|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001709492
007 cr mnu|||uuuuu
008 060508s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001365
Branch, Kathryn A.
Exploring the role of social support in heterosexual women's use and receipt of non-lethal intimate partner violence
h [electronic resource] /
by Kathryn A. Branch.
[Tampa, Fla.] :
b University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 98 pages.
ABSTRACT: The concept of social support has been found to be a protective factor in women's intimate partner violence victimization. However, little is known about the relationship between women's social support and their intimate partner violence perpetration. Research evidence demonstrates that women's perpetration of violence is surprisingly frequent, particularly in women younger than age 30. This study investigated the role of social support in heterosexual women's use and receipt of non-lethal aggression against an intimate partner among 673 female college students. The implications of these findings for research and practice are discussed.
Adviser: Christine S. Sellers, PhD.
t USF Electronic Theses and Dissertations.
Exploring the Role of Social Suppor t in Heterosexual Womens Use and Receipt of Non-Lethal Intimate Partner Violence by Kathryn A. Branch A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Criminology College of Arts and Sciences University of South Florida Major Professor: Christine S. Sellers, Ph.D. John K. Cochran, Ph.D. Kathleen M. Heide, Ph.D. Martha L. Coulter, Dr.P.H. Date of Approval: October 31, 2005 Keywords: womens aggression, college st udents, victimization, perpetration Copyright 2005, Kathryn A. Branch
Acknowledgements I would like to express my appreciation to the many people who have contributed to the preparation of this dissert ation. First, I would like to thank my Major Professor, Dr. Christine S. Sellers, for her support and encouragement. I sincerely appreciate the time you have dedicated to me both professionally and personally. I would also like to thank Dr. John Cochran, Dr. Kathleen Heide, and Dr. Marti Coulter for their invaluable input. In addition, I would like to thank Chris Gibson for his instruction on negative binomial regression. Finally, I would like to express my deepest gratitude to my family and friends. Words cannot express how much I have appreciated their love and encouragement.
Dedication This dissertation is dedicated to my son Zachary.
iv Table of Contents List of T ables ........................................................................................................vi Abstract ...............................................................................................................vii Chapter One: Introducti on....................................................................................1 Chapter Two: Womens use of aggre ssion against an inti mate partner................4 History of the study of in timate partner violen ce........................................5 Gender differences in heterosexual non-lethal intimate partner viol ence..................................................................................................10 Differences in ty pe of viol ence......................................................10 Differences in moti ve and cont ext.................................................10 Differences in consequenc e..........................................................12 Feminine gender role...............................................................................17 Chapter Three: So cial Suppor t...........................................................................20 Distinctions in so cial suppor t....................................................................20 Kinds of support ............................................................................20 Perceived vs. re ceived suppor t.....................................................21 Levels of soci al suppor t.................................................................21 Sources of soci al suppor t..............................................................21 Research on soci al suppor t.....................................................................22 Research on the role of social support in womens IPV victim ization ..........................................................................22 Research on the role of social support in womens IPV perpetra tion..........................................................................24 The present study....................................................................................26 Chapter Four: Methods.......................................................................................28 Procedur e................................................................................................28 Dependent vari ables................................................................................ 31 IPV victimiz ation............................................................................31 IPV perpetra tion............................................................................32 Independent vari ables.............................................................................. 33 Family Social Support ...................................................................33 Friends Social Support.................................................................34 Control variables: Common correlates of IPV victimization and perpet ration.....................................................................................35 History of IPV vi ctimizatio n............................................................35
v History of IPV perpetrati on............................................................36 Family history of IPV.....................................................................37 Alcohol us e....................................................................................38 Marijuana us e................................................................................38 Beliefs about men and wom en in Amer ica....................................39 Race/ethni city................................................................................40 Analytic St rategy......................................................................................44 Regression models..................................................................................46 Chapter Five: Results .........................................................................................47 Correlational Analyses .............................................................................47 IPV victimiz ation............................................................................47 IPV perpetra tion............................................................................50 Negative Binomial Regr ession Anal yses.................................................51 IPV victimiz ation............................................................................51 IPV perpetra tion............................................................................56 Summary of Results ................................................................................59 Chapter Six: Discussion ......................................................................................60 Referenc es.........................................................................................................68 Appendices .........................................................................................................80 Appendix A: Results of factor analysis for measures of IPV victim ization ....................................................................................81 Appendix B: Results of Factor Analysis for measures of IPV perpetra tion.....................................................................................83 Appendix C: Results of Factor Analysis for measures of family soci al suppor t..............................................................................85 Appendix D: Results of Factor Analysis for measures of friends soci al suppor t............................................................................87 Appendix E: Results of Factor Analysis for measures of beliefs about men and wo men in Amer ica.............................................89 About the Aut hor......................................................................................E nd Page
vi List of Tables Table 1: Description of Variabl es .......................................................................41 Table 2: Intercorrelations Between Study Va riables ...........................................48 Table 3: Negative Binomial Estimation fo r Model of IPV Vi ctimization...............52 Table 4: Negative Binomial Estimation for Model of IPV Perpetration................57
vii Exploring the Role of Social Suppor t in Heterosexual Womens Use and Receipt of Non-lethal Intimate Partner Violence Kathryn A. Branch ABSTRACT The concept of social support has been found to be a protective factor in womens intimate partner violence victimiz ation. However, little is known about the relationship between womens social support and their intimate partner violence perpetration. Research ev idence demonstrates that womens perpetration of violence is surprisingl y frequent, particularly in women younger than age 30. This study investigated the role of social support in heterosexual womens use and receipt of non-lethal aggression against an intimate partner among 673 female college students. The im plications of these findings for research and practice are discussed.
1 Chapter One Introduction In recent years a considerable body of literature has focused attention on the concept of social support. Social support has been broadly defined as information that prompts a person to be lieve she/he is cared for, loved, esteemed, and valued and is a member of a network of common and mutual obligation (El-Bassel, Gilbert, Rajah, Fol eno, & Frye, 2001, p. 247). Overall, this research shows that "social support is a valuable social commodity and those who are endowed with social support are better off in most instances than those who are not" (Hobfoll & Vaux, 1993, p. 685) Social support has been found to have a major positive effect on psychological and physical health. A significant body of research has ex amined the protective role social support plays in womens victimization by an intimate partner. Social support has been found to have a positive influence on abused womens ability to emotionally adapt to their situation or to make the decision to leave the abusive relationship (Larance, 2004). In addition, greater social support has been found to be associated with a significantly reduced ri sk of a range of mental health outcomes (e.g., depression, anxiety, PTSD) among abused women (Coker, 2003). This literature has focused primarily on clin ical samples of abused and/or drug addicted women (e.g., El Bassel, Gilbert, Rajah, Foleno, & Frye, 2004; Kocot & Goodman, 2003; Farris & Feenaghty, 2002).
2 Comparatively little is known about the impact of social support on womens use of aggression against an intimate partner. Within the past decade, the concept of social support has been applied to understanding crime perpetration and deviance (Cullen, 1994; Cu llen & Wright, 1997). Cullen (1994) suggests that social support can have a deterrent effect on motivation for crime and deviance and, therefore, research shoul d focus on the preventative effects of social support. In general, the study of womens use of aggression against an intimate partner is widely debated. Some res earchers have argued that research on female aggression may be used to blame women for instigating their own abuse, and that a focus on female aggression will dr aw attention away from mens far more lethal aggression (White & Kowalski 1994). These researchers assert that women do not initiate violence, but ra ther use it in self-defense. Proponents of studying womens use of aggression argue a different perspective. They acknowledge that male violence within the home causes or has the potential to cause the most physi cal harm; however, they propose that it is not the whole story. They argue that a failure to consider intimate partner violence in its entirety, namely by excluding female aggressors, will lead to violence that is either disr egarded or inadequately addressed. The intention of the current research is to explore the role of social support in both intimate partner violence victim ization and offending among women. This research will explore two main questions. First, does social support reduce the likelihood of womens victimization by an in timate partner? Second, does social
3 support reduce the likelihood of womens use of aggression against an intimate partner? Chapter two will examine the resear ch on womens use of aggression against an intimate partner. This chapter will discuss the debate that is currently going on regarding studying womens use of aggression and discuss why continued research in this area is nec essary. Chapter three will examine the concept of social support. This chapter will explore the multidimensional nature of the concept of social support and discuss which types of social support appear to be most important to individuals. Th is chapter will also discuss the research on the role of social support in womens victimization by an intimate partner. I will review and critique the previous res earch that has been conducted in this area and suggest that social support may also have a protective effect on perpetration. Chapter four will detail the methods used to investigate the role of social support in womens use and receipt of intimate partner violence. Chapter five will present the results of the curr ent research. Finally, chapter six will discuss the findings and implications for future research in this area.
4 Chapter Two Womens use of aggression against an intimate partner One of the most pervasive and undis puted gender stereotypes is that men are more aggressive than women. Wh ite and Kowalski (1994) describe this assertion as the myth of the nonaggressi ve woman. Assumptions of male aggression and female victimization are so taken for granted that they have influenced where researchers have looked and how researchers have decided what is known to be true about intimate partner violence. Because of the notion that aggression is a predominantly ma le attribute, researchers have disproportionately used male as opposed to female participants in their research studies on use of aggression against an intimate partner (White & Kowalski, 1994). Although gender stereotypes dictat e that anger and aggression are predominantly male domains, research does not support this claim. Research evidence suggests that womens perpetra tion of violence in the context of intimate relationships is surprisingly frequent, particularly in women younger than age 30 (Straus, 1993; Salari & Baldwin, 2002; Katz, Kuffel, & Coblentz, 2002; Makepeace, 1986; Archer, 2000; Underwood, 2003). These findings raise the question as to whether male violence against women should be the primary and/or exclusive focus of empirical inve stigation in intimate partner violence research.
5 I will begin by discussing the history of the study of intimate partner violence. Much of the current thought on intimate partner violence has to do with how it was constructed as a social pr oblem. I will then discuss the research on gender differences in intimate partner violence and conclude with a discussion of gender roles and how gender roles and socialization may influence men and womens use of aggression against an intimate partner. History of the Study of Intimate Partner Violence Intimate partner violence was once an unspoken crime. Violence against a family member (e.g., spouse, child) was considered to be a socially acceptable use of aggression. It was viewed as an e ssentially private, family matter not within the parameters of legal concern. Police frequently ignored family violence calls or purposefully delayed responding for hours. The first radical alteration of this paradigm came about in the early 1970s, through the work of Second Wave feminists. During the 1970s, the women s liberation movement began to criticize and bring attention to all types of abuse against women. Because they were concentrating on the problems of wom en--transforming what were once considered personal issues into political issues--they exposed the female victims of domestic assault. Terms such as battered wives and wife abuse were developed to name the violence that wo men were experiencing from their husbands or male partners (Frieze, 2000). This new terminology provided many abused women with a way to identify, rec ognize, and express their experiences. In addition, these terms assisted in establishing domestic violence as an
6 identifiable social problem and provided society with a way to talk about these issues. Research from the feminist perspective began with a narrow focus on the issue of wife beating, developi ng a literature that focus ed on factors specific to violence perpetrated against women by their male partners (Johnson, 1995). Theoretically, the emphasis was upon histor ical traditions of the patriarchal family, contemporary constructions of ma sculinity and femininity, and structural constraints that make escape difficult for women who are systematically beaten (e.g., Dobash & Dobash, 1977). The patriarc hal structure of society was seen as encouraging and legitimizing mens violence towards their wives, through an ideological and a legal framework. Feminist researchers gleaned their understanding of intimate partner violence from a wide variety of evidence, including clinical observations, narrative accounts of victims and batterers, the ex perience of advocates, and qualitative data taken from police and medical sources. This evidence supported the notion that domestic violence was a pattern perpetrated by men and was rooted deeply in the patriarchal traditions of the Wester n family. Clinical cases were highlighted to capture public attention and also serv ed to solidify the public perception that domestic violence was a euphemism for physical violence perpetrated by males against their female spouse. Soon after, research began to reveal that physical violence in the home actually claimed victims of both sexes. Researchers who were interested in resolution of conflict within families began to find that the victims of marital
7 violence were not always women and that women could also display physical violence toward their male partners (e.g., Straus, 1979; 1993; Gelles, 1985; Gelles & Straus, 1988). In the mid-1970s Straus and colleagues reported that women self-reported physically assaulti ng partners in marital and cohabitating relationships as often as men self-report ed physically assaulting their partners. Violence between husbands and wives, which they called spouse abuse, was viewed as part of a pattern of violence occurring among all family members. This work was particularly associated with data obtained using the Conflict Resolutions Scale (now called the Conf lict Tactics Scale) developed by Straus (1979). Family conflict studies asked about a ll possible experiences of physical violence, including minor and severe forms and violence that does not result in injury, and placed ending physical violence at the center of their agenda. Violent behavior was viewed as the central problem to be addressed. Accordingly, physical violence by women was held to be equally as problematic as physical violence by men. Rather than using limited clinical sa mples that did not offer grounds for generalizability, family violence resear ch relied primarily on representative samples of the general population to produc e estimates of prevalence. Findings suggested that rates of physical vi olence by men and women appeared to be equal. This research shifted the focus from studying only men to studying both men and women as perpetrators.
8 Reports of gender symmetry in violent assaults ran counter to what feminist researchers and the general pub lic thought they knew to be true of domestic violence. These conclusions di d not fit with their fundamental analysis of wife assault--that it was an extens ion of male political, economic, and ideological dominance over women. The response by many activists was to doubt the conclusions of family violenc e studies, criticize the study, and/or threaten the investigator.1 Female initiation and perpetration of violence was considered to be an anomaly. Accepting this anomaly as commonplace necessitated the reconstruction of prio r theory and the re-evaluation of prior facts. Many feminist scholars argued that wo men were too passive to perpetrate abusive acts against their spouses. Others suggested that men, because of their typically larger physiques, were not capable of being abused by their wives. Still others proposed that women were less c apable than men of inflicting serious harm or injury on a man and that, ther efore, physical violence by a woman against her spouse was more socially acceptable (White & Kowalski, 1994). Many feminist scholars argued that m easurement tools of family violence researchers did not explore the context of the violence (e.g., Dobash & Dobash, 1977; 2000). Feminists appeared less concerned with who was more aggressive, women or men, and more focu sed on the outcome of aggression. 1 For example, after Suzanne Steinmetz proposed the battered husband syndrome in an article published in 1978 in Victimology a speech she was asked by the ACLU to give was canceled because the organization received a bomb threat (Pearson, 1997).
9 The debate has continued to rage for the past thirty years. The discrepancies between claims of gender symmetry and claims of drastic gender asymmetry have led to significant confusion among policy makers and the general public. As a result of the cont radictory findings produced by disparate definitions and methods, increased efforts have been made by both feminist and family violence researchers to explor e the differences between men and women in the types, motives, and the psychological and physical consequences of the violence perpetration. The following review of available research will focus on studies that have investigated men and women's violence to wards their heterosexual intimate partners. Research suggests that the contexts and dynamics in same-sex and heterosexual relationships are different enough to warrant separate discussions. Thus, the review does not include the growing body of findings on intimate partner violence in same-sex couples (see Renzetti, 1992; Burke & Follingstad, 1999; Elliot, 1996; Lie & Gentlewarrier, 1991). Furthermore, this summary will concentrate on non-lethal violence in heterosexual relationships. In a lethal altercation between partners, men are predominantly the offenders and women are much more likely to be the victims (Browne, 1987; Serran & Firestone, 2004). Nevertheless, women are capable of violence and do occasionally kill their intimate partners. The majority of literature regarding womens use of lethal viol ence over the past 15 years has been concerned with women in abusive relati onships who kill their abusers (see Walker, 1979; Browne, 1987). This resear ch suggests that women generally do
10 not kill, but when they do, it is often in their own defense (Walker, 1979; Browne, 1987). Gender Differences in Heterosexual Non-Lethal Intimate Partner Violence Difference in type of violence An examination of available data provides many examples of gender differences in types of non-lethal aggression used against intimate partners. DeKeser edy, Saunders, Schwartz, and Alvi (1997) found that many of the Canadian female respondents in their survey reported using violence against their heterosexual dating partners. Only a small percentage, however, reported violence that was likely to cause serious injuries, such as beating up or using a weapon. In Makepeaces (1986) student sample, although women reported perpetrating as much psychological and physical violence as men, women repor ted being forced to have sex (24%) at much higher rates than men (3%). Swan and Snow (2002) found in their sample of women who had used aggression against an intimate partner in the past six months that the abusive behaviors that women commit are different from mens abuse. Women committed significantly mo re acts of moderate violence (e.g., throwing things and threatening to hit) agai nst their partners than their partners committed against them. The womens par tners, however, committed almost one and a half as many acts of severe physical violence against them as vice versa (e.g., choking). These results suggest that men and women use different types of aggression against an intimate partner. Differences in motive and context. Although studies have begun to pay some attention to the contexts and mo tivations of womens and mens violent
11 behavior, they tend to focus on single or very limited explanatory conditions. Specifically, studies of mens violent behavior towards intimate partners have focused on control as a primary motivation (Dobash & Dobash, 1977; Dobash, Dobash, Cavanuagh & Lewis, 1998). In contrast, studies of womens violent behavior toward intimate partners have focused on self-defense as a primary motivation (DeKeseredy & Schwartz, 1998). A review of the research indicates that neither of these provides a comp letely accurate accounting of physical violence against an intimate partner. In studies of general aggression use against another individual, qualitative and quantitative work (Campbell & Muncer 1987; Campbell, Muncer & Coyle, 1992; Campbell, Muncer & Gorman, 1993) has suggested that men (more than women) represented their aggression as an instrumental act aimed at taking control over others, whereas wom en (more than men) represented their aggression as an expressive act resulting from a temporary loss of control. Women spoke of feeling overwhelmed by arousal and anger, losing their selfcontrol, and subsequently feeling guilt y and ashamed of their behavior. Men described their aggression as an attempt to take control over a threatening or anarchic situation, emphasizing moral rectitude and subsequent mastery. This research did not specify the target of aggression as an opposite sex intimate partner. Interestingly, follow-up studies spec ifying the target of aggression as an opposite sex intimate partner have found no indication of an association between expressive beliefs and physical aggression in women but a positive association in
12 men (Archer & Graham-Kevan, 2003). For wo men, there was some indication of a positive association between instrument al beliefs and physically aggressive acts, although this correlation was weaker than for men. These findings appear inconsistent with a strictly control motivation for men and a strictly selfdefense motivation for women. Many researchers studying womens violent behavior toward intimate partners have asserted that womens main motivation is self-defense. DeKeseredy and Schwartz (1998) report t hat the majority of women in their college sample who used physical aggression toward their dating partners never initiated violence; the common motive fo r violence was self-defense. Follingstad and colleagues (1991), however, found that college men were more likely than women to report using physical violence in retaliation for being hit first. Harned (2001) found that male and female college students were equally likely to use physical violence for self-defensive pur poses. Women reported using physical violence due to anger or jealousy more o ften than men. A number of other studies point to a variety of reasons for womens assaultive behaviors that range from retaliating or punishing from past hurt, to gaining emotional attention, expressing anger, and reacting to frustrat ion as well as stress (Hamberger & Potente, 1996; Follingstad, Wright, & S ebastian, 1991; Fiebert & Gonzalez, 1997; Straus, 1999; Dasgupta, 1999; Dasgupta, 2002; Miller & White, 2003). Taken individually, the majority of these reasons would not generally meet the standards of legal or social approval as they are not executed in self-defense.
13 Differences in Consequence. Despite the fact that both men and women report using physical aggression against an intimate partner, women are more likely to sustain serious injury than ar e men. Past research has demonstrated greater negative consequences of partner violence for women relative to men (Foshee, 1996; Makepeace, 1986; Katz, Kuffel, & Coblentz, 2002). Far more men than women kill their spouses (Kimme ll, 2002). Women, on average, suffer much more frequent and more severe injury (physical, economic, and psychological) than men do (Kimmell, 2002; Dobash, Dobash, Cavanaugh, & Lewis, 1998). For the most part, legislators, poli cy makers, legal and social service professionals, and community advocates have dealt with the issue of intimate partner violence as primarily mens violence against women. Clearly, the evidence demonstrates that women are also using aggression against their intimate partners. These findings suggest that male violence against women should not be the exclusive focus of empi rical investigation on intimate partner violence. Although gender stereotypes dictate that the expression of anger and aggression are predominantly male domains, research does not support this claim (Underwood, 2003; Archer, 2000). Numerous studies have found that women are initiating aggression in intimate relationships. When women show instances of masculine forms of aggression involving direct physical confrontation, however, these are s een as pathological or due to hormonal imbalance, or their actions are unrepor ted, or seen as insignificant.
14 Implicit views about womens nat ure have influenced the way that research findings have been interpreted. There appears to be a strong desire to avoid seeing women as willful aggressors or recognizing female aggressive behavior as instrumental and intelligent (N affine, 1987). For example, Macaulay (1985) identified seven beliefs associat ed with aggression in women: women are nonaggressive, sneaky in their expressi on of aggression, unable to express anger, prone to outbursts of fury, psyc hologically distressed if they are aggressive, aggressive in defense of their children, and motivated to aggress by jealousy. Womens acts of aggression ar e thought to be the result, not of their own willful agency, but the result of hormones or abuse (Pearson, 1997). When some scholars concede the possibility of female aggression and violence, they hasten to add that women engage only in expressive aggression. Women do not, these scholars maintain, engage in instrumental aggression, the kind that is calculating. Women are constructed as victims rather than as actors in the violence they perpetrate against an intimate partner. In conceptualizing a battered woman, society has constructed her as a passive and helpless victim, who is too paralyzed by the abuse to take any actions on her own behalf. In conceptualizing a batterer, society has constructed him as a controlling and domineering pers on, who is instrumental in his aggression to achieve ultimate control of a womans life. Neither of these conceptualizations is correct as the protot ype. They fit very well with traditional beliefs about men and women. Careful analysis of research, however, suggests that these beliefs need to be re-evaluated. Historically, these conceptualizations
15 were useful to bring to light the devas tating impact of intimate partner violence and to make society aware that this problem needed and demanded attention. These conceptualizations fit well with soci etys stereotypes of men and women. For policy purposes, these ideas were easier to sell to the general public. Research has demonstrated that men and women are both using aggression within intimate partner relationships. Research findings suggest that womens violence differs from that perpetrated by men in terms of type of aggression used, motivation, and the consequences of violence. These findings thus make it impossible to interpret the violence of men and women as interchangeable. Womens use of aggression must be understood in and of itse lf, not simply in counterpoint to mens actions. The fact that violence by men has more serious physical consequences should not cause us to ignore violence by women as a topic worthy of research. To deny the fact t hat women too are violent or to hold that violence by women is unimportant or ev en justified does a grave disservice not only to the research enterprise, but ultima tely to women as well. By denying the possibility of female agency . theorists ar e with the best of intentions, actually denying women the full freedom to be human (Morrissey, 2003, pg. 102). Use of violence by women must not only be recognized but also acknowledged as a legitimate area of investigation. The majority of studies that have in vestigated womens use of aggression have examined womens aggression sec ondary to and/or in comparison with mens use of aggression. Further there has been a tendency to apply
16 explanations for male offending to fema les. While much has been written and theorized about male-on-female intimate partner violence, less is understood about female-on-male intimate partner vi olence. Generalizing male results to females implies a false sense of equality in the use of violence and leaves gaps in knowledge as to how this problem affects women specifically. In addition, focusing on differences between wom en and men without addressing overall context makes the implicit assumption that women and men operate in similar social contexts or that soci al context is irrelevant. It is well documented that there are di fferent societal expectations and social contexts for men and women in rela tion to behavior. Certain expectations and roles are assigned to men, while other s are assigned to women. Society has behavioral expectations that men are unemotional, self-focused, active, and aggressive. Society expects women to be passive, submissive, and unassertive. Male-on-female violence may be underst ood within the context of societys expectations of what men do. Men are expected to be dominant and aggressive; therefore, aggression in men is not surprising. This explanation does not work for females. The cultural norms of womens violence are quite the opposite. Cultural prescriptions for gender roles generally prohibit women from engaging in aggressive actions targeting their male partners (Dasgupta, 1999). Females are not socialized to be dominant and aggressive; conversely, females are socialized to be community-orient ed and passive. Nevertheless, some women are using aggression against their in timate partners. Researchers must resist the temptation to approach female intimate partner violence as the
17 adoption of masculinity. Womens actions must be understood in and of themselves within the context of the feminine gender role. Feminine Gender Role A gender role describes an individual or socially prescribed set of behaviors and responsibilities. In essence, a gender role comprises all the things that people do to express their individual gender identities. Gender roles are not biologically determined; they are socially constructed. The traditional feminine gender role prescribes that women ar e dependent, emotional, sexually passive, and responsible for providing the emotional support and nurturing to family members and the sick (Bem, 1983). Traditional roles for women tend to be relationship-oriented, where a womans sense of self becomes very much organized around being able to make and maintain affiliation and relationships (Shumaker & Hill, 1991). Despite womens lib and the focus on equal rights for men and women, societys expectations hav e not changed significantly over the last couple of generations. The process by which the individual is encouraged to adopt and develop certain gender roles is called socializati on. Socialization works by encouraging wanted and discouraging unwanted behavior. It is well documented that men and women experience differential socia lization (Deaux, 1984; Deaux & Major, 1987; Eagly, 1987). Research suggests that this differential socialization begins at the moment of birth (e.g., Deaux, 1984; Eagly, 1987; Bem, 1983; Bigler, 1997). Society has expectations (i.e., attitudes and beliefs) regarding appropriate male and female behaviors (Bem, 1983; Epstein, 1988). Individuals internalize
18 societal expectations and conform to gender role norms (Eagly, 1987). By age 4 or 5, most children have developed and inte rnalized gender stereotypic attitudes and beliefs (Bem, 1983; Bigler, 1997). This process has been found to continue throughout an individuals life, even in the absence of any social or institutional pressures (Eagly, 1987; Bem, 1983). The feminine gender role is associated with an expressive and communal orientation, a concern for the rela tionship between oneself and others (Bem, 1983). An exploration of the feminine gender role reveals the high salience of social support in the lives of women. Theory on gender role expectations would predict variations between males and females on the salience of social support. Research on gender differences and social support confirm this expectation, suggesting that men and women have differ ent support needs, elicit support in different ways, and that perceptions, c ontext, expectations, and the meaning of support are different for men and women (Weber, 1998). Research suggests that women receive and want more social support than men and are more likely to acknowledge the need for help or assistance, thereby explicitly fostering socially supportive relationships (Gilligan, 1982; Shumaker & Hill, 1991; Markward, McMilan, & Markward 2003). In addition, women are more likely to be informal supports than men and are also more likely to be formal supports (e.g., teachers, nurse s, social workers). Shumaker and Hill (1991) note that across the lif espan, women are more likely than men to be both support receivers and support givers.
19 The literature reveals that there are cultural reasons why women are main receivers and givers of support (Weber, 1998). Males are socialized to focus on autonomy, self-reliance, and independence and to de-emphasize the expression of feelings. This socialization process does not encourage the formation of social support networks for men. Females are so cialized to be verbally expressive and to focus on warmth and a search for intima cy; therefore, searching for social support in ones environment is a well-learned and highly valued pattern for women (Olson & Schultz, 1994). Searching for and having social support in ones environment has been linked extensively, both directly and indi rectly, to physical and mental health and well-being (Weber, 1998). For example, social support has been linked to enhanced immune function, improved copi ng with a medical condition, and reduced mortality (Weber, 1998). Therefore, it appears that searching for social support is a highly useful and highly benefic ial characteristic of the feminine gender role. The next chapter will focus mo re closely on the research on social support both in general and specifically with respect to crime and violence.
20 Chapter Three Social Support What is social support? Social support is a multidimensional construct that can involve both tangible and/or intangible aid. The broadness and complexity of the social support constr uct has required investigators to make several distinctions. Distinctions in Social Support Kinds of support. The first distinction that has received attention is the distinction among the kinds or types of support. Researchers have suggested that there are four main types of social support: emotional, instrumental, informational, and appraisal support (Weber, 1998). Emotional support involves the provision of empathy, love, trust, and caring. Instrumental support involves the provision of tangible aid and services that directly assist a person in need, such as babysitting, money, groceries, et c. Informational support involves the provision of advice, suggestions, and information that a person can use to address problems. The information that is given is not in and of itself helpful; instead, it helps people to help themselves. Appraisal support involves the provision of information that is useful for self-evaluation purposes such as constructive feedback, affirmation, and social comparison. The different kinds of social support appear to serve different f unctions to individuals. Of the four
21 forms, researchers have considered em otional support to be the primary component of social support (House, 1991; Helgeson, 1993). Perceived vs. received support. The second distinction that has been given attention is the distinction bet ween perceptions of support and actual receipt of support (Lakey & Cassady, 1990; Helgeson, 1993; Weber, 1998). In order for support to be helpful, it needs to be perceived as helpful. Therefore, support depends on the perceptions of the recipient. Weber (1998) describes perceived social support as the cognitive appraisal of being connected to others and knowing that support is available if needed. Two key dimensions of perceived social support are perceiv ed availability and perceived adequacy of supportive connections (Barerra, 1986). Received supports are the resources actually provided to the recipient. Received support has been assessed by direct observation or by asking people to indicate whether specific supportive acts have occurred. In studies that examine both perceived and received suppor t, the perception of support seems to be a better predictor of health outco mes than the actual receipt of support (Wethington & Kessler, 1986; Helgeson, 1993; Norris & Kaniasty, 1996). Levels of social support The third distinction that has been given attention is between various social levels of social support. Social support exists at several levels of society. It exists in the intermediate interactions within families and among friends and within lar ger social networks of neighborhoods, communities, and nations.
22 Sources of social support. The fourth distinction that has been given attention is the distinction among the di fferential sources of social support. Support can be delivered by either a formal agency or through informal relations. Formal support includes social support from schools, government assistance programs, and the criminal justice syst em. Informal support includes support provided through relationships with others who lack official status relative to the individual. The different sources of social s upport also appear to serve different functions to individuals. Numerous investi gations in the field of social psychology have shown that the main source of help and support when facing a problem is not formal organizations, but an individuals own informal networks (HernandezPlaza, Pozo, & Alonso, 2004). The advantages of informal social support are particularly relevant in populations with limited access to formal support. Research on social support Family ties, friendships, and involvement in social activities have been found to offer a psychological buffer against stress, anxiety, and depression (Weber, 1998). Cohen and Wills (1985) de scribe two mechanisms through which social support may work. The main effect occurs when there is a general increased level of well-being simply as a result of being part of a support network. The buffering hypothesis suggests that stress in a crisis is reduced due to the specific help that is perceived and/or provided. Social support has been examined extensively in the intimate par tner violence victim ization literature.
23 Research on the role of social suppor t in womens IPV victimization. Social support is important to the study of intimate partner violence because research suggests that less support increases womens risk of violent victimization by intimates (Feld & Str aus, 1990; Coker, 2003). Many victims of intimate partner violence indicate that they are not emotionally supported by familial and friendship ties and frequently reveal varying degrees of social isolation (Carlson, McNutt, Choi, & Rose, 2002; Larance & Porter, 2004; Dobash & Dobash, 1998). Social isolation can contribute to depression and undermine an individuals self-esteem and sense of purpose. Victims are usually secluded from supportive familial and friendship networks by their abusers (Dobash & Dobash, 1998; Coker, 2003; El-Bassel, Gilbert, Rajah, Foleno, & Frye, 2001). Research has suggested that suppor tive involvement with others can significantly reduce the risk of depre ssion and Post Traumatic Stress Disorder symptoms in abused women (Coker, 2003). In addition, having greater levels of social support has a positive influence on abused womens ability to make the decision to leave an abusive relationship (Larance, 2004). The majority of the research that has examined the effects of social support on womens experience of IPV vict imization has focused on clinical samples of abused women. For exampl e, El-Bassel et al. (2001) examined social support among women (average age of 37) in methadone treatment who had experienced partner violence. Ko cot and Goodman (2003) examined the role of social support as a moderator of the relationship between problemfocused coping and post-traumatic stress disorder and depression in low income
24 battered women. Farris and Feenaughty (2002) examined the associations between substance dependence, social isol ation, and womens experience of domestic violence in a sample of str eet recruited drug-using women (mean age 37.5). Larance and Porter (2004) examined t he process of forming social capital among female survivors of IPV. Carlson, McNutt, Choi, and Rose (2002) examined the role of social support and other protective factors in relation to depression, anxiety, and several different types of lifetime abuse in female patients (mean age of 31). Overall, this body of research suggests that social support potentially provides a buffer for abused women, pr otecting them from developing negative mental health outcomes (e.g., anxiety, depression). However, these study findings are not generalizable due to their limited focus on clinical samples of abused women. Comparatively little is known about the relationship between social support and intimate partner viol ence victimization among women in nonclinical samples. Research on the role of social support in womens IPV perpetration. Research within the past decade has begun to suggest that in addition to its buffering effect against victimization, so cial support may also have a role in preventing crime and deviance (Cullen, 1994; Cullen & Wright, 1997; Colvin, Cullen, & Vander Ven, 2002). Cullen (1994) proposes that supportive relations, beginning at birth, are essential to healthy human development. These supportive relations are in turn instrument al in the development of certain internal states such as empathy and self-contro l and create the context in which strong
25 social bonds can emerge. These inte rnal states have been found to protect against delinquent behavior, where indi viduals who have greater empathy and greater levels of self control appear to engage in fewer delinquent acts (Cullen, Wright, & Chamlin, 1999). The act of giving social support can also have a negative influence on involvement in cr ime (Cullen, 1994). Sampson and Laub (1993) found that as offenders became prov iders of emotional and instrumental support, their involvement in crime ceased. Research suggests that formation of interpersonal relationships is especially important to women (Block, 1983; Knox, Zusman, & Nieves, 1997). Female socialization encourages women toward interpersonal relationships as support receivers and support givers while males are socialized toward independence (Windle, 1992). It has been documented that social support can have a role in preventing womens victimiz ation by an intimate partner (Carlson, McNutt, Choi, & Rose, 2002; Larance & Porter, 2004; Dobash & Dobash, 1998). Can social support also have a role in preventing womens use of aggression against their intimate partner? Robbers (2004) examined whether quantit y of social support (e.g., How frequently do you have contact with your family?) moderated the relationship between strain and delinquency (e.g., stole money, hit teacher, carried a weapon, used marijuana, hit another student) in men and women ages 18 to 22. The results indicated that when females experienced certain strains (e.g., failure to achieve goals) but had high levels of social support, the likelihood of delinquency decreased. This result sugges ts that the development of social
26 support networks could play an important role in female crime. To date, there has been no research on the role of social support in womens use of aggression against an intimate partner. Cullen (1994) suggests that a caring or supportive orientation towards others facilitates connectedness and makes victimizing others incompatible. Womens traditional re sponsibility for the delivery of social support and nurturance to others may cr eate sentiments and problem-solving skills that are generally incompatible with engaging in violent and/or criminal behavior (Katz, 1988). The Present Study The present study explores the role of social support in womens IPV victimization and perpetration. The pr esent study will extend the empirical literature on the relationship between social support and intimate partner violence among women in two main ways. First, this study will be conducted with an ethnically diverse, college-based sample of women with a range of use and receipt of intimate partner violence. This sample allows for greater generalizability than would be those from research with clinically referred females. Much of the current research on the role of social support in womens IPV victimization has relied on clinical samples of abused women (e.g., Carlson, McNutt, Chot & Rose, 2002; El-Bassel, Gilbert, Rajah, & Frye, 2001; Farris & Fenaughty, 2002; Kocot & Goodman, 2003). Second, this study will examine the effects of social support on womens use of aggression against an intimate part ner. Most prior studies of social support and intimate partner violence have assessed one group, women who
27 have experienced violence. No study has been found in the extant published literature that examines t he role of social support in womens use of aggression against an intimate partner. It is expected that social s upport will be negatively associated with womens use and receipt of intimate part ner violence, even after controlling for correlates of intimate partner violence. Specifically, it is expected that women who report greater levels of social suppor t will be less likely to be victimized by their intimate partner and less likely to use physical aggression against their intimate partner.
28 Chapter Four Methods The data for this study were drawn from students who were included in a NIH-funded five-year longitudinal st udy (1990-1995) of victimization and perpetration among undergraduate college student s (White, Smith, & Humphrey, 2001). Both male and female students were assessed. It has been suggested that the use of student samples may have im plications for the study of intimate partner violence by neglecting intimate par tner violence in non-student intimate relationships (Archer, 2000). This is an im portant limitation that primarily affects studies aimed at investigating the prev alence and incidence of intimate partner violence. This limitation is of lesser c oncern to studies such as the present one that aim to test the relationships between theoretical concepts and intimate partner violence. Students are highly likely to be involved in intimate relationships and are also highly likely to be victims and perpetrators of intimate partner violence (Makepeace, 1986; Archer, 2000). Procedure Before the initial survey was adm inistered, the researchers gained permission through the university administrat ion to survey students the first day of student orientation. St udent orientation leaders were trained to administer the survey and made participation in the st udy an integral part of the student orientation activities. The student orient ation was not a requirement; therefore,
29 students who did not attend were cont acted by phone. The purpose and methods of the survey were explained, and signed consent was obtained. Surveys were administered along with cont act sheets for the purpose of followup. Surveys and corresponding contact sheets were assigned a study number to ensure confidentiality of the data. The researchers obt ained a federal Certificate of Confidentiality. Toward the end of each spring semester for four consecutive years students were contacted and asked to complete a follow-up survey during one of several sessions held at various locations around campus. Postcards were sent to remind students of the follow-up survey and to announce times and locations for sessions. These sessions were conducted by trained undergraduate psychology majors and graduate students. Students who did not attend one of the sessions were called and invited to attend. They were given the option of attending a session being held on campus, or of receiving the survey via mail. All students who participated in the follow-ups received $15 each time they participated. Students who had withdrawn from the university were also sampled. The survey was administered to two cohorts of male and female students. For the first cohort, Wave 1 of the survey was administered in Fall 1990. Waves 2, 3, 4, and 5 were administered at t he end of the Spring semester in 1991, 1992, 1993, and 1994 respectively. Likewise, for the second cohort, Wave 1 was administered in Fall 1991 with subsequent surveys administered at the end of Spring semesters in 1992, 1993, 1994, and 1995. Responses of the two cohorts
30 were aggregated at each wave of data colle ction; thus, Wave 1 consists of respondents in cohort 1 surveyed in Fall 1990 and respondents in cohort 2 surveyed in Fall 1991, and so forth. Surveys at each wave of data collection contained some identical items across a ll waves but also included items that differed from one wave to another. Be cause the items needed to measure the variables often differed from one year to the next as well as by gender of the respondent, the study was limited to heterosexual female respondents who participated in all five Waves of the survey administration.2 There were a total of 1,538 females in the original sample, of which 1,422 were heterosexual. Of these heterosex ual women, 673 (47.3%) had completed all five waves of the survey. Of these 673 female students, 76.1% were white with a mean age of 23 years old at Wave 5 (SD = .76). There were no significant 2 Social support was measured differently for males and females in the original sample. Males were assessed the quantity of social support from family and friends (e.g., number of hours spent with friends, number of times saw friends); females were assessed the quality of social support from family and friends (e.g., I can rely on my friends; I feel a strong bond with my friends). It is possible for an individual to know many people, spend time with those people, and not feel like he or she is valued by or can rely on those people. Therefore, for purposes of this research, only the female data were used. Furthermore, a complete measure of social support was av ailable only in Wave 5. Because of this, the variables that were expec ted to have contemporaneous effects with social support were taken from the data collected at Wave 5. These variables include IPV victimization and IPV per petration (as dependent variables) and alcohol and drug use as common correlates of the dependent variables. Other common correlates that are or could be time variant were taken as closely as possible to but preceding Wave 5. These correlates include history of IPV victimization (taken at Waves 1-4), hist ory of IPV perpetration (taken at Waves 14), and beliefs about men and women (taken at Wave 2). Correlates that are not time-variant were taken from Wave 1. These measures include race/ethnicity and history of family violence.
31 differences on any of the study variabl es between participants who completed all five waves and those who did not. Dependent Variables IPV victimization. Drawing from the physica l aggression items of the Conflict Tactics Scale (Straus, 1979), re spondents were asked to indicate how many times during the past year their romantic partner had (1) thrown or smashed something (but not at the res pondent); (2) threatened to hit or throw something; (3) thrown something at the respondent; (4) pushed, grabbed, or shoved the respondent; (5) hit (or tried to hit) the respondent but not with anything; and (6) hit (or tried to hit) the respondent with something hard. A romantic partner was defined as a per son whom the student was dating. Responses were initially coded from 1 to 5 (1 = never, 2 = 1 time, 3 = 2-5 times, 4 = 6-10 times, 5 = more than 10 times). The original metric for this variabl e used unnecessarily restricted ordinal response categories to measure the number of times violence was experienced. Ordinal variables can neither be added toget her to create a meaningful scale nor can they be analyzed with statistical techniques such as regression-based analyses. As a result, the six items were converted into interval-like responses by recoding ordinal values to reflect approximate counts. Never was recoded as 0 rather than 1. One time was recoded as 1 rather than 2. Two to five times was coded as 3, the midpoint of t he counts, rather than the ordinal value of 3. Six to ten times was recoded as 8, again as the midpoint of the counts, rather than the ordinal value of 4. Finally 10 plus times was arbitrarily given an
32 upper bound of 20; hence, 10 plus times was recoded as 15, the midpoint of 10 to 20, rather than 5. Under this transfo rmation of the data, the sum of the items is an approximate count of victimization experienced wit hin the past year by a romantic partner. This transformation allows for the use of regression-based techniques. An additive IPV Victimization scale was then constructed summing each of the respondents answers across the 6 item s. A principal components factor analysis of these items indicated a singlefactor solution (eigenvalue = 2.95). A Cronbachs alpha of .751 was found for the si x-item additive scale (see Appendix A). Scores on the IPV Victimization scale range from 0 to 39 with a mean of 1.43. Consistent with previous findings in the literature (Archer, 2000), 23.8% of the sample reported experiencing physical aggression from their romantic partner at least once within the past year. The additive scale is a discrete variable that is naturally left-censored at zero, artificia lly right-censored at 90, highly positively skewed, and has a standard deviation that is greater than the mean (M = 1.43, SD = 4.95). IPV perpetration. Drawing from the physica l aggression items of the Conflict Tactics Scale (Straus, 1979), res pondents were also asked to indicate how many times during the past year they had (1) thrown or smashed something (but not at their partner); (2) threatened to hit or throw something; (3) thrown something at their partner; (4) pushed, grabbed, or shoved their partner; (5) hit (or tried to hit) their partner but not with any thing; and (6) hit (or tried to hit) their
33 partner with something hard. Responses we re originally coded from 1 to 5 (1 = never, 2 = 1 time, 3 = 2-5 times, 4 = 6-10 times, 5 = more than 10 times). As with IPV victimization, the original metric used unnecessarily restricted ordinal response categories to measur e the number of times violence was experienced. Therefore, the six items were converted into interval-like responses by recoding ordinal values to reflec t approximate counts using the same transformation scheme as that descri bed above for IPV victimization. An additive IPV Perpetration scale was then constructed summing each respondents answers across the 6 item s. A principal components factor analysis of these items indicated a singlefactor solution (eigenvalue = 3.22). A Cronbachs alpha of .797 was found for the six-item additive scale. Scores on the IPV Perpetration scale range from 0 to 60 with a mean of 1.54. Consistent with previous findings in the literatur e (Archer, 2000), 26.3% of the sample reported using physical aggression against t heir romantic partner at least once within the past year (see Appendix B). The additive scale is a discrete variable that is naturally left-censored at zero, artificially right-censored at 90, highly positively skewed, and has a standard deviati on that is greater than the mean (M = 1.54, SD = 5.19). Independent Variables Family social support, measuring perceptions of emotional social support from family, is an additive scale comprised of eight items. Family Social support was measured by asking respondents to indi cate the extent to which they agree or disagree (1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly
34 disagree) with the following st atements: (1) My family cares for me very much (reverse coded); (2) My family holds me in high esteem (reverse coded); (3) I am really admired by my family (reverse coded); (4) I am loved dearly by my family (reverse coded); (5) Members of my family rely on me (reverse coded); (6) I can't rely on my family for support; (7) My family really respects me (reverse coded); and (8) I don't feel close to members of my family. An additive Family Social Support scale was constructed summing each respondents answers across the 8 items. High values on the additive scale are indicative of higher levels of perceived family social support. A prin cipal components factor analysis of these items indicated a single-factor soluti on (eigenvalue = 4.684). A Cronbachs alpha of .89 was found for the eight-item additive scale (see Appendix C). Scores on the Family Social Support scale range fr om 8 to 32 with a mean of 27.80 and a standard deviation of 4.21. Friends social support, measuring perceptions of emotional social support from friends, is an additive scale compri sed of seven items. Friends Social Support was measured by asking respondents to indicate the extent to which they agree or disagree (1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly disagree) with the following stat ements: (1) My friends respect me (reverse coded); (2) I can rely on my friends (reverse coded); (3) My friends don't care about my welfare; (4) I feel a st rong bond with my friends (reverse coded); (5) My friends look out for me (revers e coded); (6) My friends and I are really important to each other (reverse coded); and (7) My friends and I have done a lot for one another (reverse coded). An additive Friends Social Support scale was
35 constructed summing each respondents answe rs across the seven items. High values on the additive scale are indicative of higher levels of perceived social support from friends. A principal co mponents factor analysis of these items indicated a single-factor solution (ei genvalue = 4.42). A Cronbachs alpha of .90 was found for the seven-item additive scale (see Appendix D). Scores on the Friends Social Support scale range from 8 to 28 with a mean of 23.51 and a standard deviation of 3.87. Control Variables: Common Correlates of IPV Victimization and Perpetration Although the primary interest of this st udy is in the effects of friends and family social support, other variables are re lated to intimate partner violence, and ignoring these factors might produce relationships between the dependent and independent variables that are spurious. Informed by previous research, correlates of intimate partner violence t hat will be controlled for in this study are history of IPV victimization, history of IPV perpetration, family history of IPV, alcohol and drug use, stereotypic beliefs about gender roles, and race. History of IPV victimization. This variable was a combined measure of victimization from Waves 1, 2, 3 and 4. In each individual wave respondents were asked to indicate how many times in the past year their romantic partner had (1) thrown or smashed something (but not at the respondent); (2) threatened to hit or throw something; (3) thrown something at the respondent; (4) pushed, grabbed, or shoved the respondent; (5) hit (or tried to hit) the respondent but not with anything; and (6) hit (or tried to hi t) the respondent with something hard.
36 Responses were coded from 1 to 5 (1 = never, 2 = 1 time, 3 = 2-5 times, 4 = 6-10 times, 5 = more than 10 times). As previously noted, the original metric used unnecessarily restricted ordinal response categories to measur e the number of times violence was experienced. Once again the six items were converted into interval-like responses by recoding ordinal values to reflect approximate counts. An additive victimization scale was constructed for each individual wave summing each respondents answers across the 6 items. These individual wave additive victimization scales were then combined to create a History of IPV Victimization scale for Waves 1 through 4. Scores on the History of IPV Victimization scale range from 0 to 227 wit h a mean of 10.37. About 66% of the sample reported experiencing at least one form of physical aggression from a romantic partner within the past four waves. The scale was highly skewed (skewness = 4.67, kurtosis = 29.54). To reduce skewness and approach normality, .5 was added to each score and t he natural logarithm was taken of the scale scores. The .5 was added because the procedure would otherwise eliminate all cases in which the pretrans formation count = 0. Research suggests that previous victimization by an intima te partner may lead to a higher probability of subsequent victimization (Dobash & Dobash, 1979). History of IPV perpetration. This variable was a combined measure of perpetration from Waves 1, 2, 3 and 4. Respondents were asked to indicate how many times in the past year they had (1) thrown or smashed something (but not at their partner); (2) threatened to hit or throw something; (3) thrown something at
37 their partner; (4) pushed, grabbed, or shoved t heir partner; (5) hit (or tried to hit) their partner but not with anything; and (6) hit (or tried to hit) their partner with something hard. Responses were coded from 1 to 5 (1 = never, 2 = 1 time, 3 = 2-5 times, 4 = 6-10 times, 5 = more than 10 times). Once again the original metric used unnecessarily restricted ordinal response categories to measure the number of times violence was experienced. As a result, the six items were converted into interval-like responses by recoding ordinal values to reflect approximate counts. An additive perpetration scale was constructed for each individual wave summing each respondents answers across the six items. These individual wave additive perpetration scales were then combined to create a History of IPV Perpetration scale for Waves 1 through 4. Scores on the History of IPV Perpetration scale range from 0 to 252 wit h a mean of 11.01. About 68% of the sample reported using at least one form of physical aggression against a romantic partner within the past four waves. The scale was highly skewed (skewness = 4.44, kurtosis = 27.96). To reduce skewness and approach normality, .5 was added to each score and the natural logarithm was taken of the scale scores. The .5 was added because the procedure would otherwise eliminate all cases in which the pretransformation count = 0. Family history of IPV Respondents were asked to report about the period of time when they were growing up (ages 8 to 14): For an average month, indicate how often one of your parents or stepparents delivered physical blows to the other (Wave 1). Participants were asked to respond on a scale from 1 to 5
38 (1 = never, 2 = 1 to 5 times, 3 = 6 to 10 times, 4 = 11-20 times, and 5 = more than 20 times). Because of the relatively low frequency of respondents indicating that parents or stepparents had delivered blows to each other, scores on this variable were collapsed to create a dichotomous variable. Those that reported no violence were coded as 0 (91.4%) and those that reported violence were coded as 1 (8.6%). Coming from a violent home has been suggested to be a strong predictor of later IPV. Resear chers have consistently found that men exposed to marital violence are substantially more likely to be violent toward their spouse than are men not exposed to par ental violence (Carr & VanDeusen, 2002; Hotaling & Sugarman, 1986). Fema les exposed to parental aggression, however, have been found to be somewhat more likely to become victims (Doumas, Margolin, & John, 1994). Alcohol Use was assessed by asking respondents how often they drank alcohol in the past year (Wave 5). Participants were asked to respond on a scale from 1 to 5 (1 = Never, 2 = less than once a month, 3 = one to three times a month, 4 = one to two times a week, 5 = more than two times a week). The data indicated that 14.9% of respondents had never drunk, 36.1% drank less than once a month, 28.8% drank one to three times a month, 15.9% drank one to two times a week, and 4.3% drank more than two times a week. Substance use, especially alcohol, is cited frequently as a major correlate of intimate partner violence (Hotaling & Sugarman, 1986). Marijuana Use was assessed by asking respondents how often they had used marijuana within the past year (Wave 5). Participants were asked to
39 respond on a scale from 1 to 5 (1 = Never, 2 = less than once a month, 3 = one to three times a month, 4 = one to two times a week, 5 = more than two times a week). The data indicated that 78. 8% of female respondents had never used marijuana in the past year, 14.4% used it less than once a month, 3.1% used it one to three times a month, 1.5% used it one to two times a week, and 1.9% used it more than two times a week. Less than 8% of the sample reported past year use of a drug other than marijuana; therefore, other drug use was not included in the analyses. Beliefs about men and women in America This attitudinal variable was measured at Wave 2 by asking respondents to indicate the extent to which they agree or disagree (1 = strongly agree to 5 = strongly disagree) with the following statements: (1) Women are generally more sensitive to the needs of others than men are; (2) Women should take the passive role in courtship; (3) Men are more competitive than women; (4) Men are more sure of what they can do than women are; (5) Women tend to subordinate thei r own needs to the needs of others; (6) Men are more independent than women; (7) Women are more helpful than men; (8) Compared to men, women tend to be gullible; and (9) Compared to men, women are more able to devote themselves completely to others. An additive Beliefs scale was constructed summing each respondents answer across the nine items. A principal components fact or analysis of these items indicated a single-factor solution (eigenvalue = 3. 11). A Cronbachs alpha of .76 was found for the nine-item additive scale (see Appendix E). Scores on the Beliefs scale range from 10 to 45 with a mean of 28. 3 and a standard deviation of 5.18.
40 Higher scores on the Beliefs scale indicate more contemporary views of men and women in America. It has been suggested that the more contemporary a college womans attitudes on female sex-roles t he less likely she is to tolerate dating violence (Bookwala, Frieze, Smith, & Ryan, 1992). Research also suggests that mens negative beliefs regarding gender have a direct effect on their use of violence in their intimate relationshi ps (Reitzel-Jaffe & Wolfe, 2001). Race/Ethnicity. This variable is included as a statistical control variable. Race/ethnicity is measured as a dummy variable with whites as the reference category (0 = White and 1 = NonWhite). Some studies have suggested that minorities are more likely to be involved as victims and perpetrators of intimate partner violence (Barnett, Miller-Perrin, & Perrin, 1997). Table 1 provides a brief description of all variables included in these analyses.
41 Table 1 Description of Variables Variable Mean/Percent SD 1. IPV victimization (additive scale; Wave 5) 1.43 4.95 He threw something or smashed something (but not at me). .37 1.52 He threatened to hit or throw something. .31 1.56 He threw something at me. .09 .50 He pushed, grabbed, or shoved me. .45 1.79 He hit (or tried to hit) me but not with anything. .20 .97 He hit (or tried to hit) me with something hard. .02 .20 2. IPV perpetration (additive scale; Wave 5) 1.54 5.19 I threw something or smashed something (but not at him). .25 1.03 I threatened to hit or throw something. .27 1.27 I threw something at him. .13 .93 I pushed, grabbed, or shoved him. .45 1.63 I hit (or tried to hit) him but not with anything. .41 1.67 I hit (or tried to hit) him with something hard. .03 .26 3. Family Social Support (additive scale; Wave 5) 27.80 4.21 My family cares for me very much. (reverse coded) 3.79 .50 My family holds me in high esteem. (reverse coded) 3.49 .69 I am really admired by my family. (reverse coded) 3.32 .70 I am loved dearly by my family. (reverse coded) 3.67 .61 Members of my family rely on me. (reverse coded) 3.28 .72 I cant rely on my family for support. 3.49 .86 My family really respects me. (reverse coded) 3.40 .68 I dont feel close to members of my family. 3.38 .88 4. Friends Social Support (additive scale; Wave 5) 23.51 3.87 My friends respect me. (reverse coded) 3.54 .56 I can rely on my friends. (reverse coded) 3.31 .72 My friends dont care about my welfare. 3.57 .65 I feel a strong bond with my friends. (reverse coded) 3.33 .75 My friends look out for me. (reverse coded) 3.29 .68 My friends and I are really important to each other. (reverse coded) 3.39 .68 My friends and I have done a lot for one another. (reverse coded) 3.23 .82
42 Table 1 continued. Variable Mean/Percent SD 5. History of IPV victimization (addi tive scale; Wave 1-4) 10.37 22.62 He threw something or smashed something (but not at me). He threatened to hit or throw something. He threw something at me. He pushed, grabbed, or shoved me. He hit (or tried to hit) me but not with anything. He hit (or tried to hit) me with something hard. 6. History of IPV perpetration (additi ve scale; Wave 1-4) 11.01 23.57 I threw something or smashed something (but not at him). I threatened to hit or throw something. I threw something at him. I pushed, grabbed, or shoved him. I hit (or tried to hit) him but not with anything. I hit (or tried to hit) him with something hard. 7. Family History of Intimate Partner Violence (Wave 1) (0 = no history, 1 = history) No history 91.4% History 8.6% 8. Alcohol use (Likert scale; Wave 5) Never 14.9% < 1/ month 36.1% 1-3/ month 28.9% 1-2/ week 15.8% > 2/ week 4.3% 9. Marijuana use (Likert scale; Wave 5) Never 79.0% < 1/ month 14.5% 1-3/ month 3.1% 1-2/ week 1.5% > 2/ week 1.9%
43 Table 1 continued. Variable Mean/Percent SD 10. Beliefs about Men and Women in American (additive scale of Likert items; Wave 2) 28.32 5.18 Women are generally more sensitive to the needs of others than men are. 2.05 .87 Women should take the passive role in courtship. 3.68 1.00 Men are more competitive t han women. 3.05 1.17 Men are more sure of what they can do than women are. 3.83 1.03 Women tend to subordinate their own needs to the needs of others. 2.73 .85 Men are more independent than women. 3.75 1.00 Women are more helpful than men. 3.16 .88 Compared to men, women tend to be gullible. 3.26 1.04 Compared to men, women are able to devote themselves completely to others. 2.80 1.00 11. Race/ethnicity (dummy variable; Wave 1) Nonwhite (1= Nonwhite) 23.6% White (omitted category) 76.4% Note. N = 673
44 Analytic Strategy The statistical method used to analyze data may affect the relationships observed. When the assumptions of the employed statistical model are met, the observed coefficients are usually reli able and efficient (Greene, 1993); however, when these assumptions are violated, the resulting estimates may not be meaningful. This can result in the misi dentification of non-exis tent relationships (Type I errors) or the failure to discove r true relationships (Type II errors). Conventional regression models such as Ordinary Least Squares (OLS) regression are inappropriate to model t he perpetration and victimization data in the present study for several reasons. First, the data are discrete approximate counts that are non-negative (i.e, truncated at zero). The use of OLS regression on these data could lead to inconsistent and biased parameter estimates (Long, 1997; Gardner, Mulvey, & Shaw, 1995). It is also likely that the linear regression model will produce negative predicted val ues that are meaningless. Second, the distribution of the vict imization and perpetration measures is highly positively skewed, with many obser vations in the data set having a value of 0. This high number of 0s pr events the transformation of a skewed distribution into a normal one, violating OLS assumptions of normality. A skewed distribution can lead to heteroscedasticity, which can severely affect standard errors in OLS. Because for count dat a the residuals almost always correlate positively with the predictors, the esti mated standard errors of the regression coefficients are smaller than their true va lue, and thus the t-values associated with the regression coefficients are infl ated (Gardner, Mulvey, & Shaw, 1995).
45 This artificial inflation of the t-values may result in an appearance of statistical significance when, in fact, there is no stat istically significant effect. Thus, OLS regression seems prone to Type I errors for analysis of the victimization and perpetration data. For data where the dependent variabl e is a discrete non-negative count, Poisson models are a natural choice (Long, 1997). The Poisson model has a number of advantages over an OLS model, including a skew, discrete distribution, and the restriction of predicted values to non-negative numbers (Long, 1997). However, the Poisson model also has restrictive assumptions. First, the Poisson model assumes that the errors follow a Poisson, not normal, distribution. Second, the Poisson m odel assumes that the variance of the dependent variable equals its mean. Usually in practice, however, the variance of errors is larger than the mean, a condition known as overdispersion (Greene, 1993). Overdispersion causes the estimates of the standard errors to be lower than their true value, which again leads to inflated t-coefficients and potential Type I errors. Third, the Poisson regression model assumes that each occurrence is independent of the number of previous occurrences, and the expected number of occurrences is identic al for every member of the sample. Unless these assumptions are met, the Poisson model will produce incorrect estimates of its variance terms and misleading inferences about the regression. The data in the present study reveal significant variation among female respondents in IPV victimization (M = 1. 43, SD = 4.95) and perpetration (M = 1.54, SD = 5.19). In addition, it has been suggested that being victimized by an
46 intimate partner once may lead to a higher probability of a subsequent victimization (Dobash & Dobash, 1979). Both of these characteristics of the data violate assumptions of the Poisson regressi on model. As a result, it is important to consider an alternative regre ssion model for analyzing these data. The Negative Binomial regression model is a form of the Poisson regression that includes a random component reflecting the uncertainty about the true rates at which events occur for indi vidual cases (Gardner et al., 1995). This model adds an overdispersion parameter to estimate the possible deviation of the variance from that expected under Po isson (Long, 1997). The variation of this parameter can account for a vari ance that is higher than the mean. Regression Models Negative Binomial regression will be used to model womens victimization and perpetration of IPV as a f unction of social support from friends and family. The models will be run s eparately for each of the dependent variables. Model 1 will include the two social support (family and friends) scales to assess the main effects of social support (family and friends) on IPV victimization and perpetration. Model 2 will add the common correlates of IPV (history of IPV victimization, history of IPV perpetration, fam ily history of IPV, alcohol use, marijuana use, beliefs about men and women in America, and race/ethnicity) to determine the effects of social support (family and friends) when controlling for these variables. Model 2 will also include current IPV perpetration as a control in the IPV victimization models and current IPV victimization as a control in the IPV perpetration models.
47 Chapter Five Results Correlational Analyses Table 2 displays the results of Pearson product-moment correlations among dependent and independent variables. Re sults are presented separately for IPV victimization and IPV perpetration. Within these separate sections, the relationship between the independent vari ables and the measures of IPV victimization and perpetration will be exam ined first; then the relationships between the control variables and dependent variables will be examined. IPV victimization. A statistically significant relationship was found between family social support and IPV victimizati on (r = -.141 p < .01) indicating that increased perceptions of social support from family was negatively associated with IPV victimization. Those individuals who reported greater levels of perceived social support from family reported fewer IPV victimization experiences. There was also a statisti cally significant negative relationship between friends social support and IPV vict imization (r = -.123, p < .01), where those with greater perceptions of social support from friends were less likely to report IPV victimization experiences.
48 Table 2 Intercorrelations between Study Variables (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (1) IPV victimization .602** -.141** -.123 ** .299** .247** .021 .077* .075 .007 .049 (2) IPV perpetration -.155** -.196** 285** .334** .062 .010 .028 .015 .130** (3) Family social support .546** -. 115** -.074 -.092* .038 .031 .027 .001 (4) Friends social support -.152** -.123** -.077 .023 .084* .073 -.094* (5) History of IPV Victim. .754** .097* .141** .146** -.059 .014 (6) History of IPV Perp. .123** .077 .113** -.037 .088* (7) Family History of IPV .018 .047 -.006 .094* (8) Alcohol use .400** .093* -.260** (9) Marijuana use .076 -.087* (10) Beliefs -.068 (11) Race/Ethnicity Note: *p < .05, **p < .01.
49 There was a statistically significant relationship between history of IPV victimization and IPV victimization (r = 299, p < .01) such that females with a history of IPV victimizati on reported more current IPV victimization experiences. A statistically significant relationshi p was also found between history of IPV perpetration and IPV victimization (r = .247, p < .01), indicating that females with a history of using aggression against their intimate partners were more likely to report IPV victimization experienc es within the past year. No significant relationship was found between family history of IPV and IPV victimization. This finding is incons istent with previous research that has suggested that females exposed to parental aggression are more likely to become victims (Doumas, Margolin, & John, 1994). Frequency of alcohol use was significantly related to womens self reported IPV victimization ex periences (r = .077, p < .05). Females who reported drinking a greater number of alcoholic dr inks also reported a greater number of victimization experiences. This finding is consistent with past research on IPV victimization which indicates that subs tance use is a major correlate of IPV victimization (Hotaling & Sugarman, 1986). No significant relationship was found between marijuana use and I PV victimization. No significant relationship was found between beliefs about men and women in America and IPV vict imization. In addition, no significant relationship was found between race/ethnicity and IPV victimization. This finding is inconsistent with previous research that suggests that minorities are more likely
50 to be involved as victims of intimate partner violence (Barnett, Miller-Perrin, & Perrin, 1997). IPV Perpetration. A statistically significant relationship was found between family social support and IPV perpetration (r = -.155 p < .01) indicating that increased perceptions of social support fr om family were negatively associated with IPV perpetration. Those individuals who reported gr eater levels of perceived social support from family reported fewe r acts of IPV perpetration. There was also a statistically significant negative re lationship between friends social support and IPV perpetration (r = -.196, p < .01), w here those with greater perceptions of social support from friends report ed fewer acts of IPV perpetration. There was a statistically significant relationship between history of IPV victimization and IPV perpetration (r = .285, p < .01) such that females with a history of IPV victimizati on reported more use of aggression against their intimate partners. A statistically significant re lationship was also found between history of IPV perpetration and IPV perpetration (r = .334, p < .01) indicating that females with a history of using aggression against t heir intimate partners were more likely to report using aggression against an intima te partner within the past year than women without such histories. No significant relationship was found between family history of IPV and IPV perpetration. This finding is inconsis tent with previous research that has suggested that family history of IPV is a strong predictor of later IPV (Carr & VanDeusen, 2002). Inconsistent with prev ious research (Hotaling & Sugarman, 1986), no significant relationship was found between either alcohol use or
51 marijuana use and IPV perpetration. In addi tion, no significant relationship was found between beliefs about men and women in America and IPV victimization. A significant relationship was found between race/ethnicity and IPV perpetration (r = .130, p < .01) indicating that nonwhite females reported greater levels of IPV perpetration than their white c ounterparts. This finding is consistent with previous research that suggests that minorities are more likely to be involved as perpetrators of intimate partner vi olence (Barnett, Miller-Perrin, & Perrin, 1997). The correlations between family social support and friends social support (r = .546) and history of IPV victimizat ion and history of IPV perpetration (r = .754) were moderately large. As a re sult, diagnostics were run to determine if multicollinearity would be an issue in s ubsequent multivariate analyses. The variance inflation factors were well below the value of four, suggesting that multicollinearity was not a probl em (Allison, 1999; Fox, 1991). In order to clarify the most import ant predictors of I PV victimization and perpetration, Negative binomial r egression analyses were conducted. Negative Binomial Regression Analyses IPV Victimization. Because of the significant correlations between IPV victimization and the two social support va riables, it is necessary to conduct multivariate analyses to determine the stability of these findings. Table 3 presents the findings from the models asse ssing the ability of perceived social
52 Table 3 Negative Binomial Estimation for Model of IPV Victimization Model 1 Model 2 b se(b) eb b se(b) eb Family social support -.060 (-1. 72) .942 -.067 (-2.07)* .935 Friends social support -.078 (-1.92) .925 -.024 (-0.69) .976 History of IPV Victimization .480 (4.63)**1.616 History of IPV Perpetration -.141 (-.127) .868 Family History of IPV -.105 (-.27) .900 Alcohol use .098 (0.71) 1.103 Marijuana use .184 (1.18) 1.202 Beliefs about men and women .021 (0.89) 1.021 Race/Ethnicity .192 (0.68) 1.212 Current IPV Perpetration .207 (5.02)**1.230 X2 11.93** 123.48** Overdispersion 2696.38** 1441.04** Note: *p < .05, **p < .01. support (family and friends) and IPV correlate s to predict the probability of being victimized by an intimate partner withi n the past year. Model 1 in Table 3 presents the results of t he negative binomial regression analysis of the effects of the social support scales (family and fri ends) on current IPV victimization. There is significant evidence of overdispersion in model 1 (alpha = 2696.38, p < .00), theref ore, the negative binomia l regression model is appropriate and preferred to the Poisson regression model. The chi-square for Model 1 is significant (X2 = 11.93, p <.00); however, the effect of perceived social support from family on IPV victimization fail ed to attain statistical significance. Similarly, the effect of perceived social support from friends on IPV victimization also failed to attain statis tical significance. These findings are
53 inconsistent with previous research on clinical samples that suggests that increased social support reduces the like lihood of IPV victimization (Kocot & Goodman, 2003; Farris & Feenaughty, 2002; Larance & Porter, 2004; Carlson, McNutt, Choi, & Rose, 2002). In the current study, perceived social support from family is not associated with a lower level of reported IPV victimization experiences when friends support is controlled, and perceived support from friends is not related to lower victim ization when family social support is controlled; thus the first hypothesis is not supported in this data. The introduction of additional control variables into a model is done typically to ascertain whether an estimated relationship between independent variables and the dependent variables is spur ious. In the present data, however, no such relationship was found between either of the social support variables and IPV victimization when both variables are included simultaneously in the model. Nevertheless, it is still instructiv e to re-examine the relationships between social support and IPV victimization after controlling for the common correlates of victimization to determine whether they are exerting a suppressor effect on the social support-IPV victimization rela tionship (Agresti & Finley, 1997). Model 2 in Table 3 presents the results of the negative binomial regression analysis when the common corre lates of IPV are added. The data exhibit overdispersion for model 2 (alpha = 1441.04, p = .00) indicating that the negative binomial regression model remains appropriate. In addition, the model chi-square is significant (X2 = 123.48, p < .00).
54 The key finding in Model 2 is that higher levels of perceived social support from family were significantly associ ated with less frequent IPV victimization (estimate = -.067, z = -2.07, p = .038) when variables commonly associated with IPV are controlled. The change in signific ance of the relationship between family social support and IPV victimization indi cates that the relationship has been suppressed by one of the control variables. Suppression typically occurs when a control variable is positively asso ciated with the independent variable and negatively associated with the dependent vari able, or conversely, when the control variable is negatively associated with the independent variable and positively associated with the dependent va riable. An examination of the correlations among all the variables in t he model reveals that history of IPV victimization is significantly related to both family social support and IPV victimization, but in opposite directi ons. Therefore, when history of IPV victimization is allowed to vary, the relationship between family social support and IPV victimization appears to be absent, but when history of IPV victimization is controlled, a significant relationshi p between family social support and IPV victimization emerges. It is therefore r easonable to conclude that family social support does have at least a modest effect on IPV victimization when controlling for other variables. Comparatively, the relationship bet ween friends social support and IPV victimization remained nonsignificant when the control variables were added to the model. Those respondents who perceiv ed social support from friends were
55 neither more nor less likely to be a vict im of IPV than their peers who did not perceive social support from friends. As shown in Table 3, several of the common correlates of IPV produced consistent effects. History of IPV victim ization had a statistically significant effect on current victimization by an intimate partner (estimate = .480, z = 4.63, p = .00). Female respondents that reported bei ng victimized by an intimate partner within the past year were more likely to have a previous history of IPV victimization than their peers that did not report IPV vict imization within the past year. This is supportive of prior research (Dobash & Dobash, 1979). IPV perpetration within the past year also had a st atistically significant effect on IPV victimization within the past y ear (estimate = .207, z = .041, p = .00). This finding suggests that female respondents who perpetrated IPV were more likely to be victims of IPV within the past year t han their peers who were not IPV victims within the past year. As shown in Table 3, the remaining c ontrols did not produce statistically significant effects. Inconsistent with previous findings, alcohol and marijuana use did not have statistically significant effe cts on reported past year IPV victimization when social support variables were controlled. Similarly, there was no statistically significant effect for family history of IPV, race, or beliefs about men and women on reported past year IPV victim ization when social support variables were included in the model. Overall, the most robust predictors of IPV victimization in the past year were low levels of perceived social suppor t from family, having a history of IPV
56 victimization, and perpetration of IPV within the past year. My investigation now turns to the role that social support may play in perpetration of aggression by women against a male partner. IPV Perpetration Because of the significant corre lations between IPV perpetration and the two social support variables, it is nece ssary to conduct multivariate analyses to determine the stability of these findings. Table 4 presents the findings from the negative binomial regression models assessi ng the ability of perceived social support (family and friends) and IPV correlate s to predict the probability of using physical aggression against an intimate partner within the past year. Model 1 in Table 4 presents the results of the negative binomial regression analysis of the effects of the social support scales (family and friends) on current IPV perpetration. As model 1 in Table 4 indicates, there is significant evidence of overdispersion (alpha = 2619.05, p < .00), therefore, the negative binomial regression model is appropriate and preferred to the Poisson regression model. In addition, the Chi-square is significant for the overall model (X2 = 27.88, p = .00). For model 1, the central finding is that perceived social support from friends is associated with less frequent use of IPV within the past year (estimate = -.166, z = -3.96, p < .000). Comparativel y, perceptions of social support from family had no statistically significant effect on using physical aggression against an intimate partner. Therefore, perceiv ed social support from family is not
57 Table 4 Negative Binomial Estimation for Model of IPV Perpetration Model 1 Model 2 b se(b) eb b se(b) eb Family social support -.030 (-0. 87) .970 -.041 (-1.49) .960 Friends social support -.166 (-3. 96)** .847 -.128 (-4.35)** .880 History of IPV victimization -.221 (-2.35)* .802 History of IPV perpetration .572 (6.03)**1.772 Family history of IPV .224 (0.73) 1.251 Alcohol use -.164 (-1.53) .849 Marijuana use .249 (2.00)* 1.283 Beliefs about men and women .006 (0.77) 1.006 Race/Ethnicity .586 (2.54)* 1.797 Current IPV Victimization .193 (6.42)**1.213 X2 27.88** 212.00** Overdispersion 2619.05** 992.15** Note: *p < .05, **p < .01. associated with the frequency of using physical aggression against an intimate partner when controlling for friends social support. Model 2 in Table 4 presents the results of the negative binomial regression analysis when the common corre lates of IPV are added. The data exhibit overdispersion for model 2 (alpha = 992.15, p = .00) indicating that the negative binomial regression model is appr opriate and preferred to the Poisson regression model. In addition, the model chi-square is significant (X2 = 212.00, p < .00). The key finding in Model 2 was that perceived social support from friends remains associated with less frequent use of physical aggression against an intimate partner among female respondents (e stimate = -.128, z = -4.35, p <.00),
58 even when common correlates of IPV are controlled. Female respondents who perceived greater levels of social s upport from their friends used physical aggression against their intimate partner less frequently than their peers who did not perceive social support from their friends. As shown in Table 4, several of the common correlates of IPV produced statistically significant effects. History of IPV victimization had a statistically significant negative effect on current I PV perpetration (estimate = -.221, z = 2.35, p = .019). Respondents who indi cated they had a history of IPV victimization were less likely to report using physical aggression against an intimate partner within the past year. Co mparatively, histor y of IPV perpetration had a statistically significant positive effe ct on reports of current IPV perpetration (estimate = .572, z = 6.03, .000) where those respondents who reported a previous history of using aggression agains t an intimate partner were more likely to report using physical aggression against an intimate partner within the past year. Marijuana use within the past year had a statistically significant effect on IPV perpetration (estimate .249, z = 2.00, p = .045). This finding is supportive of prior research that suggests a rela tionship between substance use and IPV (Hotaling & Sugarman, 1986). Race had a st atistically significant effect on IPV perpetration (estimate = .586, z = 2.54, p = .011), where nonwhite individuals were more likely to use physical aggre ssion against their intimate partners than their white peers. Current vi ctimization also had a statistically significant effect on IPV perpetration within the past year (e stimate = .193, z = 6.42, p = .000).
59 Females who had been victims of IPV within the past year were more likely to indicate that they had used aggression agai nst a romantic partner within the past year than their peers who had not been victims. As shown in Table 4, the remaining c ontrols did not produce statistically significant effects. Family history of IPV did not have statistically significant effect on reported past year IPV perpetration. Similarly, there was no statistically significant effect for alcohol use or beliefs about men and women on reported past year IPV perpetration. Overall, the most robust predictors of IPV perpetration in the past year were low perceived social support from friends, having a history of IPV perpetration, having no long-term history of IPV victimization, marijuana use, being nonwhite, and having experienced IPV vi ctimization within the past year. Summary of Results In summary, there were two main hy potheses in the current research. First, it was expected that women who r eport greater levels of social support will be less likely to be victimized by their intimate partner. This first hypothesis was partially supported, with higher levels of perceived social support from family being associated with less frequent IPV victimization. Second, it was expected that women w ho report greater levels of social support will be less likely to use physical aggression against their intimate partners. This hypothesis was partially supported, with perceived social support from friends being associated with less frequent use of IPV within the past year.
60 Chapter Six Discussion The most important goal of this res earch was to determine whether social support played a role in college womens receipt and use of intimate partner violence. The present study utilized secondary data that collected information from female college students regarding thei r perceived levels of social support from family and friends and their use and re ceipt of intimate partner violence. The data used in the present study were derived from a NIH-funded study of college students. The current study included items measuring the respondents race/ethnicity, drug and alcohol use, family history of intimate partner violence, social support from family social support from friends, history of IPV victimization, histor y of IPV perpetration, beliefs about men and women in America, IPV victimizati on within the past year, and IPV perpetration within the past year. The dependent variables in the study were: IPV victimization within the past year and IPV perpetration withi n the past year. The two main independent variables were perceived social support from family and perceived social support from friends. Negative bi nomial regression was used to assess the effects of the independent variables on IPV victimization and perpetration. This study lends support for the argument that social support plays a protective role in college womens victim ization by an intimate partner. Results indicated that perceptions of social s upport from family were related to lower
61 reports of IPV victimization even when c ontrolling for other common correlates of IPV. The data here confirm other findings in clinical populations that social support is an important variable in determi ning the likelihood of whether or not a woman will be victimized by an intimate partner (Feld & Straus, 1990; Coker, 2003). This study also lends support to the argument that social support plays a protective role in college womens invo lvement in IPV perpetration. Results indicated that social support from friends was related to decreased use of IPV. Specifically, those who report ed greater levels of social support from their friends indicated participating in less IPV perpetration. It is interesting to contrast these findings with those of Schwartz and De Keseredy (1997), who found that male peer support was linked to greater use of aggression by males against an intimate partner. The current study findings suggest that womens peers do not provide support for the use of female-to-ma le intimate partner violence. Caution should be used in drawing solid conclusions due to the fact that women were neither asked the gender of their friends nor were they asked about the attitudes their friends had towards using aggression against an intimate partner. There are several considerations in t he present study that require certain precautions in interpreting these data. First, it is important to note that the current sample may not be representative of all women in violent partnerships. The current sample utilized women that were attending college. In 1990 (the entrance year for the women in the samp le), approximately sixty two percent of female high school graduates in the United States enrolled in college (National
62 Center for Education Statistics, 1998). Therefore, the current sample may represent a large proportion of female hi gh school graduates in the U.S., but it does not include females that did not comp lete high school or females that did not have the option to attend college. Research suggests that young women who drop out of high school have lower relative earnings, experience more unemployment during their work careers, are more likely to become pregnant at young ages, and are more likely to become single parents than those students who complete high school and/or college (Snyder & Sickmund, 1995). Dropping out of high school has also been linked to intimate partner violence (Moffitt & Caspi, 1999). Therefore, it is likely t hat individuals that do not complete high school differ from those that complete high school in the levels of IPV experienced. In addition, the college women in the current sample reported high levels of perceived emotional social support from friends and family. It is possible that women who complete high school and attend college have more support from family and friends than those that do not complete high school. Research has suggested that parents play a crucial role in keeping young people in school (Horn, 1992). It is also important to note that despi te high levels of aggression reported by the current sample, the aggression could be characterized as minor physical violence. The IPV Victimization and Per petration scales in the current study did not assess severe forms of physica l violence (e.g., choking, stabbing, shooting). Second, it is difficult to a ssess the time ordering and direction of the
63 association between intimate partner viol ence and social support from family and friends. Specifically, it is difficult to assess whether social support has a direct effect on intimate partner violence or w hether intimate partner violence has a direct effect on social support. In the cu rrent study, a complete measure of social support was only available in Wave 5. IPV victimization and IPV perpetration were expected to have contemporaneous effe cts with social support; therefore, these variables were also taken from t he data collected at Wave 5. The crosssectional approach utilized in the current st udy is limited in its ability to address time ordering and causality. Intimate partner violence is different from other forms of victimization because exposure is typically chronic rather than acute. As a result, the violence may exhaust emotional and tangible suppor t resources due to provider burnout or providers inability to continue to o ffer material resources (Thompson, Kaslow, Kingree, Rashid, Puett, Jacobs, & Matt hews, 2000). Therefore, experiencing violence may have an effect on social support which in turn may have an effect on further experiencing intimate partner violence. Longitudinal analyses are needed to clearly delineate the time ordering and causal effects. Third, the IPV victimiz ation and perpetration measures did not assess the context of IPV behavior. It was not clear, given the question format, whether perpetration and victimization were occurri ng at the same point in time. For example, when a female indicated that s he had been the victim of IPV in the past year and that she had been a perpetrator of I PV in the past year, it was not clear if the perpetration occurred in reaction to a vi ctimization or vice versa. Nor was it
64 clear whether victimization and per petration occurred within the same relationship. The current data only prov ide evidence that a female respondent had been a victim at some point wit hin the past year and/or had been a perpetrator at some point within the past year. In order to place perpetration and victimization in the appropriate context, future research in this area should assess the motivations for such behavior. Fourth, and relatedly, the current data examine past year victimization and perpetration by a romantic partner. The data for the present study do not examine the behaviors within one specific rela tionship. For example, if a female respondent reported that she was a victim of IPV five times in the past year, it is not clear how many romantic partners we re responsible for perpetrating those five acts of violence. Future research in this area could limit the IPV victimization and perpetration to one relationship. Fifth, the present study focused on I PV victimization and perpetration in heterosexual females. Research sugges ts that the contexts and dynamics in same-sex and heterosexual relationshi ps are different enough to warrant separate discussions (see Renzetti, 1992; Burke & Follingstad, 1999; Elliot, 1996; Lie & Gentlewarrier, 1991). Social support for lesbians may be different than for heterosexual women because of po ssible rejection by members of their family of origin, and discrimination fr om their community. Understanding the support networks of lesbians is impor tant because many have been rejected by their families of origin and may have dev eloped alternative support systems. It is
65 suggested that future research examine the role social support may play in homosexual intimate partner viol ence victimization and perpetration. Sixth, social support was operationa lized in the present study as the perceived emotional support provided by friends and family. There are many dimensions to social support. Research suggests that emotional support is the primary component of social support; how ever, other forms of social support (e.g., instrumental and informational) may al so serve a protective role. Future research could also include measures of instrumental support, informational support, and appraisal support. In addition, future research could compare the effects of perceived social support with t he effects of received social support on IPV victimization and perpetration to det ermine if there are any significant differences between actually receiving support and perceiving support. The present study investigated the role of in formal social support. Future research could also investigate the role of social support provided by formal agencies (schools, government, etc.). Research has suggested that these sources may be more important to isolated populations (Weber, 1998).
66 Implications Intimate partner violence is a serious public health problem in the United States and was a significant issue in t he lives of the women in the current sample. Approximately twenty four perc ent of the current sample reported being victimized by an intimate partner and appr oximately twenty six percent reported using a form of aggression against an intimate partner within the past year. The results from this study suggest t hat the effects of perceived social support from friends and family are general in that they are related to both reduced IPV victimization and reduced IPV perpetration. Research suggests that a major benefit of social support is its role in the maintenance of a positive selfesteem and self-concept (Weber, 1998). It appear s that feeling valued by friends and family reduces the likelihood that a college woman would be involved in a violent relationship either as a victim or as an offender. Young women attending college may have unique social support needs. Typically, college is the first time a young woman is away from home. The findings from the present study suggest that IPV could be reduced by creating a more supportive environment and by giving support to young college women. College campuses could offer orientation sessions for parents, caregivers, and/or family members of new college students that provide information about resources on campus, including victim advocacy pr ograms and counseling centers. In addition, the victim advocacy center on campus could provide information regarding definitions of intimate partner vi olence, risk factors for intimate partner violence victimization and perpetration, and information on what to do and who to
67 contact if an assault occurs. The resu lts from the present study suggest that maintaining the connection between the parent/caregiver and the female college student can perform an invaluable service to the female college student and potentially protect her from the effects of intimate partner violence.
68 References Agresti, A. & Finley, B. (1997). Statisti cal Methods for the Social Sciences, Third Edition. Prentice Hall. Allison, P. (1999). Multiple regression a primer. Thousand Oaks, California: Pine Forge Press. Archer, J. (2000). Sex differences in aggression between heter osexual partners: A meta-analytic review. Psychological Bulletin, 126 (5), 651-680. Archer, J. & Graham-Kevan, N. (2003). Do beliefs about aggression predict physical aggression to partners? Aggressive Behavior, 29, 41-54. Barrera, M. (1986). Distinctions between social support concepts, measures, and models. American Journal of Community Psychology, 14, 413-445. Barnett, O., Miller-Perrin, C. & Perrin, R. (1997). Family violence across the lifespan. Thousand Oaks, California: Sage Publications, Inc. Bem, S. (1983). Gender schema theor y and its implications for child development: raising gender-aschematic children in a gender-schematic society. Journal of Women in Culture and Society, 8 (4), 598-616. Bigler, R. (1997). Conceptual and methodol ogical issues in the measurement of childrens sex typing. Psychology of Women Quarterly, 21 (1), 53-69. Block, J. (1983). Differential premises aris ing from differential socialization of the sexes: Some conjectures. Child Development, 54, 1335-1354.
69 Bookwala, J., Frieze, I., Smith, C. & Ryan, K. (1992). Predictors of dating violence: A multivariate analysis. Violence and Victims, 7, 297-311. Browne, A. (1987). When battered women kill. New York: Macmillan. Burke, L. & Follingstad, D. (1999). Vi olence in lesbian and gay relationships: Theory, prevalence, and correlational factors. Clinical Psychology Review, 19 (5) 487-512. Campbell, A. & Muncer, S. (1987). M odels of anger and aggression in the social talk of women and men. Journal for the Theory of Social Behavior, 17, 489-512. Campbell, A., Muncer, S., & Coyle, E. (1992). Social representations of aggression as an explanation of sex di fferences: A preliminary analysis Aggressive Behaviors, 18, 95-108. Campbell, A., Muncer, S., & Gorm an, B. (1993). Gender and social representations of aggression: A communal-agentic analysis. Aggressive Behaviors, 19, 125-135. Carlson, B., McNutt, L., Choi, D., & Rose I. (2002). Intimate partner abuse and mental health. Violence Against Women, 8 (6), 720-745. Carr, J. & VanDeusen, K. (2002). The relationship between family of origin violence and dating violence in college men. Journal of Interpersonal Violence, 17 (6), 630-646. Cohen, S. & Wills, T. (1985). Stre ss, social support, and the buffering hypothesis. Psychological Bulletin, 98 (2) 310-357.
70 Coker, A., Watkins, K., Smith, P. & Br andt, H. (2003). Social support reduces the impact of partner violence on health: Application of structural equation models. Preventive Medicine, 37 (3), 259-267. Colvin, M., Cullen, F. & Vander Ven, T. (2002). Coercion, social support and crime: An emerging theoretical consensus. Criminology, 40 (1), 19-42. Cullen, F. (1994). Social support as an organizing concept for criminology: Presidential address to the Academy of Criminal Justice Sciences. Justice Quarterly, 11, 527-529. Cullen, F. & Wright, J. (1997). Liberating the anomie-strain paradigm: Implications for social-support theory. In N. Passas & R. Agnew (Eds.), The Future of Anomie Theory (pp 187-206). Boston: Northeastern University Press. Cullen, F., Wright, J., & Chamlin, M. ( 1999). Social support and social reform: A progressive crime control agenda. Crime & Delinquency, 45 (2), 188-207. Dasgupta, S. (1999). Just like men? A crit ical review of violence by women. In E. Pence & M. Shepard (Eds.), Coordinating community response to domestic violence: Lessons from the Duluth model (pp. 195-222). Thousand Oaks, CA: Sage Publications, Inc. Dasgupta, S. (2002). A framework fo r understanding womens use of nonlethal violence in intimate heterosexual relationships. Violence Against Women, 8 (11), 1364-1389.
71 Deaux, K. (1984). From individual differenc es to social categories: Analysis of a decades research on gender. American Psychologist, 39, 105-116. Deaux, K. & Major, B. (1987). Putting gender into context: An interactive model of gender related behavior. Psychological Review, 94, 369-389. DeKeseredy, W., Saunders, D., Schwartz, M., & Alvi, S. (1997). The meanings and motives for womens use of violence in Canadian college dating relationships: Results from a national survey. Sociological Spectrum, 17, 199-222. DeKeseredy, W. & Schwartz, M. (1998). Measuring the extent of women abuse in intimate heterosexual relationships : A critique of the Conflict Tactics Scales. Retrieved August 9, 2004, from VAWnet web site: http://www.vawnet.org/vn1/libra ry/general/ar_ctscrit/html. Dobash, R.E. & Dobash, R.P. (1977). Wive s: The appropriate victims of marital violence. Victimology, 2, 426-442. Dobash, R.E. & Dobash, R.P. (1979). Violence against wives. New York: Free Press. Dobash, R., Dobash, R., Cavanaugh, K. & Lewis, R. (1998). Separate and intersecting realities: A comparis on of mens and womens accounts of violence against women. Violence Against Women, 4, 382-414. Dobash, R.E. & Dobash, R. P. (2000). Ev aluating criminal justice interventions for domestic violence. Crime and Delinquency, 46 (2), 252-270.
72 Doumas, D., Margolin, G. & John, R. ( 1994). The intergenerational transmission of aggression across three generations. Journal of Family Violence, 9, 157-175. Eagly, A. (1987). Sex differences in social behavior: A social role interpretation. Hillsdale, NJ: Lawrence Erlbaum. El-Bassel, N., Gilbert, L., Rajah, V., Fol eno, A. & Frye, V. (2001). Social support among women in methadone treatment w ho experience partner violence. Violence Against Women, 7 (3), 246-274. Elliott, P. (1996). Shattering illusions : Same-sex domestic violence. Journal of Gay & Lesbian Social Services, 4, 1-8. Epstein, C.F. (1988). Deceptive distinctions: sex, gender, and social order. New Haven, Conn.: Yale University Press. Farris, C. & Fenaughty, A. (2002). Soci al isolation and domestic violence among female drug users. American Journal of Drug and Alcohol Abuse, 28 (2), 339-351. Feld, S. & Straus, M. (1990). Escala tion and desistance from wife assault in marriage. In M. Straus & R. Gelles (Eds.), Physical violence in American families: Risk factors and adaptations to violence in 8,145 families (pp. 489-505). New Brunswick, NJ: Transaction Books. Fiebert, M.S., & Gonzalez, D.M. (1997). Women who initiate assaults: The reasons offered for such behavior. Psychological Reports, 80, 583-590.
73 Follingstad, D.R., Wright, S., Lloyd, S. & Sebastian, J.A. (1991). Sex differences in motivations and effects in dating violence. Family Relations, 40, 51-57. Foshee, V. (1996). Gender differences in adolescent dating abuse prevalence, types and injuries. Health Education Research, 11, 275-286. Fox, J. (1991). Regression diagnostics. Newberry Park, California: Sage Publications, Inc. Frieze, I. (2000). Violence in close rela tionshipsDevelopment of a research area: Comment on Archer (2000). Psychological Bulletin, 126 (5), 681684. Gardner, W., Mulvey, E. & Shaw, E. ( 1995). Regression analyses of counts and rates: Poisson, overdispersed poiss on, and negative binomial models. Psychological Bulletin, 118, 392-404. Gelles, R. (1985). Intimate violence in families. Beverly Hills, CA: Sage. Gelles, R. & Straus, M. (1988). Intimate violence. New York: Simon & Schuster. Gilligan, C. (1982). In a different voice: Psychological theory and Womens development. Cambridge, MA: Harvard University Press. Green, W. (1993). Econometric analysis. New York: Macmillan Publishing. Hamberger, L.K. & Potente, T. ( 1996). Counseling heterosexual women arrested for domestic violence: Implicat ions for theory and practice. In L.K. Hamberger & C.M. Renzetti (Eds.), Domestic partner abuse (pp. 5376). New York, Springer.
74 Harned, M. (2001). Abused women or abused men? An examination of the context and outcomes of dating violence. Violence and Victims, 16 (3), 269-285. Helgeson, V. (1993). Two important di stinctions in social support: Kind of support and perceived versus received. Journal of Applied Social Psychology, 23 (10), 825-845. Hernandez-Plaza, S., Pozo, C., Alonso, E. (2004). The role of informal social support in needs assessment: Proposal and application of a model to assess immigrants needs in the south of Spain. Journal of Community and Applied Social Psychology, 14 (4), 284-298. Hobfoll, S. & Vaux, A. (1993). Soc ial support: Social resources and social context. In L. Goldberger & S. Br eznitz (Eds.), Handbook of stress: Theoretical and clinical aspects (2nd edition) (pp. 685-705). New York, NY: Free Press. Hotaling, G. & Sugarman, D. (1986). An analysis of risk markers in husband to wife violence: The current state of knowledge. Violence and Victims, 1, 101-124. House, J. (1991). Work stress and social support. Reading, MA: AddisonWesley. Johnson, M.P. (1995). Patriarchal te rrorism and common couple violence: Two forms of violence against women. Journal of Marriage and the Family, 57, 283-294.
75 Katz, J., Kuffel, S. & Coblentz, A, ( 2002). Are there gender differences in sustaining dating violence? An exam ination of frequency, severity, and relationship satisfaction. Journal of Family Violence, 17 (3), 247-271. Kimmel, M. (2002). Gender symmetry in domestic violence. Violence Against Women, 8 (11), 1332-1363. Knox, D., Zusman, M. & Nieves, W. (1997) Gender differences in relationship decisions among college students. College Student Journal, 31, 314-316. Kocot, T. & Goodman, L. (2003). The roles of coping and social support in battered womens mental health. Violence Against Women, 9 (3), 323346. Lakey, B. & Cassady, P. (1990). Cognitive processes in perceived social support. Journal of Personality and Social Psychology, 59, 337-343. Larance, L. & Porter, M. (2004). Obse rvations from practice: Support group membership as a process of social capital formation among female survivors of domestic violence. Journal of Interpersonal Violence, 19 (6) 676-690. Lie, G. Y., & Gentlewarrier, S. (1991). Inti mate violence in lesbian relationships: Discussion of survey findings and practice implications. Journal of Social Science Research, 15, 41-59. Long, J. (1997). Regression models for categorical and limited-dependent variables. Thousand Oaks, CA: Sage.
76 Macaulay, J. (1985). Adding gender to aggression research: Incremental or revolutionary change? In V.E. OLear y, R.K. Unger, & B.S. Wallston (eds.), Women, gender, and social psychology. Hillsdale, NJ: Lawrence Erlbaum. Makepeace, J.P. (1986). Gender differences in courtship violence victimization. Family Relations, 35, 383-388. Markward, M., McMillan, L. & Markward, N. (2003). Social support among youth. Children and Youth Services Review, 25 (7), 571-587. Miller, J. & White, N. (2003). Gender and adolescent relationship violence: A contextual examination. Criminology, 41 (4), 1207-1248. Moffitt, T. & Caspi, A. (1999). Findings about partner violence from the Dunedin Multidisciplinary health and development study (Research in brief). National Institute of Justice. NCJ 170018 Morrissey, B. (2003). When women kill: Questions of agency and subjectivity. Independence, KY: Routledge. Naffin, N. (1987). Female Crime: The construction of women in criminology. North Sydney, Australia: Allen & Unwin. National Center for Education Statistics (1998). College enrollment rates of high school graduates, by sex: 1960 to 1997 Institute of Education Sciences, U.S. Department of Educat ion: Washington, DC.
77 Norris, F. & Kaniasty, K. (1996). Received and perceived social support in times of stress: A test of the social support deterioration deterrence model. Journal of Personality and Social Psychology, 71 (3), 498-511. Olson, D. & Shultz, K. (1994). Gender diffe rences in the dimensionality of social support. Journal of Applied Social Psychology, 24 (14), 1221-1232. Pearson, P. (1997). When she was bad: Violent women and the myth of innocence. New York: Penguin Putnam Inc. Reitzel-Jaffe, D. & Wolfe, D. (2001). Predictors of relationship abuse among young men. Journal of Interpersonal Violence. Vol 16(2) 99-115. Renzetti, C. (1992). Violent betrayal: Violence in lesbian relationships. Thousand Oaks, CA: Sage. Robbers, M. (2004). Revisiting the moderat ing effect of social support on strain: A gendered test. Sociological Inquiry, 74 (4), 546-569. Salari, S. & Baldwin, B. (2002). Ver bal, physical, and injurious aggression among intimate couples over time. Journal of Family Issues, 23 (4), 523550. Sampson, R. & Laub, J. (1993). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press. Schwartz, M. & DeKeseredy, W. (1997). Sexual assault on the college campus: the role of male peer support. Thousand Oaks, CA: Sage Publications. Serran, G. & Firestone, P. (2004). Intima te partner homicide: a review of the male proprietariness and t he self-defense theories. Aggression and Violent Behavior, 9 (1), 1-15.
78 Shumaker, S. & Hill, D. (1991). Gender differences in social support and physical health. Health Psychology, 10 (2), 102-111. Snyder, H.N., and Sickmund, M. (1995). Juvenile Offenders and Victims: A National Report. Washington, DC: U.S. Depar tment of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention. Stanko, E. (1985). Intimate intrusions. London: Routledge & Kegan Paul. Straus, M. (1979). Measuring intrafam ily conflict and violence: The conflict tactics scales. Journal of Marriage and the Family, 41, 75-88. Straus, M. (1993). Physical assault by wives : A major social problem. In Gelles, R.J. & Loseke, D. (Eds.), Current controversies on family violence. New Brunswick, NJ: Transaction. Straus, M. (1999). The controversy over domestic violence by women: A methodological, theoretical, and sociology of science analysis. In Arriaga, X. and Oskamp, S. (Eds.), Violence in Intimate Relationships. Thousand Oaks, CA: Sage. Swan, S.C. & Snow, D.L. (2002). A typol ogy of womens use of violence in intimate relationships. Violence Against Women, 8 (3), 283-319. Thompson, M., Kaslow, N., Kingree, J., Rashid, A., Puett, R., Jacobs, D. & Matthews, A. (2000). Partner vi olence, social support, and distress among inner-city African American Women. American Journal of Community Psychology, 28 (1), 127-143.
79 Underwood, M. (2003). Social aggression among girls. New York, NY: The Guilford Press. Walker, L. (1979). The battered woman. New York: Harper and Row. Weber, M. (1998). She stands alone: A review of the recent literature on women and social support. Prairie Womens Health c entre of Excellence: ISBN 09684540-3-8. Wethington, E. & Kessler, R. (1986). Perceived support, received support, and adjustment to stressful life events. Journal of Health and Social Behavior, 27 (1), 78-89. White, J. & Kowalski, R. (1994). Decons tructing the myth of the nonaggressive woman. Psychology of Women Quarterly, 18, 487-508. White, J., Smith, P., & Humphrey, J. Longitudinal study of violence against women: Victimization and perpetration among college students in a statesupported University in the United St ates, 1990-1995. [Computer file]. ICPSR version. Greensboro, NC: The University of North Carolina at Greensboro [producer], 2001. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2002.
81 Appendix A: Results of Factor Analysis for measures of IPV victimization
82 Appendix A (continued) Results of Factor Analysis for measures of IPV victimization IPV Victimization Scale Items Factor Loadings Item-to-Scale Factor 1 Correlation 1. He threw or smashed something (but not at me). .69 .74 2. He threatened to hit or throw something. .74 .84 3. He threw something at me. .70 .72 4. He pushed, grabbed, or shoved me. .85 .82 5. He hit (or tried to hit) me but not with anything. .75 .75 6. He hit (or tried to hit) me with something hard. .40 .58 Eigenvalues: 2.95 Cronbachs Alpha: .75
83 Appendix B: Results of Factor Analysis for measures of IPV perpetration
84 Appendix B (continued) Results of Factor Analysis for measures of IPV perpetration IPV Perpetration Scale Items Factor Loadings Item-to-Scale Factor 1 Correlation 1. I threw or smashed something (but not at him). .77 .68 2. I threatened to hit or throw something. .52 .78 3. I threw something at him. .81 .82 4. I pushed, grabbed, or shoved him. .88 .82 5. I hit (or tried to hit) him but not with anything. .81 .85 6. I hit (or tried to hit) him with something hard. .52 .68 Eigenvalues: 3.22 Cronbachs Alpha: .80
85 Appendix C: Results of Factor Analysis for measures of family social support
86 Appendix C (continued) Results of Factor Analysis for m easures of family social support Family Social Support Scale Items Factor Loadings Item-to-Scale Factor 1 Correlation 1. My family cares for me very much. .78 .76 2. My family holds me in high esteem. .82 .79 3. I am really admired by my family. .84 .82 4. I am loved dearly by my family. .82 .80 5. Members of my family rely on me. .59 .62 6. I cant rely on my family for support. .64 .68 7. My family really respects me. .86 .84 8. I dont feel close to members of my family. .73 .76 Eigenvalues: 4.68 Cronbachs Alpha: .89
87 Appendix D: Results of Factor Analysis fo r measures of friend social support
88 Appendix D (continued) Results of Factor Analysis for measures of friend social support Friend Social Support Scale Items Factor Loadings Item-to-Scale Factor 1 Correlation 1. My friends respect me. .65 .64 2. I can rely on my friends. .86 .85 3. My friends dont care about my welfare. .71 .71 4. I feel a strong bond with my friends. .89 .88 5. My friends look out for me. .87 .86 6. My friends and I are really important to each other. .91 .89 7. My friends and I have done a lot for one another. .63 .67 Eigenvalues: 4.77 Cronbachs Alpha: .90
89 Appendix E: Results of Factor Analysis for measures of beliefs about men and women in America
90 Appendix E (continued) Results of Factor Analysis for measur es of beliefs about men and women in America Beliefs about Men and Women Scale Items Factor Loadings Item-to-Scale Factor 1 Correlation 1. Women are generally more sensitive to the needs of others than men are. .53 .52 2. Women should take the passive role in courtship. .47 .51 3. Men are more competitiv e than women. .61 .63 4. Men are more sure of what they can do than women are. .60 .61 5. Women tend to subordinate their own needs to the needs of others. .41 .44 6. Men are more independent than women. .67 .66 7. Women are more helpful than men. .63 .60 8. Compared to men, women t end to be gullible. .67 .65 9. Compared to men, women are more able to devote themselves completely to others. .65 .64 Eigenvalues: 3.11 Cronbachs Alpha: .76
About the Author Kathryn Branch received a Bachelors Degree in Psychology with a minor in Criminology from the University of South Florida in 1998 and a M.A. in Criminology from the University of Sout h Florida in 1999. She started teaching undergraduate Criminology courses while in the Masters program and entered the Ph.D. program at the Universi ty of South Florida in 2000. While in the Ph.D. program at the Univ ersity of South Florida, Ms. Branch worked in the Research and Development Division of the Hillsborough County Sheriffs Office and taught as an adjunct instructor in Criminology at the University of South Florida and the Un iversity of Tampa. She has also coauthored three publications and made seve ral paper presentations at national conferences of the American Societ y of Criminology and the Academy of Criminal Justice Sciences.