem, S. L. (1974). The measurement of psychological androgyny. J. Consult. Clin. Psychol. 42: 155162. Brinkerhoff, M. B., and Lupri, E. (1988). Interspousal violence. Canadian J. Soc. 12(4): 407434.

28
Bronfenbrenner, U. (1979). Ecology of Human Development, Harvard University Press, Cambridge, MA. Brownridge, D. A., and Halli, S. S. (2000). Living in sin and sinful living: Toward lling a gap in the explanation of violence against women. Aggr. Viol. Behav. 5(6): 565583. Buvinic, M., Morrison, A. R., and Shifter, M. (1999). Violence in the Americas: A framework for action. In Morrison, A. R., and Biehl, M. L. (eds.), Too Close to Home: Domestic Violence in the Americas. Inter-American Development Bank, John Hopkins University Press, Washington, DC, pp. 334. Carlson, B. E. (1984). Causes and maintenance of domestic violence: An ecological analysis. Soc. Service Rev. 58(4): 569587. Castro Martin, T. (2002). Consensual unions in Latin America: Persistence of a dual nuptiality system. J. Comp. Fam. Studies 33(1): 3555. Coleman, D. H., and Straus, M. A. (1990). Marital power, conict, and violence in a nationally representative sample of American couples. In Straus, M. A., and Gelles, R. J. (eds.), Physical Violence in American Families, Transaction, New Brunswick, NJ, pp. 287 300. Counts, D., Brown, J., and Campbell, J. (1992). Sanctions and sanctuary, Westview, Boulder, CO. Cuellar, I., Arnold, B., and Gonzalez, G. (1995). Cognitive referents of acculturation: Assessment of cultural constructs in Mexican Americans. J. Comm. Psych. 23: 339356. Economic Commission for Latin America and the Caribbean (2002). 20012002 Social Panorama of Latin America, United Nations, New York. Ellsberg, M. C., Pe na, R., Herrera, A., Liljestrand, J., and Winkvist, A. (1999). Wife abuse among women of childbearing age in Nicaragua. Am. J. Pub. Hlth. 89(2): 241244. Ellsberg, M. C., Pe na, R., Herrera, A., Liljestrand, J., and Winkvist, A. (2000). Candies in hell: Womens experiences of violence in Nicaragua. Soc. Sci. Med. 51(8): 15951610. Ellsberg, M. C., Winkvist, A., Heise, L., Pe na, R., and Agurto, S. (2001). Researching domestic violence against women: Methodological and ethical considerations. Stud. Fam. Plann. 32(1): 116. Farrington, K. (1977). Family violence and household density: Does the crowded home breed aggression? Association paper, Society for the Study of Social Problems Whitman College, Walla Walla, WA. Gelles, R. J. (1974). The Violent Home, Sage, Beverly Hills, CA. Gelles, R. J. (1997). Intimate Violence in Families. Sage, Thousand Oaks, CA. Giraldo, O. (1972). El machismo como fen omeno psicocultural. Revista Mexicana de Psicologia 3: 350354. Gonzales de Olarte, E., and Gavilano Llosa, P. (1999). Does poverty cause domestic violence? Some answers from Lima. In Morrison, A. R., and Biehl, M. L. (eds.), Too Close to Home: Domestic Violence in the Americas, Inter-American Development Bank, John Hopkins University Press, Washington, DC, pp. 3550. Goode, W. J. (1971). Force and violence in the family. J. Marr. Fam. 33(4): 624636. Greene, M. E. (1991). The importance of being married: Marriage choice and its consequences in Brazil. Doctoral Dissertation, University of Pennsylvania. Heise, L. (1998). Violence against women: An integrated, ecological framework. Viol. Against Women 4(3): 262290. Hoffman, K. L., Demo, D. H., and Edwards, J. N. (1994). Physical wife abuse in a non-western society: An integrated theoretical approach. J. Marr. Fam. 56(1): 131146. Holtzworth-Munroe, A., Bates, L., Smutzler, N., and Sandler, E. (1997). A brief revision of research on husband violence: Part I: Maritally violent versus nonviolent men. Aggr. Viol. Behav. 2(1): 6599. Hotaling, G. T., and Sugarman, D. B. (1986). An analysis of risk markers in husband-to-wife violence: The current state of knowledge. Viol. Vict. 1(2): 101123. Ingoldsby, B. B. (1991). The Latin American family: Familism vs. machismo. J. Comp. Fam. Stud. 22(1): 5762.

Flake and Forste


Jewkes, R. (2002). Intimate partner violence: Causes and prevention. Lancet 359(9315): 14231429. Kim, J. Y., and Sung, K. (2000). Conjugal violence in Korean American families: A residue of the cultural tradition. J. Fam. Viol. 15(4): 331 345. Larrain, S. (1993). Estudio de frecuencia de la violencia intrafamiliar y la condici on de la mujer in Chile. Pan American Health Organization, Santiago, Chile. Madrigal, E. (1998). Latin America. In Grant, M. (ed.), Alcohol and Emerging Markets: Patterns, Problems, and Responses, Taylor and Francis, Ann Arbor, MI, pp. 223262. Martin, S. E. (ed.) (1993). Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives. National Institute on Alcohol Abuse and Alcoholism (NIAAA) Research Monograph No. 24. National Institute of Health Publication No. 933496. Rockville, MD, pp. vxi. Martin, S. L., Tsui, A. O., Maitra, K., and Marinshaw, R. (1999). Domestic violence in northern India. Am. J. Epid. 150(4):417426. McCord, W., and McCord, J. (1960). Origins of Alcoholism, Stanford University Press, Stanford, CA. Messing, U. (1999). Introduction. In Morrison, A. R., and Biehl, M. L. (eds.), Too Close to Home: Domestic Violence in the Americas, Inter-American Development Bank, John Hopkins University Press, Washington D.C., pp. xixiii. Mosher, D. L. (1991). Macho men, machismo, and sexuality. Ann. Rev. Sex. Research 2: 199247. Murray, C., and Lopez, A. (eds.) (1996). The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020, Vol. 1, Harvard University Press, Cambridge, MA. Nelson, E., and Zimmerman, C. (1996). Household survey on domestic violence in Cambodia. Ministry of Womens Affairs and the Project Against Domestic Violence, Phnom Penh, Cambodia. Nock, S. L. (1995). A comparison of marriages and cohabiting relationships. J. Fam. Issues 16(1): 5376. OBrien, J. E. (1971). Violence in divorce prone families. J. Marriage Fam. 33: 692698. Okun, L. (1986). Woman Abuse: Facts Replacing Myths, State University of New York Press, Albany. Pan, H. S., Neidig, P. H., and OLeary, K. D. (1994). Predicting mild and severe husband-to-wife physical aggression. J. Consult. Clin. Psych. 62(5): 975981. Parker, R. N., and Rebhun, L. (1995). Alcohol and Homicide: A Deadly Combination of Two American Traditions, State University of New York Press, Albany. Paz, O. (1961). The Labyrinth of Solitude: Life and Thought in Mexico, Grove Press, New York, pp. 2930. Population Reference Bureau (2002). 2002 World Population Data Sheet. Population Reference Bureau, Washington, DC. Rettig, K. D. (1993). Problem-solving and decision-making as central processes of family life: An ecological framework for family relations and family resource management. Marr. Fam. Rev. 18(3/4): 187222. Riding, A. (1985). Distant Neighbors, Vintage, New York. Rivera, J. (1998). Domestic violence against Latinas by Latino males. In Delgado, R., and Stefancic, J. (eds.), The Latino/a Condition: A Critical Reader, New York University Press, New York, pp. 501507. Roizen, J. (1997). Epidemiological issues in alcohol-violence. In Galanter, M. (ed.), Recent Developments in Alcoholism, Vol. 13: Alcoholism and Violence, Plenum Press, New York, pp. 740. Stevens, E. (1973). Machismo and marianismo. Society 10: 5763. Stith, S. M., and Farley, S. C. (1993). A predictive model of male spousal violence. J. Fam. Viol. 8(2): 183201. Straus, M. A. (1990). Social stress and marital violence in a national sample of American families. In Straus, M. A., and Gelles, R. J. (eds.), Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families. Transaction, New Brunswick, NJ, pp. 181201.

Domestic Violence in Latin America


Straus, M. A., Gelles, R. J., and Steinmetz, S. K. (1980). Behind Closed Doors: Violence in the American Family, Anchor Press, New York. Villemez, W. J., and Toughey, J. C. (1977). A measure of individual differences in sex stereotyping and sex discrimination: The Macho Scale. Psych. Reports 41: 411415.

29
Walker, L. E. (1984). The Battered Woman Syndrome, Springer, New York. Yllo, K. A. (1993). Through a feminist lens: Gender, power and violence. In Gelles, R. J., and Bogard, M. (eds.), Feminist Perspectives on Wife Abuse, Sage, Newbury Park, CA, pp. 4762.

Das könnte Ihnen auch gefallen