VIOLENCE AGAINST WOMEN is a major social problem that requires well-informed, empirically grounded policy and practice responses. The Bureau of Justice Statistics of the U.S. Department of Justice (1998) indicates that approximately 1 million violent crimes are committed annually against persons by their current or former spouses or partners and women are victims in 8 of 10 of the cases. Slightly more than half of these women report having children under the age of 12 at home (Bureau of Justice Statistics, 1998).
Several contemporary researchers identify stereotypes and challenge early research on battered women (Gelles, 1997; Herman, 1997; Pagelow, 1997; Ross & Glisson, 1991; Stark & Flitcraft, 1996). These authors maintain that previous studies often mistakenly focused on the victim's deficits and pathology, described the woman's hysteria, emphasized myths and stereotypes (Pagelow, 1997), and blamed the victim for her abuser's violence (Ross & Glisson, 1991). These pathological perspectives viewed battered women as having a contradictory set of attributes including personalities that are masochistic, immature, inadequate, incomplete, and sexually perverse (Herman, 1997; Pagelow, 1997; Stark & Flitcraft, 1996). Today, while most of the psychological literature is free from blatant untruths (Herman, 1997), victim blaming is common (Ewing & Aubrey, 1987). Public opinion about battered women is saturated with victim-blaming beliefs and negative stereotypes. Ross and Glisson (1991) find that social workers, like others, have judgmental attitudes toward battered women. One of these misconceptions may involve a critique of the victim's ability to distinguish "right" from "wrong" about her relationship. A popular belief is that battered women chose to be in relationships with violent partners (Gelles, 1997) and therefore make morally immature decisions. Because violence between adults is considered culturally "wrong," a popular belief ensues: battered women who stay in violent relationships condone the aggression, make the "wrong" choices to stay in their relationships, and therefore must be morally underdeveloped. Genes (1997) writes, "quite a few people believe that battered wives are somehow culpable, and their culpability is enforced by their decision not to leave. Nothing can be further from the truth" (p. 9). Although these beliefs are widespread, there is little empirical research to support these opinions.
Domestic violence research suggests that a woman's perceived choices are shaped by the degree to which she feels isolated, unsupported, physically weak, and confused by the trauma in her life (Johnson, 1992). Moreover, research suggests that the more controlled a woman is by her batterer, the more likely she is to feel stuck in the relationship (Johnson, 1992). Ultimately, these factors alter her perception of and ability to make a decision to leave an abusive relationship. Consequently, the public's judgment of a battered woman's moral development often surrounds the victim's decision to stay in her violent relationship. Research investigating public opinion about the stay/leave decisions of battered women indicates that many people believe that a battered woman can "simply leave" her abusive relationship and if she stays she is "emotionally disturbed" (Ewing & Aubrey, 1987, p. 261). While these results imply beliefs that battered women have choices, the truth may be that limited resources and a lack of viable alternatives to the abusive relationship may restrict their stay/leave choices.
In leaving an abusive relationship, a battered woman must balance her own and her children's safety and needs, including considerations of economics (homelessness and joblessness), community and family resources, and the danger of her abuser's retaliation (Gelles, 1997). Other studies investigated factors contributing to a battered woman's stay/leave decision and determined that a complex and often interdependent set of variables influences a woman's decision to stay or leave the violent relationship. These factors include social, cultural, religious, and economic factors and levels of support; physical violence; and fear (Baker, 1997; Benton, 1986; Choice & Lamke, 1997; Davis & Srinivasan, 1995; Gelles, 1976, 1997; Johnson, 1992; Strube & Barbour, 1983). Despite the fact that numerous research studies have investigated the stay/leave decision for battered women and have offered reasonable explanations, there is little empirical evidence and a notable lack of attention to the moral reasoning influencing a battered woman's stay/leave judgment. And, because most battered women stay in their relationships for at least some period of time, some conclude that battered women, in general, are morally immature individuals.
Research exploring myths held about battered women has provided substantial evidence to discredit many untruths (Pagelow, 1997). With this aim, this study investigates the moral judgments of battered and non-battered women with regard to their stay/ leave decisions and discusses implications for social work education. The purpose of this study was to (a) investigate differences in moral reasoning between battered women and non-battered women and (b) to investigate differences in moral reasoning between women who had left their intimate partner relationships and women who had not. Interaction effects were also evaluated.
Moral behavior is defined as making decisions about "right" and "wrong" in the context of socially acceptable standards (Guralnik, 1982). Generally accepted standards imply that members of society regard some actions and decisions as "right" or "more right" than other decisions. The converse is also true; generally unacceptable standards suggest that certain actions and decisions are viewed as "wrong." Moral reasoning is the supposition or ability to draw good and right inferences or conclusions from known or assumed facts with regard to others (Elliott, 1997).
Given the high prevalence of domestic violence, it is likely that social workers will encounter battered women who need their assistance. Uneducated social workers, as with the general public, may hold victim-blaming stereotypes about battered women as truth. Except for the occasional elective course on domestic or family violence, most social work students receive little education or skill-based training to competently respond to women facing intimate partner violence. Investigating the moral development of battered women may prove helpful in the development of fully informed practice and policy interventions to be taught in social work education programs.
This study employed a nonequivalent comparison group design (Rubin & Babbie, 1997) of battered and non-battered women. A 2 x 2 factorial design was used: history of domestic violence (with two levels--battered and non-battered women) and status of the relationship (with two levels--women who had left their intimate relationships and women who had not left). All the women who participated in this study indicated that they previously experienced or were currently experiencing relationship discord. Relationship discord was defined as a conflict in which they had considered or were considering ending their intimate partner relationships.
Participants were assigned to groups based on their responses to two questions. First, women were asked to self-report on their histories of domestic violence in their current or recent past intimate partner relationships. The question required a dichotomous "yes" or "no" response. The question asked women if, in their intimate partner relationships, they had ever experienced physical abuse that is "pushing, shoving, hitting, twisting, punching, kicking, grabbing, choking, and the use of a weapon" (Pence, 1987, p. 8), sexual abuse or rape, or the threat of physical violence, including the "jeopardy of harm by using loud voice, gestures, smashing things, and destroying property" (Pence, 1987, p. 8). Participants were also asked whether they were currently living with, intimately involved with, or separated or divorced from their intimate partners. Battered women were asked to specifically report if they were currently involved with their violent partners and non-battered women were asked to report if they were currently involved in their relationships with discord. Intimate partner was defined as a person whom the study participant is or had been married to, cohabititating with, or dating.
Given the difficulty of reaching non-sheltered battered women, this study used convenience and snowball sampling procedures (Rubin & Babbie, 1997). Recruitment targeted community support groups including support groups for battered women, parents without partners, and other self-help groups for women. At the end of the group meetings, group facilitators were asked to announce the study and distribute information to interested women. No data were collected on women declining participation in this study, and therefore any differences between women choosing to participate and those choosing to decline is unknown. Because of the high level of crisis for battered women residing in safe homes, they were excluded from recruitment procedures. Data were collected from women who live in nine counties in South Carolina, North Carolina, and Georgia. Prior to completing the Defining Issues Test each woman was given a $15 Wal-Mart gift certificate. Data were collected from December 1999 through May 2000 by the researcher or a trained research assistant. Participation in the study was completely voluntary and informed consent was obtained. The committee for the protection of human subjects at the University of South Carolina approved this project.
Ninety-two women participated in this study. Approximately equal proportions of battered and non-battered women were achieved. The summary in Table 1 indicates that slightly more than half (n=49, 53%) of the study participants were battered women and slightly fewer (n=43, 47%) were non-battered women. Among the battered women, it was difficult to achieve an equal distribution of women who had left and not left their intimate relationships. Of the battered women participants, 73% (n=36) had left their relationship while 27% (n=13) indicated that they were still involved with their intimate partners. However, among non-battered women participants, an approximately equal proportion of women who had left (n=21) and not left (n=22) their intimate relationships was achieved (49% and 51%, respectively).
The typical study respondents self-identified as being of lower socioeconomic status and reported a strong Protestant faith. Half of the women (n=46) reported having children at home; the other half did not report having children living at home. Overall, there was variability in the educational level of the participants. However, the non-battered women participants were approximately twice as likely to have completed 4 years of college or graduate school (n=31) than the battered women participants (n=16). The sample was composed mostly of Black, non-Hispanic (n=28) and White, non-Hispanic women (n=60; 31% and 66%, respectively). Hispanic women were underrepresented (n=3) and Asian and Native American women were not represented. The average age of the battered women participants was slightly higher than the average age of the non-battered women participants (42 years and 36 years, respectively). Sixty-four percent (n=59) of the study participants reported earning less than $40,000 per year. Battered women (n=11) were four times more likely than non-battered women (n=3) to have an annual income of less than $10,000 (12% and 3%, respectively). Table 2 provides additional information about the characteristics of the sample.
Defining Issues Test (DIT)
The Defining Issues Test (DIT) is the most widely used measure of moral development (Moon, 1986) and is considered a reliable and valid measure of moral reasoning (Rest & Narvaez, 1998; Thoma, 1986). Rest (1979) developed the DIT, a computer-scored instrument that quantifies the moral development of individuals, based on the earlier work of Lawrence Kohlberg. The DIT has been used in over 1,000 research studies involving thousands of participants in over 40 countries (Rest & Narvaez, 1994). The DIT generates a P score that indicates the relative importance the participant gives to stages 5 and 6 of principled moral reasoning (Rest, 1979, 1993). The P score is given as a percentage and ranges from 0 to 95 with higher scores indicating higher principled moral reasoning. The DIT also calculates an M or Meaningless score as an internal consistency check by determining if the participant followed the directions.
Participants are given six dilemma stories (including the well-known Heinz, his ill wife, and the unaffordable drug dilemma) and asked to determine what decision the character should make. Participants respond to three sets of questions for every dilemma. The first section requires the participant to decide if the character in the dilemma (a) should take action or (b) should not take action. Participants are also given option (c) cannot decide. The second set of questions contains 12 forced-choice questions about the dilemmas. The forced-choice responses are based on 5-point Likert-type ratings of "great," "much," "some," "little," and "no." The final section requires the participants to rank the four most important statements in the previous section. The DIT assumes an 8th-grade reading level and takes approximately 35-40 minutes to complete.
The DIT has an .80 test-retest reliability (over a period of several weeks) and an internal reliability (alpha coefficient) of .80 (Rest, 1979, 1986). The DIT has undergone extensive evaluation including determination of its relationship with age, education, and gender. Both age and education correlate highly with DIT scores (Rest, 1993; Rest & Narvarez, 1998). Of these variables, education has consistently been determined to be the more powerful predictor (Rest, 1994). The DIT has also undergone extensive validity and reliability testing with regard to gender. Women score slightly higher than men on level of moral judgment, and gender accounts for a very small percentage of the variance of DIT scores (Rest, 1994; Thoma, 1986).
Power estimates for main and interaction effects were calculated based on a sample size of 97 (Lauter, 1978). Power analysis indicated median effect sizes for main effects (p=.71) and interaction effects (p=.56).
Assumptions of the Analysis of Covariance (ANCOVA)
Before the inferential statistical data analysis was conducted, the data were evaluated to determine the degree to which the assumptions of analysis of covariance (ANCOVA) were met. All the assumptions were tested at the .01 level. T tests for skewness and kurtosis were performed to determine whether the sample was drawn from a normal distribution. It was concluded that the DIT scores were drawn from a normally distributed population (t kurtosis=-.058, t skewness=1.08, t critical value=[+ or -] 2.67, df=91). The null hypothesis of normality was retained. The assumption of homogeneity of variance was evaluated using Levene's test of equality of error variance. Results indicated that the data met the assumption of homogeneity of variance and equal variance was assumed (F=2.58, df=3, p=.059). The third assumption of ANCOVA is the linearity of regression slopes. Two analyses were performed. First, an examination of the scatterplots indicated that the slopes were parallel. A failure to reject the null hypothesis of parallel slopes indicated that the slopes were homogeneous. Second, a statistical test of homogeneity of regression slopes was performed to test the interactive term in a factorial ANCOVA in which the independent variables and the covariate education were evaluated (SPSS, 1988). The regression of the covariate education on the dependent variable DIT was similar for battered and non-battered women (F=2.210, p=.093). The ANCOVA assumption of homogeneity of regression slopes is met.
A total of 97 women completed the DIT. However, the scores of five participants were purged from the final statistical analysis after individual subject internal reliability checks, provided by the scoring service, detected inconsistencies or too many non-discriminating items. A 5% loss of participant data is well below the 15% average experienced nationally (Rest, 1993). Therefore, 92 participant DIT scores passed the internal consistency checks and these scores were used in the final analysis. Internal reliability analysis was performed (Cronbach's alpha=.68).
Overview of Research Questions
Research questions were nondirectional queries evaluated using an ANCOVA (p=.05). Both main and interaction effects were evaluated. The first research question focused on whether there were differences in main effects on the dependent variable principled moral reasoning--between battered women and non-battered women. Research question 1 asked, "Is the judgment of battered women different from non-battered women who have experienced relationship discord? If differences exist, what are they?" The second research question focused on whether there were differences in main effects on the dependent variable--principled moral judgment--between women who had left their intimate partner relationships and women who had not left. Research question 2 asked "Do battered and non-battered women who have experienced relationship discord and left their intimate partners differ in their moral development from battered and non-battered women who have experienced relationship discord and not left their intimate partners? If differences exist, what are they?"
Interaction effects were also evaluated with two research questions. The interaction effects addressed whether there were differences in the level of principled moral reasoning between battered women and non-battered women who had left their intimate partner relationships and battered women and non-battered women who had not left their intimate partner relationships. Research question 3 asked, "Do battered women who have left their abusive relationships differ in their principled moral development from non-battered women who have left their relationships? If differences exist, what are they?" Research question 4 asked, "Do battered women who have not left their relationships differ in their moral development from non-battered women who have not left their relationships? If differences do exist, what are they?"
Table 3 summarizes descriptive statistics of principled moral judgment by the two independent variables at both levels. Unadjusted and adjusted means are provided. For the sample of battered women, the adjusted mean scores are slightly higher than their unadjusted mean scores while the adjusted mean scores of non-battered women are slightly lower than their unadjusted mean scores. The changes in adjusted mean scores of both subgroups can by the explained the covariate-level of education. Non-battered women had higher educational levels than the sample of battered women.
The analysis indicates that overall the sample (M=33.15, SD=13.91) was comparable to national norms of senior high school (M=31.03, SD=13.90) and college students (M=43.19, SD=14.32) on principled moral judgment. The scoring service at the Study for Ethical Research at the University of Minnesota performed t tests to evaluate differences between the study's subsamples and norm groups. The subgroup of battered women was compared to norm test scores of high school students and the subgroup of non-battered women was compared to norm group scores of college students. These norm group scores were selected for comparison because of the reported overall education levels of the two subgroups in this study.
The t test for differences of scores of battered women and norm scores for high school seniors show that the scores were not significantly different (t=.12, df=305, p=.60). Similarly, statistical differences were not found for non-battered women and norm scores of college students (t=-1.52, df=305, p=.13). Therefore, one can assume that the study's subsamples scored comparably to national samples.
Analysis of Covarience Results
When a factorial analysis of variance was conducted, the main effect for history of domestic violence was significant (F=7.97, df=1, p=.01). The proportion of explained variance ([N.sup.2]=.08) was modest. Approximately 8% of the variance in the dependent variable--moral judgment--can be explained by the independent variable, history of domestic violence. Battered women scored lower on principled moral judgment (x=30.24, SD=11.91) than non-battered women (x=37.33, SD=15.31). Statistical significance of differences in mean scores of the DIT were not found for women who had left and not left their intimate partner relationships (F=2.25, df=1, p=.137) nor was the interaction significant (F=.001, df=1, p=.98).
When age and education were simultaneously utilized as covariates, main effects (battered and non-battered: F=3.80, df=1, p=.054; left and not left: F=2.77, df=1, p=.10) and interaction effects were not significant (F=.004, df=1, p=.95). The importance of these covariates was discussed above.
The two covariates age and education were also evaluated separately for their effect. When a factorial analysis of covariance was conducted with age as the covariate, the main effect for history of domestic violence was significant (F=11.07, df=1, p=.001). There were no other significant main or interaction effects. However, when education level was used as a covariate there was a different result. Main effects (battered and non-battered: F=2.18, df=1, p=.144; left and not left: F=3.15, df=1, p=.08) and interaction effects were not significant (F=.001, df=1, p=.98). When controlling separately for education, the statistical differences between battered (x=30.90, SD=11.91) and non-battered (x=35.33, SD=15.17) women on principled moral reasoning disappeared. The adjusted mean scores for battered women improved slightly while the adjusted mean scores for non-battered women declined. Table 4 summarizes the factorial analysis of covariance results.
Therefore, it appears that education level is the more influential covariate. Moreover, evidence about population differences in moral judgment between battered women and non-battered women appear to be more accurately associated with education level than "true" differences in moral reasoning. It is probable, in fact, that level of education is the variable explaining moral reasoning rather than a history of domestic violence.
There are several drawbacks to this study. This study employed a nonprobability, convenience sample of battered and non-battered women. It is clear that this sample is not representative of all battered women and therefore results may not be applicable to other groups of battered women. However, results may be generalized to populations similar to this sample. In addition, there may be other explanations for these findings.
Another limitation of this study relates to a somewhat narrow definition of battered women. This study defines the term of "battered woman" as one subjected to physical violence or the threat of physical harm; it excludes emotional, psychological, and economic abuse. Domestic violence experts often advocate against this narrow and myopic definition of intimate partner abuse. However, this definition was used because of the difficulty in operationalizing emotional, psychological, and economic abuses. There are two limitations of this study in regard to the use of the DIT. First, it excluded women with low reading levels because of the required 8th-grade reading level. Second, it may not be the best measurement to evaluate the moral judgment of women. Feminist scholars and morality researchers may argue that women's moral decision-making processes cannot be pigeonholed into a single quantified score. Rather, moral decision making by women is best understood within their personal context. While this may be true, the DIT has been highly scrutinized for gender bias and Thoma (1986) suggests that it is free from gender bias.
Based on the results of the analysis reported above, the following responses to the research questions are made. Question 1: When education level is controlled, there are no overall differences in the level of moral development between battered and non-battered women. Question 2: When education level is controlled, there are no differences in the level of moral development between women who have left their intimate partner relationships and women who have not left their intimate partner relationships. There were no significant main effects or interactions. Therefore, it must be concluded that, after controlling for the effects of education, battered and non-battered women did not differ in their level of moral judgment. In response to research questions 3 and 4, when compared separately, no differences occurred between those who had left their intimate partners and those who had not. When drawing conclusions from these findings it is important to consider the limitations mentioned previously.
Implications for Social Work Education
Social work intervention often seeks to lessen suffering, discomfort, oppression, or discrimination either directly, through interventions and personal service programs, or indirectly, through policies and research. Woman battering is a significant social problem. The risk of intimate violence for women and children in this country and the consequential psychological, emotional, and physical effects substantiate and demonstrate the need for social work involvement. Moreover, these risks legitimize training and skill-based education for social work students in the area of intimate partner violence. Social workers must be prepared to intervene and counsel battered women and advocate on their behalf in larger societal systems. Although many social work education programs provide an elective course on domestic violence for interested students, content on domestic violence may also be readily infused in the four primary areas of the social work curriculum. This study provides information relevant to these four theoretical domains: human behavior in the social environment (HBSE), social work practice, social policy, and social science research.
Infusion of domestic violence content into the HBSE curriculum might require students to develop a comprehensive framework for understanding battered women within their social environments. This framework may include attention to relevant theories that seek to explain the development and behavior of battered women within a human biological-psychological-social context, systems that promote and deter battered women's maintenance or attainment of optimal health and well-being; the impact of discrimination, economic deprivation, and oppression on battered women; and the values and ethics related to understanding this population. As this present study illustrates, students may be asked to challenge traditional developmental theories (such as theories on moral development), societal myths and stereotypes, and social constructions that do not fully or adequately explain this vulnerable group. Assignments should provide the opportunity for students to critically examine human behavior models to determine shortcomings, biases, and inaccuracies. At the outset, an untrained observer may believe that the obvious "right" decision for a battered woman is to leave her violent relationship, and this observer may conclude that when she does not leave her violent relationship she is immoral, making the "wrong" decision. However, a critical examination of decision making, using an ecological model, may lead the observer to later conclude that the "right" decision should be individually determined and is, at best, unclear.
As this current study suggests, practitioners need a clear understanding of the complexity of a battered woman's situation to be able to offer a thoughtful intervention approach. Competent practice can be achieved only if students are made aware of emerging theories and related practices. This study provides information for increased competency with battered women. For example, an understanding of the moral reasoning of battered women will improve our practice with women in regard to assisting them in their stay/leave decisions. Teaching social work students about the challenges and barriers battered women face will increase their practice competency. In practice courses, social work students may survey the benefits and limitations of various practice approaches or faculty may choose to teach about empowerment perspectives as an appropriate intervention for battered women (Busch & Valentine, 2000). It is necessary for social work students to learn how to work with battered women in a process that provides empowering choices for battered women that are consistent with what we have learned is good and "morally consistent" social work practice. Furthermore, Rhodes (1985) suggests that it is useful for social workers to examine moral development models within the context of ethical dilemmas that might occur in our direct practice with clients.
In the area of policy, assignments should be developed so that students analyze laws, regulations, standards, and politically driven initiatives that affect battered women and their children. In the context of domestic violence, social work students could explore determinants of private issues versus public problems. Policy, program, and service development and the evaluation of each may also be a class requirement. For example, appropriate policy responses might include the development of programs that provide battered women with resources to support their choices in their relationships.
This study challenges existing attitudes, such as victim blaming, which can result in policies and programs that hold battered women accountable for their circumstances. One example of this practice takes place in child protective agencies when children are removed from their non-offending mothers for failure to protect them against their batterers' violence. These policy actions have developed in part because battered women are viewed as culpable for the violence in their homes (Gelles, 1997; Tilden, 1989). These processes often result in increased danger for battered women and their children and an intensified fear of retribution from their abusers. Programs and services should be reevaluated for victim-blaming nuances and redesigned so that there is recognition of leaving as a process that requires safety planning and support. Therefore, it is critical for students to challenge values embedded in policies and programs that perpetuate myths and stereotypes of vulnerable populations.
Advocates for battered women suggest that early and perhaps current domestic violence research reflects our negative cultural values and assumptions about this vulnerable population (Ferraro, 1996). Social work research courses could explore the ways in which battered women have been studied, the appropriateness of these methodologies in understanding this vulnerable population, and the value of the knowledge added by more fitting research studies. As a course requirement, students may develop a research proposal that contributes to our knowledge base on domestic violence. This current research suggests queries such as, "How do we help protect battered women?" and "How can we empower battered women?" These questions presuppose that support, resources, and alternatives are essential components for ensuring battered women's safety. Corollary inquiries such as "Why is he violent?" and "How do we prevent and detain a batterer's violent and harmful behavior?" are also more useful questions in the pursuit to end intrafamily violence than the proverbial victim-blaming questions once examined. Promoting domestic violence as a research agenda priority benefits the community by underscoring issues that affect the lives of millions of women and children.
Battered women victims are subjected to intimidation and threats, emotional, verbal, physical, and sexual abuse and economic deprivation by their intimate partners. Victims repeatedly face decisions of morality, particularly during the acute battering phase of their cycle of violence. As a result, many abused women experience continuous and extreme cognitive dissonance. Their struggles stem from competing and polarized thoughts about their sobering situation. Most battered women admit that the violence is "wrong" (Busch, 2000), but many battered women, like other women, are committed to their intimate partnerships, children, and marriage vows and hope that rehabilitative counseling is a viable option. The "right" decision is often not clear. Moreover, most battered women need additional resources in order to escape the violence and realize that leaving does not necessarily mean an end to the violence.
This study argues that domestic violence is a serious and complex social problem that requires competent and empirically supported policy and practice responses by social workers. It also makes recommendations for the inclusion of domestic violence content across the curriculum.
TABLE 1. Number of Women Participants by History of Domestic Violence and Status of the Relationship (N=92) History of Domestic Violence Subgroup Total Sample Status of Relationship % n % n Battered (n=49) Left relationship 73 36 39 36 Not left relationship 27 13 14 13 Non-battered (n=43) Left relationship 49 21 23 21 Not left relationship 51 22 24 22 TABLE 2. Demographic Characteristics of Participants (N=92) Battered Non-battered (n=49) (n=43) % n % n Race Black, non-Hispanic 13 12 18 16 White, non-Hispanic 40 36 26 24 Hispanic -- -- 3 3 Income Less than $10,000 12 11 3 3 $10,001-$25,000 17 16 12 11 $25,001-$40,000 9 8 11 10 Greater than $40,000 14 13 19 17 Education High school and less 21 19 8 7 Completed 2 years college/ vocational school 15 14 5 5 Completed 4 years college 11 10 20 18 Completed graduate school 6 6 14 13 Number of children at home None 24 22 26 24 One 12 11 11 10 Two 15 14 4 4 Three or more 2 2 5 5 Religion Protestant 38 35 31 28 Roman Catholic 3 3 7 6 No preference 1 1 4 4 Other 10 9 5 5 Self-rating of faith Strong 35 32 34 31 Somewhat strong 15 14 10 9 Not very strong 1 1 1 1 Not applicable 1 1 2 2 Note. For the variables race, religion, and faith, N=91; for the variable income, N=89. TABLE 3. Mean and Standard Deviation Scores for P Values on the Defining Issue Test by History of Domestic Violence and Status of Relationship History of Domestic Violence Status of n Unadjusted SD Adjusted SD Relationship M M Battered Left 36 31.78 12.67 33.35 12.67 Not left 13 25.97 8.54 28.45 8.54 Non-battered Left 21 38.97 14.47 37.84 14.47 Not left 22 35.76 16.00 32.81 16.00 TABLE 4. Analysis of Covariance of Level of Moral Development by History of Domestic Violence and Status of Relationship, Controlling for Education Source SS df MS F p Education 2251.35 1 2251.35 14.2 -- History 345.41 1 345.41 2.18 .144 Status 498.66 1 498.66 3.15 .080 History by Status 8.72E-02 1 158.54 0.001 .981 Error 13793.01 87 158.54 Total 121203.31 91
The author thanks the Lady Sunshine Foundation, the Twigg Foundation, and the South Carolina Coalition Against Domestic Violence and Sexual Assault for their support of this research.
Baker, P. L. (1997). And I went back: Battered women's negotiation of choice. Journal of Contemporary Ethnography, 26(1), 55-74.
Benton, D. (1986). Battered women: Why do they stay? Health Care for Women International, 7, 403-411.
Bureau of Justice Statistics of the U.S. Department of Justice. (1998, March). Retrieved May 10, 2003, from http://www.ojp.usdoj.gov/bjs
Busch, N. B., & Valentine, D. (2000). Empowerment practice: A focus on battered women. Affilia, 15(1), 82-95.
Busch, N. B. (2000). Battered women's moral reasoning: Conceptions and considerations of "right" and "wrong." (Doctoral dissertation, University of South Carolina, 2000). Dissertation Abstracts International, 62(2), 772.
Choice, P., & Lamke, L. K. (1997). A conceptual approach to understanding abused women's stay/leave decision. Journal of Family Issues, 18(3), 290-314.
Davis, L., & Srinivasan, M. (1995). Listening to the voices of battered women: What helps them escape the violence. Affilia, 10(1), 49-69.
Elliott, D. (1997). Universal values and moral development theories. In C. Christians & M. Traber (Eds.), Communication ethics and universal values (pp. 68-83). Thousand Oaks, CA: Sage.
Ewing, C., & Aubrey, M. (1987). Battered women and public opinion: Some realities about myths. Journal of Family Violence, 2, 257-264.
Ferraro, K. J. (1996). The dance of dependency: A genealogy of domestic violence discourse. Hypatia, 11(4), 76-91.
Gelles, R. (1976). Abused wives: Why do they stay? Journal of Marriage and the Family, 38, 659-668.
Gelles, R. J. (1997). Intimate violence in families (3rd ed.). Thousand Oaks, CA: Sage.
Guralnik, D. B. (Ed.). (1982). Webster's new world dictionary (2nd college ed.). New York: Simon and Schuster.
Herman, J. (1997). Trauma and recovery: The aftermath of violence--From domestic abuse to political terror. New York: Basic Books.
Johnson, I. (1992). Economical, situational, and psychological correlates of the decision-making process of battered women. Families in Society, 73(1-4), 168-176.
Lauter, J. (1978). Sample size requirement for the T2 test of MANOVA (Tables for one-way classification). Biometrical Journal, 20, 389-406.
Moon, Y. (1986). A review of the cross-cultural studies on moral judgment development using the Defining Issues Test. Behavioral Sciences Research, 1-4, 147-177.
Pagelow, M. (1997). Battered women: A historical research review and some common myths. Journal on Aggression, Maltreatment, & Trauma, 1, 97-116.
Pence, E. (1987). In our best interest: A process for personal and social change. Duluth, MN: Minnesota Program Development, Inc.
Rest, J. (1979). Development in judging moral issues. Minneapolis, MN: University of Minnesota Press.
Rest, J. (1986). Moral development: Advances in research and theory. New York: Praeger.
Rest, J. (1993). Guide for Defining Issue Test: How to use the optical scan form and the Center's scoring service. Minneapolis, MN: University of Minnesota.
Rest, J. (1994). Background: Theory and research. In J. Rest & D. Narvaez (Eds.), Moral development in the professions: Psychology and applied ethics (pp. 1-27). Hillsdale, NJ: Lawrence Erlbaum Associates.
Rest, J., & D. Narvaez (1998). Supplement to guide for DIT-1. Minneapolis, MN: University of Minnesota.
Rest, J., & D. Narvaez (1994). Moral development in t he professions. Hillsdale, NJ: Lawrence Erlbaum.
Rhodes, M. (1985). Gilligan's theory of moral development applied to social work. Social Work, 2, 101-105.
Ross, M., & Glisson, C. (1991). Bias in social work intervention with battered women. Journal of Social Science Research, 14(3/4), 79-104.
Rubin, A., & Babbie, E. (1997). Research methods for social work (3rd ed.). Pacific Grove, CA: Brooks/Cole.
SPSS-X user's guide (3rd ed.). (1988). Chicago: SPSS, Inc.
Stark, E., & Flitcraft, A. (1996). Women at risk: Domestic violence and women's health. Thousand Oaks, CA: Sage.
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.
Tilden, V. (1989). Response of the health care delivery system to battered women. Issues in Mental Health Nursing, 10, 309-320.
Thoma, S. (1986). Estimating gender differences in the comprehensive and preference of moral issues. Developmental Review, 6, 165-180.
Noel Bridget Busch
University of Texas at Austin
Noel Bridget Busch is assistant professor, School of Social Work, University of Texas at Austin.
Address correspondence to Noel Bridget Busch, School of Social Work, University of Texas at Austin, 1925 San Jacinto Boulevard, San Antonio, TX 78712-1203; email: email@example.com.