Sunteți pe pagina 1din 31

An Evidenced-Based Approach to a Theoretical Understanding of the Relationship Between Economic Resources, Race/Ethnicity, and Woman Abuse

Kameri Christy-McMullin, PhD

ABSTRACT. Within the context of advocating for evidence-based practice in the field of social work, the need also exists to test or systematically review studies that support or refute theories used in social work practice. The purpose of this paper is to present a systematic review of domestic and international studies that examine the relationship between economic resources, race and ethnicity, and woman abuse. Findings of these studies are connected to their implications for exchange, feminist and asset-building theories. Lamentably, few studies include race or ethnicity. While the use of divergent measures for economic resources and woman abuse result in mixed findings, one consistent finding is that economic assets are negatively related to woman abuse. [Article copies available for a fee from
The Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address: <docdelivery@haworthpress.com> Website: <http://www.HaworthPress.com> 2006 by The Haworth Press, Inc. All rights reserved.]

KEYWORDS. Woman abuse, economic resources, race

Kameri Christy-McMullin is affiliated with the School of Social Work, University of Arkansas, 106 ASUP, Fayetteville, AR 72701 (E-mail: kmcmull@uark.edu). Journal of Evidence-Based Social Work, Vol. 3(2) 2006 Available online at http://www.haworthpress.com/web/JEBSW 2006 by The Haworth Press, Inc. All rights reserved. doi:10.1300/J394v03n02_01

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

The need for women to achieve economic security or independence, and in turn be able to leave an abusive relationship, is a theme commonly found in the literature on woman abuse (Davis, 1991; Davis & Hagen, 1992; Kurz, 1998; Gondolf, 1988; Shepard & Pence, 1988). Feminists conceptually build on this premise to hypothesize that economic resources are so important because they bestow women with more power both within their relationship and in society (Dobash & Dobash, 1979; Levinson, 1989; Peterson, 1980; Shepard & Pence, 1988). Exchange theorists also make a conceptual connection between having resources or power and freedom from woman abuse (Gelles, 1982). However, an inherent limitation of both feminist and exchange theorists is their tendency to conceptualize economic resources too narrowly. Asset-building theory adds a potentially crucial dimension to the way in which we think about economic resources or status (Conley, 1999; Oliver & Shapiro, 1995; Sherraden, 1991). This paper systematically reviews studies that examine the complex relationship between economic resources, race/ethnicity, and woman abuse in order to evaluate the appropriateness of using feminist-exchange and asset-building theories to increase our understanding of woman abuse. THE ROLE OF POWER IN ABUSIVE RELATIONSHIPS Feminists assert that violence against woman is a symptom of patriarchy and theorize that men use assorted tactics of abuse to obtain and keep power/dominance over women (Davis & Hagen, 1992; Dobash & Dobash, 1979; Koss, Goodman, Browne, Fitzgerald, Keita, & Russo, 1994; Yllo, 1993). One method of acquiring additional power within a relationship is to have more access to and control over financial resources than the other person. In this country, economic and social structures have historically made women more vulnerable to economic inequities (Davis & Hagen, 1992; McCue, 1995), with women of color being the most at risk (Abramovitz, 1996). Drawing from feminist exchange theory, Riger and Krieglstein (2000) contend: husbands with more marital power than their wives can afford to be abusive . . . men use violence as a means of control because they can get away with it; women have too few resources to affect the balance of power in the relationship. (p. 636)

Kameri Christy-McMullin

This has led some womens advocates to predict that increasing womens independent economic security will result in a decrease in woman abuse because their partners will have more to lose economically if they use to violence (Farmer & Tiefenthaler, 1997; Riger & Krieglstein, 2000). On the other hand, there is concern that if women have their own economic resources, men might become more abusive in an attempt to maintain their dominance over women (Kalmuss & Straus, 1982; Riger & Krieglstein, 2000; Stout, 1992). This phenomenon is often referred to as the backlash hypothesis (Avakame, 1999; Riger & Krieglstein, 2000). INCLUDING RACE/ETHNICITY IN OUR CONCEPTUAL FRAMEWORK Historic economic and political oppression faced by members of various racial and ethnic groups, as well as the close relationship between economic resources and woman abuse, highlight the importance of including race/ethnicity in the conceptual framework regarding woman abuse. However, perhaps in an attempt to obtain public support for the problem of woman abuse, feminists have typically amalgamated all women together (Crenshaw, 1994; Kanuha, 1996). For example, it has been commonly asserted that all women are vulnerable to woman abuse, regardless of socioeconomic status, race, ethnicity, age, religious background or sexual orientation (Crenshaw, 1994; Kanuha, 1996). Consequently, white, middle-income, heterosexual women have become the reference group; with the unique experiences of women outside this group remaining relatively unexamined (Ho, 1990; Kanuha, 1996). Additionally, due to the oppression marginalized groups have endured, some women from these communities remain silent regarding their experiences of abuse, due to feelings of being unsafe (Allen, 1982; Crenshaw, 1994; Ginorio & Reno, 1986; White, 1986) or of being disloyal to their community (Crenshaw, 1994; Richie, 1987; White, 1986). In light of these barriers, Kanuha (1996, p. 35) notes, often discussed but never well-elucidated, analyses of domestic violence in the relationships of women and men of color have been disappointingly meager. Given these gaps in theory and research, scholars need to move beyond a superficial examination of comparative data between different racial and ethnic groups (Lockhart, 1985). In regards to economic resources, it may not be adequate to perfunctorily state that

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

woman abuse is more common among women of color than it is for White women because as a group, have lower income and less prestigious jobs than Euroamericans have. A deeper examination may also reveal that economic measures do not benefit or affect minority women in the same way as they do White ones. Any conceptual framework about woman abuse needs to include an explanation about the shared effects of race/ethnicity and economic resources. CONCEPTUALIZATING ECONOMIC RESOURCES When conceptualizing social class or socioeconomic status, by and large, social scientists have relied on the measures of income, education and/or employment (Conley, 1999). While poverty theorists and researchers in the United States has focused almost exclusively on income, Sherraden (1991) observes that very little theorizing has actually been done to justify the use of income as the measure of economic welfare. However, in the past decade, a few scholars have sought to broaden our thinking about economic status by including measures of wealth, or economic assets (e.g., Conley, 1999; Oliver & Shapiro, 1995; Sherraden, 1991). Sherraden (1991) theorizes that economic assets may have social benefits and outcomes separate from, or in addition to, the more traditional measures of economic resources. He hypothesizes that assets improve household stability; connect people with a viable, hopeful future; stimulate development of human and other capital; enable people to focus and specialize; provide a foundation for risk taking; yield personal, social, and political dividends; and enhance the welfare of offspring (p. 148). Given the recent formulation of asset-building theory, Sherraden acknowledges that there is a need to test the extent to which each proposition is true, and under what circumstances, and possible interrelationships among the propositions (p. 148). Within the past two decades, researchers have begun to explore the relationship of assets to such measures of well-being as physical health (Robert & House, 1996), stress (Rocha, 1994), life satisfaction and perceived self-esteem (Roche & Stegman, 1994), civic involvement in neighborhood associations (Perkins, Florin, Rich, Wandersman, & Chavis, 1990), community volunteerism (Thompson, 1993), wife abuse (Levinson, 1989; Peterson, 1980), womens social status (Noponen, 1992), education attainment of children (Green & White, 1997; Hill & Duncan, 1987), and teen pregnancy (Green & White, 1997). However, it is important to con-

Kameri Christy-McMullin

duct additional research to explore whether these propositions are true for specific groups of people, especially those that are most vulnerable socially, economically, and politically. REVIEW OF THE LITERATURE This paper critiques sixteen quantitative inquiries that explore the relationship between woman abuse and economic resources (see Table 1). In order to be included in this review, the studies were required to have at least one measure for both economic resources and woman abuse. It would have been preferable to also have included race/ethnicity as one of the criteria for inclusion; this would have severely limited the number of articles available for review. Additionally, studies examining the relationship between economic assets and woman abuse were purposively included in this review of existing research in order to assess Sherradens (1991) theory that asset building is related to social wellbeing (e.g., safety from woman abuse). Research on Economic Resources and Femicide Two of the studies examined the relationship between economic resources and the most serious type of woman abuse; the murder of a woman by a current or former intimate male partner (Avakame, 1999; Stout, 1992). Both studies also explored macro-level, as well as micro-level economic variables. Avakame (1999) Avakame (1999) analyzed data from the 1990-1994 Supplemental Homicide Reports (N = 6,533). The independent variable, economic resources, is operationalized as percent of females active in the civilian workforce and economic deprivation (proportion of families earning less than $15,000 in 1989). Femicide is measured as the murder of a wife, common-law-wife, girlfriend, or ex-wife by an intimate partner. One of the control variables is race, which is a dichotomous variable (white or minority). The findings from multivariate analyses demonstrate that an increase in the female employment rate, and an increase in the number of people experiencing economic deprivation are significantly associated with a decrease in femicide rates (p = .03 and .02, respectively). While the results also indicate, minority females are more

6 TABLE 1. Summary of the Overall Findings of the Studies Reviewed


Study Avakame, 1999 Measure of Economic Resource(s) Percent of females in civilian labor force; and economic deprivation (proportion of families earning less than $15,000 in 1989). Measure(s) of Woman Abuse Statistically Significant Finding(s) Femicide (women killed by current or Increase in female employment rate former partners). and in number of people experiencing economic deprivation is negatively associated with femicide rates. Farmer & Omaha: man working for pay; woman Tiefenthaler, 1997 working for pay; womans educational level; level of mans non-wage income; and level of womans non-wage income Charlotte: mans monthly earnings; womans monthly earnings; and whether the woman graduated from high school. Kalmuss & Straus, 1982 Womans economic dependency (womans employment, children under 5, and percent of womens income to mans). Family income; educational level of both partners; and occupational status of both partners. Economic inequality: inheritance norms of valuable property; ownership of wealth; ownership of a dwelling; & control over the distribution of material products. Family income; income disparity between the man and woman; total family occupational prestige; and disparity between the man and womans occupational prestige.

Omaha: number of fights in past 6 months resulting in physical injury. Charlotte: number of times in past 6 months of physical abuse.

Omaha: men and women working for pay, and the level of womens non-wage income have a negative relationship to physical injury. Level of mens non-wage income has a positive correlation. Charlotte: womans monthly earnings have a negative association with physical abuse.
Severe physical abuse is positively related to economic dependency.

Physical abuse (minor and severe subscales of the CTS) in the past year.

Lambert & Firestone, 2000 Levinson, 1989

Number of different types of physical Family income has a negative abuse, in the past year (CTS). association with the number of different types of physical abuse. Wife beating: frequency and intensity Female control of all four asset-based of physical assaults. variables is negatively related to wife beating.

McCloskey, 1996

Woman abuse (verbal aggression & violence subscales of CTS, and one question about sexual assault), in the past year.

As income disparity between men and women increases, woman abuse increases. The lower the combined occupational status, the more likely the man will abuse the woman.

Oropesa, 1997

Ever been hit by husband. Male and female educational attainment; number of years of female employment (before and during marriage); employment status of both partners; national female employment rate; and female-male employment ratio. Womans educational attainment, employment status, income, and income in relation to poverty line; household income; savings; and homeownership. Male income; male education; male employment stability; male occupational status; and homeownership. Marital conflict, ever (either person openly disagrees, argues heatedly, or shouts); and marital violence, ever (either person hits, throws things, or becomes physical during an argument).

Negative association between womans educational attainment and being hit; and a positive relationship between number of years a woman is employed after being married, and both partners being unemployed with being hit. Controlling for womans independent economic resources and household income, homeownership is negatively related to marital violence.

Page-Adams, 1955

Peterson, 1980

Ever been physically abused (hit, Homeownership is negatively slapped, beat, burnt, choked, stabbed, correlated with ever being physically or cut). abused. Participation in a credit program is negatively associated with being beaten in the past year. Both having some schooling and participating in a credit program are negatively associated with being beaten in the past year. Women from households where women are employed and the men are unemployed are at highest risk of injury. Employed women with employed partners have the lowest risk of injury.

Schuler & Hashemi, 1994 Schuler, Hashemi, Riley & Akhter, 1996

Womens exposure to credit programs Beaten in the past year. after one year (development of a small business). Participation in a credit program; female ever attended school; economic level of household; and womans financial contribution to household. Employment status of men and women. Having been beaten in the past year.

Schwartz, 1990

Battery/physical injury (NCS).

8 TABLE 1 (continued)
Study Smith, 1990 Measure of Economic Resource(s) Family income; educational level of both man and woman; occupational status of both partners; and employment status of both partners. Measure(s) of Woman Abuse Any physical abuse, ever; any physical abuse in past year; severe physical abuse, ever; and severe physical abuse past year (CTS). Statistically Significant Finding(s) Both family income & mans educational level are negatively associated with severe physical abuse (ever and in past year). Mans educational level is negatively correlated with any physical abuse, ever. Stets & Strauss, 1989 Stout, 1992 Educational level of both partners; and Rate of physical violence; and severity Physical violence occurs more occupational status of both partners. of physical assaults (CTS). frequently among blue-collar workers than among white-collar workers. Femicide Percent of women employed in civilian labor force; type of female employment; type of male employment; and state laws enacted to protect women in workforce. Higher female, as compared to male, unemployment has a positive relationship to femicide; increased femicide when more women than men have managerial or administrative employment; and decreased femicide in states with protective legislation for female employees. Straus, 1994 3 subscales of the Gender Equality Any physical abuse (CTS), in the past GEI is negatively associated with any Index (economic, political, and legal); year. physical abuse. and income inequality between men and women at the state level (Gini index).

Kameri Christy-McMullin

likely to be killed than their White counterparts (p. 283), analyses were not conducted to test for an interaction between minority status and economic resources. There are several concerns about the variables used in this inquiry to measure economic resources. Not only does Avakame (1999) use just two indicators of economic resources, it is not clear why labor force participation is limited only to those in civilian employment. The selection of a dichotomous variable for economic deprivation, rather than using a richer (perhaps continuous) variable of family income is also a limitation. Lastly, creating a variable of economic deprivation based solely on family income is counter to the more common method of including both family size and household income when measuring poverty. Given the seemingly contradictory findings regarding female workforce participation and economic deprivation, using these findings to expand our conceptual framework regarding the relationship between a specific type of woman abuse, femicide, and economic resources is problematic. These conflicting results may be due, in part, to the use of somewhat arbitrary and limited criteria to operationalize economic measures. Stout (1992) In a study drawing upon an ecological framework, Stout (1992) conducted a secondary data analysis of statewide data from 1980-1982 (N = 50). The average rate of women whose deaths were classified as murder or non-negligent manslaughter and who were killed by male intimate partners (p. 34) is the dependent variable. It is likely that this study includes more women killed by an intimate partner than does the Avakame (1999) study, since Stout includes non-negligent manslaughter in her data set. However, a limitation of this inquiry is the lack of a list of all of the independent variables. Stout (1992) reports that the economic variables of women being employed in the civilian workforce (p = .8) and percent of women employed in professional and technical occupations, compared to men in those positions (p = .2), are not significantly correlated with femicide. The findings suggest that when women are disproportionately unemployed, relative to men, there is an associated increase in the femicide rate (p = .02, r2 = .11). There is also a relationship between an increase in the femicide rate and a higher percent of women than men being employed in managerial and administrative positions (p = .02, r2 = .11). Other significant finding of this study include a negative correlation between femicide and states having fair employment laws (p = .0005),

10

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

equal pay laws (p = .0001), and laws against sexual discrimination in the area of housing (p = .02). Bivariate analysis of race (white or black) and femicide indicate that black women are disproportionately represented, comprising 37 percent of the women who were killed by an intimate partner. Information regarding other races or ethnic groups was not provided. Additionally, Stout did not test for an interaction between race/ethnicity and economic resources in relation to femicide. This study is useful because numerous measures of economic resources were included, many of which pertain to macro or structural variables. Inclusion of these macro-level variables allowed for a comparison of femicide rates between those states that have enacted various economic/political interventions with those that have not done so. However, total reliance on bivariate analyses yields less fruitful information than could be gained if multivariate analyses had been conducted. Additionally, including the Pearsons Product Coefficient for the structural variables would have allowed the calculation of the effect size of those variables in relation to femicide. Research on Economic Resources and International Studies While it is questionable whether or not the findings of this body of research can be generalized to poor women in this country, these studies are included because few domestic studies have tested for the relationship between economic assets and woman abuse. Therefore, these studies were included to demonstrate that the research being conducted in other countries operationalize a broader conceptualization of economic resources. These findings may assist scholars in the U.S. to more fully understand the relationship between economic resources and woman abuse. These international studies may also provide important direction for future innovative research in the U.S. Schuler and Hashemi (1994) In a Bangladeshi study, Schuler and Hashemi (1994) analyzed ethnographic data from a large randomly selected sample of 1,305 rural, married Bangladeshi women who were interviewed in 1991 and again 18 months later. The authors randomly selected women from four different samples: women participating in the Grameen Bank program, women participating in the Bangladesh Rural Advancement Committee (BRAC) program, women living in villages served by Grameen but who did not

Kameri Christy-McMullin

11

participate in the program (comparison group I), and women who lived in villages that were not served by either program (comparison group II). While programming for both the Grameen Bank and BRAC are very similar, unlike the Grameen Bank, BRAC targets all eligible residents (the landless poor) from entire villages. Additionally, the Grameen Bank tends to enroll fewer women, recruits a slightly more well off selection of the very poor and uses a more structured format for its meetings. Participation or not in a credit program was the independent variable of interest, and the dependent variable of interest was whether the woman had been beaten in the past year. The authors found that after the credit programs had been running for one year, women in both comparison groups I and II reported having been beaten 1.4 and 1.8 times more often in that past year than women in the BRAC programs and 2.3 and 3 times more often than women participating in the Grameen programs. A more sophisticated analysis of the data on woman abuse and economic resources was not conducted. While the authors conceptualized the independent variable as womens exposure to credit programs, this may be misleading. It may actually be the accumulation of small, independent business assets that help strengthen womens economic roles (Schuler & Hashemi, 1994), and help explain the decrease in physical abuse. While the quasi-experimental design attempts to provide this inquiry with more internal validity, the lack of baseline data makes it difficult to assess if these differences existed among the four groups prior to the intervention. However, the researchers include women from villages served by the Grameen Bank, but who do not participate in the program (comparison group I) to address internal validity concerns about nonparticipants inadvertently benefiting from the effects of an intervention. Given that the women in the Grameen Bank program experienced less physical abuse than women participating in the other credit program, it may be that the differences in the programs themselves explain at least part of the variation in the abuse rates. In addition to using an asset measure in the conceptualization of economic resources, another potential strength of this inquiry is the use of data collectors who were ethnographers. These researchers lived in the villages for an extended period of time, which may have increased the likelihood that the women reported whether they had been beaten, and thereby decreased the occurrence of underreporting that is problematic in most studies of woman abuse.

12

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

Schuler, Hashemi, Riley, and Akhter (1996) In another Bangladeshi study, Schuler and her colleagues (1996) re-examine their earlier data to specifically test for the relationship between participating in credit programs (independent variable) and physical abuse (dependent variable). Other independent variables include: whether the woman ever attended school, economic level of the household, and the womans financial contribution to the household. The findings indicate that women with some schooling are about 30 percent less likely to have been beaten than women who had no schooling. The authors also found that, women who participate in either Grameen Bank or BRAC are less likely to be beaten by their husbands than women in villages without credit programs (p. 1737). Neither the economic level of the household nor the womans financial contribution to the support of the family were significantly related to physical abuse. As with the earlier study (Schuler & Hashemi, 1994), this inquiry measured woman abuse in a very limited way, using only the indicator of whether a woman was beaten in the past year. A final limitation with both studies is that selection bias might explain the differences in abuse rates among the different groups of women. It may be that women who experience less physical abuse are more likely to join a credit program to start their own small business. Levinson (1989) From the field of cultural anthropology, Levinsons (1989) worldwide comparative study of 90 small-scale and peasant societies examined the broad phenomenon of family violence. Wife beating was operationalized as frequency and intensity of physical assaults. The results indicated that each of the four asset-based indicators had a statistically significant (at the .05 level or less), but weak relationship to wife beating. The results for these variables were: inheritance norms of valuable property favors females (Spearmans rho = .12), women owning their own independent wealth (Spearmans rho = .21), female ownership of a dwelling (Spearmans rho = .22), and female control over the distribution of monetary/material products (Spearmans rho = .25). Of the social structural variables, the only predictor of wife beating was whether women participate in exclusively female work groups (Spearmans rho = .3). Levinson hypothesized that this finding was a result of the social support and economic independence that membership in these groups provide women.

Kameri Christy-McMullin

13

One of the unique contributions this study made to the professional literature was the use of four asset measures and collection of data regarding womens independent assets. As with Stouts (1992) inquiry, this study examined macro-level data and demonstrated that societies/ communities with structural supports for womens economic security are also less likely to practice physical abuse of women. Another strength was that intercoder reliability for the four asset and one wife beating measures ranged from 73 to 94 percent agreement. Levinson also utilized Spearmans rho coefficients to demonstrate the degree to which two variables were related. However, as with the other two international studies, physical abuse was operationalized narrowly, encompassing only the most severe types of abuse. Oropesa (1997) Oropesa (1997) analyzed data from a 1992 national survey in Mexico. A multistage probability sampling procedure was used to survey 60 localities, resulting in a sample of 794 married and cohabiting women in Mexico. The dependent variable was measured by the question Has your husband or partner ever hit you? (p. 1299). Approximately 19 percent of the women reported ever being hit. Educational attainment of the woman, the disparity in education attainment between the partners, number of years the woman has been employed both before and during the relationship, the employment status of both partners, the employment rate for women in the county, and the female-male employment ratio were among the independent variables used in this inquiry. The findings suggested a negative relationship between a womans educational attainment and being hit, a positive relationship between the number of years a woman was employed after being in the relationship and being hit, and both partners being unemployed and being hit. There was no significant relationship between the other economic variables and being hit. The author concluded, The risk of violence is affected less by the creation of dual-earner families than by the creation of no-earner families (Oropesa, 1997, p. 1312). Limitations included the exclusion of never married, separated, and divorced women and the use of only one indicator for woman abuse. Additionally, there is a fallacy in asking if a woman has ever been hit and then trying to relate that to the level of her current economic resources (or those of the county in which she resides). For instance, a woman may be currently employed but she may have been hit 3 years ago, when she was unemployed. However, this distinction cannot be as-

14

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

certained from the data, resulting in the findings indicating women who have been hit are also employed. Research on Economic Resources and Non-Standardized Measures of Woman Abuse The following research articles are grouped together because they examined the relationship between economic resources and physical abuse of women by an intimate partner. However, a commonality of these inquiries is their use of a single indicator, and/or non-standardized measures in the operationalization of physical abuse. Peterson (1980) Petersons (1980) study was groundbreaking in that it appears to be the earliest inquiry that includes an asset measure (homeownership) in operationalizing economic resources. He collected data via a telephone survey in 1977 and 1978 of 602 randomly selected married, divorced or cohabiting women in the state of Maryland. The relationship between a woman having ever been abused (hit, slapped, beat, burnt, choked, stabbed, or cut) and the economic resources of the household (homeownership, and the mans income, education, occupation, and employment stability) were analyzed in this inquiry. When reviewing the findings, Peterson asserted that all five of the economic measures were related to woman abuse. One problem with his assertion is that of the five economic measures, only homeownership was significantly related to ever being abused (x2 = 39.33, p = .001). Furthermore, Peterson stated, women who rent report an incident of ever being abused that is nine times greater than that reported by those who own their own homes (p. 398). However, the numbers provided in the findings tables report that 19 homeowners and 32 renters reported abuse, indicating that renters are 1.7 times more likely than homeowners to report being abused. Additionally, multivariate analyses were not conducted in order to test for the effects of one variable (e.g., income) on another variable (e.g., homeownership). Much like Oropesas study (1997), another shortcoming of this study is that Petersons (1980) measurement of physical abuse renders these findings very suspect. While a woman was asked whether she had ever been abused by her husband, she was asked about their current housing status (as well as current employment, income, etc). So, a woman may have been physically abused 5 years earlier (when the couple rented

Kameri Christy-McMullin

15

their housing), but at the time of the interview they were homeowners. This would result in a misrepresentation of the findings; the couple is classified as homeowners with the woman being abused, when they should be classified as renters with the woman being abused. Schwartz (1990) Using data from the National Crime Survey (NCS), Schwartz (1990) explores whether the relationship between the employment status of women and men, and experiences of battery and physical injury. He compared the data among three subsamples of women; those who were abused/injured by either a spouse or ex-spouse (N = 667), women who reported abuse/injury by a stranger (N = 832), and women who reported experiencing no abuse/injury during the same time frame (N = 850). It should be noted that Schwartz did not explicitly state the size of each subsample; the numbers provided here are estimates. The findings indicated that the vast majority of men who batter their current or former wives are currently employed; only eight percent are unemployed. However, women abused by their partners are more likely to be unemployed. While this difference is significant (p = .0001), the effect is weak (Cramers V = .155). As for injury, 40 percent of the women who were the sole breadwinner reported receiving medical care, as compared to 19 percent of women whose husbands were the sole breadwinner and 21 percent of the women who lived in households with two breadwinners (p = .02, Cramers V = .173). A potentially confounding issue with the last finding is that women who are sole breadwinners may in fact be separated or divorced from their partners. However, Schwartz did not control for marital status in the data analyses. One of the main limitations of this study is its brevity. Consequently, much of the relevant information, such as how battery is operationalized, was lacking. It appears that injury was operationalized as whether the woman sought medical treatment. Since not all women seek medical care after being abused, this will result in an underreporting of injuries. There are numerous questions in the NCS that ask about different types of physical abuse and injury, but Schwartz did not specify which ones he used for his analyses. Additionally, information regarding the year or years of data collection, and how duplicated observances were handled was not provided. A final limitation is the lack of attention that was given to the possible interaction between race/ethnicity and economic resources.

16

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

Page-Adams (1995) Page-Adams (1995) conducted a secondary data analysis with a large subsample (N = 2,827) from the National Survey of Families and Households (NSFH, 1986).The NSFH researchers employed random sampling in obtaining both male and female respondents, and collected data via face-to face interviews. The variable of martial conflict was operationalized by two indicators; the respondents were asked whether they ever argue heatedly or shout at each other during disagreements (p. 36), and whether they ever openly disagreed with each regarding household tasks, money, spending time together, sex, having a(nother) child, in-laws, or the children. Marital violence is measured by two indicators; whether the couple ever hits or throws things during a disagreement, and whether either partner had become physical during an argument. In addition to the more traditional measures for economic resources (highest level of education completed by the woman, womans employment status, womans income, womans income in relation to the federal poverty line, and household income), two asset measures (savings and homeownership) were also included in the operationalization of economic resources. The large sample size allowed for multivariate analysis to be run, thereby statistically controlling for the effects of one variable (e.g., income) on another (e.g., savings). Page-Adams (1995) reported, the only economic resource that has a consistently negative effect on marital violence is household assets in the form of homeownership (p. 93). This result held true even when controlling for the effects of household income and the womans independent economic resources. Additionally, none of the economic measures had significant effects on such violence when controlling for the effects of homeownership. A significant limitation of this study is that the dependent variables are marital conflict and marital violence. It is not possible to differentiate between male and female victims in this study. Consequently, the findings are not generalizable to women who experience abuse by an intimate partner, but only to couples who report experiencing conflict and/or violence. Farmer and Tiefenthaler (1997) In an article by Farmer and Tiefenthaler (1997), data were collected from police call responses in Omaha, Nebraska from 1986-1987 (N = 560 observations) and from the Charlotte (North Carolina) Spouse Rep-

Kameri Christy-McMullin

17

lication Project, 1987-1989 (N = 340 observations). The Omaha data were collected at baseline, and at six and twelve month follow-up interviews. A severe limitation of this study is that the authors treated the responses of the three different interview times as if different women were interviewed each time. While the authors do not address this concern, the sample may contain women who were interviewed two or even three times. This potentially skews the analyses, and hinders the credibility of the findings of this study. Only information collected at baseline was provided for the Charlotte data, but it is unclear from the article if there might have been duplicated observances in this data set as well. There are a number of significant differences between the two data sets that make the flow and comprehension of this study by Farmer and Tiefenthaler (1997) somewhat challenging. The dependent variable for the Omaha data was number of fights in the past six months that resulted in the woman being physically injured, while for the Charlotte data it was the number of times in the past six months that the woman was physically abused. The Omaha data set contains the independent variables of whether the woman worked for pay, the level of the womans monthly non-wage income, years of education for the woman, whether the man works for pay, and the level of the mans monthly non-wage income. The independent variables for the Charlotte data included whether the woman completed high school, womans monthly take-home earnings and mans monthly take-home earnings. The results of the Omaha data demonstrate that both men and women working for pay, as well as the level of womens non-wage income, have a negative relationship with physical injury. Conversely, the level of mens non-wage income is positively associated with physical injury. Only the womans monthly earnings were significantly correlated (negatively) with physical abuse in the Charlotte data. Along with the other limitations of this inquiry, the limited geographic areas from which the data were collected and obtaining participants via police records greatly restrict the ability to the generalize the findings to other abused women. Critique of the Conflict Tactic Scales (CTS) to Measure Abuse Although surrounded by much controversy, Richard J. Gelles and Murray A. Straus are considered to be leaders in the study of family violence. The Conflict Tactic Scale (CTS) was originally developed by Straus and Gelles in the 1970s and is the most commonly used instrument for collecting data regarding physical violence in families

18

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

(Straus & Gelles, 1990). The CTS is a 19-item scale that falls into three general modes of behaviors used to settle differences: Reasoning (rational discussion), Verbal Aggression (verbal or nonverbal acts that symbolically hurt another), and Violence (physical aggression). At times, Straus and others have divided the Violence subscale into (1) minor and severe violence (Kalmuss & Straus, 1982; Stets & Straus, 1989) or (2) ordinary and severe violence (Straus, 1990c). These 19 items are asked in a hierarchical order of less severe behaviors to most severe behaviors (Straus, 1990a). The internal reliability of this instrument is reported to range from .83 to .88 for husband-to-wife violence (Straus, Gelles, & Steinmetz, 1988). The main strength of the CTS is its use of a Guttman-type scale, which orders the questions from the least to most aggressive acts. Theoretically, this increases the rapport between the researcher and participant; thereby increasing the likelihood that the respondent will answer truthfully (Straus et al., 1988). The use of specific terminology, such as slapped or hit with an object, rather than vague terms, such as abuse or violence may also help respondents know what kinds of behavior the researchers are interested in; thereby increasing the validity of the scale. However, it can also be argued that if the behavior(s) the woman has experienced is (are) not part of the questionnaire, she may not report that she has been abused. Feminists have seriously questioned the validity of the CTS. The theoretical premise of the CTS is that conflict exists in all social units, including the family, and violence or aggression occurs in attempts to resolve those conflicts (Kalmuss & Straus, 1982; Straus, 1990a; Straus et al., 1988). The purpose of this scale is to measure violence in the family by asking about the means used to resolve conflicts of interest among family members (Straus et al., 1988, p. 26). This alleged ability to broadly apply the CTS to all types of family violence and to all family members is one of the criticisms raised by feminists (Dobash & Dobash, 1982). They contend that conflict and personal interests occur along lines of gender and generational entitlement (Yllo, 1990, p. 52). Therefore, it is inappropriate to conceptualize this process of resolving conflict as being gender or age neutral. While the work of Straus and Gelles has been helpful in moving the focus of woman abuse from individual psychopathology to an examination of social forces that contribute to violence, The patriarchal nature of these social forces was not adequately incorporated into the theories or empirical measures (Yllo, 1990, p. 50). Another concern levied against the CTS is that it dominates the family violence field to an ex-

Kameri Christy-McMullin

19

tent rarely matched by other scales in the field . . . [and] has been adopted by a wave of researchers who have never explored, questioned, tested, or often, shared its underlying theoretical assumptions (Yllo, 1990, p. 52). Furthermore, neither Straus nor Gelles have ever explained, why violence is best conceptualized as a conflict tactic (Yllo, 1990, p. 52). Feminists hypothesize that violence is better conceptualized as a tactic men use to maintain their control and power both within the relationship (Koss et al., 1994; Kurz, 1990, Yllo, 1990), and within society (Koss et al., 1994). A final problem with the CTS is that Straus and Gelles (1988) have defined violence as an act carried out with the intention, or perceived intention of causing pain or injury to another person (p. 15). However, as noted by Koss and her colleagues (1994, p. 64), the parameters of those definitions are not measured by the CTS (e.g., intent). Research on Economic Resources and the Use the CTS The six studies discussed in this section of the research critique all measure woman abuse as physical abuse only, and all use the violence subscale of the CTS. These inquiries demonstrate four different ways in which the CTS can be employed in research. In the first three studies, Kalmuss and Straus (1982), Smith (1990), and Stets and Straus (1989) divide the violence subscale into minor or any violence and severe violence to explore whether different levels of violence have the same or divergent association with economic resources. However, Straus (1994) used the overall violence subscale in his analyses. While Lambert and Firestone (2000) also used the overall violence subscale, they converted the level of measurement for this dependent variable into a continuous variable by counting the number of different types of victimizations a woman reported having experienced. In a somewhat unique study, McCloskey (1996) expanded the operationalization of woman abuse to include not only physical but also verbal and sexual abuse in operationalizing woman abuse. She did this by using both the verbal aggression and violence subscales of the CTS, as well as by adding a question regarding sexual assault to the interview instrument. Minor versus Severe Abuse Kalmuss and Straus (1982). Using a national area probability sample of 1,183 married or cohabitating women interviewed in 1976, Kalmuss and Straus (1982) examined the relationship between a womans eco-

20

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

nomic dependency and physical abuse (the minor and severe violence subscales of the CTS). Economic dependency was indexed by summing the scores of three dichotomous variables; whether the woman was employed, whether she had children 5 years or younger, and whether her partner earned 75 percent or more of the familys income. The internal reliability of this index is poor (.59), and severely jeopardizes confidence in the findings. Another limitation is combining all abused women together as being homogeneous, rather than testing to see if there might be differences among various groups of women (such as age, marital status, or race/ ethnicity). The results indicated that while minor violence was not significantly related to economic dependency (p = .15), severe violence did have a positive relationship to economic dependency (p = .02). The authors suggest, wives who are highly dependent on marriage are less able to discourage, avoid, or put an end to abuse than are women in marriages where the resources between husbands and wives is more nearly equal (p. 282). However, this conclusion seems premature in that two of three dimensions (having children age 5 or younger and female employment) used to create the economic dependency index do not measure equal access to resources. Smith (1990). In a study with Toronto women, Smith (1990) analyzed data collected in 1987 via random digit dialing of 604 currently or formerly married/cohabiting. The formerly married/cohabiting women had to have been in a relationship within the past two years to be included in the survey. The response rate of 56.4 percent is problematic in that selection bias may influence the results of the analyses. As compared to Kalmuss and Straus (1982), who used two measures for violence, Smith (1990) operationalized violence as (1) any abuse, past year, (2) any abuse, ever, (3) severely abused, past year, and (4) severely abused, ever. Smith also used more measures of economic resources; (1) total family income, (2) highest level of education completed for both partners, (3) occupational status of both partners, and (4) employment status of both partners. The findings of these two studies were comparable, in that fewer economic resources (familys income and mans education) seem to be associated more with severe abuse, regardless of when the abuse occurred. One variable (mans education) was negatively related to any abuse, ever (p = .03). Interestingly, none of the other variables, specifically measures of the womans economic resources, were significant when the multivariate analyses were conducted. However, a limitation of the results of any abuse, ever and severe abuse, ever is the same one experi-

Kameri Christy-McMullin

21

enced by Peterson (1980) and Oropesa (1997); asking about any history of abuse when the current measures of economic resources are being used is likely to yield inaccurate findings. Stets and Straus (1989). In a study with a slightly different focus, Stets and Straus (1989) tested for the interaction between marital status, economic resources, and violence. Consequently, the authors were primarily interested in comparing economic differences among dating, cohabiting and married couples, rather than testing for potential protective economic variables against woman abuse. The data analyzed for this inquiry came from two samples. The first was a probability sample of non-married students at a large midwestern university (N = 526) in 1987. Only students who had dated during 1986 were allowed to participate. The second sample was comprised of married (N = 5,005) and cohabiting (N = 237) couples from the 1985 the National Family Violence Resurvey. The education level of both partners was measured categorically, and occupational status of both partners was operationalized as either blue or white-collar worker. Marital status was operationalized as dating, cohabiting, or married, and violence was operationalized as rate of physical violence and severity of physical assaults. The findings suggested that, after controlling for age, education, and occupational status, cohabiting couples have both the highest rate of violence and the most severe violence compared to dating or married couples. While violence occurred more frequently with blue-collar workers (p = .05), education was not significantly related to the rate of violence (Stets & Straus, 1989). One limitation of this study is that the number of cohabiting persons was extremely small, only four percent of the entire sample. Such an uneven distribution of the marital status variable might affect the statistical analyses. Drawing the dating sample from one university is another drawback, in that this greatly limits the generalizability of the findings. A final limitation is that analyses of race and/or ethnicity are noticeably absent from this study. Overall Violence Straus (1994) study tested the relationship between the Gender Equality Index (GEI) and income inequality of the 50 states, and the percentage of married/cohabiting women who reported experiencing any physical abuse by her partner within the past 12 months. The data regarding physical abuse were also drawn from the 1985 National Family Violence

22

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

Resurvey. The GEI contains three spheres; economic (comprised of seven indicators), political (comprised of four indicators), and legal (comprised of 13 indicators). The internal reliability of the GEI was reported as being marginal at .65. Income inequality is measured by the 1980 Gini index. Control variables included percent of the state population that was African American, and the unemployment rate for the state. The author reported that gender equality, as measured by the GEI, is negatively related to physical violence. However, physical abuse is not significantly associated with income inequality. Given the comprehensiveness of the GEI variable, it is not possible to determine whether specific indicators of the GEI have a stronger relationship to overall physical abuse. It is difficult to imagine that all 24 indicators of the GEI are equally powerful in explaining the variance of physical abuse. Additionally, the low internal consistency of the GEI, as with the economic dependency index used by Kalmuss and Straus (1982), make the credibility of these findings questionable. Number of Physical Assaults Lambert and Firestone (2000) also analyzed data from the 1985 Physical Violence in American Families (known also as the 1985 National Family Violence Resurvey). Data were collected from both men (N = 413) and women (N = 569) regarding physical male violence against the intimate female partner. The authors hypothesize that it was the number of different types of physical assaults that was related to various economic measures (family income, highest level of education completed for both partners, occupational status of both partners, and disparity between both the man and womans educational attainment and their occupational status). Control variables included race (white or black) and ethnicity (Hispanic or non-Hispanic). The average number of abuses reported by the women was 2.14 and the average number reported by the men was 1.75. The findings suggest that there is a negative relationship between family income and the total number of different types of physical abuse reported by women (p = .01). However, family income accounted for about only two percent of the variance in the total number of different physical abuses reported by women. Expanding the Operationalization of Woman Abuse McCloskey (1996) recruited 365 women from domestic violence and other community agencies to explore the relationship between eco-

Kameri Christy-McMullin

23

nomic resources and woman abuse. Woman abuse was operationalized by using both the verbal aggression and violence subscales of the CTS and by adding one question about sexual assault. The number of abusive incidents was combined to create a 7-point scale, resulting in a Cronbachs alpha for these combined items of .95. The economic variables were (1) total family income, (2) income disparity between the man and woman, (3) total family occupational prestige, and (4) disparity between the man and womans occupational prestige. While household income was not significantly related to woman abuse, a woman having more income than her partner was related to increased woman abuse (p = .007). Conversely, disparity of occupational prestige was not significantly associated with women abuse, but lower combined occupational status of the couple was related to increased woman abuse (p = .024). CONCLUSIONS Given the wide discrepancies in how economic resources are measured, it is very difficult to extract any overall patterns within the research reviewed for this article. For example, Smith (1990) and Lambert and Firestone (2000) operationalized income as household or family income, while McCloskey examined both independent male and female income, as well as the disparity between the two. An additional barrier to unearthing any general trends is that some studies use only one or two measures of economic resources (Avakame, 1999; Schwartz, 1990). This makes it impossible to know if other economic measures would have also been statistically significant. Furthermore, some studies used only bivariate analyses (Levinson, 1989; Peterson, 1980; Schuler & Hashemi, 1994; Stout, 1992), which restrict our ability to know if one economic measure serves as a proxy for another. Consequently, the findings throughout the literature are mixed; income may not be significantly related to woman abuse in some studies (Page-Adams, 1995; Peterson, 1980), while there is a statistically significant relationship between these two variables in other studies (McCloskey, 1996; Lambert & Firestone, 2000; Smith, 1990). However, the one consistent finding in the literature is that asset measures have a negative correlation with woman abuse. This holds true, whether the assets are operationalized as homeownership (Page-Adams, 1995; Peterson, 1980); womens ownership of a small business (Schuler & Hashemi, 1994; Schuler et al., 1996); womens independent

24

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

ownership of wealth, material products, or housing; or social norms promoting female inheritance of property (Levinson, 1989). Critically assessing such issues as measurement, sampling, and research design, makes it easier to discern why, on the surface, there appears to be mixed findings regarding the relationship between economic resources and woman abuse. For example, at first glance it may appear that results demonstrating that employed women with unemployed partners have the highest risk for woman abuse (Schwartz, 1990) contradict Smiths (1990) findings that the significant relationship is between woman abuse and mens education, not mens employment. However, Schwartz was looking specifically at injury, while the dependent variable in Smiths study was physical abuse. Additionally, Schwartz findings pertained to a disparity between the employment status of the couple and Smiths did not. Unfortunately, it seems that most of the studies discussed have significant limitations. This decreases the ability to have much confidence in their findings. In turn, the ability to develop a conceptual framework of how economic resources, race/ethnicity, and woman abuse are or are not related to each other is greatly constrained. Measurement The studies by Peterson (1980), Smith (1990), and Oropesa (1997) all used unrelated time frames with their measures for woman abuse and economic resources, which rendered the findings questionable. The poor internal reliability of the economic measures used in the studies by Kalmuss and Straus (1982), and Straus (1994) hindered the credibility of their results. Another potential limitation is that the selection of only one or two economic indicators (as done in Avakame, 2000 and Schwartz, 1990) may be too narrow a measure for such a broad and complex concept as economic resources. While Page-Adams (1995) inquiry is interesting in that she used innovative measures of economic resources, the operationalization of conflict and violence as pertaining to both people equally in the relationship contributes little to our understanding of the relationship between woman abuse specifically and economic resources. A glaring gap in the literature is the lack of empirical knowledge being generated regarding the interaction between race/ethnicity, economic resources and woman abuse. Of the 16 studies reviewed, only one (Lambert & Firestone, 2000) includes both race and ethnicity, and a second inquiry (Avakame, 1999) includes race.

Kameri Christy-McMullin

25

Sampling In conjunction with measurement problems, Smith (1990) had the additional drawback of a poor response rate (56.4 percent). Farmer and Tiefenthaler (1997) had two problems pertaining to their sampling design. The first was the probability that there were duplicated observances unaccounted for in the sample. Additionally, generalizability is compromised by the fact that they drew the samples from residents of Omaha, Nebraska and Charlotte, North Carolina. Similarly, McCloskeys (1996) use of a non-randomized sample of women from one geographic locality significantly restricts the generalizability of her findings. Research Design and Data Analyses Since all the studies reviewed are correlational, it is not possible to know whether higher levels of economic resources cause women to experience less physical abuse or if those women who do not experience physical abuse are more likely to accumulate economic resources. It may be that other variables, such as age (Gelles & Cornell, 1990) or marital status, influence both woman abuse and the amount of economic resources a woman has. However, the use of comparison groups helps increase the internal validity of two international studies (Schuler & Hashemi, 1994; Schuler et al., 1996). Additionally, exclusive reliance on bivariate analyses (Levinson, 1989; Schuler & Hashemi, 1994; Stout, 1992) provides only a very basic (and perhaps only a partially correct) understanding of the relationship between the dependent and independent variables. IMPLICATIONS FOR SOCIAL WORK PRACTICE The limitations of previous studies, the variation between the inquiries and the resulting mixed findings makes it problematical to ascertain implications for direct and policy practice. However, the consistent, negative correlations between assets and woman abuse suggest that social workers need to expand their conceptualization of economic resources to include wealth (e.g., savings accounts, homeownership, retirement/ pension funds, and small business ownership). Consequently, social workers need to be educated about and involved in the newly emerging asset-building movement that seeks to assist low-income and lowwealth individuals and families accumulate assets in order to move out

26

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

of poverty. Furthermore, social workers need to ensure that abused women are included in these anti-poverty policies and programs (ChristyMcMullin, 2000). Social work involvement includes lobbying for and promoting increased funding from both the private and public sectors, making appropriate referrals for low-income individuals to asset-building programs, developing asset-building programs that are culturally sensitive and meet the needs of a diverse population, as well as implementing and evaluating the outcomes of asset-building programs. IMPLICATIONS FOR FUTURE RESEARCH Of the domestic studies, two stand out as providing the most guidance for future inquires. While her sampling design is a limitation, McCloskey (1996) operationally expanded upon woman abuse to include not only physical, but verbal and sexual assaults as well. This is a very unique characteristic of this study, and the internal reliability of these combined items was very high (.95). However, rather than having one dichotomous indicator for woman abuse, if possible, it would be instructive to test for relationships between different types of woman abuse and different types of economic resources. She also employed four measures of economic resources, although two pertained to income and the other two were concerned with occupational prestige. There are also several strengths in Lambert and Firestones (2000) inquiry. The use of a randomly selected, national study helps increase confidence in generalizing the findings. The national data set provides a large enough sample to allow for multivariate analyses, which is particularly beneficial when studying the complex relationship between woman abuse, economic resources, and race/ethnicity. Additionally, the operationalization of physical abuse included both minor and severe indicators of physical abuse. Not only did the authors use seven indicators for economic resources, they also controlled for the effects of race and ethnicity on their findings. Unfortunately, the findings of the two domestic studies that used some type of economic asset measure are tenuous, at best. While the results of the three international studies are more promising, the basic statistical analyses of Levinson (1989), and Schuler and Hashemis (1994) inquiries limit a more comprehensive understanding of the potential relationships between various economic asset measures and woman abuse. While these international studies suggest that economic assets may also have a negative association with woman abuse in this country, there has

Kameri Christy-McMullin

27

yet to be empirical support for this hypothesis. A final gap in the asset literature is that the operationalization of woman abuse in each of the five studies was narrow and somewhat ambiguous. These limitations highlight the need for future studies that (1) use a randomly selected, national sample, (2) include a variety of measures for economic resources, specifically asset measures, (3) include measures for race and/or ethnicity, (4) measure woman abuse more broadly to include emotional, physical, economic, and sexual abuse, (5) conduct multivariate analyses to control for potential effects of one economic measure on another, and (6) test for an interaction between economic resources and race/ethnicity, and woman abuse. In conjunction with expanding policies and practice to include asset development for abused women, social workers need to be actively involved in testing the outcomes for abused women in such programs. In order to have a strong, empirical foundation to support or refute feminist-exchange and asset-building theories, it is imperative that future studies take into consideration the above recommendations so that there will be consistency among studies. This will facilitate our ability to determine which interventions are most germane for an extremely vulnerable population, women who experience abuse by an intimate partner. REFERENCES
Abramovitz, M. (1996). Regulating the Lives of Women (Rev. ed.). Boston: South End Press. Allen, P.G. (1986). Violence and the American Indian woman. In M.C. Burns (Ed.), The Speaking Profits Us: Violence in the Lives of Women of Color (pp. 5-7). Seattle, WA: Center for the Prevention of Sexual and Domestic Violence. Avakame, E.F. (1999). Females labor force participation and intimate femicide: An empirical assessment of the backlash hypothesis. Violence and Victims, 14 (3), 277-291. Christy-McMullin, K. (2000). An analysis of the Assets for Independence Act of 1998 for abused women. Violence Against Women, 6 (10), 1066-1084. Coley, S. M., & Beckett, J. O. (1988). Black battered women: Practical issues. Social Casework, Oct., 483-490. Conley, D. (1999). Being Black, Living in the Red. Los Angeles, CA: University of California Press. Crenshaw, K.W. (1994). Mapping the margins: Intersectionality, identity politics, and violence against women of color. In M.A. Fineman and R. Mykitiuk (Eds.), The Public Nature of Private Violence: The Discovery of Domestic Abuse (pp. 93-118). New York: Routledge. Davis, L.V. (1991). Violence and families. Social Work, 36 (5), 371-374.

28

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

Davis, L.V., & Hagen, J.L. (1992). The problem of wife abuse: The interrelationship of social policy and social work practice. Social Work, 37 (1), 15-20. Dobash, R.E., & Dobash, R. (1979). Violence Against Women: A Case Against the Patriarchy. New York: The Free Press. Farmer, A., & Tiefenthaler, J. (1997). An economic analysis of domestic violence. Review of Social Economy, 55 (3), 337-359. Gelles, R. (1982). An exchange/social control theory. In D. Finkelhor, R. J. Gelles, G. T. Hotaling, and M. A. Straus (Eds.), The Dark Side of Families: Current Family Violence Research, pp. 151-165. Beverly Hills, CA: SAGE Publications. Gelles, R.J., & Cornell, C.P. (1990). Intimate Violence in Families (2nd ed.). Newbury Park, CA: SAGE Publications. Gelles, R.J., & Straus, M.A. (1990). Appendix A: Sample design and comparability of the two national surveys. In M.A. Straus & R.J. Gelles (Eds.), Physical Violence in American Families (pp. 529-534). New Brunswick, NJ: Transaction Publishers. Ginorio, A., & Reno, J. (1986). Violence in the lives of Latina women. In M.C. Burns (Ed.), The Speaking Profits Us: Violence in the Lives of Women of Color (pp. 13-16). Seattle, WA: Center for the Prevention of Sexual and Domestic Violence. Gondolf, E.W. (1988). The effect of batter counseling on shelter outcome. Journal of Interpersonal Violence, 3, 3, 275-289. Green, R.K., & White, M.J. (1997). Measuring the benefits of home owning: Effects on children. Journal of Urban Economics, 41, 441-461. Hill, M.S., & Duncan, G.J. (1987). Parental family income and the socioeconomic attainment of children. Social Science Research, 16, 39-73. Ho, C.K. (1990). An analysis of domestic violence in Asian American communities: A multicultural approach to counseling. Women & Therapy, 9 (1-2), 129-150. Kalmuss, D.S., & Straus, M.A. (1982). Wifes marital dependency and wife abuse. Journal of Marriage and the Family, 44, 277-286. Kanuha, V. (1996). Domestic violence, racism, and the battered womens movement in the United States. In J.L. Edleson & Z.C. Eisikovits (Eds.), Future Interventions with Battered Women and Their Families (pp. 34-50). Thousand Oaks, CA: SAGE Publications, Inc. Koss, M.P., Goodman, L.A., Browne, A., Fitzgerald, L.F., Keita, G.P., & Russo, N.F. (1994). No Safe Haven: Male Violence Against Women at Home, at Work, and in the Community. Washington, DC: American Psychological Association. Kurz, D. (1993). Physical assaults by husbands: A major social problem (pp. 88-103). In R.J. Gelles & D.R. Loseke (Eds.), Current Controversies on Family Violence. Newbury Park, CA: SAGE Publications, Inc. Kurz, D. (1998). Women, welfare, and domestic violence. Social Justice, 25 (1), 105-122. Lambert, L.C., & Firestone, J.M. (2000). Economic context and multiple abuse techniques. Violence Against Women, 6 (1), 49-67. Levinson, D. (1989). Family Violence in Cross-Cultural Perspective. Newbury Park, CA: Sage. Lockhart, L.L. (1985). Methodological issues in comparative racial analysis: The case of wife abuse. Social Work Research and Abstracts, 21, 35-41. McCloskey, L.A. (1996). Socioeconomic and coercive power within the family. Gender & Society, 10 (4), 449-463.

Kameri Christy-McMullin

29

McCue, M.L. (1995). Domestic Violence: A Reference Handbook. Denver, CO: ABCCLIO, Inc. Noponen, H. (1992). Loans to the working poor: A longitudinal study of credit, gender, and the household economy. International Journal of Urban and Regional Research, 16, 234-251. Oliver, M.L., & T. M. Shapiro. (1995). Black Wealth/White Wealth: A New Perspective on Racial Inequality. New York: Routledge. Oropesa, R.S. (1997). Development and marital power in Mexico. Social Forces, 75 (4), 1291-1317. Page-Adams, D. (1995). Economic Resources and Marital Violence. Unpublished doctoral dissertation, Washington University, St. Louis, MO. Perkins, D.D., Florin, P., Rich, R.C., Wandersman, P., & Chavis, D.M. (1990). Participation and the social and physical environment of residential blocks: Crime and community context. American Journal of Community Psychology, 18, 83-115. Peterson, R. (1980). Social class, social learning, and wife abuse. Social Service Review, 54, 390-406. Richie, B. (1987). Battered black women: A challenge for the black community. The Black Scholar, 16 (2), 40-44. Riger, S., & Krieglstein, M. (2000). The impact of welfare reform on mens violence against women. American Journal of Community Psychology, 28 (5), 631-647. Robert, S., & House, J.S. (1996). SES differentials in health by age and alternative indicators of SES. Journal of Aging & Health, 8, 359-388. Rocha, C.J. (1994). Effects of Poverty on Working Families: Comparing Models of Stress by Race and Class. Doctoral dissertation, Washington University, St. Louis. Roche, W.M., & Stegman, M.A. (1994). The effects of home ownership on the self-esteem, perceived control, and life satisfaction of low-income people. Journal of the American Planning Association, 60, 173-184. Schuler, S.R., & Hashemi, S.M. (1994). Credit programs, womens empowerment, and contraceptive use in rural Bangladesh. Studies in Family Planning, 25, 65-76. Schuler, S.R., Hashemi, S.M., Riley, A. P., & Akhter, S. (1996). Credit programs, patriarchy, and mens violence against women in rural Bangadesh. Social Science and Medicine, 43 (12), 1729-1742. Schwartz, M.D. (1990). Work status, resource equality, injury, and wife battery: The national crime survey data. Free Inquiry in Creative Sociology, 18 (1), 57-61. Shepard, M., & Pence, E. (1988). The effects of battering on the employment status of women. Affilia, 3 (2), 55-61. Sherraden, M. (1991). Assets and the Poor: A New American Welfare Policy. Armonk, NY: M.E. Sharpe. Smith, M.D. (1990). Sociodemographic risk factors in wife abuse: Results from a survey of Toronto women. Canadian Journal of Sociology, 15 (1), 39-58. Stets, J.E., & Straus, M.A. (1989). The marriage license as a hitting license: A comparison of assaults in dating, cohabiting, and married couples. Journal of Family Violence, 4 (2), 161-180. Stout, K. (1992). Intimate femicide: An ecological analysis. Journal of Sociology & Social Welfare, 19 (3), 29-50.

30

JOURNAL OF EVIDENCE-BASED SOCIAL WORK

Straus, M.A. (1994). State-to-state differences in social inequality and social bonds in relation to assaults on wives in the United States. Journal of Comparative Family Studies, 25 (1), 7-24. Straus, M.A. (1990a). Measuring intrafamily conflict and violence: The Conflict Tactic (CT) Scales. In M.A. Straus & R.J. Gelles (Eds.), Physical Violence in American Families (pp. 29-47). New Brunswick, NJ: Transaction Publishers. Straus, M.A. (1990c). Ordinary violence, child abuse, and wife beating: What do they have in common? In M.A. Straus & R.J. Gelles (Eds.), Physical Violence in American Families (pp. 403-424). New Brunswick, NJ: Transaction Publishers. Straus, M.A., & Gelles, R.J. (1990). How violent are American families? Estimates from the National Family Violence Resurvey and other studies. In M.A. Straus & R.J. Gelles (Eds.), Physical Violence in American Families (pp. 95-112). New Brunswick, NJ: Transaction Publishers. Straus, M.A., Gelles, R.J., & Steinmetz, S.K. (1988). Behind Closed Doors: Violence in the American Family. Beverly Hills, CA: SAGE Publications. Strube, M.J., & Barbour, L.S. (1983). The decision to leave an abusive relationship: Economic dependence and psychological commitment. Journal of Marriage and the Family, 45, 785-793. Thompson, A.M. III. (1993). Rural emergency medical volunteers and their communities: A demographic comparison. Journal of Community Health, 18, 379-392. White, E.C. (1986). Life is a song worth singing: Ending Violence in the black family. In M.C. Burns (Ed.), The Speaking Profits Us: Violence in the Lives of Women of Color (pp. 11-13). Seattle, WA: Center for the Prevention of Sexual and Domestic Violence. Yllo, K.A., & Straus, M.A. (1990). Patriarchy and violence against wives: The impact of structural and normative factors. In M.A. Straus & R.J. Gelles (Eds.), Physical Violence in American Families (pp. 383-399). New Brunswick, NJ: Transaction Publishers. Yllo, K.A. (1993). Through a feminist lens: Gender, power, and violence. In R.J. Gelles & D.R. Loseke (Eds.), Current Controversies on Family Violence (pp. 47-62). Newbury Park, CA: SAGE Publications, Inc.

S-ar putea să vă placă și