AN EXAMINATION OF THE EFFECTS THAT RELATIONSHIPS WITH INTIMATE PARTNERS HAVE ON WOMEN OFFENDERS By Dondrea Jackson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Criminal Justice 2012 ABSTRACT AN EXAMINATION OF THE EFFECTS THAT INIMATE RELATIONSHIPS HAVE ON WOMEN OFFENDERS By Dondrea Jackson The women offenders’ intimate relationships with males have demonstrated differences in effects on criminal involvement when compared to the effects of men’s intimate relationships. Research has indicated that women’s romantic relationships with men influence their participation in crime. The purpose of this thesis was to examine the influence that intimate relationships have on the recidivism of women under community supervision. A secondary analysis of an existing dataset was used to determine whether having a criminal intimate partner was a significant predictor of recidivism, and for a subsample of women who said their partners had contributed to criminality or violations of conditions of supervision, to present the women’s explanations of how men got them in trouble. Bivariate and multivariate analyses revealed relationships between criminal partners and women’s recidivism were not significant. However, based on the perceptions of women offenders, the qualitative analyses indicated multiple ways that male partners got some women in trouble. Copyright by DONDREA JACKSON 2012 Dedicated to my mother, Michele Jackson, and my grandfather, Matthew Currie. iv ACKNOWLEDGEMENTS It is an honor for me to show my appreciation to several people who encouraged me and assisted me in the preparation and completion of this thesis. First and foremost I would like to thank my Lord and Savior, Jesus, for ordering my steps and never leaving my side during this process. I extend my deepest gratitude to Dr. Merry Morash. I cannot express enough how much I respect you as a researcher and phenomenal advisor. It has truly been an honor for me to receive your guidance throughout this journey. You have taught me so much and I cannot thank you enough. I would also like to thank Dr. Cobbina for always extending an open door and being patient with me as I completed this study. I appreciate your encouraging words and sincere support. I would like to express my appreciation to Dr. Nalla. I will never forget your genuine support. Thank you. To my family, friends, and fellow New Mount Calvary members, I cannot thank you enough for your prayers and kind words that helped make the completion of this thesis a reality. To Dr. George Wilson, I thank you for simply for believing in me and always being a phone call away. To Sheena Horsford, thank you for helping me get the ball rolling. You have cheered me on from day one. Without your support my days would have been much longer. I would like to thank Jaemia Pratt and Simone Terfa for “shooting with me in the gym.” Our days together in the library were essential to this thesis no longer being a work in progress. I especially thank Tia Stevens and Mim Northcutt for all the feedback, comments, and help with statistics. I really appreciate you allowing me to borrow your books and bug you with questions. Mim I really value the friend that I have found in you and all the advice you have given me. v To my mother, Boot, no words or things can define how thankful I am for you. Your endless love and support has kept me going. I thank you for all the cards and motivational words that you have shared with me. You always knew how to renew my drive and you never stopped motivating me. Without you this would not be possible. You are greatly appreciated. vi TABLE OF CONTENTS List of Tables................................................................................................................................viii Chapter 1..........................................................................................................................................1 Chapter 2..........................................................................................................................................5 2.1 pathways to criminality..................................................................................................5 2.2 differences between romantic relationships and friendships.........................................13 2.3 the effects of abusive partners.......................................................................................19 2.4 overview of prior research and contribution of the present study...............................23 Chapter 3........................................................................................................................................24 3.1 sample..........................................................................................................................26 3.2 quantitative data...........................................................................................................26 3.3 coding the qualitative data...........................................................................................29 3.4 analysis of quantitative data.........................................................................................31 Chapter 4........................................................................................................................................33 4.1 description of the sample for quantitative analysis......................................................33 4.2 bivariate analyses of the relationship of criminal partner and abuse with recidivism.36 4.3 multivariate analyses....................................................................................................43 4.4 qualitative analysis.......................................................................................................44 Chapter 5........................................................................................................................................52 5.1 limitations....................................................................................................................53 5.2 recommendations.........................................................................................................54 Appendices.....................................................................................................................................57 Appendix A........................................................................................................................57 Appendix B........................................................................................................................68 Appendix C........................................................................................................................71 References......................................................................................................................................74 vii LIST OF TABLES Table 2.1 This shows Richie’s (1996) pathways that lead women to offending………............7 Table 2.2 This is Daly’s (1992) pathways that lead women to offending…...……...................9 Table 3.1 Multicollinearity Regression Diagnostics……………………………….....………32 Table 4.1 Descriptive Information for all Variables Used in the Analyses…………………..35 Table 4.2 Bivariate Analysis to Compare Women with Criminal Partners on Official Outcomes of Recidivism………………………………...........……………………37 Table 4.3 Bivariate Analysis to Compare Women with Criminal Partners on Composite Indicator of Recidivism……………………………………….……………………37 Table 4.4 Bivariate Analysis to Compare Women with Different Outcomes of Composite Recidivism on Abuse……………………………………………...………….....…38 Table 4.5 Mean Comparison of LSI Score and Crimnal Partner…………………..…………39 Table 4.6 Bivariate Relationships: Correlations between Criminal Partners, Forced Crime, Abuse, Gender Responsive Supervision, and LSI Score…………………………..42 Table 4.7 Logistic Regression Predicting Recidivism from Criminal Partner, Abuse, Gender Responsive, and LSI Score………………………………………..............43 Table 4.8 Bivariate analysis: Comparison of Criminal Partners on Specific Trouble Outcomes……………………………………………………………………..........48 Table 4.9 Bivariate analysis: Comparison of Criminal Partners on General Trouble Outcomes……………………………………………………………….................51 Table 6.1 Summary of Literature Indicating how Men get Women in Trouble.......................57 viii CHAPTER 1: THE NEED FOR THE RESEARCH In recent years, women have become the fastest growing population in the prison system and in community correction’s populations. Despite their rapid increase in correctional systems, for the most part women offenders tend to be overwhelmingly invisible or ignored in research and policy (Salisbury and Van Voorhis, 2009). Like their male counterparts, the majority of incarcerated women will return to their former communities and families after imprisonment. Post-incarcerated women reenter society facing a number of challenges that can hinder successful re-entry and desistance from crime; women on probation also face many of the same difficulties. One hypothesized negative influence on women offenders is intimate partners who support illegal behavior. However, there is limited research indicating how romantic relationships have impacted women offenders and their lawbreaking, and showing whether having criminal partners predicts women’s recidivism. In the United States, there are over 1 million women under correctional and community supervision (Bloom, Owen, and Covington, 2004; Salisbury and Van Voorhis, 2009), and women constitute 6.9% of all individuals incarcerated (Harrison and Beck, 2005). A positive relationship is present between the number of women incarcerated and on parole. As the number of women incarcerated rises, there is an increase in women’s presence in parole populations (Morash, 2010). Women’s representation in adult probation and parole populations has increased rapidly since the 1990s (Morash, 2010). The criminal justice system supervises approximately 85% of convicted women in the community (Bloom, Owen, and Covington, 2004). In 2001, women accounted for 22 % of the adult probation population (Glaze, 2002). During 2010, more than 712,000 women were on probation, making up 24% of adults on probation, indicating a 20% 1 increase in their population compared to 1990 (Glaze, 2002; Glaze and Bonczar, 2011). Women constituted 12% of adults on parole in 2001, accounting for more than 90,000 women in total (Glaze, 2002; Glaze and Bonczar, 2011). By 2010, there were more than 100,000 women in the adult parole population. This increase is largely due to more apprehension and harsher penalties resulting from the war on drugs. Thus, the majority of women on parole and probation are supervised for committing non-violent property or drug offenses (Greenfield and Snell, 1999; Richie, 2001). Consequential to the war on drugs, women have received tougher and longer sentences, right along with men, when they are found guilty of violating drug laws (Bloom et al., 2004; Covington, 2008; Morash, 2010; Richie, 1996). A growing body of research on women has tried to uncover the pathways they take into criminality. Their roads to criminal lifestyles or isolated instances of breaking the law differ from the pathways of male offenders (Bloom et al., 2004; Salisbury and Van Voorhis, 2009). Scholars emphasize that women’s involvement in crime commonly results from social relationships, coercion, economic strain, histories of sexual or physical abuse, and/or drug addiction (Covington, 2008; Daly, 1992; DeHart, 2008; Leverentz, 2006; Richie, 1996; Rumgay, 2004; Salisbury and Van Voorhis, 2009). Especially relevant to the present thesis, some research has shown that intimate partners often play key roles in crime producing relationships through pathways associated with coercion, economic strain, victimization, and substance abuse (Daly, 1992; Richie, 1996; Salisbury and Van Voorhis). Recent research conducted by Salisbury and Van Voorhis (2009) examines gender-specific pathways women take into criminal activity. The researchers’ relational model tests how dysfunctional intimate relationships influenced recidivism. Dysfunctional relationships were conceptualized as harmful, unsupportive, unsatisfying intimate relationships. The results 2 from their multivariate analysis indicated that dysfunctional relationships were not directly related to prison admission. Unhealthy relationships indirectly placed women at risk of participating in criminal behaviors through adult victimization, reduced levels of self-confidence, and issues of depression and substance abuse. Salisbury and Van Voorhis (2009) found that women experiencing more dysfunction in their intimate relationships were more likely to experience economic strain, and that employment and financial difficulties had direct effects on imprisonment (Salisbury and Van Voorhis, 2009). The findings suggest that women’s criminality and recidivism is influenced by gender-specific causes that will be further explained in the following chapter. Furthermore, the authors suggest how pathways to criminality are created by unhealthy intimate relationships. In addition to Salisbury and Van Voorhis (2009), other criminologists have identified how romantic relationships influence women’s criminality. Through relationships with men, many women are exposed to opportunities that start them on a path to criminal behaviors. It has been argued that some women are coerced into violating the law by their male partners through the use of abuse and manipulation (DeHart, 2008; Leverentz, 2006). Additionally, some women’s dependence on male companions guides them into situations in which they break the law and their freedom is ultimately forfeited. These and other effects of unhealthy relationships on females are explained by proponents of the pathways theory (Reisig, Holtfreter, and Morash, 2006; Richie, 1996). Moreover, researchers have argued that intimate relationships have effects on women that significantly differ from effects on men (De Li and MacKenzie, 2003; Giordano, Cernkovich, and Rudolph, 2002; Laub, Nagin, and Sampson, 1998; Laub and Sampson, 2001; Leverentz, 2006. One specific reason why the difference occurs is because the selection pool of prosocial 3 romantic partners for women offenders provides limited choices, especially of prosocial candidates (De Li and MacKenzie, 2003; Leverentz, 2006). However, the selection pool for male offenders more often includes prosocial women, because fewer women commit crimes compared to men. Thus women offenders may be at especially high risk for involvement with male partners who support or promote the women’s criminality. Not only have women’s unique pathways to crime been understudied, but also programming for male offenders often has been applied to women without attention to the unique pathways leading to women’s illegal activity. For example, as Covington (2008) notes, much research examining the needs of drug addicted offenders has focused on male samples. The needs of women battling chemical dependency differ from the needs of men. Despite this knowledge, research designed for men is generally applied to the programming for women in treatment and recovery. The application of findings of research on men to women may result in interventions that do not address the causes of women’s drug use and other criminality. Offender reform is most beneficial when its design is gender responsive (Rumgay, 2004). Thus, the present study will shed light on how and to what extent programming for women offenders should consider relationships with romantic partners. More specifically, the purpose of this current study is to examine how romantic relationships impact women offenders. Some of the women were recently released from prison on parole, and others were currently on probation. The study describes how the women explain the influences of their partners and tests hypotheses linking partner characteristics to outcomes of recidivism. 4 CHAPTER 2: THE EXISTING LITERATURE ON INTIMATE PARTNERS AND WOMEN’S CRIMINAL BEHAVIOR This chapter starts by reviewing evidence that females have much lower levels of crime than do males, and explains women’s pathways to criminality. Additionally, the chapter discusses different types of crime females commit when they do break the law. The reasons for females having different levels and types of crime than males include limits on women’s access to networks that provide opportunities for crime (Steffensmeier, 1980). This information shows the importance of understanding unique influences on women’s lawbreaking. Another topic covered in this chapter is the difference between friendships and romantic relationships. This is relevant to understanding the effect of male intimate partners on women’s criminality. Key research that highlights women’s vulnerability to men’s influences on women’s criminal involvement will be discussed. Also discussed is conflicting information that raises questions about the effects of intimate partners on women’s illegal activity, and thus signals the need for the present study. Additionally, this chapter presents evidence that crimes committed with male partners lead to women receiving longer and more punitive sentences. A failure to understand men’s influence on women’s criminality contributes to this problem of harsher treatment of women and men who break the law together. Pathways to Criminality Especially relevant to this thesis are the unique pathways that researchers have found that lead women to participate in crime. As previously explained, Salisbury and Van Voorhis (2009) used a relational model to describe how dysfunctional relationships indirectly influence women’s recidivism. Specifically, they found that dysfunctional relationships directly increased women’s risk of adult victimization and reduced women’s levels of self-efficacy. Consequentially, 5 women’s methods of coping with depression and anxiety without the use of substances were less likely. Ultimately, this established women’s participation in crime. Additionally, Beth Richie (1996) identified 6 distinctive pathways women followed into crime. Richie’s study focused on African American battered women incarcerated in prison. The research sample included life history interviews retrieved from 37 women recruited from Rikers Island Correctional facility from 1991-1992. Her sample included White and African American women with and without prior experiences of abuse. The results of Richie’s (1996) study found that women followed the following pathways in crime: 6 Table 2.1 Richie’s (1996) Pathways that Lead Women to Offending This pathway was associated with women who were abused by the men who killed their children. These women were held hostage and isolated with their male partner’s use of violence. The women following this pathway were arrested and found guilty for the deaths of their children that their abusers committed. This pathway represented women who committed violent acts against men who posed as symbolic beings of their past abusers or made an attack against other men as a response to past abuse. This pathway involved women who were incarcerated for illegal sex work. Many of these women had significant rates of sexual abuse in their past. The women in this pathway used sex to make money and support them. Additionally, some women were forced to submit to illegal sex work by their abusers. This pathway included women charged for crimes committed against their batterers during an abusive event. Women Held Hostage (Child Murder) Battered African American n=4 Projection and Association (Assaulted other Men) Battered African American n=4 Sexual Exploitation (Illegal Sexual Acts) Battered African American n=6 Battered White n=4 Fighting Back (Crime During Assault) Battered African American n=3 Battered White n=2 Poverty (Economic Crime) This pathway was associated with women who committed economically motivated crimes, including, but not limited to property offenses. Specifically, some of these women were lead to commit such offenses because their partners used abuse to coerce them. The abuse endured had impacts on the women’s work performance which lead to a need for extra finances, or hindered them from going to work all together. Battered African American n=5 Non-battered African American n=1 7 Table 2.1 (cont’d) This pathway included women that abuse drugs as a coping mechanism. Specifically, some of these women identified their drugs use began after issues of abuse occurred in their relationship. Others began using drugs as a method to improve various categories of their relationships with their abusers. (Belknap, 2007; Richie, 1996) Addiction (Illegal Drug Activity) Battered African American n=4 Non-battered African American n=4 Battered White n=1 8 Daly (1992) found 5 unique pathways that lead women to offending. Daly (1992) analyzed a group of forty women’s biographies retrieved from a larger study on equality and justice in criminal courts. Initially, information was gathered on the women from New Haven felony court from 1981-1986. Daly (1992) used women’s presentence investigation reports to develop a statistical profile and to write an individual biography for each woman. The statistical profiles and biographies were used to identify the women’s pathways to crime. Table 2.2 Daly’s (1992) Pathways that Lead Women to Offending This pathway was associated with women who had a history of abuse, neglect, or being negatively labeled as a child. Ultimately, these women’s harmful experiences as a child were displayed in their harming acts that they committed against others. Additionally, these women suffer from a mental illness and used abuse drugs or alcohol to cope. This pathway involved women who were in relationships with abusive men. These women ultimately fight back against their abusive intimate partners and were charged. This pathway involved women who left home at a young age due to some form of abuse or conflicts with their parents independently. To support themselves these women committed common street crimes such as prostitution and theft. These women’s lives on the streets lead them to using drugs. In turn, their substance abuse habits were supported by committing additional financial driven street crimes. This pathway is associated with women who used or sold drugs at the expense of their relationships with intimate partners or family members. These women were arrested for drug-related offenses. This pathway included women who did not fit into the previously described pathways. Additionally, these women did not have any prior arrests, issues with drug or alcohol use, and were less likely to have issues as a child related to abuse. Harmed and Harming Women n=15 Battered Women n=5 Street Woman n=10 Drug-Connected Women n=6 Other n=4 9 Women Breaking the Law with Men Regardless of whether women violate the law alone or with other women, the crimes they commit differ substantially from the crimes they have access to when co-offending with males (Alarid, Marquart, Burton, Cullen, and Cuvelier, 1996; Becker and McCorkel, 2011). Women who offend without men are more likely to commit female “traditional” crimes that include prostitution and shoplifting (Alarid et al., 1996). Some women prostitute and shoplift to support drug habits (Daly, 1992). Women are more likely to display patterns of criminality that are less profitable, be non-violent, and typically less serious than offenses committed among men (Alarid et al., 1996; Becker and McCorkel, 2011). This can be explained by their specific pathways to criminality. Access to Criminal Networks. In addition to criminal pathways, the pattern of women’s crime can be explained by the level of access to opportunities for crime that differs between males and females. Women are likely to be involved in fewer crimes and a smaller range of crimes when they do not offend alongside of men (Becker and McCorkel, 2011). However, when men grant women entrance into criminal networks, gender gaps in arrests narrow for serious crimes such as robbery and drug sales (Alarid et al., 1996; Becker and McCorkel, 2011). Alarid, Marquart, Burton, Cullen, and Cuvelier (1996) interviewed 104 women and found that when women worked alongside of men, they were more likely to contribute to male dominated illicit activities compared to when they worked alone. Offending with men placed some women in secondary roles as accessories to criminal offenses (Alarid et al., 1996). Women’s secondary roles in crime with men commonly involve women concealing drugs or weapons as well as playing minor roles in serious offenses (Alarid et al. 1996). Specifically, 10 Alarid and his coauthors found that when offending alone, 19% of the women committed a drug offense and none of the women committed a burglary or aggravated assault. In comparison, when offending with a man who initiated the crime, 43% of the sample committed a drug offense, 13% committed burglary, and 10% committed aggravated robbery. It should be noted, however, that despite these tendencies, some women who take a leadership role in breaking the law feel that men neither encourage nor coerce them into crime (Alarid et al., 1996). It can be argued that such women made person decisions to offend at their own free will. Becker and McCorkel (2011) conducted a study to explain how co-offending impacted women’s participation in crime. The researchers used data reported to NIBRS-participating police departments from 2002 to 2008 to test their hypothesis. They found that 37.76% of women participated in co-offending acts of crime (Becker and McCorkel, 2011). Additionally, 21.51% co-offended with one or more men, compared to about 16% of women co-offending with other women. The results of the researchers’ study indicated women were more likely to work alone for most crimes except robbery (56.74% worked with men, compared to 24.43 worked alone, and 18.82% worked with other women), gambling (54.19% worked with men, compared to 36.34 worked alone and 9.47% worked with other women), weapon law violations (52.25% worked with men, compared to 39.80% worked alone and 7.95% worked with other women), and drug sales (57.76% worked with men, compared to 34.46 worked alone and 7.78% worked with other women) (Becker and McCorkel, 2011). Additionally, compared to men, women made up the smallest proportion of offenses for the majority of crimes (Becker and McCorkel, 2011). After calculating differences in ratios of offending alone, with other women, or with men the results of Becker and McCorkel’s (2011) study indicated having a male co-offender increases a 11 woman’s likelihood of participating in most violent crimes, burglary, motor vehicle theft, drug offenses, gambling offenses, weapon law violations, and stolen property offenses. More specifically, compared to women, men have higher levels of access to criminal networks and resources (i.e., tools and information) required for committing some illicit activities (Becker and McCorkel, 2011; Leverentz, 2006). To protect their interests and concerns, men commonly exclude women from networks that would allow women to engage in illicit markets that provide significant financial benefits, produce fewer risks, and increase their status (Becker and McCorkel, 2011). For example, “financially rewarding” offenses such as burglary, robbery, and drug dealing are typically a part of a criminal market men control that denies women access (Alarid et al., 1996; Becker and McCorkel, 2011). Some men protect their interests and concerns by denying women access because women endure increased levels of attention through their social networks (Alarid, Burton, and Cullen, 2000; Alarid et al., 1996; Becker and McCorkel, 2011; Hagan, Simpson, and Gillis, 1979). However, in some cases opportunities to commit offenses become available to women when partners give them access to these networks, at least on a temporary basis. Previously, evidence for this was provided in the description of Alarid and his coauthors’ (1996) study that showed women offend with men who are more likely to commit profitable and serious crimes. It is not uncommon for male intimates to grant women access in efforts to use them as pawns to take the blame for crimes (Alarid et al., 1996; Jones, 2008; Morash, 2010; Richie, 1996). This allows males to avoid convictions by criminal justice officials. Alarid, Marquart, Burton, Cullen, and Cuveilier’ s (1996) study showed half of the women who played secondary roles in crimes with men were held accountable for the crimes while the men were not prosecuted. Thus, when women commit crimes with men, the range of crime they engage in broadens (Alarid et al., 12 1996; Becker and McCorkel, 2011; Leverentz, 2006), but they become more at risk of apprehension (Alarid et al., 1996). Economic Stressors and Crime. It is relatively well known that crime is often motivated by pressure and stress resulting from economic tension (Steffensmeier, 1980). Approximately 37% of women under correctional supervision have monthly incomes that are less than $600 (Greenfield and Snell, 1999; Richie, 2001). Women who experience more economic strain showed increased levels of participation in crime compared to women without financial stress (Leverentz, 2006; Steffensmeier, 1980). Also, Griffin and Armstrong (2003) interviewed 195 female drug-abusing probationers and found that women were less likely to sell drugs if they were employed compared to those who were unemployed. Just economic strain, however, does not provide women with opportunities to engage in profitable offenses dominated by men (Griffin and Armstrong, 2003). As shown in research reviewed above, often men must allow access to criminal opportunities in order for women to commit crimes (Becker and McCorkel, 2011; Griffin and Armstrong, 2003). Yet, in research on the effect of criminal intimate partners on women’s offending, it would be important to control for alternative explanations, such as economic strains to control for a spurious relationship. Differences between Romantic Relationships and Friendships There are obvious distinctions between friendships and romantic relationships that are explained by affectionate and sexual aspects of human nature (Carcedo, Perlman, Orgaz, Lopez, Fernandez-Rouco, and Faldowski, 2011). In comparison to friendships, romantic relationships are more inclined to affect individuals’ emotional, cognitive, and sexual systems (Carcedo et al., 2011). Specifically, within romantic relationships, connections are more likely to be formed that 13 accommodate emotional, financial, and social needs (Carcedo et al., 2011). Thus, the presence and depth of romantic relationships can have powerful impacts on women, for example by affecting their experience of stress, which may in turn influence drug use or other illegal activity (Carcedo et al., 2011). Effects of Romantic Relationships on Offenders. The intimate partner relationships most male offenders form are with prosocial women who influence them to desist from committing crime (Giordano, Cernkovich, and Rudolph, 2002; Laub, Nagin, and Sampson, 1998; Leverentz, 2006). Therefore, these heterosexual relationships that males form contribute to them not taking advantage of criminal capital and illicit markets (Giordano et al., 2002; Laub et al., 1998; Leverentz, 2006). More specifically, several studies have found that good marriages lead to desistance from crime for males (Giordano et al., 2002; Laub et al., 1998; Laub and Sampson, 2001). Good marriages are characterized by strong marital (attachment) bonds (Giordano et al., 2002; Laub et al., 1998; Laub and Sampson, 2001). Marriage to a person who is not involved in crime or drug use plays the role of altering routine activities (Giordano et al., 2002; Laub and Sampson, 2001). This is likely to occur because time dedicated to roles within the marriage is removed from time spent with deviant peers and engaging in subsequent offending behavior (Laub and Sampson, 2001; Leverentz, 2006; Warr, 1998). For males, female romantic partners often act as stakes in conformity and may exert control in the relationship (Laub and Sampson, 2001; Leverentz, 2006). Women provide increased levels of surveillance over their male partners. The watchful eye of women exercises control over males’ access to opportunities to commit unlawful acts (Laub and Sampson, 2001; Leverentz, 2006). 14 Warr (1998) examined the impact of marriage on desistance using data from the National Youth Survey. The youth survey is a continual longitudinal study. The dataset has a sample size of 1,725 individuals who were first interviewed between the ages of 11 and 17. Questioning the idea that marriage leads to desistance from crime, Warr (1998) hypothesized that marriage is attractive to those who have discontinued participation in crime. His study did not differentiate between males and females. The results of his study did not support his hypothesis that relationships with peers involved in crime end before former offenders get married to both males and females. The findings of his study showed that marriage is commonly followed by substantial declines in time spent with delinquent peers. Specifically, decreases in time spent with delinquent peers between the fifth and sixth interview were substantially larger among those respondents who married, compared to the unmarried group (Warr, 1998). Compared to male offenders, women offenders are more likely to have criminal partners. As noted in the introductory chapter, one reason is that there is a limited pool of partners for women offenders to select from because their pools are disproportionately concentrated with non-prosocial males (Carbone-Lopez and Kruttschnitt, 2010; Leverentz, 2006; Rhule-Louie and McMahon, 2007). After terms of incarceration or while on probation, most women live in the same communities where they broke the law and that are filled with crime and offenders (Carbone-Lopez and Kruttschnitt, 2010; Leverentz, 2006; Richie, 2001). Consequentially, women have little chance of establishing romantic relationships with men having no associations with crime (Carbone-Lopez and Kruttschnitt, 2010). In addition to the unavailability of prosocial men, there may be a pull to form relationships with men who break the law. Research has shown that women bond with men whose lives are similar to their lives (Rhule-Louie and McMahon, 2007). Typically, women form 15 relationships with men they are able to identify with (Carbone-Lopez and Kruttschnitt, 2010). The relationships they build are often centered on similarities based on social economic status, community orientation, and criminal histories (Rhule-Louie and McMahon, 2007). Similarities within the relationship give women a sense of comfort, acceptance, and being understood (Leverentz, 2006; Rhule-Louie and McMahon, 2007). Consistent with this general pattern, women offenders are more likely to develop romantic relationships with men battling drug addiction and resisting desistance (Leverentz, 2006). As a result, marriage or residing with a significant other has extremely different effects on offending for women and men. Specifically, for men marriage hinders their involvement in crime (Laub and Sampson, 1998). Research indicates that pairing with men from similar backgrounds often reinforce women offenders’ criminal and addictive behaviors (Rhule-Louie and McMahon, 2007; Leverentz, 2006). Leverentz’s (2009) research shows the pattern in which women offenders become involved in and are negatively influenced by criminal men. Leverentz (2006) interviewed 49 women residing in a halfway house on parole and probation. Constituting 33% of the sample, many of the women developed a new romantic relationship with a partner who had a history of substance addition. The author’s (2006) qualitative data found that many women met their new partners through their participation in recovery groups and programs. In addition, many other women encountered their new intimate companions in environments heavily populated with exoffenders. Leverentz (2006) describes the outcome of one woman’s relationship with a new partner she met at a recover meeting as destructive. The woman relapsed following her partner’s relapse, continual drug use, and possession of drugs around her (Leverentz, 2006). Criminally-involved men are often able to influence women to take responsibility for breaking the law because of the romantic nature of the relationship. Many women who offend 16 due to their intimate relationships find personal fault for the arrest of their partners and themselves (DeHart, 2008; Welle and Falkin, 2000). Some women may express guilt for being apprehended and feel they were the cause of their male companions getting arrested. Specifically, some men use control and manipulation to influence vulnerable women to feel responsible for the offense (Welle and Falkin, 2000). Love and loyalty are among the feelings that may explain how men are able to influence the women to take the blame for crimes they did not commit (Welle and Falkin, 2000). This pattern of “taking the rap” for offenses committed by intimate partners is very different from women with non-romantic partners in crime who often express concerns that their arrests have been the result of betrayal from their non-intimate counterparts or being “framed” by the police (Welle and Falkin, 2000). Also showing the negative effects of having criminal partners on offenders, Schroeder, Giordano, and Cernkovich (2007) suggest the quality of the romantic relationships has less influence on offenders’ involvement in crime than the decency of the partner. The quality of the relationship is measured by perceived happiness with the romantic partner. Respondents were asked to rate the level of happiness in their relationship from 1= extremely unhappy to 7= perfect (Schroeder et al., 2007). The authors’ results indicate the decency (level of criminal involvement) of the partner is a more critical factor that promotes criminal behavior (Schroeder, Giordano, and Cernkovich, 2007). Lending further support that intimate partners negatively influence women offenders, some women who avoid romantic relationships have greater chances of discontinuing criminal behaviors (Leverentz, 2006). Leverentz (2006) found women who were not in a relationship during recovery made up 41% of the sample. Some of these women avoided romantic relationships because they recognized previous relationships with men influenced them to offend 17 (Leverentz, 2006). One woman in particular decided she preferred to be by herself after being with partners who relapsed (Leverentz, 2006). In contrast, for some people romantic relationships with both individuals attempting to abstain from criminality and recover from substance addictions sustains a support system of exchanged personal accounts and empathy (Bui and Morash, 2010; Giordano et al., 2002; Leverentz, 2006). In such cases, partners in the relationship share their histories of drug use and crime while providing systems of support. Questions about the Influence of Intimate Partners. Despite studies showing the influence of romantic partners on women’s criminality, questions remain about how important an influence these partners are on whether women recidivate. Studies have indicated that not all women fall victim to pressures to commit crime from romantic partners (Morash, 2010). Some women participate in crime of their own free will in which they exercise personal agency (Alarid et al., 1996; Morash, 2010). This is apparent when women continue to offend when their romantic partners are not present (Morash, 2010). Additionally, some women believe they motivate other men or women to commit crimes (Alarid et al., 1996). Raising further questions about the influence of intimate partners on women’s offending, although some women offenders feel men are more trustworthy than women (Alarid et al., 1996), others partner with other women to break the law, for they view them as “good” as and more trustworthy than men (Welle and Falkin, 2000). They sometimes describe the women they offend with as individuals who have fallen prey to deceitful and overpowering men who take advantage of them (Welle and Falkin, 2000). Welle and Falkin (2000) state some women commit crimes alongside other women as a response to being in a support system. The authors (2000) describe this as a form of gender- 18 based mutual assistance found among women taken into custody with their non-intimate partners in crime. It is common for women to come to the aid of female associates when their access to illicit markets has been blocked by men. A woman explains how supporting another woman to commit a crime caused her to get in trouble: The whole thing came out where an associate of mine, I was coming out of O.T.B. (Off Track Betting) and this girl asked me to go purchase drugs for her and I told her to do it yourself, but the guys, they would not give her anything because she is very loud and boisterous, so she took the money from the undercover and gave it to me. I went and purchased the drug. I came back and gave her the drug and she gave it to the man, so there was my Catch 22 right there. (Welle and Falkin, 2000) In sum, studies of cooperation among women offenders and with non-romantic partners cast some doubt on how much influence male intimate partners have on women’s recidivism. The Effects of Abusive Partners Fifty-seven percent of women on probation have intimate partners in the past or at the present time that abused them (Harlow, 1999). As previously detailed in Daly (1992) and Richie’s (1996) research on pathways to crime, many women’s incarceration results from them being victims of intimate violence (Bui and Morash, 2010; DeHart, 2008; Richie, 1996; Welle and Falkin, 2000). Bui and Morash (2010), Morash (2010), and Richie (1996) conducted interviews with women who described instances when the abuse from their partners indirectly or directly influenced them to break the law. Welle and Falkin’s (2000) qualitative data also included some women with abusive romantic co-defendants commit crimes to support their partners, accept accountability for crime(s) their partners committed, and offend as responses to being in love and devoted (Welle and Falkin, 2000). However, Welle and Falkin did not specify the proportions of women affected in each of the different ways. 19 Some relationships with abusive men who pressure women to commit crimes have been described as a form of entrapment (Morash, 2010; Richie, 1996). Richie (1996) uses gender entrapment to explain myths and complications in African American battered women’s lives that lead them to offend. To avoid abuse from violent partners, some women take complete responsibility for legal violations committed or organized by their significant others (Morash, 2010; Richie, 1996). Dehart (2008) found intimate partner violence in combination with drug addiction played a substantial role in women’s involvement with crime. She conducted open ended interviews with 60 women incarcerated in a maximum security state prison. Her qualitative analysis indicated that some men used threats and abuse to force women to shoplift, rob, prostitute, cash fraudulent checks, or commit homicides. Additional qualitative results in her study showed that some men pressured women to take the blame for crimes in which they did not commit at all or did not commit alone with the use of persuasion, manipulation, threats, or abuse (also see Richie, 1996). For example, a 44 year old woman who was interviewed explained a situation in which she received a ticket and 2 points on her license for a car accident in which her husband was the responsible party (Dehart, 2008). The woman’s partner falsely notified law enforcement that she was driving (Dehart, 2008). The woman explained that she decided not contradict her husband’s story because she was afraid of being abused (Dehart, 2008). Some women in Dehart (2008) and Richie’s (1996) studies frequently suggested that their substance addictions and abuse from their partners left them vulnerable to influences from men. As another example, a woman in Richie’s (1996) study described her relationship with her partner who abused her. The abusive partner would intravenously give her drugs to numb her 20 pains from the abuse (Richie, 1996). Her intimate partner portrayed himself as her caretaker (Richie, 1996). In addition, he would tell her he beat her because he cared about her (Richie, 1996). The woman admitted to subsequently being addicted to drugs and adopting the behavior of purchasing the drugs on her own (Richie, 1996). There is additional evidence of the combined effects of drug addiction and abuse on women’s offending. Many women commit crimes as a means to support personal drug habits and their partners’ addictions (Morash, 2010). Moreover, supporting partners’ drug addictions helps some women avoid being battered (Richie, 1996). Numerous women that abuse drugs and/or alcohol experience distress from significant histories of abuse, including current partner abuse (Covington, 2008; Leverentz, 2006; Morash, 2010; Richie, 1996). Drugs and alcohol are often used to alleviate trauma resulting from abusive relationships or to act as filters in adverse relationships (Covington, 2008; Salisbury and Van Voorhis, 2009). Survivors of sexual, emotional, and physical abuse may experience relapses activated by memories of the past and not feeling secure within themselves or environments (Covington, 2008). The literature reviewed thus far shows how some male intimate partners victimize women when they are paroled and also when they are on probation (DeHart, 2008; Richie, 2001; Welle and Falkin, 2000). Some qualitative results from Welle and Falkin’s (2000) study indicate that treatment might help women who use drugs avoid further crime, but some close relationships with intimate partners hinder their receiving treatment. Specifically, some partners tell women what to do and say in court in efforts for the women to take the blame for crimes (Welle and Falkin, 2000). In some cases, this can cause women to be viewed as uncooperative and thus undeserving of options of alternative sanctions. Through these dynamics, power and 21 control that coercive male partners possess over convicted women limits treatment and additional alternatives to incarceration they may obtain (Welle and Falkin, 2000). Some abusive partners cause women to not obtain needed substance abuse treatment, and in some cases to accept incarceration instead. Specifically, qualitative results have indicated some women opt out of participating in programs that provide alternatives to incarceration in efforts to avoid continual harassment from abusive partners (DeHart, 2008; Welle and Falkin, 2000). It has been detailed that some women refuse treatment in the community because their abusers continue the victimization because they assume women are working with the court system (Welle and Falkin, 2000). Additionally, it is not uncommon for abused women to view sentences of imprisonment as a means to escape their crime stimulated relationships (Welle and Falkin, 2000). In such occasions, issues of abuse, which fuel the crime partnership, become mitigating factors in court or during prison intakes that is not formally documented by justice officials (Cannici, Glick, and Garmon, 1989; Welle and Falkin, 2000). As previously stated, women offending with male romantic partners are more likely to commit violent and serious offenses (Becker and McCorkel, 2011; Leverentz, 2006). The crimes women commit influence the forms of punishment they receive. Their broaden range of criminality decreases their opportunities to gain access to alternative community sanctions (Welle and Falkin, 2000). Welle and Falkin (2000) make the argument that due to the seriousness of women’s offenses, longer sentences are ordered from the courts and drug treatment as well as additional alternative sanctions is less likely to be provided. Women’s involvement in violent crimes affects their legal outcomes specified by criminal justice officials. Welle and Falkin (2000) also point out that some women’s romantic relationships with 22 criminally-involved men opens their access to criminal markets and hinders their chances of receiving treatment necessary to fight substance addictions and help rehabilitate them. Overview of Prior Research and Contribution of the Present Study A significant limitation is present in the literature because many of the studies examined provided qualitative results that did not indicate the proportions of their samples influenced in specific ways by romantic partners to be involved in crime. The tables in Appendix A summarize the findings and study design of prior research on how male romantic partners affect women’s offending. In this study, I first conducted a qualitative analysis to understand whether and how women on probation and parole describe similar or unique ways that men influence their lawbreaking, violations of probation and parole conditions, or official justice system involvement. I generated information on the proportion of women describing each type of influence by criminal romantic partners. Responding to questions the literature raises about how strong a predictor criminal romantic partners are of recidivism, I also conducted a quantitative analysis to test the hypothesis that women offenders are more likely to reoffend due to the effects of romantic relationships with a man who breaks the law. 23 CHAPTER 3: METHODOLOGY For the purpose of this current study, an available data set is used that includes women participating in gender-responsive supervision and women who experienced probation and parole supervision under a traditional approach. Gender responsive programs were designed to target the unique needs of women (Morash, 2010). Traditional programs provided supervision that did not differentiate the needs of women from men (Morash, 2010). The Women on Parole and Probation study (Morash, 2010) was conducted from 1997 to 1999 and included 402 women who were interviewed at the beginning of their involvement in their community supervision program. The women were referred to research staff from a parole or probation agent or a reporting center in two counties of a Midwestern state and two counties in a state in the Northwest. The average interview lasted 60 to 90 minutes. Women who participated in the study were compensated with $25 for each interview. A 12 month follow up interviews asked women about their social interactions, participation in their program, quality of life, and involvement in criminal activity since the first interview. Interviews were composed of open- and closed-ended questions. To obtain the qualitative data used in this thesis, women were asked the same questions about a partner getting them in trouble with the law at the beginning of the study, 6 months later, and then 6 more months later. The question was “in the last 6 months (or the last year if the second interview was skipped), (in reference to relationships with men) did you find that a date, husband, or partner got you in trouble with the law or with probation/parole violations?” “If yes, how many times?” “What was the most serious result that happened?” “For that time, how did the person get you in trouble?”. Information was also gathered on the most serious offense of partners (spouses and live-in boyfriends) during the initial interviews and in the follow up interviews. For women whose partners physically abused them, the qualitative data also included 24 responses to a question about the most serious injury received, which was asked at the beginning of the study, and during follow up interviews over the course of a year. Women convicted of at least one felony offense were eligible to participate in the study. Women’s offense patterns were identified from multiple sources of information, including women’s self-reports, supervising officers’ case notes, and surveys of supervising officers (Morash, 2010). Sixty-five percent (n=241) of the women’s dominant crime was substancecentered. These women would commit crimes in order to financially support their substance addictions; they were also convicted of possession of an illegal substance (Morash, 2010). Women who committed economic-only offenses made up 17.1% (n=63) of the complete sample. The remaining women’s offenses were limited to cultivation of marijuana (3.3%, n=12), child abuse (2.4%, n=9), violence but no drug involvement (3.5%, n=13), or drug production or distribution but no use (0.8%, n=3). Economic crimes include shoplifting and various forms of fraud (Morash, 2010). Thirteen women had crimes dominated by violence. Most of these women acted out of anger just once (Morash, 2010). Marijuana cultivation was the dominant crime for 12 women. Nine women in the sample only violated the law by maltreating children. Less than 2% of the women’s dominant crime was solely partner influenced; drug manufacture and trade; or felony driving offenses. Six women aided and abetted a criminal intimate partner (called partner-influenced). Three women were convicted for drug manufacture, distribution, or selling (called drug manufacture and trade). Only two women in the sample had crimes limited to felony driving. National statistics show that two thirds of women involved in the justice system have histories of drug use (Morash, 2010); thus the sample appears to be similar in type of criminality to the national profile of women offenders. 25 Sample For the quantitative analysis, of the 402 women who completed the first interview, a total of 77 women did not complete the final interview, so no outcome could be determined. For the purpose of this thesis, these women are omitted from the analysis to predict outcome. An additional woman was omitted from the sample because she did not provide a response indicating whether or not she was in a romantic relationship with a criminal partner. Consequentially, to test the hypothesis linking having a criminal partner to outcome, analysis focuses on a sample of 324 women. The subsample for the qualitative analysis of how men influenced women’s contact with the justice system included women who indicated at one or more interviews that a romantic partner “got you in trouble” with the law or with violations of conditions of probation and parole. The number of women in the subsample was 322. Quantitative Data Variables Used to Describe the Sample Women self-reported their age and race at the time of the first interview. The average age for women was approximately 31 years old. Race was operationalized as 1=Asian, 2=Black or African American, 3=White or Caucasian, 4=Native American or Indian, and 5=other/answered more than one. Women reported their marital status at the time of the first interview as well. The variable was categorized as 1=never married, 2=divorced, 3=separated, 4=live with significant other but married or divorced, 5=married, and 6=widowed. th Women also reported their highest education level, which was categorized as 6=6 grade, th th th th th th 7=7 grade, 8=8 grade, 9=9 grade, 10=10 grade, 11=11 grade, 12=12 grade, 13=some th th technical school, but no 12 grade, 14= GED certificate, 15=some technical school and 12 26 grade, 16=some college, and 17=bachelors degree. The average women in the sample did not th complete the 12 grade. Women self-reported whether or not they resided in a neighbor with high crime. The variable is categorized as 0=no and 1=yes. Dependent Variables Two dependent variables are used to measure recidivism. One variable indicates whether or not any source of information – official records, interviews, or probation/parole officer case notes – showed that women committed any new crimes or continued drug use while under supervision. This composite variable is operationalized as 0=no, 1=yes. Morash (2010) categorized the women based on all available information. The second dependent variable is based on a 4-item scale that indicates a combined measure of official records of recidivism and violations. If official records indicated a violation, an arrest, a conviction or a revocation of community supervision, a woman was categorized as breaking the law according to official records. This was coded as 1 for breaking the law and 0 for not breaking the law Independent Variables The main independent variable is women’s self report that their partner was a criminal. The variable was categorized as 0=no and 1=yes. Several variables in the data set were selected for analysis to control for known predictors of recidivism to determine whether a relationship between a criminal partner and recidivism is spurious. The abuse variable indicates a report of being physically assaulted by a partner. The variable is categorized as 0=no and 1=yes. An additional predictor variable measures what type of probation/parole program women were involved in. Gender responsive programs are designed to target the unique needs of women (Morash, 2010). Traditional programs provide supervision in a form that does not differentiate the needs of women from men (Morash, 2010). The variable is categorized as 0=no and 1=yes. 27 Since the type of supervision could influence outcome, it was important to use this as a control variable. The Level of Supervision Inventory-Revised (LSI-R) is a classification tool used in the original study to measure offenders’ risks of reoffending and to identify treatment needs (Reisig et al., 2006). The tool is gender-neutral and often used in correctional setting. Research has found the LSI-R misclassifies women offenders, as gendered pathways to crime are not taken into account (Reisig et al., 2006). Although this risk assessment is not ideal when applied to women, it was the best available instrument when the data were collected, and it does have some validity with women (Smith, Cullen, and Latessa, 2009). Practitioners use the quantitative assessment measurement to determine what level-high, medium, and low categories of recidivism risk-offenders are classified as (Reisig et al., 2006). The standardized tool uses 54 risk-need items to identify “Static” and “Dynamic” risks (Reisig et al., 2006). “Static” risks address aspects of a person’s life that are unchangeable, such as race, age, gender, etc. (Reisig et al., 2006). “Dynamic” risks points out personal experiences that frequently change as time goes on, such as employment, education, etc. (Reisig et al., 2006). Static and dynamic risks are classified through a 10 item subscale for the LSI-R. The original study uses a modified 10 item subscale to measure risk. The following presents the questions used to construct the scale. They are closed-ended questions with the response categories, yes or no. The LSI-R subscale items are included Appendix C. The composite score from the LSI-R can be used to indicate whether women who have men in their lives who get them into trouble were already at risk. The LSI-R scale and subscales will be used as a control variable in the logistic regression to predict outcomes from having a male intimate partner who influences outcome. 28 Coding the Qualitative Data In order to analyze the qualitative data I identified specific themes that indicate the various ways men get women into “trouble.” For the purpose of indicating how criminallyinvolved male partners got women in trouble, the unit of analysis has been established by using the unitization procedure developed by Guetzkow (1950). My advisor and I met and reviewed 25 cases retrieved from the qualitative data set. We both identified units in each case. We then discussed the number of units found in the cases reviewed. The decision was made not to count text as a unit when it was unclear if the partner caused woman to violate the law or conditions of supervision or to come into contact with the justice system. We also agreed to code a passage as a unit if the partner used substance (drugs or alcohol) with a woman or if the woman used substance in the presence of her partner. We recognized that using drugs with a person or being with a person who drinks or uses drugs does not necessarily influence a person to use substances, so this theme will be analyzed separately. Guetzkow’s U was .000 (Guetzkow, 1950), signifying agreement in the number of units identified the cases read. Based on Guetzkow’s U our agreement was perfect. Previous to meeting with my advisor again, I constructed themes and rules to code the different ways partners got women into trouble. The themes were constructed by reading each case and identify similarities in the women’s descriptions of how men got women in trouble with the law. Various similarities were identified when reviewing the cases, which aided in establishing rules that defined specific themes. During the meeting, the themes and rules were read and discussed. In the discussion of the first 5 cases, there were no disagreements over how the partners got the women into trouble. The next 10 cases were reviewed independently. Coding reliability was measured using Cohen’s Kappa (Cohen, 1960). The coding reliability for the 10 29 cases was .784. The strength of the agreement is considered to be good, but it was agreed that higher reliability was necessary. We discussed areas of disagreement and improved the codebook by defining a few specific themes. After improving the codebook, 10 more cases were independently coded. The Cohen’s Kappa was then assessed at 1. Then, I completed the coding by following the rules outlined in the codebook. A higher level set of themes where then identified that indicated general ways men get women into “trouble.” A research assistant and I met and reviewed 10 cases retrieved from the qualitative data set. We both identified units in each case. Guetzkow’s U was .043, which indicated an acceptable level of agreement (Guetzkow, 1950). Previous to meeting with the research assistant, I constructed themes and rules to code the different general ways partners got women into trouble. This set of themes was constructed by reading each case and identifying general similarities in the women’s descriptions of how men got women in trouble with the law. Common similarities and subthemes were identified when reviewing the cases, which aided in establishing rules that defined the general themes. During the meeting, the themes and rules were read and discussed. In the discussion of the first 10 cases, there were a few disagreements over how the partners got the women into trouble and themes were amended. The next 22 cases were reviewed independently. Coding reliability was measured using Cohen’s Kappa (Cohen, 1960). The coding reliability for the 22 cases was .766. The strength of the agreement is considered to be good, but it was agreed that higher reliability was necessary. This led us to discussed areas of disagreement and improved the codebook by redefining a few general themes. 30 After improving the codebook, 10 more cases were independently coded. The Cohen’s Kappa was then assessed at 1. Then, I completed the coding by following the rules outlined in the codebook. Analysis of Quantitative Data For this thesis, women who complete the initial interview are described with the use of means and standard deviations for continuous variables. Categorical variables are examined with frequency distributions. Correlations are used to show the interrelationships between having a criminal partner, having an abusive partner, and based on the qualitative data, having a partner who brought women into conflict with the law. Bivariate analysis is also used to compare women who do and do not have criminal partners on the measures of recidivism. For bivariate analysis, the Chi-square statistic is used to examine the relationship between nominal level independent variables and recidivism. Contingency tables are used to present the joint frequency distribution. Independent samples ttests are used to examine the relationships between continuous independent variables and recidivism. The effects of having a criminal partner on women’s likelihood of recidivating were modeled using binary logistic regression. Each measure of recidivism was regressed on criminal partner, abuse, type of community supervision, and the LSI-R scale. Abuse, type of community supervision, and the LSI-R scale are included because they, may make the relationship of criminal partner to outcomes spurious, are abuse, type of community supervision, and the LSI-R scale. Before the logistic regression analysis was done, regression diagnostics were run to check for problems with multicollinearity. Table 3.1 shows all Variance Inflation Factors (VIF) were 31 less than 4 as well as all Tolerance Statistics were well over .6. The results of the regression diagnostics indicate the logistic regression analysis will not produce inaccurate statistics based on the independent variables being highly correlated. Table 3.1 Multicollinearity Regression Diagnostics Criminal Partner Collinearity Statistics Tolerance .848 VIF 1.180 Abuse .791 1.265 Gender Responsive .996 1.004 LSI Score Dependent Variable: Composite Recidivism .912 1.097 32 CHAPTER 4: RESULTS Description of the Sample for Quantitative Analysis Table 4.1 presents the frequency distributions for all categorical variables. Additionally, the table displays the means and standard deviations for all continuous variables in the present study. The women in this study were about 31 years of age, undereducated, and poor. The majority of the women in this thesis were White or Caucasian. Specifically, white or Caucasian women constituted 73.1% (n=234) of the sample. African American or Black women followed with a total of 45respondents accounting for 14% of the sample. Approximately 6% were Native American or Indian (n=20), or reported another race or more than one race (n=19). Less than 1% (n=2) of the women in the study were Asian. Race was missing for four women. In regards to employment status the majority of the women were not unemployed and not a homemaker. Women who were not unemployed and not homemakers made up approximately 76% (n=245) of the sample. Women who were unemployed and not homemakers comprised almost 24% (n=79) of the sample. The majority of the women (58.3%, n=189) reported living in a community that did not have high crime. Nearly 42% (n=135) of the women interviewed reported living in an area high in crime. Additionally, many (39.6%, n=128) of the women in the sample reported they had never been married. Twenty-two percent (n=71) were divorced. Approximately 20 % (n=66) of the women stated they were married during their first interviews. Nearly, 12% (n=39) of the women were separated from their partner. During the initial interviews 5.9% (n=19) of the women reported they lived with a significant other. Marital status was missing for one woman. Furthermore, the majority of women did not have a criminal partner. Women interviewed, constituting 82.1% (n=266) of the sample, reported their partner was not involved in crime. 33 Almost 18% (n=58) of the woman had a criminal partner. The majority (83%, n=269) of the woman reported they had not been assaulted by their partner. Seventeen percent (n=55) of the respondents reported they were assaulted by a partner. At 51.9% (n=168), majority of the women interviewed were in a traditional probation program. Women in a gendered responsive program constituted 48.1% (n=156) of the sample. Based on the official records the results also showed that approximately 52% (n=114) of the women did not recidivate; 47.7% of the women interviewed recidivated (n=104). More women lacked official data than the more comprehensive assessment of outcome. Official records were missing for 106 women in the sample. In contrast, when all available information was used to determine recidivism, 55.9% (n=181) of the women interviewed used drugs and/or committed new crimes. Making up 44.1% of the sample, 143 women did not commit any new crimes and did not use drugs. In regards to reoffending most of the women’s risk level was moderate. The mean score th of the LSI-R is 13.56 (SD=5.05). Women with scores falling in the top 20 percentile are th classified as high. Score corresponding to the bottom 20 percentile are classified as low risk. th Women who had scores between the upper and lower 20 percentiles are classified as having medium risks to recidivate. 34 Table 4.1 Descriptive Information for all Variables Used in the Analyses (n=324) Valid % 2 45 .6 14.1 234 20 73.1 6.3 19 5.9 320 4 100 128 71 39 19 39.6 22 12.1 5.9 66 323 1 20.4 100 Total 245 79 324 75.6 24.4 100 Total Age Highest Education Level Highest Earnings Made in a Week 189 135 324 58.3 41.7 100 Total Missing Mean 31.49 13.06 1.88 N Total Missing Descriptive Variables 114 104 218 106 52.3 47.7 100 SD 8.30 2.67 1.62 Race Asian Black or African American White or Caucasian Native American or Indian Other/Answered More than One Total Missing Marital Status Never Married Divorced Separated Lives with Significant Other but Married or Divorced Married Unemployed and not a Homemaker No Yes High Crime No Yes Dependent Variables Official Recidivism No Yes 35 Table 4.1 (cont’d) Valid % 143 181 44.1 55.9 324 100 Total 266 58 324 82.1 17.9 100 Total 269 55 324 83 17 100 Total Mean N Total Descriptive Variables Composite Recidivism No Yes 168 156 324 51.9 48.1 100 SD Independent Variables Main Variable: Criminal Partner No Yes Predictor Variables: Abuse No Yes Gender Responsive Program No Yes LSI-R Score 13.56 5.05 Bivariate Analyses of the Relationship of Criminal Partner and Abuse with Recidivism Table 4.2 and table 4.3 indicate that having a criminal partner does not have a significant effect on recidivism as indicated by official measures. 36 Table 4.2 Bivariate Analysis to Compare Women with Criminal Partners on Official Outcomes of Recidivism Official Recidivism Criminal Partner No Total Yes No 90 51.7% 24 54.5% 114 52.3% 84 48.3% 20 45.5% 104 47.7% 174 100% 44 100% 218 100% Yes Total Note: Chi-square = .11, p ≥ .05 Table 4.3 Bivariate Analysis to Compare Women with Criminal Partners on Composite Indicator of Recidivism Composite Recidivism Criminal Partner No Total Yes No 122 45.9% 21 36.2% 143 44.1% 144 54.1% 37 63.8% 181 55.9% 266 100% 58 100% 324 100% Yes Total Note: Chi-square = 1.80, p ≥ .05 Table 4.4 shows women who are abused are more likely to recidivate than those who are not abused to recidivate based on any reports of recidivism. Approximately 69% of the women who were abused by their partners had recidivated either from using drugs or continuing other criminal activities, where as nearly 53% of the women who were not abused by their partners 37 had recidivated based on any reports of recidivism. Based on these percentages there is moderate variation, and there is a relationship between the composite measure of recidivism and abuse. Additionally, table 4.4 indicates Abuse is significantly related to Composite Recidivism at the bivariate level (X2 =4.70, p < .05). The Pearson Chi-square value (62.274) has a significance level of .030. Because the probability is less than .05 and the chi-square value is larger than the critical value (3.84) the relationship is found to be significant. This means there is a relationship in the population between women’s experience of abuse and having any reports of recidivism. Table 4.4 Bivariate Analysis to Compare Women with Different Outcomes of Composite Recidivism on Abuse Composite Recidivism Abuse No Total Yes No 126 46.8% 17 30.9% 143 44.1% 143 53.2% 38 69.1% 181 55.9% 269 100% 55 100% 324 100% Yes Total Note: Chi-square = 4.70, p < .05 Table 4.5 presents the results of the independent samples t-test for women with a criminal partner who are more likely than those women without a criminal partner to score higher on the LSI test based on official recidivism. It shows that there is a relationship between having a criminal partner and LSI score. Women with criminal partners score higher on the LSI test. This relationship is significant indicating women with criminal partners are at higher risk because they have higher scores. The mean score of the LSI test for women with a criminal partner is 38 higher than the mean score of the LSI test for women without a criminal partner. The average woman with a criminal partner in the sample has a LSI score of 14.86, while the average woman without a criminal partner in the sample has a LSI score of 13.27. The calculated value of F is 3.65, with a significance of .057. The significant level of F is greater than .05 indicating the variances between the variables are equal. The equal variance assumed estimate of t is used and it indicates a calculated value of -2.19 with a significance level of .029. Specifically, the mean score of the LSI test for women without a criminal partner is equal to the mean score of the LSI test for women with a criminal partner. It is concluded that the means are equal in the population. The t-test mean comparison of LSI Score and Criminal Partner is significant based on the calculated value -2.19 with a significance level of .029. Table 4.5 Mean Comparison of LSI Score and Criminal Partner Criminal Partner Yes 14.86 (N= 266) LSI Score No 13.27 (N= 58) Note: * F= .057, t= -2.19 Table 4.6 presents the bivariate correlations between the independent variables used in the multivariate analysis. The Forced Crime variable, which indicates the partner got the woman in trouble, is also included in the bivariate correlations. The Pearson coefficient for the relationship between Criminal Partner and Forced Crime is .39. The relationship is positive and its strength is moderate. The significance level of the correlation (.000) is less than .05 indicating the statistical Pearson correlation is significant. This indicates having a criminal partner is significantly correlated with being forced or influenced by a partner to commit crime in the 39 population. The Pearson coefficient for the relationship between Criminal Partner and Abuse is .39. The relationship is positive and its strength is moderate. The significance level of the correlation (.000) is less than .05 indicating the statistical Pearson correlation is significant. This indicates having a criminal partner is significantly correlated with being abused in the population. Moreover, the Pearson coefficient for the relationship between Forced Crime and Abuse is .24, indicating the relationship is positive and its strength is weak. The significance level of the correlation (.000) is less than .05 indicating the statistical Pearson correlation is significant. This suggests being forced or influenced by a partner to commit crime is significantly correlated with being abused in the population. The bivariate correlations among Criminal Partner, Forced Crime, and Abuse point out interrelationships among them. The interrelationships found suggest women who have been abused by a partner and woman who have been forced or influenced by a partner to commit crime are in interrelationships with a criminal partner. Additionally, the Pearson coefficient for the relationship between Criminal Partner and LSI Score is .15, and it is positive. Women who have a criminal partner have LSI scores indicating risk. The coefficient indicates the relationship is positive, and its strength is weak. The significance level of the correlation (.007) is less than .05 indicating the statistical Pearson correlation result is significant. This indicates being in a relationship with a criminal partner is significantly correlated with having a LSI score suggesting some level of risk. Furthermore, the Pearson coefficient for the relationship between Abuse and LSI Score is .29, and it is positive. Women who have indicated abuse from their partner have LSI Scores indicating risk. The coefficient indicates the strength of the relationship is weak. The significance level of the 40 correlation (.000) is less than .05 indicating the statistical Pearson correlation result is significant. This indicates being abused is significantly correlated with having a LSI Score suggesting some level of risk. 41 Table 4.6 Bivariate Relationships: Correlations between Criminal Partners, Forced Crime, Abuse, Gender Responsive Supervision, and LSI Score Criminal Partner Forced Crime Abuse Gender Responsive LSI Score Criminal Partner Pearson Correlation Sig. 1 Pearson Correlation Sig. .39* .000 Sig. .39* .000 1 Forced Crime Abuse Pearson Correlation .24* .000 1 Gender Responsive Pearson Correlation Sig. .03 .682 .02 .682 .15* .007 .041 .457 .106 .06 .29* 1 LSI Score Pearson Correlation Sig. -.004 .936 .000 Note: * p < .05 42 1 Multivariate Analyses Table 4.7 presents the results of the binary logistic regression of official and composite recidivism on criminal partner, abuse, participation in gender responsive community supervision, and LSI-R score. Shown for each predictor are the unstandardized logistic regression coefficients and odds ratios. The results indicate that, controlling for abuse, type of community supervision, and risk score, having a criminal partner did not significantly affect women’s likelihood of recidivating. LSI score was the only variable that had significant effects on recidivism at the p < .05-level. Specifically, each one-unit increase in LSI score increased the odds of official recidivism by a factor of 1.06 and composite recidivism by a factor of 1.15. Interaction effects were examined to determine whether the effects of having criminal partner on recidivism depended on abuse, participation in gender responsive community supervision, or LSI-R score (results not shown). Interaction terms were not significant. Table 4.7 Logistic Regression Predicting Recidivism from Criminal Partner, Abuse, Gender Responsive, and LSI Score Predictor Official Recidivism b Composite Recidivism EXP (b) b EXP (b) Constant -.948 .388 -1.857 .156 Criminal Partner -.303 .739 .183 1.201 Abuse .115 1.122 .069 1.071 Gender Responsive LSI Score .214 1.239 .340 1.405 .141* 1.152 .058* 1.060 5.10 Model Chisquare 43 38.86* Qualitative Analysis Although criminal partner did not significantly predict women’s outcomes of recidivism, the qualitative data shows how some romantic partners influence some women’s recidivism. The analysis of the qualitative data includes responses from 322 women with a romantic partner, of whom 61 indicated their partner influenced them to break the law or be in trouble with the law. Two women did not provide any response at all indicating their partner’s role in their reoffending behaviors. These women omitted from the qualitative analysis. Table 4.8 and 4.9 provides a complete analysis of the specific and general themes reflecting, how in women’s eyes, romantic partners get some women in trouble. Specific Themes Making up 6.5% of the sample with a romantic partner, 21 women indicated they recidivated because of the following: partner provided woman with drugs to manufacture, trade, or hold; woman was with her partner when drugs were manufactured, traded, or possessed; woman manufactured, traded or possessed drugs on her property or at a residence with approval or instruction from her partner. As an example, respondent 531 indicated that her husband was a drug dealer and she was caught in the home with drugs and firearms. Almost 3% (n=8) of the women were reported on to the authorities by their romantic partners, most (5) of whom were criminal. Respondent 1000’s husband told her parole officer that she was using drugs. Another woman’s fiancé lied on her to the police while she was on parole. The lie that was told was not specified, but it caused her to get in legal trouble. Seven women constituting 2.2% of the sample stated that their partners asked them to forge checks or influenced them to forge checks to support their drug addictions. Specifically, respondent 559 stated that her criminal partner made her sign stolen checks. Another woman 44 stated her partner taught her how to forge checks. Consequentially, she got caught. In addition, respondent 1504 stated that she tried to stop taking drugs, but her criminal partner kept bringing drugs home. Ultimately, she forged checks to support her substance addiction. Almost 2% (n=6) of the women had a romantic partner who supported their drug habits, forced/influenced them to use drugs, or influenced them to commit a crime to support their drug addiction. All but one of these women told the interviewer that their partners were criminal. Specifically, woman 1022 stated that her criminal husband received $5000 and used it to buy drugs, so that they both relapsed. Another woman simply said her partner got her back on drugs. Exemplifying the Stealing or Theft theme, five women (1.6%) had a criminal partner who influenced them to rob, burglarize, or commit retail fraud; committed robbery, burglary, or retail fraud with a criminal partner; or were charged with robbery, burglary, or retail fraud that their criminal partner was responsible for committing. One woman in particular stated her partner convinced her to rob her mother’s house. Less than 1% of the women’s romantic partners impacted their recidivism through the following means: Domestic Violence (0.9%, n=3), Abscond or Avoid Legal Apprehension (0.9%, n=3), Withholding Information or Evidence from Law Enforcement (0.9%, n=3), Stolen Car (0.6%, n=2), Accept Responsibility for Charge (0.6%, n=2), Economic Strain (0.6%, n=2), Car Illegally (0.3%, n=1), or Prostitution (0.3%, n=1). Thirteen (4%) women’s re-offense could not be identified within a theme or it was unclear how their intimate partner affected their recidivism. General Themes Some (5.3%, n=17) women got in trouble while under community supervision as a direct result of a crime their criminal partners committed. Furthermore, if their partners had not 45 committed a crime, attention from law enforcement could have been avoided. One woman in particular stated that her criminal partner was driving a stolen truck when the “cops” stopped them, and realized she had a warrant for her arrest. Another woman was charged with theft of DVDs after her male partner put DVDs in her backpack and she carried them out of a store. In what I have called the “Any Crime” theme, 4.7% (n=15) of the women committed a crime in partnership with their criminal partner. These women did not specify being forced or influenced to offend. Specifically, woman 515 stated that she and her partner were dealing drugs. She also indicated more of the drugs were his than hers. Eleven (3.4%) women had a criminal partner who influenced them or forced them to reoffend. For instance, a woman said her partner gave her drugs and drug money to hold on to. Consequentially, she was arrested for it. Another woman stated that her criminal partner started her on heroin, which led to her stealing things until she was finally caught. The Told on by Partner general theme applied to seven (2.2%) women. Four of these women’s partners were themselves offenders. Respondent 615 stated that her partner made a police report for violence, but the case was dismissed. Ultimately, the woman still got in trouble with her PO (probation or parole officer) because of it. Six (1.9%) women assaulted or fought with their partners, four of whom were also offenders, while under supervision. One woman in particular said she and her partner had been fighting and she was arrested for domestic abuse. Other woman provided similar responses. Almost 2% (n=5) of the women committed a crime that prevented or protected their criminal partners from getting into trouble or was held responsible for a crime that prevented or protected their criminal partners from getting into trouble. One woman stated that she was with her boyfriend, who was on probation, when he was driving without a license, they eluded the 46 police, but when they were stopped she “took the wrap” and was arrested. Another woman said her partner stole a woman’s wallet and made her forge checks and cash them. Ultimately, they both were caught, but her criminal partner did not come forward and say he forced her to commit the crime. Almost 6% (n=18) additional woman with romantic partners provided responses indicating how their partner impacted their recidivism. However, how the partners influenced them was unclear or did not fit into a general theme. 47 Table 4.8 Bivariate analysis: Comparison of Criminal Partners on Specific Trouble Outcomes Specific Themes Criminal Partner Counts and Percents No Yes Total Told on by Partner No 258 3 Yes 98.9% 1.1% 56 5 91.8% 8.2% 314 8 97.5% 2.5% Total 261 Car Illegally No 261 0 Yes 100% 61 100% 322 100% 100% 0% 60 1 98.4% 1.6% 321 1 99.7% 0.3% Total 261 Stolen Car No 261 0 Yes 100% 61 100% 322 100% 100% 0% 59 2 96.7% 3.3% 320 2 99.4% 0.6% Total 261 100% 61 100% 322 100% No 260 1 Yes 99.6% .04% 56 5 91.8% 8.2% 316 6 98.1% 1.9% Total 261 Drug M.T.P* No 261 0 Yes 100% 61 100% 322 100% 100% 0% 40 21 65.6% 34.4% 301 21 93.5% 6.5% Total 261 Stressful Relationship No 260 Yes 1 100% 61 100% 322 100% 99.6% .04% 60 1 98.4% 1.6% 320 2 99.4% 0.6% 100% 61 100% Total 261 *M.T.P. stands for manufacture, trade, or possession of drugs. 322 100% Drug Use 48 Table 4.8 (cont’d) Specific Themes Criminal Partner Counts and Percents No Yes Total Domestic Violence No 260 1 Yes 99.6% .04% 59 2 96.7% 3.3% 319 3 99.1% 0.9% Total 261 Prostitution No 261 0 Yes 100% 61 100% 322 100% 100% 0% 60 1 98.4% 1.6% 321 1 99.7% 0.3% Total 261 Stealing or Theft No 261 0 Yes 100% 61 100% 322 100% 100% 0% 56 5 91.8% 8.2% 317 5 98.4% 1.6% Total 261 100% 61 100% 322 100% No 261 0 Yes 100% 0% 54 7 88.5% 11.5% 315 7 97.8% 2.2% Total 261 Avoid Apprehension No 261 0 Yes 100% 61 100% 322 100% 100% 0% 58 3 95.1% 4.9% 319 3 99.1% 0.9% Total 261 Withhold Information No 261 0 Yes 100% 61 100% 322 100% 100% 0% 58 3 95.1% 4.9% 319 3 99.1% 0.9% Total 261 100% 61 100% 322 100% Forging Checks 49 Table 4.8 (cont’d) Specific Themes Criminal Partner Counts and Percents No Yes Total Accept Responsibility No 261 0 Yes 100% 0% 59 2 96.7% 3.3% 320 2 99.4% 0.6% Total 261 Economic Strain No 261 0 Yes 100% 61 100% 322 100% 100% 0% 59 2 96.7% 3.3% 320 2 99.4% 0.6% Total 261 100% 61 100% 322 100% No 255 6 Yes 97.7% 2.3% 54 7 88.5% 11.5% 309 13 96% 4% Total 261 100% 61 100% 322 100% Other 50 Table 4.9 Bivariate analysis: Comparison of Criminal Partners on General Trouble Outcomes General Themes Criminal Partner Counts and Percents No Yes Total Any Crime No 261 0 Yes 100% 0% 46 15 75.4% 24.6% 307 15 95.3% 4.7% Total 261 Force/Influence Crime No 261 0 Yes 100% 61 100% 322 100% 100% 0% 50 11 82% 18% 311 11 96.6% 3.4% Total 261 Prevents Partner No 261 0 Yes 100% 61 100% 322 100% 100% 0% 56 5 91.8% 8.2% 317 5 98.4% 1.6% Total 261 Partner’s Crime No 261 0 Yes 100% 61 100% 322 100% 100% 0% 44 17 72.1% 27.9% 305 17 94.7% 5.3% Total 261 100% 61 100% 322 100% No 258 3 Yes 98.9% 1.1% 57 4 93.4% 6.6% 315 7 97.8% 2.2% Total 261 Domestic Violence No 259 2 Yes 100% 61 100% 322 100% 99.2% 0.8% 57 4 93.4% 6.6% 316 6 98.1% 1.9% Total 261 100% 61 100% 322 100% No 253 8 Yes 83.6% 3.1% 51 10 83.6% 16.4% 304 18 94.4% 5.6% Total 261 100% 61 100% 322 100% Told on by Partner Other 51 CHAPTER 5: DISCUSSION Unfortunately, the influences on women’s desistance from crime have received limited study. Many studies focus on males and their paths of desistance. Women’s differential needs and realities are often neglected and overlooked (Bui and Morash, 2009; Laub and Sampson, 2001; Rumgay, 2004; Salisbury and Van Voorhis, 2009). Although some literature on romantic relationships’ impact on women’s offending has indicated various effects on their criminality, these studies have commonly relied on samples of women outside of Midwestern and Northwestern states. Furthermore, the research findings do not provide estimates of the proportions of women in the samples who were affected by partners who influenced their breaking the law or getting in trouble with the authorities. Additionally, no study to date has examined the effect of romantic partners on women’s recidivism using their male partners’ involvement in crime in a predictive model. This thesis attempted to discover more information about the possible effects of romantic partners on women’s recidivism. It was first hypothesized that there would be a positive relationship between criminal partners and women’s recidivism. Previous research has indicated unhealthy intimate relationships have indirect effects on women’s imprisonment (Daly, 1992; Dehart, 2008; Morash, 2010; Richie, 1996; Salisbury and Van Voorhis, 2009; Welle and Falkin, 2000). At the bivariate and multivariate levels, relationships between criminal partners and women’s recidivism were not statistically significant. Additional analyses to look at whether the effects of having a criminal partner depended on other independent variables (i.e., analysis of interaction terms) showed that this was not the case. Although the first hypothesis was not supported, the bivariate and multivariate analyses found significant relationships between LSI score and women’s recidivism. This suggests that 52 regardless of a women’s romantic relationship, the risk factors measured by the LSI put her at risk of reoffending. In addition, further analysis indicated that women with a criminal partner scored higher on the LSI. The relationship was found to be significant, suggesting that women with a criminal partner are at higher risk to start with, for reasons other than the characteristics of their partners. Additionally, at the bivariate level it was found that women who are abused by their partners are more likely to recidivate even if there are not official reports of them reoffending. Results also showed women who have experienced abuse from their partners scored higher on the LSI. This compliments Richie’s (1996) findings on abused women’s pathways into incarceration suggesting abuse is a predictor for offending. It was further hypothesized that women on probation and parole have similar or unique ways that men influence their lawbreaking, violations of probation and parole conditions, or official involvement in the justice system. The qualitative analysis found several different ways that, according to the women, men influenced their conflicts with the law. The primary purpose for the present study was to examine how romantic relationships impacted women offenders. Thus, it was hypothesized that women offenders are more likely to reoffend due to the effects of romantic relationships with a man who breaks the law. Although significant relationships were not found, the qualitative analyses described how some criminal partners could influence some women’s recidivism. Limitations The first limitation of this study is its sample of minority women. Less than 30% of the sample was minority women. Minorities’ relationships with their romantic partners may have different effects on outcomes of recidivism. Secondly, in regards to the various ways identified 53 that male partners influence women to reoffend; women’s personal agency was not considered. Some women may have independently made decisions to reoffend. In such cases male romantic partners did not influence their continued involvement in crime, even if the women perceived that they did. In addition, the present study does not assess whether women stay in relationships with the men who get them into trouble during the study period. The limitations presented could explain why statistical support for the relationship between criminal partners and women’s recidivism was not found. Recommendations The limitations of this study establish a need for future research examining the relationship between romantic partners who break the law and women’s recidivism. Thus, future research should assess the relationship with a larger sample of minority women. This could indicate whether or not minority women offenders’ relationships with romantic partners have significant effects on recidivism that challenge the results of this current study. Additionally, it is important for future research that analyzes male romantic partners’ impact on women offenders to include a measurement for personal agency. Women’s personal agency can help explain why relationships with male criminal partners do not have a significant impact on women offenders. Considering that the samples of women in the present study were drawn from two counties in a Midwestern and two counties in a Northwestern state, future research should add to the literature by analyzing the affects of intimate relationships on outcomes of recidivism in a different geographic location. Also, it is important for future research to more fully study the women with criminal partners who often score high on LSI assessments. Additional analysis could add to the literature by explaining how criminal partners contribute to risk for women who for other reasons are likely to break the law. In conclusion, although the present quantitative 54 findings of this thesis suggest there is not a significant relationship between women’s romantic relationships with criminal partners and them reoffending, it is yet apparent further research on the subject is needed. Despite the limitations of this study, in distinct ways, some women view themselves as being impacted by their romantic relationships to reoffend. Future research that incorporates a measure of personal agency, focus on minority women, and samples from a different geographical location, and that identify women’s most recent relationship status with romantic partners that got them into trouble, will offer the possibility of alternative results explaining how romantic partners involved in law breaking impact women offenders. 55 APPENDICES 56 Appendix A T able 6.1 Summary of Literature Indicating how Men get Women in Trouble How do men get women in trouble? Alarid, Marquart, Burton, Cullen, and Cuvelier, 1996 How does researcher know this? Eg. Official records, interviews? Manipulate women to Interview take the blame for Official records crime. Use women to aid in crimes (robbery). Provide women with access to illegal market such as drug offenses, larceny, robbery, and burglary. Teach women how to commit crimes (sell drug’ robberies) Use women to hold drug money or weapons. Force women to commit crimes with abuse. Introduced women to drug use. Sample How take the size? sample? 104 Recruited women convicted of felony residential court ordered boot camp program 57 Dates of study? Place of study? What about women places them at risk for men’s influence on them? July 1992 Large -Attraction to – southern the fast life September metropolitan -Gang 1993 area membership -Need for excitement -Race (African American) -Abuse -Drug use - Need for shelter protection, drugs, and money -Looking for acceptance from friends (trying to be cool) Table 6.1 (cont’d) Alarid, Burton, and Cullen, 2000 Reinforce women’s criminality. Force women to commit crimes with abuse. Paper and pencil questionnaire 1153 (122 women ) Recruited felon men and women sentenced to a residential court ordered boot camp program June 1992 Harris – county, September Texas 1993 Becker and McCorkel, 2011 Provide women with tools and raw materials, information, specialized skill sets, and distribution channels to offend (robbery, drug trafficking, burglary, gambling, kidnapping, and weapons offenses). Provide access to criminal markets. Analysis of incident, offense, and offender data reported to NIBRSparticipating police agencies 26% of U.S. Populat ion Data reported to NIBRSparticipating police agencies from 2002-2008 20022008 58 United States -Abuse -Escape from family victimization -Relationships with criminally unconvention al men (criminally involved) -Need for access to criminal markets Table 6.1 (cont’d) Bui and Morash, 2010 Cannici, Glicks, and Garmon, 1990 Men encourage drug use and criminal involvements. Encourage women to aid and support their criminal behaviors (robbery). Force them to offend with abuse and threats. Introduced women to drug use and selling. Women assume the identity of their male partner. Interviews 20 Analysis of egostrength subscale (ES) of the MMPI Women recruited from parole program 120 2007 Recruited male NA and female (60 inmates at a male/60 Federal females Correctional ) Institution 59 Mid size city in southern state -Abuse -Drug use Fort Worth, Texas -Low levels ego strength (self esteem) Table 6.1 (cont’d) ChesneyLind and Rodrigue z, 1983 Introduce women to crime (prostitution) Manipulate women’s needs for acceptance and security Interviews 16 Recruited long-term women convicts at Oahu Community Correctional Center 1982 Hawaii Covingto n, 2008 Women use addictive substances to maintain relationship. Addictive substances used to fills voids in relationship. Drugs/alcohol used to alleviate pains from abuse in relationship. Rational model Integration of addiction, trauma, and women’s psychological development perspectives NA NA NA NA 60 -Economical strain -Unstable living conditions (running away from home) -Working in bar-related female professions -Damage selfesteem -Abuse as a child -Drug use -Need for acceptance and security -Unhealthy relationships -Drug use Table 6.1 (cont’d) Dehart, 2008 De Li and MacKenzie, 2003 Force them to offend with abuse and threats. Pressure them to take the blame for crime. Cause women to retaliate violently to abuse. Abuse family in which woman is implicated for child abuse. Vandalize or steal from home causing economical strain. Drugs used as coping mechanism to abuse Force them to commit crimes when they are living together. Force them to offend with abuse and threats. Open ended interviews 60 Recruited women 1870 in maximum security state prison NA NA -Having incurable medical conditions (HIV) -History of child abuse -Abuse in relationship -Low self esteem -Drug addiction -Self blame for abuse Interviews 125 Recruited male and female probationers September 1994 – March 1996 Northern Virginia -Being sympathetic to males unconventional behaviors -Relationship maintenance (crime to please or strengthen relationship -Abuse -Drug use -With criminal past hard to find conventional male 61 Table 6.1 (cont’d) Griffin and Armstrong, 2003 -Motivate women to commit crimes to maintain the relationship -Reinforce women’s criminality -Provide access to drug markets Interviews 195 Recruited women with substance abuse problems from Maricopa County’s jail facility 62 January 1999 – August 1999 Arizona -Unstable living conditions -Relationship maintenance -Weak social bonds -Living with partner (drug dealing) -Drug use -economic strain Table 6.1 (cont’d) Jones, 2008 Force them to Interviews offend (robbery, stealing, and drug use) with abuse and threats. Introducing women to drugs. Manipulation and pressure to use drugs. Men express ideas that they were taking care of them by giving them drugs. Women offending to escape abuse from partners. Persuade women to take blame for crimes. Women offend to support t and help men. Manipulation of love and care women has for them. 50 Recruited incarcerated women 21 and over with codefendant 63 December England 2004December 2005 -Drug Use -Isolation from social networks -Abuse -Ideas of love and loyalty -Homelessness -Wanting to be loved and settling in relationship -Emotional bond with partner Table 6.1 (cont’d) Leverentz, Reinforce criminal Interviews 2006 and addictive behaviors of women. Initiate crime criminal behavior. Use of woman as partner accomplice in crime. Women commit crime on behalf of men. Initiate crime criminal behavior. Morash, 2010 Rhule-Louie and McMahon, 2007 Women commit crimes to support partner’s drug habits. Force them to offend with abuse and threats. Pressure them to take the blame for crime. Force and pressure them to use drugs to avoid abuse. Reinforce criminal and addictive behaviors of women. Initiate crime criminal behavior. 49 Women recruited from current or past in 14 bed halfway house 1 year Chicago Interviews Staff surveys Case notes Official record checks 439 Recruited women on parole and probation (gendered responsive county and traditional county) 19971999 Two counties of a Midwester n state and two counties of a Northweste rn state. A review of literature NA NA NA NA 64 -Selection of nonprosocial partner (drug addictions, criminal histories) -Being released into communities with poor selection pools -Multiple sources of strain (racial/gender discrimination, victimization, abuse, financial) -Poor social bonds in relationships -Drug Use -Abuse -Having problem behaviors (antisocial behaviors) -Selecting partners based on relatedness Table 6.1 (cont’d) Richie, 1996 Women commit Life-history crimes to support interviews partner’s drug habits. Force them to offend (prostitution, burglaries) with abuse and threats. Pressure them to take the blame for crime. Force and pressure them to use drugs to avoid abuse. Women use addictive substances to maintain relationship. Addictive substances used to fills voids in relationship. Use women’s drug addictions to manipulate them Drugs/alcohol used to alleviate pains from abuse in relationship. Battered and pressure them to neglect children. Aiding men dealing drugs and other crimes. 37 Recruited women at Rikers Island Correctional Facility 65 19911992 Rikers Island, NY -Drug use -Abuse -Economical strain -Race/ethnicity (Black/African American) Table 6.1 (cont’d) Richie, 2001 Influence low (Series of qualitative research projects) 42 NA NA NA interviews Salisbury and Van Voorhis, 2009 Influence low self efficacy, depression and anxiety, substance abuse, and adult victimization. (Multivariate Analysis) Interviews and survey data 313 Recruited women probationers 20042005 Missouri Schroeder, Giordano, and Cernkovich , 2007 -Being with criminally involved romantic partner. Data drawn from Ohio Life Course Study First wave 254 (N) 127 female Recruited adolescents from juvenile correctional institute in 1982 1982 1995 2003 Ohio Second wave 210 (N) -Low self confidence -Economic conditions -Drug use -Depression -Abuse as a child and adult -Drug use -socialization Third wave 152 (N) 77 female Note: Schroeder et al. (2007) did not report the number of females included the 2 wave of the Ohio life-course study. 66 Table 6.1 (cont’d) Steffensmier, Provide women with 1980 access to illicit markets Analysis of national arrest statistics of the Uniform Crime Report U.S. population (rates at per 100,000) Welle and Falkin, 2000 (Ethnography study) -Open ended interviews -Observation of treatment activities 60 Force them to offend with abuse and threats. Teach women how to sell and distribute drugs. Encourage women to rob and steal. Control women’s economic activity and drug use. Women’s efforts to flee partner because of abuse. 67 National arrest statistics of adult males and females of the Uniform Crime Report Recruited women in drug treatment programs 19651977 United States -Need for access to illicit markets -Economical strain 2 years Portland, Oregon and New York City -Uneven distribution of authority in relationship -Drug Use -Abuse -Isolation from Family -Parole and Probation restriction -Ideas of loyalty and love Appendix B Specific Themes 1. Told on by Partner Partner informs probation/parole agent or law enforcement about woman’s illegal activities. Partner informs probation/parole agent or law enforcement about woman’s illegal activities that she denies committing. 2. Driving or Riding in Car Illegally Woman was driving or riding in car with expired tags with partner. Woman was driving or riding in car with expired tags, owned by partner. Woman riding with partner whose license is suspended. Woman driving car while license is suspended. 3. Driving or Riding in Stolen Car Woman was driving or riding in vehicle that was stolen by partner. Woman was driving or riding in stolen vehicle with partner. 4. Drug use Partner brings drugs around woman influencing her to use. Partner forces woman to use drugs. Partner provides money for drug use. Partner influences crime to support drug use. Woman uses drugs in the presences of partner. 5. Stressful Relationship Influencing Strain Woman use drugs to cope with stress coming from partner. Woman does not participate in substance abuse programs due to stressful relationship. Woman commits crime resulting from stressful partner. 6. Drug Manufacture or Trade or Possession Partner provides woman with drugs to sell, manufacture, deliver, or hold. Woman was with partner when drugs are being sold, manufactured, delivered, or possessed. Woman sold, manufactured, delivered, or possessed drugs on property or at residence with approval or instruction from partner. 7. Domestic Violence Woman assaults or fights with partner. 68 8. Prostitution Partner influences woman to prostitute. Partner is woman’s pimp. 9. Stealing or Theft Partner influences woman to rob, burglarize, or commit retail fraud. Woman commits robbery, burglary, or retail fraud with partner. Woman is charged with robbery, burglary, or retail fraud that partner committed. 10. Forging Checks Partner asks woman to forge checks. Partner influences to forge checks to support drug habit. 11. Abscond or Avoid Legal Apprehension Partner influences woman not to report to probation/parole agent. Partner influences woman not to turn herself in for arrest warrant. Woman avoids legal apprehension with partner. 12. Withholding Information or Evidence from Law Enforcement Woman aids and abets a partner. 13. Accept Responsibility for Charge Woman takes the blame for a charge on partner’s behalf. Woman takes the blame for a charge to avoid abuse. 14. Economical Strain Partner leaves woman financially unstable. 15. Other Ways the women indicate they got into trouble that is unclear or does not fit into detailed themes. 69 General Themes 1. Any Crime in Partnership Woman commits any crime in partnership with partner. This does not include woman being forced or influenced to commit crime. 2. Crime Forced/Influenced by Partner Woman is forced or influenced to commit crime with or for partner. 3. Prevents Partner’s Trouble Woman commits crime that prevents or protects partner from getting in trouble. Woman is held responsible for a crime that prevents or protects partner from getting in trouble. 4. Partner’s Crime Partner’s crime directly causes the women to get caught or apprehended by law enforcement. 5. Told on by Partner Partner informs probation/parole agent or law enforcement about woman’s illegal activities. Partner informs probation/parole agent or law enforcement about woman’s illegal activities that she denies committing. 6. Domestic Violence Woman assaults or fights with partner. 7. Other Ways the women indicate they got into trouble that is unclear or does not fit into detailed themes. 70 Appendix C LSI-R Scale Items The “Criminal History” item scale consist of “Are you presently charged with or serving time for 3 or more offenses?” “Have you ever escaped custody?” “Have you ever been charged with assault or robbery with violence?” and “Have you ever been punished for institutional misconduct?” The items that form the “Education/Employment” scale are “Were you working when you were arrested?” “Were you employed more than six months in the year before your arrest?” “Have you ever been employed for a full year?” and “Have you ever been fired from a job?” Homemaker was considered as employment. The “Education” item was assessed with “Have you ever been expelled or suspended from school?” The items used to construct the “Work and Education” scale are “Did you like school or the job you were doing before you were arrested?” “Did you get along with fellow students or workers?” and “Did you get along well with your boss or teacher?” The questions used to measure the “Financial” item scale are “Was your financial situation good before you were arrested?” and “Before your arrest, did you rely on social assistance such as welfare?” The items used to construct the scale for “Family/Marital” are “Do you have a good relationship with your partner or are you satisfied being single?” “Do you have (or, if they are deceased, did you have) a poor relationship with your parents?” “Do you have a good relationship with other relatives?” and “Has your marital partner or any member of your family been convicted of a crime?” “Alcohol/Drug Problems” is an additional subscale that was constructed “Has alcohol ever been a problem?” “Have drugs ever been a problem?” “Did you ever break the law while under the influence of drugs or alcohol?” “Did drugs or alcohol ever cause any marital or family problems?” “Did you ever have problems in school or at work because of drugs or alcohol?” and “Have you ever suffered from loss of control from drugs or alcohol intake such as drinking or taking drugs until unconscious – sneaking drinks or tokes – or blackouts?” The “Neighborhood” scale consists of “Did you like where you were living before you were arrested?” “Did you have 3 or more address changes in the 12 months before you were arrested?” and “Did you live in a high crime neighborhood?” 71 The items used to construct the “Leisure/Recreation” scale are “Did you belong to an organized club in the 12 months before you came onto supervision?” and “Did you have any hobbies or play any sports teams?” “Mental Illness” consists of items “Have you ever received an assessment from a psychologist or psychiatrist?” “Have you ever been a resident of a mental health hospital or program?” “Are you presently receiving services from a psychologist or psychiatrist?” “Before your involvement in this program, have you ever been prescribed medicine to take care of a mental illness?” and “Have you ever attempted suicide?” 72 REFERENCES 73 REFERENCES Alarid, L., Burton, V., and Cullen, F. (2000). Gender and crime among felony offenders: Assessing the generality of social control and differential association theories. Journal of Research in Crime and Delinquency, 37(2), 171–199. doi:10.1177/0022427800037002002 Alarid, L., Marquart, J., Burton, V., Cullen, F., and Cuvelier, S. (1996). Women’s roles in serious offenses: A study of adult felons. Justice Quarterly, 13(3), 431–454. doi:10.1080/07418829600093041 Becker, S. and McCorkel, J. (2011). The gender of criminal opportunity: The impact of male cooffenders on women’s crime. Feminist Criminology, 6(2), 79– 110. doi:10.1177/1557085110396501 rd Belknap, J. (2007). The invisible woman: Gender, crime, and justice. 3 ed. Belmont, CA: Thomas Wadsworth. Bloom, B., Owen, B., and Covington, S. (2004). Women offenders and the gendered effects of public policy. Review of Policy Research, 21(1), 31–48. doi:10.1111/j.15411338.2004.00056.x Bui, H., and Morash, M. (2010). The impact of network relationships, prison experiences, and internal transformation on women’s success after prison release. Journal of Offender Rehabilitation,, 49(1), 1–22. doi:10.1080/10509670903435381 Cannici, J., Glick, S., and Garmon, S. (1989). Ego-strength changes associated with incarceration in males and females. Journal of Offender Counseling Services Rehabilitation, 14(1), 31– 35. Carbone-Lopez, K., and Kruttschnitt, C. (2010). Risky relationships? : Assortative mating and women’s experiences of intimate partner violence. Crime and Delinquency, 56(3), 358– 384. doi:10.1177/0011128709333727 Carcedo, R., Perlman,, D., Orgaz,, M., López,, F., Fernández-Rouco, N., and Faldowski, R. (2011). Predictor of loneliness, sexual satisfaction, and quality of life heterosexual romantic relationships inside of prison : Partner status as predictor of loneliness, sexual 74 satisfaction, and quality of life. International Journal of Offender Therapy and Comparative Criminology, 55(6), 898– 924. doi:10.1177/0306624X10373593 Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37–46. Covington, S. (2008). Women in Prison. Women and Therapy, 21(1), 141–155. De Li, S., and MacKenzie, D. (2003). The gendered effects of adult social bonds on the criminal activities of probationers. Criminal Justice Review, 28(2), 278–298. doi:10.1177/073401680302800205 DeHart, D. (2008). Pathways to prison : Impact of victimization in the lives of incarcerated women. Violence Against Women, 14(2), 1362–1381. doi:10.1177/1077801208327018 Giordano, P., Cernkovich, S., and Rudolph, J. (2002). Gender, crime, and desistance: Toward a theory of cognitive transformation1. American Journal of Sociology, 107(4), 990–1064. doi:10.1086/343191 Glaze, L. (2002). Probation and parole in the united states, 2001. Washington: Bureau of Justice Statistics. Glaze, L., and Bonczar, T. (2011). Probation and parole in the united states, 2010. Washington: Bureau of Justice Statistics. Greenfield, L., and Snell, T. (1999). Women offenders. Washington: Bureau of Justice Statistics. Griffin, M., and Armstrong, G. (2003). The effect of local life circumstances on female probationers’ offending. Justice Quarterly, 20(2), 213–239. doi:10.1080/07418820300095511 Guetzkow, H. (1950). Unitizing and categorizing problems in coding qualitative data. Journal of Clinical Psychology, 6, 47–58. 75 Hagan, J., Simpson, J., and Gillis, A. R. (1979). The sexual stratification of social control: A gender-based perspective on crime and delinquency. The British Journal of Sociology, 30(1), 25–38. Harlow, C. (1999). Prior abuse reported by inmates and probationers. Washington: Bureau of Justice Statistics. Harrison, P., and Beck, A. (2005). Prison and jail inmates at midyear 2004. Washington: Bureau of Justice Statistics. Laub, J., Nagin, D., and Sampson, R. (1998). Trajectories of change in criminal offending: good marriages and the desistance process. American Sociological Review, 63(2), 225–238. Laub, J., and Sampson, R. (2001). Understanding desistance from crime. Crime and Justice, 28, 1–28. Leverentz, A. (2006). The love of a good man? Romantic relationships as a source of support or hindrance for female ex-offenders. Journal of Research in Crime and Delinquency, 43(4), 459–488. doi:10.1177/0022427806293323 Morash, M. (1999). A consideration of gender in relation to social learning and social. Theoretical Criminology, 3(4), 451–462. doi:10.1177/1362480699003004005 Morash, M. (2010). Women on probation and parole. Hanover and London: University Press of New England. Reisig, M., Holtfreter, K., and Morash, M. (2006). Assessing recidivism risk across female pathways to crime. Justice Quarterly, 23(3), 384–405. doi:10.1080/07418820600869152 Rhule-Louie, D., and McMahon, R. (2007). Problem behavior and romantic relationships: Assortative mating, behavior contagion, and desistance. Clinical Child and Family Psychology Review, 10(1), 53–100. doi:10.1007/s10567-006-0016-y Richie, B. (1996). Compelled to crime: The gender entrapment of battered black women. New York: Routledge. 76 Richie, B. (2001). Challenges incarcerated women face as they return to their communities: Findings from life history interviews. Crime and Delinquency, 47(3), 368–389. doi:10.1177/0011128701047003005 Rumgay, J. (2004). Scripts for safer survival: Pathways out of female crime. The Howard Journal of Criminal Justice, 43(4), 405–419. doi:10.1111/j.1468-2311.2004.00338.x Salisbury, E., and Van Voorhis, P. (2009). Gendered pathways: A quantitative investigation of women probationers’ paths to incarceration. Criminal Justice and Behavior, 36(6), 541– 566. doi:10.1177/0093854809334076 Schroeder, R., Giordano, P., and Cernkovich, S. (2007). Drug use and desistance processes. Criminology, 45(1), 191–222. doi:10.1111/j.1745-9125.2007.00076.x Smith, P., Cullen, F., and Latessa, E. (2009). Can 14,737 women be wrong? A meta-analysis of the LSI-R and recidivism for female offenders. Criminology and Public Policy, 8(1), 183–208. doi:10.1111/j.1745-9133.2009.00552.x Snyder, H. (2011). Arrest in the united states, 1980-2009. Washington: Bureau of Justice Statistics. Steffensmeier, D. (1980). Sex differences in patterns of adult crime, 1965-77: A review and assessment. Social Force, 58(4), 1080–1108. Warr, M. (1998). Life-course transitions and desistance from crime. Criminology, 36(2), 183– 216. Welle, D., and Falkin, G. (2000). The everyday policing of women with romantic codefendants. Women and Criminal Justice, 11(2), 45–65. doi:10.1300/J012v11n02_03 77