UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Female offenders : attachment & parenthood Swihart, Gayla 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-854124.pdf [ 10.25MB ]
Metadata
JSON: 831-1.0091293.json
JSON-LD: 831-1.0091293-ld.json
RDF/XML (Pretty): 831-1.0091293-rdf.xml
RDF/JSON: 831-1.0091293-rdf.json
Turtle: 831-1.0091293-turtle.txt
N-Triples: 831-1.0091293-rdf-ntriples.txt
Original Record: 831-1.0091293-source.json
Full Text
831-1.0091293-fulltext.txt
Citation
831-1.0091293.ris

Full Text

FEMALE OFFENDERS: ATTACHMENT & PARENTHOOD by GAYLA SWIHART B.A., The U n i v e r s i t y of B r i t i s h Columbia, 1996 M.A., The U n i v e r s i t y of B r i t i s h Columbia, 1998 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES (Department of Psychology) We accept t h i s t h e s i s as conforming to the reqjuired standard THE UNIVERSITY OF BRITISH COLUMBIA J u l y , 2002 © Gayla Swihart, 2002 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Many women in prison are likely to have children, but for various reasons may not be playing the role of primary caregiver prior to incarceration, and may be unable to fulfill this role when released. This research was designed to investigate the interpersonal difficulties that may contribute to the parenting problems of this unique population. Female offenders have an increased likelihood of personality pathology and history of abuse, creating a set of circumstances that affect these women's ability to have satisfactory relationships. This may be manifested in an orientation towards relationships that could be classified as an insecure attachment style, although it is not the relationship per se that is classified as an insecure attachment style; rather, the insecure attachment style (or pattern) refers to the individual's view of themselves and others, and their attachment style may then influence their responses to others in intimate relationships (e.g., their warmth, caregiving, jealousy, ability to trust, etc). The female offender population provides a distinctive opportunity for investigating the relationships between attachment styles and incarcerated women's personality pathology, experiences with their primary caregivers, as well as their own experiences as mothers. The attachment styles of one hundred eleven incarcerated female inmates were assessed using the Relationship Questionnaire (RQ; Bartholomew & Horowitz, 1991) and the Experience in Close Relationships Inventory (ECRI; Brennan, Clark, & Shaver, 1998), and a sub-sample of these women were interviewed using the Peer Attachment Interview (Bartholomew & Horowitz, 1991). Offenders were also assessed for quality of relationships with parents using the Egna Minnen Betraffande Uppfostran (EMBU; Perris, Jacobsson, Lindstrom, von Knorring, & Perris, 1980), as well as with children using the Parental Stress Scale (PSS; Berry & Jones, 1995) and a Parental Bonding Interview (created specifically for this research to gather information about parent-child relationships, the level of caregiving provided to the child, and the mother's subjective account of her parenting experiences). The SCID-Il Clinical Interview (First, Gibbon, Spitzer, Williams, & Benjamin, 1997) was used to assess for personality disorder characteristics, which have been found in previous research to be prevalent in female offender populations. It was found that three-quarters of the women in the sample had given birth to children, however, 88.4% were no longer playing the role of primary caregiver upon incarceration. Nearly 61% of respondents claimed that social services had taken at least one child away, and 38.2% claimed that they had voluntarily given their children away. Reasons for no longer having custody varied, but the majority of women (70%) reported that drug and alcohol abuse was a major factor. Further, it was found that women in the sample were characterized by insecure attachment styles. Characteristics of personality disorders were present in over 73% of the sample. Results showed a high rate of abuse in childhood by primary caregivers, high levels of parental rejection, and low levels of parental warmth. In addition, participants reported unstable relationships with the fathers of their children. Further, the presence of an insecure attachment style was related to higher mean number of personality disorder characteristics, higher reported means of maternal and paternal rejection, and higher reported means of emotional and physical abuse by fathers. Impression management was found to be an issue with the respect to the majority of the measured utilized in the research. The goal of this research was to attain a greater understanding of the interpersonal difficulties of the female offender population, and how family, romantic, and peer relationships are related to parenting difficulties for these women. These issues, as well as limitations surrounding the use of current attachment measures, are discussed herein. Table of Contents A B S T R A C T ii T A B L E O F C O N T E N T S IV L I S T O F T A B L E S VIII D E D I C A T I O N I X A C K N O W L E D G M E N T S X P R E F A C E X I I N T R O D U C T I O N 1 Female Offenders 2 History of Abuse 3 Female inmate as a parent 7 ATTACHMENT 9 History of Attachment P Working models 16 ADULT ATTACHMENT 22 Anxiety & Avoidance 26 A Reconceptualization of Adult Attachment Theory 27 Continuity and Stability of Attachment Styles Across the Lifespan 39 Intergenerational transmission 45 Trauma and Abuse 48 PERSONALITY DISORDERS 54 CAREGIVING & PARENTAL BONDING TO CHILDREN 61 USEFULNESS OF ATTACHMENT THEORY IN RESEARCH ON FEMALE OFFENDERS 69 FEMALE OFFENDERS AND ATTACHMENT 70 SUMMARY 75 M E T H O D 78 PARTICIPANTS 78 MEASURES 79 SCID-II Personality Questionnaire 79 Background Information Questionnaire 80 Relationship Questionnaire 80 Experiences in Close Relationships Inventory 81 Egna Minnen Betraffande Uppfostran 81 Abuse 82 Parental Stress Scale 82 Balanced Inventory for Desirable Responding 83 Parental Bonding Interview 84 SCID-II Clinical Interview for DSM-IV Axis II Personality Disorders 85 Peer Attachment Interview 86 INTERVIEWERS 87 PROCEDURE 88 Representativeness of the Sample 91 Demographics 93 Reliability and Validity of Measures 99 Impression Management 101 ATTACHMENT STYLES 104 The Relationship Questionnaire 104 Experiences in Close Relationships Inventory 108 Peer Attachment Interview HO PERSONALITY DISORDERS I l l Attachment styles and Personality Disorders 114 Recollections of Early Childrearing 114 REPORTS OF ABUSE BY PRIMARY CAREGIVERS 116 INMATES AS PARENTS 119 Child Custody 119 PARENTAL STRESS SCALE 126 D I S C U S S I O N 130 Recapitulation of Main Findings 130 Representativeness of Sample 131 Veracity of reports 132 Impression Management 134 New Admissions 138 T H E WOMEN 139 ATTACHMENT STYLES 144 INMATES AS PARENTS 152 LIMITATIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH 161 vi 169 CONCLUSION I U REFERENCES 1 7 2 APPENDIX 1 1 8 8 APPENDIX 2 1 9 0 APPENDIX 3 1 9 4 APPENDIX 4 1 9 6 APPENDIX 5 2 0 0 APPENDIX 6 2 0 2 APPENDIX 7 2 0 5 APPENDIX 8 2 0 8 APPENDIX 9 2 0 9 APPENDIX 10 2 2 0 APPENDIX 11 • 2 2 2 vn List of Tables TABLE 1- Index Offenses of Present Sample Compared to Index Offenses of All Admissions to BCCW During Study Period. TABLE 2-Summary of Demographic Information for this Sample versus Turnbull's (1996) Sample. TABLE 3- Impression Management Scores and the EMBU, RQ, ECRI, & PSS. TABLE 4- Relationship Questionnaire Across Various Samples. TABLE 5- Comparisons Between the Current RQ Dimension Scores, Those From Another Forensic Sample, and Those From a Non-forensic Sample. TABLE 6- Experiences In Close Relationships Inventory Compared To Relationship Questionnaire for Present Sample and Non-forensic Norms. TABLE 7- Comparison Between the Current Personality Disorder Scores and Those From Another Forensic Sample, A Study Of Four Clinical Samples, and a Non-clinical Swedish Sample. TABLE 8- Results of EMBU in present sample compared to Control sample. TABLE 9- Reports of Abuse by Primary Caregivers. TABLE 10- Reasons Given By Female Inmate Sample Regarding Loss Of Children. TABLE 11-Location Of Fathers Of Female Inmates' Offspring. TABLE 12- Number of Fathers of Female Inmates' Children (for those with more than one child). TABLE 13- Primary Caregiver for Inmates Who No Longer Have Custody of Children. TABLE 14- Inmates Reports of Who Child is Closest To. TABLE 15- How Often Inmate Sees Children When Not In Prison. Dedication To my mother, Gladys Payne, for your unstinting love, support, and belief in me, and for your kind and generous spirit; To my other mother, Rosaline Beekhan, for your courage and love, and for the sacrifice you made as a teenage girl so that I could become the person that I am today; And to John DeHart, my best friend, my love, for showing me that dreams do come true and prayers are answered, and for making it safe for me to explore the boundaries of my heart and soul. Your sweet and gentle ways have turned my world into a playground, a rose garden, and a day at the beach come snow © or come shine.... You're double rainbows beside a setting sun. ix Acknowledgments This research project would not have been possible without the assistance of many people. First and foremost, I would like to thank all the women at BCCW who agreed to talk to us about your innermost thoughts and fears, faults and strengths, hopes and dreams. It took a great deal of courage to let us into your lives, even for so brief a time, and I am forever grateful. Many thanks to John Yuille for supporting, encouraging, and advising me throughout this research project (and my graduate career), to Don Dutton for your support and friendship, and Patricia Conrod, for your attention to detail and for your contribution to this thesis. Also, thanks to Liisa Galea for your editing suggestions, and to Mark Schaller for your kindness and for helping me see the 'big picture'. Thank-you to Kim Bartholomew and Daniel Perlman for your assistance in the planning stages of this project, and thank-you to Donna McLean, Tammy McGee, Robin Kelly & the other Programs and Security personnel at BCCW for facilitating this research at your institution. Your helpfulness was greatly appreciated. Thank-you to SSHRC for partially funding this research through a grant to John C. Yuille. A very special thank-you to the interviewers and coders who gave so unselfishly of your time; Samantha Richer, Teresa Howell, & Alicia Spidel—you three are awesome! Carolin Klein, I truly appreciate your dedication. Thank-you also to Jennifer Lang & Karen Parhar for all the time you devoted to this project. I am also indebted to Jason Winters for your advice on various aspects of this research, to Amelie Petitclerc for your editorial comments on an early draft, and to Mike Papsdorf for saving me on a number of occasions. Diane Sirkia, many thanks to you for your editorial comments, coding, and advice on this project. I am honoured to call you my friend. Your warmth and compassion are a constant comfort to me. Tracy Lavin, for your statistical expertise, editorial comments, and most importantly, your friendship. Friends like you are very hard to come by. I would also like to thank Darcy Taylor for your unwavering support throughout my university career, and Garret Swihart, Sr. (Dad) for your encouragement and support. Thank-you to my brother, Garret Swihart, Jr., my sister-in-law, Marianne, my amazing nephew, Ty, Rob Kercher, Lisa Furlan, Lisa Robinson, Anastasia Spalding, Tim Joynt, Dave Retson, & Janet Branch. As well, my appreciation to Eric Eich, for all your good advice and for your support and encouragement, Hugues Herve for your help and your friendship, and everyone else in the forensic area for making the journey interesting. There are many other people, friends and family, who have had a hand in helping me reach this point; I trust that you know who you are. I am who I am, in part, because of who you are and because of our experiences together. Finally, thank-you to Bruce and Margie DeHart for bringing such a wonderful son into the world, for raising him to be secure and confident, kind and gentle; and Margie, for bringing him to the beach one sunny June afternoon. Thank-you to you, and Uncle Robert, Neall, and Nancy, for so warmly accepting me into your family. x Preface When I embarked upon this research project, I did so with the intention of investigating female offenders who were utilizing a recently introduced mother-child live-in program at a local prison. The purpose of this program was to allow low -risk inmates to maintain their roles as primary caregivers for their very young children (under two years of age) in apartment-style housing within the "Open Living Unit" on the prison property. I was interested in ascertaining whether there were differences in attachment styles or other interpersonal variables between mothers who were participating in this program and mothers who chose not to participate in the program. However, I soon discovered that there were only four apartments designed for this program at the institution, and further, these were rarely used. In fact, their most common use was for weekend-only visits between mothers and their children, not the stable living arrangement that I had intended to study. It became apparent that the fundamental issue that I wished to address could not be studied. Previous research (MacLean, 1997) had suggested that offenders who were mothers had enough to worry about (e.g., dealing with their addictions) and needed to address these and other problems before attempting to resume their roles as primary caregivers to their children, and these women were aware of this fact. In other words, it appeared that the women were choosing not to have their children live with them in prison, even though overall offender sentiment towards the mother-child program was found to be positive. I wanted to find out how these women saw their roles as mothers, as their perceptions were clearly different from those of the prototypical mother in the general population. Assumptions about mothers in the general population generalizing to inmate mothers might not be valid. Although attachment research has focused primarily on undergraduate and/or white middle class samples, attachment theory was developed by Bowlby (1969) to address issues surrounind interpersonal difficulties within at-risk populations; hence it was determined to be the appropriate theoretical foundation for this research. x i Introduction The Canadian Centre for Justice Statistics reports that women account for 5 % of all inmates in the Canadian correctional system (Trevethan, 1999). Many women in prison are likely to have borne children, but for various reasons were not playing the role of primary caregiver immediately before being incarcerated (MacLean, 1997). One recent study found that 75% of female offender participants had given birth to at least one child, and less than 25% of these women were playing the role of primary caregiver prior to incarceration (MacLean, 1997). One goal of this research was to explore reasons for this. Female offenders are likely to be more disadvantaged than an average citizen and many have endured difficult life circumstances including family disruption and abuse (Johnson & Rodgers, 1993). This, combined with an increased likelihood of personality pathology, creates a set of circumstances that may affect these women's ability to have satisfactory interpersonal relationships, including relationships with their children (Blanchette & Motiuk, 1996). These relationship difficulties may be related to an insecure attachment style on the part of the mothers, as well as an inability for these women to function as secure attachment figures for their children (Turnbull, 1996). Therefore, the relationship styles of female inmates with their parents, as well as towards romantic partners and peers, was investigated. The investigation of relationship styles is informed by attachment theory, an established theory that describes an individual's characteristic interpersonal style, and provides the basis for this research. An individual's current attachment orientation is derived largely from previous experiences in close relationships, particularly with their primary caregiver(s), although subsequent romantic relationships may also have a significant effect. 1 The general prediction of this research is that insecure attachment styles are greater in the female offender population than in the general population (or the subsections of it that have been studied to date), that these insecure attachment styles are related to more personality pathology, worse experiences with parents, unstable relationships with the fathers of their children, and higher levels of parenting stress. The purpose of this research was to investigate the interpersonal relationships of female offenders, particularly their experiences as parents. As this research examined a specific population, a brief review of the literature on female offenders follows to provide a clearer understanding of their exceptional circumstances. Female Offenders* The average Canadian federal2 female offender is characterized as an unmarried woman between twenty and thirty-four years old who has committed manslaughter or robbery in circumstances that would likely only occur once-in-a-lifetime: for example, killing an abusive spouse in self-defense. Further, the majority of female offenders have had children, and many of these women were caring for their children before they were incarcerated (Adelberg & Currie,1987). Recent U.S. statistics are similar (Greenfield & Snell, 1999). However, these statistics are misleading, as they do not include provincially charged (or in the U.S, State-1 There is an extensive literature on female offenders and offense types, theories of female offending, and controversies surrounding female offending. However, for the purposes of this dissertation, this discussion will be limited to those issues that are directly relevant to female offenders, their attachment styles, personality disorders, and relationships with intimate others. For those interested in reading more about the types of crimes women commit: see Johnson (1987), Robertson, Bankier & Schwartz (1987); Keeney & Heide (1994) discuss gender differences in serial murderers; Mathews, Matthews, & Speltz (1990) research female sexual offenders; Rosenblatt & Greenland (1974) & Sommers and Baskin (1993) consider female crimes of violence; Straus (1978) theorizes husband-beating; Sommer, Barnes, & Murray (1994) investigate husband beating; Widom (1978a,b) classifies female offenders into types of psychopaths; and Wilbanks (1983) discusses U.S. female homicide offenders. 2 A federal sentence is given for serious offenses, usually violent, and is of at least 2 years duration. A provincial sentence is given for less serious offenses (e.g., Break & Enter) and spans from 1 day to 2 years-less-a-day. Offenders being held on remand are awaiting trial or sentencing. 2 charged) offenders, nor do they include those women being held on remand, who when combined, make up the majority of female offenders. Further complicating this picture in the Lower Mainland of British Columbia is the fact that all of categories of female offenders are held at the same institution. Unfortunately, statistics for non-federal offenders are not published, and therefore painting an accurate picture of the representative female offender is not as clear-cut as the above statistics imply. For example, the majority of non-federal female offenders are recidivists and many have returned to prison on several occasions. Further, recent research has shown that many female offenders do not have custody of their children prior to incarceration (MacLean, 1997). This research attempted to explore the needs of the female offender population, one such consideration being an awareness and understanding of previous abuse, particularly within the family-of-origin, and the effect that being the victim of abuse can have on interpersonal relationships. History of Abuse Female offenders comprise a small proportion of'all women', but according to Johnson & Rodgers (1993), are over-represented as victims of physical and sexual abuse. These authors estimated that one half of all women will be exposed to unwanted sexual acts during their lifetime, and sadly, as may often be the case with the general population of women, female offenders may be reliant upon their abusers for emotional and financial support. In 1991, the Canadian Task Force on Federally Sentenced Women found that 80% of female federal inmates reported physical and/or sexual abuse in their lifetime, usually by intimate others (Shaw, Rodgers, Blanchette, Hattem, Thomas, & Tamarack, 1991; see also Lake, 3 1993). The U.S. Department of Justice reports that 44% of female inmates have experienced sexual and/or physical abuse in their lifetimes (Greenfield & Snell, 1999). The cause of the discrepancy in reported sexual abuse between Canadian and U.S. female offenders is unclear, and neither report includes how this item was assessed, that is, how it was defined, nor whether the information was gathered through self-report questionnaire or interview. According to McClellan, Farabee & Crouch (1997), female inmates report high levels of childhood maltreatment, and for women the rate of being victims of maltreatment increases as they grow into adults. Further, females young and old are more frequently and more seriously victimized than boys and men (Finkelhor, 1986). As well, many female offenders may leave an abusive environment for an environment that is even more dangerous, although some move to safer environments. Clearly, abuse and the problems it produces play a part in the unfortunate circumstances in which many female offenders find themselves. Sultan & Long (1988) reviewed several studies that indicated that abuse and its resulting personality disturbances leads to an increased likelihood of criminality in women. As many as 70% of prostitutes surveyed in several large studies connected early sexual exploitation and physical and emotional abuse to their choice to become a prostitute (Lukianowicz, 1972; Peters, 1976; Silbert & Pines, 1981; as found in Sultan & Long, 1988). Childhood victimization was found to be strongly correlated to severe substance abuse (Benward & Densen-Gerber, 1975; as found in Sultan & Long, 1988), and in another drug-user sample, 44% reported histories of incest (Van Buskirk & Cole, 1983; as found in Sultan & Long, 1988). Sultan, Long, Kiefer, Schrum, Selby, & Calhoun (1984) surveyed 61 violent or other serious offenders at the North Carolina Correctional Center for Women and found that 75% of their sample reported experiencing some type of serious abuse at some point in their lives. 4 More recently, research has found that as juveniles, abused and neglected females tended to be at greater risk for being arrested for violent offenses (N= 451) than control females (women who were not abused or neglected; N= 333), but this was not the case for adult females (Rivera & Widom, 1990). However, the lack of significant differences in risk of arrest for violent offenses between abused/neglected adult females and control females is not surprising given the low base rates of violent offending in adult females; in their sample only 3.4 % of females sampled were violent offenders (Rivera & Widom, 1990). Further, there are inherent limitations involved in conducting research that addresses abuse issues. For instance, when using officially reported cases of abuse or neglect, there is the possibility that the report led to disruptions in caregiving and labeling of the child and parents. These in turn may create a self-fulfilling prophecy (Pagelow, 1982; as found in Rivera & Widom, 1990). Further, these reported cases cannot be generalized to unreported cases, or cases that were handled unofficially. In addition, there is no guarantee that the individuals in the control group were not the victims of abuse or neglect, only that these were not reported (Rivera & Widom, 1990). Because there is a high incidence of childhood sexual abuse in the lives of female offenders (Tien, Lamb, Bond, Gillstrom, & Paris, 1993), it is anticipated that these victims may experience difficulties in parenthood. Studies have found that mothers who were abused by family members as children may push their children toward premature autonomy, possibly believing that the child needs to learn to take care of him- or her- self at an early age (Cole & Woolger, 1989). Incest survivors may also have less confidence as parents, feel less control, lack organization and consistency, and make fewer demands for maturity on their children (Cole, Woolger, Power, & Smith, 1992). Further exacerbating their difficulties is the finding that 5 experiencing abuse may lead to alcoholism and/or drug addiction, which research has found contributes to feelings of incompetence in mothers (Harmer, Sanderson, & Mertin, 1999). Native women have an increased risk of victimization. It is estimated that one in three Native children spend part of his or her childhood as a legal ward of the state (Fournier & Crey, 1997). Many of these children were physically and/or sexually abused in educational institutions, foster homes, or in their natural homes. Fournier and Crey (1997) suggest that the crimes of self-injury committed by Aboriginal women is a reaction to this violence, and note that for decades, more than half of female offenders in Canada have been Aboriginal women whose children are separated from parents, perpetuating the cycle of becoming government wards or delinquents themselves. The authors also note that many abused natives, like their non-native counterparts, turn to drugs and alcohol in an attempt to escape the trauma and emotional distress they experience. An unfortunate consequence is that rampant alcohol abuse leads to higher incidences of Fetal Alcohol Syndrome and Fetal Alcohol Effects (Short & Hess, 1995), which further compromises already strained family circumstances (Havens, Mellins, & Pilowski, 1996). Drug abuse leads to higher incidences of babies born HIV+ or addicted to drugs (Matheson, Schonbaum, Greenberg, & Pliner, 1997). Research has shown that a disproportionate number of FAS and FAE cases involve Aboriginal children (Fournier & Crey, 1997). This is particularly true in Vancouver's downtown east side, where many female offenders in British Columbia reside when not incarcerated. In Vancouver's downtown east side, it has been estimated that half of all babies born are damaged by alcohol or drugs and are usually apprehended from mothers at birth (Fournier & Crey, 1997). These authors add that the proper care of a baby born with FAS will cost at least $1.5 million over his or her lifetime. 6 As can be seen, female inmates, particularly those who have been sexually abused in the past and/or have addictions to drugs or alcohol, face significant obstacles (besides being incarcerated) to the task of successfully parenting their own children. Issues surrounding this topic are explored next. Female inmate as a parent The U.S. Department of Justice (Snell & Morton, 1991) found in their survey of state prisons that more than 75% of female inmates had children, and that the majority of these mothers had children under the age of 18. Interestingly, of the women who had children under 18, only one quarter of them reported that the father was caring for the child, while 90% of incarcerated men with minor children said that the child(ren) was living with the mother. This highlights an important difference between the male and female offender- the male offender can often rely upon the mother of his children to assume the role of primary caregiver, whereas the female offender usually cannot rely upon the children's father to care for the children while she is in prison. Over 50% of the women in U.S. Department of Justice survey reported that their children were living with grandparents. Approximately 10% of the women had children in a foster home, agency, or institution (Snell & Morton, 1991; see also Greenfield & Snell, 1999). It was also estimated that 46% of women with minor children had telephone contact with them at least once a week, and 45% reported mail contact at least once a week. Only 9% received prison visits from their children (Snell & Morton, 1991). The U.S. Federal Bureau of Prisons' Office of Research and Evaluation found that the percentage of women who received visits from their children is low. This may be largely due to the fact that many of these women are housed in institutions that are not in proximity to their 7 children (Gwinn, 1992). Further, it is unlikely that these families can afford to travel long distances for brief mother-child visits, and some mothers have found that their children were uncomfortable in prison visiting rooms, which were too crowded and restricted privacy and children's activities (Gwinn, 1992). Although these factors may limit mother child visits, a study by MacLean (1997) found that over 75% of the female offenders in her sample who were mothers did not live with their child before being incarcerated, and many did not intend to live with their child once released. MacLean (1997) notes that for most of the female inmates in her sample, substance abuse and living in dysfunctional families (not incarceration) were among the factors to blame for separation from their children. Further, nearly half of the mothers in her sample had at least one of their children taken away by a social service agency, and many other mothers relinquished custody of their children to the child's father, grandparents, or other relatives or friends. This fragmentation and inconsistency in availability of the mother to the child unquestionably affects the mother-child relationship. While it may be argued that a woman's relationship with her child is her most salient interpersonal relationship, an incarcerated female may have many obstacles impinging upon success in this realm. For instance, she may be having difficulties in her other relationships. She is at least temporarily displaced from her support network of friends and family, if indeed she has such a network. Further, her support network may consist of other law-breaking individuals, both male and female, and many may have a make-shift support network within the prison that may not provide the quality of support she needs. It is also likely that she has a drug or alcohol problem that prevents her from prioritizing her interpersonal relationships. She may in fact have an insecure attachment style that makes it difficult for her to have satisfying relationships in her 8 life, even with her children. More importantly, she may not be providing the quality of caregiving required to raise healthy, happy, and well-adjusted children who can function in and contribute to society. As attachment theory forms the theoretical foundation of this thesis, a review of the literature on this approach follows. Attachment Attachment theory is a developmental perspective focusing on the importance of interpersonal relationships throughout the lifespan. This includes parent-child relationships as well as peer and romantic relationships. Because it is a theory that provides a paradigm for explaining the development, maintenance, and dissolution of close relationships as well as addressing personality development, emotion regulation, and psychopathology (Fraley & Shaver, 2000), there is an extensive body of research on attachment theory . The following •section provides a detailed history of attachment theory for the reader's interest only. The core of the introduction resumes on page 18 with the section entitled "Working Models". History of Attachment Central to Freud's developmental theory is the conception that childhood relationships with parents influence later interaction styles with peers, romantic partners, and children (Ricks, 1985). Attachment theory, developed by John Bowlby and operationalized by Mary Ainsworth, similarly points to how quality of parenting is affected by one's own childhood experiences and importance of continuity between generations (Bowlby, 1979). John Bowlby graduated from the 3 The following history section includes Bowlby's early theorizing and the early research conducted by Ainsworth using the Strange Situation, which is thoroughly described herein. This, and its derivatives, are the methods most commonly used to infer attachment styles of infants and young children. 9 University of Cambridge in 1928, and while volunteering at a residential school for maladjusted children, he encountered two children who sparked his interest in how parental characteristics affect children's interaction styles. One of these was an adolescent who had been expelled from school for stealing, and was described as remote and affectionless. The other was an eight-year-old who was particularly anxious and followed Bowlby around like his shadow (Ainsworth & Bowlby, 1991; Bremerton, 1992). Bowlby began his first research experiment with 44 juvenile thieves at the London Child Guidance Clinic, a school for maladjusted children. He used a matched control group and found that the experimental group, the thieves, had experienced prolonged separation from their mothers or were deprived of maternal care much more often and were more likely to have been diagnosed as 'affectionless' than were the adolescents in the non-offending control group. Bowlby published a paper on this research entitled " Forty-four Juvenile Thieves: Their Characters and Home Lives" (Bowlby, 1944; as found in Bremerton, 1992). Mary Salter Ainsworth studied at the University of Toronto, both for her undergraduate and graduate degrees. As an undergraduate, she took a course on "security theory" (Blatz, 1966, as found in Ainsworth & Bowlby, 1991) from William Blatz, who stressed that infants and young children need to feel securely dependent on their parents in order to explore their environments. Ainsworth's dissertation was entitled "An Evaluation of Adjustment Based on the Concept of Security" (Salter, 1940; as found in Bremerton, 1992). After WWII, Mary Salter married Leonard Ainsworth and they moved to London where she was hired at the Tavistock Clinic to do research on the effect of separation from the mother in early childhood on personality development. It was there that she met John Bowlby, who was head of the Children's Department in the Clinic at the time (Ainsworth & Bowlby, 1991; Bretherton, 1992). 10 Bowlby then became interested in ethology, particularly fascinated by Lorenz's (1935) experiences with imprinting in geese, who found that social bonding was not necessarily tied to feeding (Bremerton, 1992). Bowlby also studied Harlow's (1976) work with infant monkeys, particularly noting that proximity to, or close bodily contact with, their mothers was more indicative of attachment than was the provision of food. Harlowe's (1976) research on rhesus monkeys found that infant monkeys prefer a soft cuddly mother without a feeding nipple to wire frame mother with a feeding nipple, particularly when frightened. Bowlby proceeded to write several papers pertaining to this, and contrasted his view of the active nature of attachment with the passive conception of dependence considered inevitable by traditional psychoanalytic theories. Meanwhile, Ainsworth had moved to Uganda with her husband to begin an extensive study with a sample of 28 unweaned Ganda babies and their mothers. Using an interpreter as an assistant, Ainsworth proceeded to interview these mothers and observe their interactions with their infants every two weeks over a nine-month period. Throughout this research, she was struck by the active nature of the babies' participation in the interactions. She noticed that there were differences between mother-infant pairs in formations of attachment, and these were distinguishable by whether or not there was distress upon separation, or threat of separation from the mother. These differences could also be discerned by how the infant greeted the mother upon her return from an absence. Importantly, these differences appeared to be related to maternal sensitivity. Mothers who were sensitive to infant signals provided spontaneous detail during the interviews and had securely attached infants who did not cry much and were happy to explore when the mother was nearby. Less sensitive mothers did not seem to be in tune with the needs and behaviours of their infants, and had insecurely attached infants who explored little and cried 11 a lot, even when their mothers held them. Mothers of these infants were classified as insensitive. There was a third pattern of attachment displayed by the infants, and this was characterized as not-yet attached- that is, these infants did not display any differential behaviour toward the mother (Ainsworth & Bowlby, 1991; Bremerton, 1992). In 1963, Ainsworth began an ambitious research project in Baltimore, where she had a position at Johns Hopkins. She and her research team interviewed and observed 26 families over 18 home visits, averaging 4 hours each, beginning in the first month of a baby's life, and ending when the infant was 54 weeks-old. The purpose of the lengthy interviews was to enable the families to be comfortable with the researchers' presence in order to facilitate true-to-life naturalistic observation. As a result, there was approximately 72 hours of data collected for each family. In this research, Ainsworth emphasized the meaning of behaviour patterns within the context in which they occurred, rather than merely counting the number of times a specific behaviour occurred, as though they occurred within a vacuum (Ainsworth & Bowlby, 1991; Bretherton, 1992). During the first three months, feeding, face-to-face interaction, crying, attachment-exploration, obedience, close bodily contact, approach behaviour, and affectionate contact were explored, and there were striking differences in how responsive mothers were, as well as how smoothly interactions transpired (Bretherton, 1992). Sensitivity of the mother during the first quarter was related to the quality of the mother-infant interaction during the fourth quarter and was also related to the behaviour of the infant in a now well-known procedure called the Strange Situation (Ainsworth & Bowlby, 1991; Bretherton, 1992). This procedure consists of a series of incidents, as described below: 12 Series of Incidents Used in Strange Situation Mother and infant together in a strange room with toys Mother and infant joined by female stranger Mother leaves the infant with the stranger Mother returns and the stranger leaves Mother leaves infant alone Stranger returns Mother returns SOURCE: Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Wall, S. (1978). Ainsworth found that timely and appropriate close bodily contact in response to infant signals such as crying, led to infants who cried less, explored more, were more readily comforted, and responded more positively when they were put down. Securely attached infants were distressed when their mother left the room, especially if they were in an unfamiliar situation, but became less so over time as they realized that the mother would return. Also, securely attached infants responded positively when the mother returned. On the other hand, insecurely attached infants were not easily soothed, did not seem to think that their mother was available when she was out of sight, greeted the mother grumpily or with a cry when she returned, and cried a lot in general. Some of these insecurely attached infants appeared indifferent when their mother left the room and avoided her when she returned- that is, they become rather detached, presumably as a defense mechanism. These avoidant infants had mothers who were rejecting or simply insensitive to the signals of their infants. This research illuminated the differences between the securely and insecurely attached infants, as well as between avoidant insecure infants and ambivalent-resistant insecure infants (Ainsworth & Bowlby, 1991). Note that these differences among attachment patterns in infants were related to 13 the mothers' behaviours, not to temperament or other characteristics of the infant (Crowell, Fraley, & Shaver, 1999; Crowell & Treboux, 1995). Infant temperament may be related to the quality of later attachment with the mother, but research on this is inconclusive (Bretherton & Waters, 1985; Vaughn & Bost, 1999). Research findings seem to indicate that a mother's personality characteristics and behaviour are more strongly related to child's attachment style than temperament, in that a fussy child can be calmed or further upset by the patterns of responding by the mother. In other words, a responsive mother may, in time, alter an infant's temperament from fussy to calm, and a mother with a difficult personality could perpetuate a baby's negative emotionality, or even cause a calm baby to become more irritable and fussy (Scher & Mayseless, 1997). While Ainsworth was working on the Baltimore project, Bowlby was undertaking his Attachment & Loss trilogy (Bowlby 1969, 1973, 1980). In the first volume, Attachment, Bowlby (1969) expanded on ethology and evolutionary theory and the relationship between these and attachment theory. He described the function of working models, to be discussed in detail in the next section. In Volume Two, Separation, Bowlby (1973) defined attachment behaviour as any form of proximity-attaining behaviour towards a particular individual who is usually perceived to be stronger and/or wiser. He summarized attachment theory in these three propositions: 1) confidence in the availability of an attachment figure is inversely related to an individual's level of chronic or intense fear; 2) this confidence is developed during infancy, childhood, and adolescence, and the expectations developed throughout this time tend to persist throughout life; and 3) these expectations are fairly accurate reflections of the actual experiences of these individuals. 14 In Volume Two, Bowlby also expanded upon his working model theory discriminating between working models of self and other and how these are complementary to each other. He noted the adaptiveness of attachment behaviour, and theorized that it had evolved through natural selection. Further, he hypothesized that the attachment system was one of several organized behavioural systems that were designed to ensure procreation as well as protection and survival; these included exploration, caregiving, and sexual mating. Ainsworth expanded on this conceptualization, stating that the maintenance of felt security is the goal of the attachment system (Bartholomew, 1990, Feeney & Noller, 1996). However, insecure attachment can also be adaptive in some cases. For example, if an infant has a dismissing mother who does not respond to the infant's crying, it is more adaptive for the infant to be avoidant and autonomous than to attempt to garner attention from the mother. Similarly, if an infant has a preoccupied mother whose behaviours cannot be predicted, the infant seeks to reduce anxiety about the mother's availability, should she be needed, by becoming ambivalent. These reactions are strategic and functional within this relationship, but hamper exploratory behaviour. Rather than being able to assume that the mother will be there if needed, thus allowing exploration of the world and of the other people in it, the infant's attachment system is activated and, as a result, behaviour is focused on ensuring proximity to the mother (Crowell et al, 1999; Haft & Slade, 1989; Crittenden & Ainsworth, 1989; Ainsworth, 1967). In Loss (Volume Three), Bowlby (1980) drew upon information-processing theories to account for the apparent stability of internal working models, and therefore of attachment styles. He also discussed his conceptualization of the structural organization of this dynamic system (Ainsworth & Bowlby, 1991; Bretherton, 1992). Later in his career, Bowlby concentrated his 15 efforts toward extending attachment theory out of a primarily theoretical and research domain into a more clinical and applied domain (Ainsworth & Bowlby, 1991). Attachment theory has adopted concepts from several disciplines, including psychoanalysis, object relations theory, systems theory, information processing theory, ethology, evolutionary theory, cognitive theory, and personality and social psychology among others (Ainsworth & Bowlby; 1991, Feeney & Noller, 1996). All of these areas led to the conceptualization of attachment theory that centres on the importance of expectations about the availability and responsiveness of attachment figures. These expectations become incorporated into working models, also referred to as scripts or schemas, or early representational models. It is possible that some female offenders have come from homes in which caregivers were not available or responsive to them. This may have resulted in the development of insecure attachment styles, which may have, in turn, influenced their availability and responsiveness to their own children. This lack of availability and responsiveness affects what and how much these women and children expect in their interpersonal relationships- that is, these deficits affect their internal working models. Working models Working models of attachment are developed in infancy based on experiences with caregivers, and continue to develop throughout childhood and adolescence (Bowlby, 1973). These working models are cognitive or mental representations, like maps, that allow individuals to interpret and predict the behavioural and emotional responses of their attachment figures. Early in life, children incorporate their attempts to gain comfort and security, as well as the success of these attempts, into these mental representations. Such experiences are usually 16 reasonably stable throughout childhood given a consistent pattern of caregiving, even if the consistent pattern is that of inconsistent or unstable parenting, and as a result, models are more or less solidified through repeated experience, evolving into generalized beliefs and expectations. When children do not experience consistent caregiving, these generalized beliefs and expectations may become confused. For example, Haft & Slade (1989), using the Adult Attachment Interview (AAI), found that mothers' attachment orientations were highly correlated with their attunement to infant affect. Secure mothers consistently attuned to a range of infant affect, positive and negative, while dismissing mothers only attuned to positive affect and preoccupied mothers attuned randomly to positive and negative affect. They concluded that infants learn which emotions mothers do not attune to, or attune to incorrectly, and therefore come to deny or reject these feelings, because they cannot be shared with the mother. Therefore, when a mother is defensive, confused, or in denial, this may consistently affect how attuned she is to her infant's emotional states, and if she is uncomfortable with certain kinds of affect, the infant may learn to ignore his or her own affect at a very early age. When this occurs, the infant feels helpless to influence the mother's response because the mother's responses are determined by her own internal experience, which causes interactions to be unpredictable. When a child experiences something on a regular basis it may be encoded as a schema, and although the initial experiences that led to the script may be inaccessible to memory, the actual information encoded in the script is likely to be interpreted and remembered as factual (Colin, 1996). It has been theorized that this memory is largely based on the affect associated with the memory, rather than on causes and consequences of the experience (Crittenden, 1997). For example, a child may find that his mother is sometimes not available or responsive when he is sick, depending on whether or not she is drinking alcohol. At other times, when she is sober, 17 she may be there for him when he needs her. After experiencing this several times, the boy comes to learn that people can be counted on sometimes but not other times. He may not be able to discern the difference between these incidents - that is, he may not realize that the mother's responsiveness is dependent on the presence or absence of alcohol. However he knows that there are things beyond his control, and as far as he knows, beyond anyone's control, and, further, that he is unable to determine whether or not he can expect sensitive responding from others, particularly important others. In addition, as he becomes older he may not recall specific incidences of his mother not being available to him, but he is likely to recall how he felt when he needed comfort, and his resulting disappointment when comfort was not forthcoming. As a consequence, he now has a generalized belief that others cannot be counted on to be there for him when he needs or desires it. This becomes his working model or map of interpersonal relationships. In times of stress when the attachment system is activated, these working models or scripts, like other social schemas, act automatically as a sort of heuristic or shortcut, to provide the individual with missing information and guide automatic actions and reactions to the stressor or situation. In this way, the relative danger or safety of a situation does not need reassessment upon every occurrence (Bretherton, 1985). Also, it has been suggested that the more threatened an individual feels, the less likely she is to fully evaluate the information or her response to the threat (Crittenden, 1997). Further, according to Bretherton (1987), the accuracy of detail of the internal working models may be less important than their consistency with the reality they represent. How easily a working model is accessed will depend on the amount of experience on which the working model is based (Collins & Read, 1994). A problem exists in that often these 18 working models are outdated but are resistant to change. This is due in part to a self-fulfilling prophecy (Bowlby, 1973); individuals attract and are attracted to situations and other people that confirm their already existing views, and the processing of new information is guided by existing scripts (Colin, 1996). As well, people are more likely to remember instances of agreement with their already existing expectations and previous experiences, than those instances that are outside their area of 'expertise' (Collins & Read, 1994), and these memories also serve to confirm existing views. However, Bowlby adopted the term "internal working model", originally used by Craik in 1943, because he liked the fact that it implied the ability to change and construct or reconstruct the model (Bretherton, 1985; Bretherton & Waters, 1990). Bowlby (1973) stressed that working models must change as circumstances change and as new experiences disconfirm existing models. If working models are not revised, then an individual may be using an outdated model that will misguide behavior. Models must be a reasonably accurate representation of present circumstances if they are to lead to adaptive behavior in relationships (Bowlby, 1988; Kobak&Hazan, 1991). Parental protection is the set-goal of the attachment behavioural system, and although there are different attachment strategies, the behavior is organized in order to attain the set-goal (George, 1996). According to this assumption, it is not whether the attachment style is secure or insecure that ensures the child's ability to have a functional relationship with the parent. As long as the representational system is organized, the child will be able to adapt to their parents and use feedback to monitor their behavior in order to achieve the set goal of proximity to the parent when necessary (Bowlby, 1969; George, 1996). 19 Bowlby (1973) theorized that confidence in the availability and responsiveness of a caregiver is contingent upon whether the individual judges the attachment figure to be someone who can be relied upon in times of need and whether the individual judges the self as someone towards whom others, particularly the attachment figure, is likely to respond. He suggested that these models of self and other would complement and mutually confirm one another. Bretherton (1990) also discussed the importance of dyadic transactional patterns, which are models of self and others. A child begins to develop internal working models of self and others based on his or her history of experiences with attachment figures. These are the expectations of availability and support from the caregiver, and what this availability, or lack thereof, means to the child in terms of self-esteem, self-efficacy, acceptability, and lovability (Bretherton, 1987). If a primary attachment figure is unresponsive or unavailable, this could lead to a model of others as unavailable, and a model of self as unlovable or unworthy of attention or affection. However, if the caregiver is sensitive and emotionally available, this could lead to a model of others as dependable, and a model of self as lovable and worthy of attention. These expectations bias the individual's interactions in new relationships, and may remain relatively stable throughout the lifespan, although this is not necessarily the case (Bretherton, 1987). However, what is functional in a child's relationship with her parent may not be what is functional in her relationships as an adult with other adults. In childhood, attaining the set-goal is a matter of survival, but in adulthood, relationships take on more complex dynamics. The stability of working models and possibility for change is discussed in more detail later. As can be seen from the preceding review, the purpose of the working model is to allow at least the perception of control over one's environment, and this is accomplished in a largely non-conscious or automatic way. Through experience the individual learns what to expect during 20 interpersonal interactions with others, whether or not others will be responsive, and how best to elicit a response from others when desired or needed. In infancy and childhood these expectations develop primarily within the context of the parent-child relationship. In the previous example, the young boy whose mother only responds to him when she is sober may come to learn that it best never to rely on other people. He may have found that it was easier to rely only on himself rather than be disappointed by others. Further, he may have interpreted his mother's lack of attention as an indication that he was not worthy of love, and as a result, he grows up to be a man who does not feel loveable and does not feel that others can be counted on in times of need. As an adult, this man is likely to have difficulty participating in romantic relationships, as he does not think that anyone would really love him and feels that if he did embark on a relationship, he would likely have constant worry that his partner is going to let him down- how could he/she not? He is not loveable, nor can others be trusted, even those others who are supposed to love him the most. Because drug and alcohol abuse is rampant in the female offender population, as well as other problems such as poverty and abuse, the stage is set for a lack of maternal responsiveness to children who are at a vulnerable stage in the development of their working models of attachment. If individuals do not have positive relationship experiences from which to draw when developing their working models, and if they do not as adults experience sensitive and consistent responding from their romantic attachment partners, it will be more difficult to revise faulty and maladaptive working models. According to attachment theory this has implications for treatment (West & Sheldon-Keller, 1994). Although attachment theory and research originated with infants and children, Bowlby (1979) said that attachment styles are part of our relationships "from the cradle to the grave" (p. 129). Aptly, there has been a burgeoning interest 21 in adults and attachment over the past two decades, and a brief review of this literature is considered in the next section. Adult Attachment George, Kaplan, & Main (1985) designed the Adult Attachment Interview (AAI), a structured interview that elicits information about the individual's childhood relationships with his or her parents, as well as current evaluations of these early experiences. Importantly, scoring of the AAI depends more on how the individual describes her childhood than on the content of the interview. The language used is considered to cast light upon the adult's state of mind regarding attachment relationships (Crowell & Treboux, 1995), in what Main, Kaplan, and Cassidy (1985, as found in Crowell et al, 1999) called "a move to the level of representation". There are four attachment patterns that can be identified from the interview transcripts, including autonomous or secure (feels comfortable talking about childhood, is objective, coherent, and values relationships), dismissing (has difficulty recalling specific incidents in childhood, has idealized global impressions that are contradicted by specific memories that are recalled, and does not value relationships), preoccupied (is incoherent when recounting childhood experiences, is still dependent on parents and struggles to please them, is entangled with family), and unresolved (has not adequately dealt with loss or trauma). Main and Goldwyn (1985) found links between these adult attachment styles as assessed by the AAI and the attachment type of their children, in research that took place 6 years earlier using the Strange Situation (see page 24 of this document). Secure infants generally had secure or autonomous parents, avoidant infants were more likely to have dismissing parents, anxious-ambivalent infants usually had preoccupied parents, and disorganized infants were more likely to have 22 parents who were categorized as unresolved using the AAI. As can be seen, this research has implications for intergenerational transmission of attachment styles and potential relationship difficulties. This is addressed in more detail below. Other researchers have focused on adults' relationships with their romantic partners. This was a logical transition, as attachment theory is a developmental theory that functions across the lifespan, and, beginning in adolescence, parents are replaced by peers and/or romantic partners as primary attachment figures (Fraley & Shaver, 2000). Hazan and Shaver (1987) theorized that there was a connection between attachment experiences as a child and later attachment styles within adult romantic relationships when they noticed that many adults who were either distant or overly engrossed in their relationships reported troubled childhood relationships with their parents. Hazan and Shaver (1987) noted that there are differences between child-parent attachment and adult-adult attachment, namely that adult-adult attachment usually involves reciprocal caregiving and sexuality. However, they argued that there are similarities between individuals' attachment systems in adult relationships and Ainsworth's (1978) infant attachment styles (secure, ambivalent, and avoidant). Hazan & Shaver (1987) developed a brief self-report questionnaire based on these infant attachment styles. Individuals were asked to choose which of the following best described their feelings: Avoidant: I am somewhat uncomfortable being close to others; I find it difficult to trust them completely, difficult to allow myself to depend on them. I am nervous when anyone gets too close, and often, others want me to be more intimate than I feel comfortable being. Secure: I find it relatively easy to get close to others and am comfortable depending on them and having them depend on me. I don't worry about being abandoned or about someone getting too close to me. Ambivalent: I find that others are reluctant to get as close as I would like. I often worry that my partner doesn't really love me or won't want to stay with me. I want to get very close to my partner and this sometimes scares people away. 23 Hazan & Shaver (1987) used this measure in two studies. The first was a broadly based sample (N = 620) who answered a 'love quiz' published in a local newspaper, and the second was a sample of undergraduate students (N=108). Findings from both samples were similar to those assessing American infants based on Main's classification system (Campos, Barrett, Lamb, Goldsmith, & Sternberg, 1983), i.e., 60% securely attached, 15% anxious-ambivalent, 25% avoidant. However, there is a difference in the conceptualization of the avoidant/dismissing categories between Main and Hazan & Shaver. Main found that a dismissing individual denies experiencing distress and does not value attachment needs, while Hazan & Shaver found that these individuals do in fact experience subjective distress (Bartholomew, 1990). Hazan and Shaver's (1987) newspaper sample (average age = 36) were found to be 56% secure, 19% anxious-ambivalent, 25% avoidant, and their undergraduate sample (average age =18) was found to be 56% secure, 20% anxious-ambivalent, and 23% avoidant. Note that the proportions of the various styles are very similar across three age ranges- infancy (-12 months), adolescence (18 years), and adulthood (36 years). Hazan & Shaver (1987) predicted that adults in the various attachment categories would differ in their descriptions of their childhood relationships with their parents, as well as in their . relationships with others and romantic love experiences (Shaver & Hazan, 1988; Feeney & Noller, 1996). All predictions were supported. Interestingly, it was found that insecure subjects' relationships lasted half as long on average as did secure subjects' relationships, they had a significantly higher divorce rate, and scored significantly higher on trait loneliness measures. Of course, it is also important to note that people who read newspapers are not necessarily representative of the general population, nor are university undergraduates. The authors caution that this research was used to focus on relationship qualities that differentiate the three 24 attachment styles, which may give it a trait-like appearance, but that obviously situational factors such as partner characteristics and other circumstances are likely to influence relationship characteristics. This state versus trait issue is important to bear in mind when reviewing attachment constructs and measurement tools, noting that the clear and important differences reported by Hazan & Shaver (1987) were found when dichotomizing the three attachment styles into two-secure and not secure- categories. It is argued that perhaps Hazan & Shaver's (1987) two insecure categories, preoccupied and avoidant (similar to Bartholomew's [1991] fearful) are functionally more similar than not, and the use of the secure/not secure dichotomy was most appropriate for their research. Nevertheless, Feeney & Noller (1996) purport that this research by Hazan & Shaver (1987) provided a normative account of the typical processes of adult romantic relationships. Further, they note that this research laid the groundwork for applying attachment theory to individual differences in adult relationship styles in a measurable way, bridging the infant attachment theory to that of adult romantic relationships. Hazan & Shaver's (1987) measure was modified by Levy & Davis (1988) to include a Likert scale used to indicate to what degree each description characterized the individual's feelings and experiences in relationships. This allowed the experimenter to assess the degree to which a style characterizes a person, and did not assume that the three attachment styles were mutually exclusive. However, this measurement tool was further refined and developed into multiple-item attachment scales (see Feeney, Noller, & Callan, 1994). In these scales, Hazan & Shaver's multi-sentence categories were broken down sentence-by-sentence into a number of statements to which subjects respond using a rating scale. This addressed the issue of having to 25 choose and rate a multi-statement description in an all-or-nothing style. These multi-item scales varied somewhat between researchers, and consensus on the items and important underlying dimensions remains rather elusive (Feeney & Noller, 1996; see also Mikulincer, Florian, & Tolmacz, 1990; Simpson, 1990; Simpson, Rholes, & Nelligan, 1992; Collins and Read, 1990; West, Sheldon, & Reiffer, 1982; 1987; West & Sheldon-Keller, 1992). Anxiety & Avoidance One thing that most researchers do agree upon is that there are two underlying dimensions which emerge when using multiple-item measures based on Hazan & Shaver's (1987) three categories: comfort with closeness (avoidance) and anxiety (Feeney & Noller, 1996; Griffin & Bartholomew, 1994; Brennan et al, 1998, Shaver & Clark, 1994; Fraley & Shaver, 2000; Bartholomew, Kwong, & Hart, 2001). Attachment-related anxiety involves feelings of discomfort stemming from fears of abandonment, rejection, and separation. Attachment-related avoidance or closeness involves emotional and/or behavioral reaction to attachment anxiety- that is, whether or not the individual seeks reassurance or defensively avoids attachment figures (Bartholomew et al, 2001). An individual with a low propensity for attachment anxiety combined with a low propensity for avoidant behavior in response to attachment anxiety is likely to have a secure attachment orientation. Individuals with high attachment anxiety and/or avoidance are likely to fall into one of the insecure attachment patterns. Some researchers propose that adult attachment styles may be completely accounted for by the dimensions of anxiety and avoidance (Fraley & Waller, 1998). Fraley & Shaver (2000) argue that the anxiety dimension corresponds to the feelings about self, and the avoidance dimension corresponds to feelings about others, and that this functional definition is more practical than the attachment 26 classifications in categorizing interpersonal relationship styles. This research utilizes the Experiences in Close Relationships Inventory (ECRI; Brennan, Clark, & Shaver, 1998), a multi-item measure of attachment which employs the anxiety/avoidance dimensions of attachment in assessing attachment style. The authors of the ECRI found that the test-retest reliability of the measure over a four-week time span was .70 (Brennan, personal communication, April. 2002). A Reconceptualization of Adult Attachment Theory Bartholomew (1990) did an extensive review of the literature on adult attachment and noticed that although Main's AAI and Hazan & Shaver's questionnaire categorized people as secure, avoidant, or ambivalent, the two measures were not assessing the same things. In particular, Main et al (1985) found that dismissing individuals denied experiencing distress and did not value attachment needs, while Hazan & Shaver found that these individuals did in fact experience subjective distress. It was apparent to Bartholomew (1990) that there were two types of avoidant individuals- a dismissing type that was motivated to maintain self-sufficiency, and a fearful type that was consciously afraid of anticipated rejection by others. She proposed that the two underlying dimensions of adult attachment measures could be conceptualized as model of se//" (positive versus negative) and model of others (positive versus negative), and that the combinations of these dimensions yielded four attachment orientations. The model of self was associated with how dependent the individual is on acceptance from others, the more dependent individual being characterized as having a negative model of self. The model of other was associated with how avoidant of relationships the individual is, the more avoidant individual being characterized as having a negative model of others. As a result, the four models were described as this: 27 Secure: Positive model of self/positive model of other Low level of dependence/low level of avoidance These individuals are comfortable with intimacy and autonomy Preoccupied: Negative model of self/positive model of other High level of dependence, low level of avoidance These individual are overly dependent and would be classified as preoccupied by Main, and as ambivalent by Hazan & Shaver. Dismissing: Positive model of self/negative model of other Low level of dependence/high level of avoidance These individuals deny attachment and are counter-dependent. This category aligns with Main's category of the same name. Fearful: Negative model of self/negative model of other High level of dependence/high level of avoidance These individuals display fear of attachment and are socially avoidant. This category aligns with Hazan's 'avoidant' category. Source: Bartholomew (1991). Bartholomew and Horowitz (1991) developed the Relationship Questionnaire in order to test Bartholomew's four-category model. It is similar to Hazan & Shaver's prototype questionnaire, however, it includes the two avoidant styles and instructions include a forced-choice format as well as a rating scale for each format. The following are the prototypical descriptions of the four attachment styles: Secure: It is relatively easy for me to become emotionally close to others. I am comfortable depending on others and having others depend on me. I don't worry about being alone or having others not accept me. Dismissing: I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. Preoccupied: I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. Fearful: I am somewhat uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely or to depend on them. I sometimes worry that I will be hurt if I allow myself to become too close to others. Bartholomew & Horowitz, (1991). Bartholomew's four-category, two-dimensional model is compatible with many of the other adult attachment approaches. For example, Hazan & Shaver's three categories, namely secure, ambivalent, and avoidant, closely correspond conceptually to Bartholomew's secure, 28 preoccupied, and fearful categories (Bartholomew & Shaver, 1998; Brennan, Shaver, & Tobey, 1991), and Main's avoidant category corresponds conceptually to Bartholomew's (1991) dismissing category. This four-category model also aligns with research on infant attachment, which parallels the secure, ambivalent (preoccupied), and avoidant (dismissing) types, and has also added a category called disorganized-disoriented which corresponds to Bartholmew's fearful-avoidant group (Brennan et al, 1991). Feeney et al (1994) found that fearfully avoidant individuals were characterized by extreme insecurity, which is also a primary characteristic of the disorganized-disoriented attachment style, thus lending indirect support to Brennan et al's (1991) proposition. Four measurement instruments have been developed to assess Bartholomew's four-category model. Two of these were used in the present research. The Relationship Questionnaire (RQ; see above; Bartholomew & Horowitz, 1991) is a self-report measure that, when used dimensionally, measures the correspondence between the individual and each of the attachment prototypes4. The Peer Attachment Interview (PAI; Bartholomew & Horowitz, 1991) is a one-hour semi-structured interview that measures an individual's attachment style. The PAI investigates feelings about close relationships, including romances and friendships. The interviews are audio-taped and later rated on a 9-point scale by trained coders according to how closely each interviewee parallels each of the four attachment prototypes (Griffin & Bartholomew, 1994). Bartholomew (1991), using the PAI and the Relationship Questionnaire with an undergraduate population (N = 77), found that 47% were secure, 14% were preoccupied, 18% 4 The RQ can be used in two ways- the participant can choose which of the hour paragraphs is the best description of his/her interpersonal relationship style, or the participant rates EACH of the four descriptions according to how well the description characterizes his/her relationships style. Both methods are used in the present research. 29 were dismissing-avoidant, and 12% were fearful-avoidant. Recall that the breakdown for Main's (1985) research was 60% secure attached, 15% anxious-ambivalent, 25% avoidant, and Hazan and Shaver's (1987) undergraduate sample was found to be 56% secure, 20% anxious-ambivalent, and 23% avoidant. Bartholomew's (1991) findings were similar. She found fewer securely attached individuals (47% versus 60 or 56%), slightly fewer preoccupied individuals (14% versus 19 or 20%) and slightly more avoidantly attached individuals when combining the two avoidant styles (30% versus 25%). The authors did not indicate reasons for the difference between the number of secure individuals in their study and the number of secure individuals found in Main's (1985) and Hazan & Shaver's (1987) research. It is possible that the differences in the number of securely attached individuals in the various studies is related to either content differences or types of measures used. Fraley & Waller (1998) note that there are no clear or natural category boundaries between attachment styles, rendering all such classifications arbitrary. Scharfe & Bartholomew (1994) reported test-retest correlations for females over an 8-month period of .65, indicating moderate stability of attachment styles over time. The four-category model found that, as proposed, attachment ratings that are diagonally opposed in the model (see Figure 1) showed significant negative correlations. That is, secure ratings, which are indicative of a positive model of self and a positive model of others are negatively correlated with fearful attachment ratings, which are characterized by a negative model of self and a negative model of others. Further, preoccupied ratings, which are characterized by a negative model of self and a positive model of others, were negatively correlated with dismissing attachment ratings, which are characterized by a positive model of self and a negative model of others. As well, rating in adjacent positions showed low or no correlations, with correlations 30 ranging from -.32 to .06. In other words, secure ratings (+ model of self and - model of other) were not highly correlated with preoccupied ratings (- self and + other), and dismissing ratings (+ self and - other) were not highly correlated with fearful ratings (- self and + other). Figure 1. Positive (Low) Model of Other (Avoidance) Negative (High) Model of Self (Dependence) Positive Negative (Low) (High) Secure Comfortable with intimacy and autonomy Dismissing Dismissing of intimacy Counter-dependent Preoccupied Preoccupied with relationships Fearful Fearful of intimacy Socially avoidant Figure I. Model of Adult Attachment, Scharfe & Bartholomew (1991). The validity of the four attachment styles has been demonstrated in a number of studies. Brennan, Shaver, & Tobey (1991) assessed the attachment styles in a sample of 840 college students using both the Hazan & Shaver (1987) three-category measure and four-category Relationship Questionnaire, and found significant relationships between the classifications on these measures. Of those who classified themselves as secure on the RQ, 82% also classified themselves as secure on the three-category measure, of those who classified themselves as preoccupied on the RQ, 57% classified themselves as anxious-ambivalent on the 3-category measure, of those who classified themselves as fearful on the RQ, 61% classified themselves as avoidant on the three-category measure, and of those who classified themselves as dismissing on 31 the RQ, 43% classified themselves as avoidant and 45% classified themselves as secure on the three-category measure. There is no classification on the three-category measure that closely resembles the RQ dismissing style, so the participant must choose between fearful (avoidant tendencies) and secure (autonomous tendencies), accounting for the latter results. Brennen et al (1991) also found that all correlations between corresponding ratings (e.g., fearful (RQ) and avoidant (three-category) were highly significant, ranging from r = .46 to .55, and all correlations between corresponding ratings were higher than found among any correlations for non-corresponding ratings. However, the authors do not report these latter correlations or whether any of these were significantly related to one another, only that these were lower than those found for corresponding ratings. Bartholomew & Horowitz (1991) used interviews, self-reports, and peer reports to assess the attachment styles of 77 university undergraduates (40 females, mean age = 19.6). These three sets of ratings produced similar results, demonstrating convergent validity, and results further showed that each attachment type was distinctive (as previously described) and were consistent with the 2-dimensional positive and negative models of self and other, thus demonstrating discriminant validity. As well, the authors found that self-concept measures, as assessed using the Rosenberg Self-Esteem Inventory (Rosenberg, 1965), the Fey Self-Acceptance Scale (Fey, 1955), and the Inventory of Interpersonal Problems (IPP; Horowitz, Rosenberg, Baer, Ureno, & Vilasenor, 1988; Horotwitz, Rosenberg, Ureno, Kalehzan, & O'Halloran, 1989) differentiated attachment styles with respect to model of self only, and the sociability means, as assessed by the Sociability Scale (Cheek & Buss, 1981), differentiated attachment styles with respect to the model of other only, thus demonstrating construct validity. Bartholomew (1991) provided preliminary support for her two-dimensional model of self and model of other approach. Using 32 self-report, peer report, and expert ratings of a semi-structured interview, individuals who had negative models of self (preoccupied or fearful) had significantly lower scores on self-concept than did individuals with positive models of self (secure or dismissing). Further, individuals who had positive models of others (secure or preoccupied) had significantly higher sociability scores and fewer reported interpersonal problems than those with a negative model of others (dismissive or fearful). In a second study, using 69 undergraduates (33 female, mean age = 19.5), Bartholomew & Horowitz (1991) found significant correlations between ratings of peer attachment interviews and family attachment interviews (r = .29 to .66), that is, attachment styles with family members tended to transfer to attachment styles with peers. This suggests that there is a meaningful relationship between attachment styles within the family of origin and attachment styles in adulthood, but it also clearly demonstrates that adult attachment styles are not merely representations of childhood experiences, otherwise correlations would be expected to approximate 1.00. Bartholomew & Shaver (1998) also compared the four-category self-report measure (RQ) to the Peer Attachment Interview (PAI) and the Family Attachment Interview (FAI), and found that the FAI was more related to the PAI (r = .29 to .66), and the PAI was more related to the RQ (r = .24 to .45), than the FAI was related to the RQ (r = .0 to .35). These differences were reportedly due to the similarities or differences between the methods of assessment (interview versus self-report) and domains (family versus romantic relationships). That is, the FAI and PAI were related because they are both in interview format, and the PAI and the RQ are related because they both assess romantic relationships. It was also found that the associations for corresponding ratings (e.g., dismissing on RQ and dismissing on FAI/PAI) were stronger than 33 those for non-corresponding ratings (e.g., dismissing on RQ and fearful on FAI/PAI), while none of the non-corresponding ratings were significantly positively associated with one another. These findings suggest both convergent and discriminant validity. This group of findings appears to bridge the gap between attachment styles as assessed by the AAI for childhood memories of relationships with parents and attachment styles as assessed through self-report of recent romantic relationships with adults. The Family Attachment Interview assesses childhood experiences and is significantly related to the Peer Attachment Interview, which is significantly related to the Relationship Questionnaire. This suggests that there may be a core set of interpersonal styles underlying the various attachment measures that are compatible with attachment theory's proposition that adult attachment styles are largely established during childhood through experiences with important attachment figures. Of course, the research reported thusfar was conducted on a young undergraduate sample, and therefore issues of generalizability arise. Fortunately, research has been conducted on other samples. Clinical interviews of 30 women (mean age not reported) were independently rated using both Adult Attachment Interview experts and Peer & Family Attachment Interview experts (Bartholomew & Shaver, 1998). The authors found that the classifications obtained by the three AAI categories (secure, preoccupied, dismissing) were significantly related to Bartholomew's four categories. Further, the three corresponding PAI/FAI (Bartholomew's) categories were compared to the AAI categories, producing a significant relationship, which is not surprising given that the AAI preoccupied and dismissing categories are very similar to Bartholomew's preoccupied and dismissing categories. However, the secure categories of the AAI and PAI/FAI did not correspond as well, in that of the eight women classified as secure on the AAI, only four were classified as such on the PAI/FAI. Three of the remaining four AAI secure women were 34 classified as dismissing and one was classified as preoccupied. This may be due to the fact that the AAI does not have a fourth (fearful) category, and therefore individuals who may have been classified as fearful were likely distributed among the remaining classifications. Regardless, considering the small sample size and the fact that the 30 women in this study were not interviewed using either the AAI or the PAI/FAI (according to the authors, the clinical interview used was not similar to either interview) the correspondence between the two interviews is respectable and suggests evidence for convergence. In the present research, the Relationship Questionnaire was used categorically as well as dimensionally. When the dimensional method is used with this measure, participants rate the degree to which each description fits the individual rather than requiring a decision of the best-fit description. Also, the four-category method uses prototypes, and this categorization approach "allows for the complex patterns of individual differences that may define 'types' of persons, while also recognizing that not all members of a group are equally good exemplars of that group" (Griffin & Bartholomew, 1994, p.23). For example, an individual who is primarily securely attached, and secondarily dismissing would differ in important and obvious ways from an individual who is primarily securely attached and secondarily preoccupied (Griffin & Bartholomew, 1994; Bartholomew, 1997). However, this can also be accomplished by using a multi-item scale, such as the Experiences in Close Relationships Inventory (ECRI; Brennan, Clark, & Shaver, 1998). Whereas the RQ, the PAI, and the FAI directly assess the fit between the individual and the four attachment prototypes, the ECRI is an indirect measure, in that it derives the fit between the individual and the four attachment prototypes from computed mean scores of groups of items. Rather than being forced to choose which paragraph most closely describes her attachment characteristics, which is a cruder measure and results in loss of 35 information, use of the ECRI allows an individual to choose the various components embedded within each descriptive paragraph (e.g., the RQ four paragraphs describing each attachment style) to be assessed individually. These responses are then analyzed to form a composite profile of each individual's attachment styles, as well as where they fall on the attachment/avoidance dimensions. Brennan et al (1998) found that the anxiety and avoidance scales of the ECRI were almost uncorrelated with one another, r = .11, and each correlated highly with its parent factor, r = .95 in both cases. Figure 2 demonstrates the correlations of the ECRI anxiety and avoidance scales with the subscales of the Relationship Questionnaire: Figure 2. RQ subscale Anxiety Avoidance Dismissiveness .39 -.29 Fearfulness .81 .32 Preoccupation -.13 .73 Security -.70 -.46 Brennan, Clark, and Shaver (1998) Figure 2. Correlations of the ECRI anxiety and avoidance scales with the subscales of the Relationship Questionnaire. The patterns of the anxiety and avoidance scores on the ECRI were conceptually related to Bartholomew's four attachment styles. For example, individuals who scored low on both anxiety and avoidance on the ECRI were likely to be classified as secure on the Relationship Questionnaire. Other research has found that secure attachment, as assessed using the RQ accounted for a significant amount of variance in associations between the avoidance scale of the Impact of 36 Events Scale and abuse characteristics (e.g., age of onset of abuse, type of abuse, degree of coercion used, and relationship with perpetrator; Alexander, 1993), another demonstration of the conceptual similarities between the RQ and the avoidance dimension of the ECRI. Brennan et al (1998) also compared the four attachment style categories assessed with the Bartholomew four-category measure and the ECRI using cross-tabulation and found there was substantial similarity between the two assessment schemes. However, they found that their measure (ECRI) was more conservative than the RQ in classifying people as secure, claiming that it is because the ECRI can discriminate more precisely between various levels of insecurity. That is, the RQ is more likely to categorize a person who is only mildly insecure as secure, whereas the ECRI would categorize that same person as insecure. Further, Brennan et al (1998), using a sample of 1082 university undergraduates (682 females, mean age = 18), compared the results from the ECRI with measures of attachment related to emotions, thoughts, and behaviours regarding touch and sexuality in romantic relationships. Results showed that the ECRI produced stronger relationships with these measures than the Relationship Questionnaire, suggesting that the ECRI is a more sensitive indicator of these aspects of romantic relationships. This research includes both the RQ and the ECRI in order to compare the utility of each, and more specifically to assess whether they measures demonstrate convergent validity within a female offender sample. It must be noted that there are limitations to existing attachment research. Using interviews and self-reports to investigate attachment history is problematic because the report given is retrospective, and therefore the memory may simply be a reconstruction of an event, and not an accurate memory. It has been suggested that experiences may be misremembered according to present cognitive-affective filters that have developed over time and out of 37 experiences in other relationships (Sperling & Berman, 1994; van Ijzendoorn, 1992; Waganaar, 1986). However, even an inaccurate memory of an event can be informative because it is the current representation of an individual's attachment-related experience (Collins & Read, 1994), and although it is assumed that there is continuity of development over the lifespan (Bowlby, 1969), according to Bowlby (1988) there may be a disruption in the link between early attachment experiences and adult attachment relationships (lack of continuity) due to later positive attachment experiences with a partner or a therapist. Therefore, the present research assessed romantic and peer attachment styles in adulthood rather than relying on retrospective accounts of childhood relationships with parents. However, there are also issues with regards to the stability of attachment across time, and these will addressed in the following section. Finally, it appears that disagreement exists across measures regarding the precise breakdown of the attachment styles, and research shows that individuals have varying degrees of characteristics of more than one attachment style across time and within and across relationships (Bartholomew & Horowitz, 1991). It may be the case that it is more useful to simply categorize individuals as either secure or not secure until measures can be refined to more accurately discriminate between the various insecure attachment styles and between primary and secondary attachment characteristics. To summarize, Bartholomew's (1991) four-category model distinguishes between two types of avoidant attachment- dismissing and fearful, in addition to secure and preoccupied styles. Avoidant behaviour may present similarly across situations, but cleverly, Bartholomew (1991) discovered that the motivation behind the avoidance may be a lack of interest or positive regard of others. Conversely, the avoidance may be based on fear and a lack of positive regard for the self. This two-dimensional, four-category model of self and model of others shows both 38 divergent and convergent validity, at least with undergraduate samples as well as with a sample of traumatized women, and has been used in previous research on a female offender population (Turnbull, 1998); its use here allows an attempt to replicate previous findings, as well as allowing a comparison between the Relationship Questionnaire and the Experiences in Close Relationships Inventory. The ECRI is a multi-item attachment questionnaire that assesses for the dimensions of anxiety and avoidance, includes within group variance, and provides a more detailed profile of each individual's attachment styles. However, it has been used less extensively in previous research, and therefore comparisons are difficult to make. This research utilizes Bartholomew & Griffin's (1991) Relationship Questionnaire (RQ), a self-report measure, to assess attachment orientations for both romantic and peer (platonic friendships) relationships. It was predicted that female offenders would be more likely to endorse a secure attachment style in their platonic friendships than in their romantic relationships. As well, the Experiences in Close Relationships Inventory (Brennan et al,1998) was used as an additional self-report measure of attachment orientation, in an attempt to address measurement issues and to obtain a more comprehensive assessment of attachment orientations. It was predicted that women in this sample would report higher levels of insecure attachment than found in other, non-offender samples, and that attachment security, or insecurity, would be related to a number of other interpersonal variables in this sample. Continuity and Stability of Attachment Styles Across the Lifespan As noted previously, working models of attachment-related interactions begin early in life, usually develop within the context of family settings, and tend to endure throughout the lifespan. Although some fluctuation is expected in attachment styles across time, these styles 39 appear to be reasonably stable. Assuming that such continuity exists, it was anticipated that there would be a relationship between the female offenders' early experiences with their parents and their current attachment styles. Because these working models, once formed, are thought to function largely outside conscious awareness, they should be resistant to change. However, Bowlby (1980) suggested that changes in working models are possible. He proposed that people may or may not be satisfied with their attachment orientation, and if they are not, they may be motivated to consciously attempt to change their interpersonal styles. Further, life experiences may affect the stability of attachment styles. Another possibility is that working models may become outdated and no longer effectively predict the behaviour and reactions of others, and therefore may need to be altered. Some researchers have suggested that working models may be updated in the context of later relationships outside the family. Individuals who previously developed an insecure attachment style in the context of their family of origin may later become involved in a relationship that challenges that interaction style, thus facilitating an alteration in working models about relationships (Ricks, 1985; Sroufe & Fleeson, 1986; Cohn, Silver, Cowan, Cowan, & Pearson, 1992). These findings suggest that therapy that deals with attachment issues, specifically working models, whether or not they are still adaptive, and whether or not there is evidence to disconfirm present working models, may be effective in altering an individual's tendency to behave in ways that result in self-fulfilling prophecies. This may provide an opportunity for revising and updating representational structures and may lead to more secure interactions with others (Rothbard & Shaver, 1994). Change may also occur outside of the therapeutic context. Ricks (1985) suggested that improving relationships with one's parents might assist change. Fraley & Shaver (1998) found that attachment issues such as anxiety about separation or 40 abandonment are reduced the longer that romantic partners remain together, and that partners' security becomes more similar to one another over time. Main & Goldwyn (1993) believe that a trusted friend, spouse, or therapist could aid an adult in "earning" security, by providing a secure base from which to work through unfavourable childhood experiences. This was indirectly supported by Rutter, Quinton, & Hill (1990), who found that the children who were reared in institutions, and therefore had many attachment relationships, were able to become competent parents if they had a supportive, non-deviant, warm and confiding spouse in early adulthood. Hazan & Shaver (1987) propose that most people have several friendships and romantic relationships and each of these provide opportunities to revise working models of self and others (see also Crittenden & Ainsworth, 1989). If female offenders address their attachment issues during and after incarceration, their potential for becoming involved in relationships that promote secure attachment may be enhanced, or the dynamics in their current relationships may be altered. Further, their abilities to provide responsive and consistent caregiving to their children may be improved. The precise degree to which attachment styles are stable remains unknown, as most longitudinal research to date has assessed attachment orientation across short periods of time (see Ainsworth, 1979; Waters, 1978; Main & Weston, 1981; Main, Kaplan, & Cassidy, 1985). However, Baldwin & Fehr (1995) examined the stability of attachment styles across several studies, including their own, and found that that although the proportions of attachment styles was consistent across time, approximately 30% of all subjects changed their attachment style classifications over time spans ranging from one week (Pistole, 1989) to one year (Senchak & Leonard, 1992; as found in Baldwin & Fehr, 1995). This occurred across many measures, including the Hazan & Shaver (1987) measure. Baldwin & Fehr (1995) compared their sample 41 of 221 undergraduates (mean age = 20.5) and found that 32.6% changed their reported attachment styles over a three- to four- month time period. Far fewer individuals who initially classified themselves as securely attached at Time 1 changed their classification at Time 2 (19.5%), compared to those who initially rated themselves as avoidant and Time 1 and changed their classification at Time 2 (42.5%), or those who initially rated themselves as anxious-ambivalent at Time 1 and changed their classification at Time 2 (68%), suggesting that those who initially report secure attachment orientations are more stable, at least among this undergraduate sample. Baldwin & Fehr (1995) analyzed a number of other studies and found similar rates of instability, ranging from 19.8% to 28.9% overall, and again found that individuals who classified themselves as anxious-ambivalent at Time 1 were the most likely to change their classifications at Time 2, with rates of change ranging from 42.4% to 75%. These authors then pooled data across several samples and found that the overall rate of changes in self-reported attachment style from Time 1 to Time 2 was 28%; specifically, 17.2% of secure individuals at TI, 33.5% of avoidant individuals at TI, and 55% of anxious-ambivalent individuals at TI changed their self-reported attachment classifications at T2. Baldwin & Fehr (1995) claim that their changes in classifications could not be attributed solely to change in relationship status between TI and T2, and speculate that these changes are likely largely due to problems with the measures. In Scharfe & Bartholomew's (1994) sample of participants who were in relationships of at least two years duration, it was found that 37% of females and 44% of males changed their attachment styles at T2 (8 months later), which indicates more attachment instability than when using the three-category criteria (Baldwin & Fehr, 1995). Interestingly, of the participants who endorsed one of the three insecure attachment styles on Bartholomew's four-category model at TI, 53% 42 switched to a different style at T2. It appears that 30% of the women who initially classified themselves as secure later changed to an insecure attachment classification, and 46% of the women who initially classified themselves as one of the insecure attachment styles later changed their attachment classification when assessed using the RQ. Further, 62.5% of insecurely attached women who changed their reported attachment style at T2 changed to the secure attachment style. However, when interview measures were used, Scharfe & Bartholomew (1994) found that less than 16% of those women who initially classified themselves as secure later classified themselves in one of the insecure categories, and 33% of women who initially classified themselves in one of the insecure attachment categories later changed to another attachment style. Of this latter group, more than 61% changed to a secure attachment category at T2. It appears that interview measures have higher test-retest correlations than self-report measures. As overlap between self-report and interview ratings of attachment is not absolute, a measurement issue arises that needs to be addressed (Bartholomew & Horowitz, 1991). Interestingly, when attachment was assessed using categorical (three and four category) and continuous measures (PAI) concurrently, they found that 96.5% of all subjects were correctly classified, reducing the likelihood that the differences across time were due to measurement error (Baldwin & Fehr, 1995). These researchers found that when these measures were administered at different time, only 62.6% of all subjects were correctly classified. Another explanation for the lack of stability in attachment classifications over time is that people who change their relationship status between the first and second testing might also experience a change in attachment style(e.g., Feeney, Noller, & Callan, 1994; Kirkpatrick & Hazan, 1994). However, this was not the case in either the Baldwin & Fehr (1995) or Scharfe & Bartholomew (1994) samples. Scharfe and Bartholomew (1994) suggest that interview and 43 continuous measures are superior to one-item categorical measures and that internal consistency of measures should be increased in order to increase test-retest correlations. However, even when using lengthy interview measures, instability is still a problem. Baldwin & Fehr (1995) propose that this instability might not necessarily be due to problems with the measures, rather the underlying construct of attachment that these measures are trying to assess. These authors suggest that maybe it is not only significant life experiences that cause alterations in adult attachment styles, but also an individual's momentary attachment orientation based on a movie just watched (romantic versus scary, for example) or on the state of a current relationship, and that these state characteristics might influence which attachment-related schemas are activated at the time of testing. Finally, these authors state that they will continue to use both the categorical and continuous measures in their research, but will view these assessments as describing individuals' current attachment orientations rather than as a stable personality trait that has remained unchanged since childhood. That is, they will assess for attachment styles at the same time as they assess for other variables of interest, as even a one-week time lag seems to produce some attachment instability, thus leading to increased error in measuring relationships between variables. They also note the importance of indicating which interpersonal relationship is being examined, as an individual may very well have differing attachment styles for her various types of relationships (e.g., friends versus romantic partners). Further research needs to be conducted to determine whether these fluctuations in reported attachment styles are temporary (state), or whether, on a third or fourth testing, and after short and long intervals, individuals evidence a consistent pattern in their attachment styles (trait). For example, the types of populations being studied should be taken into consideration. For example, two of the above-mentioned studies were assessing the attachment styles of newlyweds (Pistole, 44 1989; Feeney et al, 1994). Of course, the Pistole (1989) sample was re-tested only one week after the original test, but rather than assuming that this was due to an overall instability in attachment style, another consideration warrants attention. The instability of attachment styles for newlyweds may be due to the recent significant life change. That is, it is possible that the first year or so in the life of a newly wed is an unstable period, that is, a state fluctuation rather than trait fluctuation. To summarize, attachment styles appear to be generally stable across time, although approximately 30% of individuals assessed in various studies have altered their attachment styles over short periods of time (one week to one year). This may be due to the inaccuracies of measurement tools, or may indicate an underlying conceptual problem, or some combination of these two explanations. However, assuming that attachment styles are reasonably stable across time and are developed out of childhood experiences with primary caregivers as attachment theorists posit, it follows that the women's experiences with their parents will affect their interactions with their peers and romantic partners, as well as with their children. There is a body of literature that discusses the possibility that interpersonal relationship styles can be transmitted from mother to child, and will be briefly reviewed next. Intergenerational transmission The concept of working models proposes that internal representations of self and others develop out of relationships with primary caregivers. Bowlby (1973) suggested that patterns of parenting tend to be transmitted across generations, and that it is likely that parents who are rejecting of their children were themselves rejected as children. In other words, it is likely that a woman's relationship with her parents affects her current attachment style, which will in turn 45 affect the attachment style of her children. A mother who is herself insecurely attached may be defensive and may misinterpret infant signals and fail to react sensitively, repeating past patterns of behavior, patterns experienced in her own childhood. She may also be uncomfortable with certain emotional states, and this may affect how comfortable her infant is with these same emotional states. Alexander (1992) suggests that the mother's caregiver may also have been uncomfortable with similar emotions, and although some mothers may be conscious of their limitations and intend to avoid repeating their parents' mistakes, working models are not always consciously available, and these attempts to alter patterns may fail. A review of the literature supports the proposition that there appears to be a tendency to transmit attachment orientations across generations. A number of researchers have found an 80% correspondence between parental attachment classifications and infant Strange Situation classifications (Zeanah, Benoit, Barton, Regan, Hirshberg, & Lipsett ,1993; Main, Kaplan, & Cassidy, 1985; Fraiberg, Adelson, & Shapiro, 1975; Fonagy, Steele, & Steele, 1991; Benoit & Parker, 1994). Benoit & Parker (1994) found a 75% concordance between mothers' and grandmothers' three AAI classifications (unresolved not included). Ward & Carlson (1995) predicted maternal sensitivity assessed at three and nine months from prenatal classes of adolescent mothers, and these prenatal assessments also predicted child attachment status at 15 months. Other researchers have found that mothers classified as securely attached showed greater attunement to their infant's affect (Haft & Slade, 1989), were more sensitive to and understanding with their infants, and found that 86% of mothers who were classified as securely attached had infants who were classified as securely attached (Grossman et al, 1988). It has also been found that preschool children of secure mothers are more affectionate than other children, 46 and children of preoccupied mothers are more negative than other children (Crowell & Feldman, 1988). Although there is a tendency to transmit attachment styles and behaviours across generations, these findings demonstrate that there is not a 100% concordance rate between the attachment styles of mothers and the attachment styles of their children. Reasons for this include the possibility that treatment and/or experiencing one or more supportive relationships may alter not only one's own attachment style, but also the manner with which one interacts with offspring, inviting the possibility of positive change and breaking a cycle of insecure attachments (Egeland et al, 1988). In addition, the other parent has his or her own attachment style, and this may affect the attachment style of both the mother and the child (Cohn, Cowan, Cowan, & Pearson, 1992). This research investigates inmates' memories of their relationships with their parents, as assessed by the Egna Minnen Betraffande Uppfostran (EMBU; Perris, Jacobsson, Lindstrom, von Knorring, & Perris, 1980). The EMBU assesses for perceived warmth and rejection by both their male and female primary caregivers. It was predicted that women who had an insecure attachment style would be more likely to report higher levels of parental rejections and lower levels of parental warmth. Relationships between the women's interpersonal relationships with three generations of intimate others - their parents, their peers and their romantic partners, and their children- are the focus of this research. The female offender population experiences unique difficulties and disruptions in their interpersonal relationship circumstances by virtue of the fact that many of these women are separated from their children as well as their romantic partners and peers, oftentimes on an ongoing basis. Further, as discussed previously, there is a high incidence of 47 abuse among the female offender population. Many female offenders have been abused in childhood as well as in their adult romantic relationships and this is thought, for obvious reasons, to affect attachment orientations in this population. A brief review of this literature follows. Trauma and Abuse Main & Solomon (1986/1990) added a category to Ainsworth's three-category model of attachment, which included secure, anxious-ambivalent, and avoidant attachment styles. This fourth category was called the D category, and was created to account for a disorganized/disoriented attachment style. Some children, upon the return of the parent displayed unusual behaviours such as extreme approach followed by sudden avoidance, fear of the parent's return, approaching the parent with an averted head, and dazed or disoriented expressions. It has been found that, in at least some cases, the parent of a disorganized child has experienced trauma such as sexual abuse or a significant loss during their own childhoods (Main & Cassidy, 1988), and as a result may somehow frighten or be frightened by children (Main & Hesse, 1990). Ainsworth & Eichberg (1991) investigated infant disorganization and mother's failure to resolve the loss of an attachment figure in 45 mother-infant dyads. They found that all of mothers who had not resolved their loss had babies who were classified as D (10/10), only ten percent of mothers who had resolved their loss had babies who were classified as D (2/20), and twenty percent of mothers who had not experienced loss had babies who were classified as disorganized (3/15). Although the sample size is small, these findings support the Main & Hesse (1990) results that suggested that it is not loss per se, rather the resolution of loss, or lack thereof, that is associated with disorganized infant attachment behaviour. Specifically, mothers who had experienced loss were assessed on the " U " ("unresolved") scale of the Adult Attachment 48 Interview, which examines statements that are indicative of continual mental disorganization and processes, lapses in metacognitive monitoring of reasoning processes, discourse processes, and reports of exceedingly disorganized or disoriented behavioural responses to a death with no accompanying convincing evidence of later successful resolution of mourning ( Main & Hesse, 1990). The mother's experience of loss was only associated with disorganized infant attachment behaviour when the mother was categorized as "unresolved" with respect to that loss. Alexander (1992) theorized that, in cases of abuse, a disorganized child does not have a reliable working model with which to predict the behaviour of attachment figures upon separation and reunion because the attachment figure is in a position to both cause and relieve the child's anxiety. Based on previous research, it was predicted that many female offenders would likely have experienced abuse (Tien et al, 1993), so it is meaningful to note that evidence suggests that daughters of sexually abused women are more likely to become abused themselves (Goodwin, McCarthy, & DiVasto, 1981). Although this intergenerational transmission of abuse is not a universal occurrence, it raises the possibility that the female children of the inmates may be at a heightened risk themselves. According to an attachment perspective, it is likely that the parent-child attachment relationship is a mediating factor in vulnerability to abuse. That is, mothers who have been abused may have difficulties relating to their children, therefore transmitting an insecure attachment style, which may subsequently make the child vulnerable to later abusive relationships. Child abuse influences children's behaviours and these behaviours may be carried into adulthood, thus affecting how they treat their own children. For example, Main & Goldwyn (1984) found that battered children were more likely to avoid friendly overtures from both parents and peers, did not show concern or sadness when peers were crying, and sometimes even reacted to it with anger or aggression. Further, when a child is rejected by his or her mother, the 49 child may (adaptively) not focus attention on attachment figures and experiences; this lack of focus on close relationships may also be carried into adulthood. This avoidance of attention may lead also to difficulties in remembering one's childhood (Egeland, Jacobovitz, & Sroufe, 1988; Main & Goldwyn, 1984). According to Alexander (1992), attachment experiences involving trauma or loss are thought to merge into existing working models of relationships, which are now distorted, and these distortions affect perceptions and expectations of relationships in ways that become self-perpetuating and self-fulfilling. Individuals who were rejected or abused during childhood tend to believe that they are not worthy of love and expect to be rejected by others (Goldberg, 1991). In some cases, the child is told that she is bad and is being punished for her own good, implying that the mother is only concerned about the child's best interests, resulting in a model of other (i.e., mother) as good, and of self as bad (Crittenden & Ainsworth, 1989). Kobak & Sceery (1988) suggest that one's principles for arranging distress-related affect develop gradually from related experiences with parents, that is, whether or not the parent responds to signals of distress. These working models of attachment figures are assumed, according to attachment theory and research, to be associated with how the individual presently regulates distress. Some abused children, usually boys, display aggressive behaviors with their peers, and this may be caused by a tendency to expect others to be aggressive towards them, and possibly to misinterpret the behaviour of others. If a boy responds to his misinterpretation with aggression, the aggressive responses of others will confirm his biases. Abused girls, on the other hand, may engage in promiscuous behaviour as a way of coping with feelings of incompetence and lack of control. Whereas males more be more likely to attempt to cope with these feelings by taking what they want, females may be more likely to cope with similar feelings by giving others what they want 50 (Crittenden & Ainsworth, 1989). However, it is also likely that females may act aggressively (e.g., violent female offenders) or that males may cope by giving others what they want, and further, abused males and females may display both or neither behaviours. Other research has found that when an abused child has an attachment to a non-abusive parent, and the non-abusive parent is supportive after the abuse, the long-term effects of sexual abuse may be ameliorated (Bauserman & Rind, 1997; Gold, 1986; Wyatt & Mickey, 1987). Egelend & Sroufe (1981) found that, among a group of maltreated children ages 12 to 18 months, the presence of a supportive family member was associated with a secure attachment orientation. Further, Egeland et al (1988) found that mothers who were able to break the abusive cycle had a supportive non-abusive parent, participated in therapy at some point in their lives, and experienced a non-abusive, stable, emotionally supportive, and satisfying romantic relationship. In cases where the parent is not able to break the abusive cycle, attachment theory would suggest that she is unable to incorporate new non-confirmatory information into her existing working model (Crittenden & Ainsworth, 1989). Of course, it is also probable that some women never receive new non-confirmatory information that challenge their existing working models of attachment. An example would be a woman who relies on prostitution as a means of earning income. Her pool of available romantic partners might not include someone who would be able provide her with consistent and responsive caregiving, in that a partner who was emotionally capable of such commitment may not be able to accept her occupation. Further, she would be receiving confirmatory information about her existing models of attachment every time she went to work in that she would be servicing married men, men with little respect for women, and/or men who mistreat her. As well, she may encounter traumatic experiences by virtue of her dangerous lifestyle. 51 Steele & Steele (1994) reported that their research using the Adult Attachment Interview with a middle class sample of 200 parents yielded 30-40 percent of the autonomous group reporting past or current experiences of significant trauma or loss, now resolved, presumably through both the absence of irrational thought processes concerning the trauma, and the presence of the ability to discuss that influence of the trauma in an organized and spontaneous way. They found that individuals with a dismissing attachment style had indirect memories of unsupportive parents, and of being unloved and neglected and/or rejected, although the significance of these memories was often denied, or not acknowledged at all. For individuals classified as preoccupied, a role reversal in the parent-child relationships was frequently reported, placing the child in the unenviable position of having to offer emotional support to the parent (Fonagy, Steele, & Steele, 1991; Steele, Steele, & Fonagy, 1993). However, Alexander (1993), in her community sample of 112 women who had experienced inter-familial abuse (mean age = 37) found that secure attachment was associated with lower levels of avoidance, as assessed by the avoidance subscale of the Impact of Events Scale, indicating a willingness to confront traumatic memories. To summarize, research suggests that a history of trauma is related to a variety of interpersonal difficulties, including insecure attachment orientation, and it has been theorized that although there is a possibility of ameliorating the effects of abuse through a positive relationship with a trusted other, these difficulties tend to persist and affect a variety of interpersonal relationships. The women in the present sample were assessed for the presence or absence of emotional, physical, or sexual abuse by parents or primary caregivers. This was accomplished by adding three Likert-type questions to the end of another questionnaire. Although the potential for 52 dishonesty on a self-report measure was recognized, it is also possible that these women may find it easier to disclose abuse during the course of a written questionnaire than to verbally disclose it to an interviewer. It was predicted that that female offenders in this sample would report higher rates of abuse at the hands of their caregivers than has been reported by women in the general population, and it was further predicted that women who report abuse at the hands of their caregivers would be more likely to endorse an insecure attachment style in their romantic relationships. Attachment theory outlines the likely outcomes, for both mothers and children, of attachment styles in women who have experienced abuse or other trauma. There is no cause-and-effect relationship between abuse and specific insecure attachment styles. Rather, there appears to be an increased likelihood of developing an insecure attachment orientation if an individual is abused, does not receive sensitive and responsive caregiving, or experiences unpredictable caregiving. Further, there are factors that could exacerbate the effects of trauma and abuse on an individual's ability to form secure attachment relationships, personality pathology being one of these. Attachment orientation and personality are linked to one another in that personality likely precedes attachment, and perhaps the type of attachment experienced is dependent upon differences in dimensions of personality. Specifically, personality disorders appear to map onto normal personality dimensions, but some personality pathology is an unexplained variant. It is conceivable that personality pathology develops out of a lack of secure attachment, or is, at a minimum, affected by it; the more specific nature of the disorder is dependent upon individual personality style. Regardless, individuals with secure attachment orientations share similarities 53 with individuals who have personality disorders in that both appear to have difficulty connecting with other people in the traditional ways, causing dysfunction in their interpersonal relationships. The relationship between personality pathology and attachment styles was addressed in this female offender sample. A brief review of the literature on psychopathology and attachment style follows. Personality disorders Research has generally found that there is a fairly high rate of psychopathology, including personality disorders, among female inmates. For example, Tien, Bond, Lamb, Gillstrom, Paris, & Worsfold, (1993) found that 74% of female inmates (N =75) were diagnosed with at least one personality disorder, as assessed by the Structured Clinical Interview for DSM-III-R Personality Disorders (SCID-II, First, Gibbon, Williams, & Spitzer, 1991). Tien et al (1993) found that the largest Axis II diagnostic category among their female offender sample was Antisocial Personality Disorder (49%), followed by Borderline (16%), Avoidant (15%), Paranoid (12%), Disorder Not Otherwise Specified (12%), Narcissistic (5%), Dependent (3%), Schizoid (3%), Passive-Aggressive (2%), Schizotypal (2%), Histrionic (1%), and Obsessive Compulsive (1%) (see also Herjanic, Henn, & VanderPearl, 1977; D'Orban, 1979; Daniel & Harris, 1982; 1981; Brownstone & Swaminath,1989; Daniel, Robins, Reid, & Wilfley, 1988; Maden, Swinton, & Gunn,1994; Ingram-Fogel,1991; Salekin, Rogers, & Sewell, 1997; Salekin, Rogers, Ustad, & Sewell, 1998; Blanchette & Motiuk, 1996). In some cases, insecure attachment orientations may result in functional problems within interpersonal relationships (i.e., personality disorders), although other potential problems that may result from an insecure attachment style include low self-esteem, increased risk of 54 unhappiness, and maladaptive responses (Colin, 1996). Individuals with personality disorders also have difficulty feeling secure in relationships, and it is possible that certain personality disorders arise out of unfavourable childhood experiences with caregivers. It follows that these experiences may be associated with or result in an insecure attachment orientation. That is, some individuals may be insecurely attached but do not present with personality disorders, while others may have a secure attachment orientation but do present with personality disorders. Of course, having both an insecure attachment style as well as a personality disorder, or concordantly, having a secure attachment style and no personality disorder are other possible alternatives. Individuals with personality disorders have difficulty feeling secure in relationships. The DSM-IV (American Psychiatric Association, 1994) defines a personality disorder as: An enduring pattern of inner experience and behavior that deviates markedly from the expectations of the individual's culture, is pervasive and inflexible, has an onset in adolescence or early adulthood, is stable over time, and leads to distress or impairment (p. 629). Zimmerman & Coryell (1989) reported that the presence of personality disorders appeared to be related to low rates of marriage, and high rates of marital discord, separation, or divorce. These authors found in their non-patient sample that 70% of individuals with antisocial personality disorder, 78.6% of those with dependent personality disorder, and 100% of those with borderline personality disorder had a lifetime history of being separated or divorced. Attachment theory provides a basis for speculating about the kinds of personality disorders that could be associated with the various attachment styles. For example, histrionic and borderline personality disorders would be expected to be associated with a preoccupied attachment style, and avoidant and dependent personality disorders would be expected to be associated with a fearful attachment style. On the other hand, a dismissing attachment style may be more adaptive, and although relationships satisfaction may be low, these individuals have high self-esteem and 55 experience low levels of subjective distress and depression (Bartholomew et al, 2001; Bartholomew, 1997; Scharfe & Bartholomew, 1995; Bartholomew & Horowitz, 1991). Van Ijzendoorn et al (1997) note that problems with attachments are apparent in most DSM-III-R personality disorders, yet there has been little research to test the relationship between personality disorders and attachment style. Ainsworth et al (1978) reported that mothers with rigid, perfectionistic, and compulsive personalities in the Baltimore study had infants that were avoidant. These mothers appeared overcontrolled and mechanical in their body movements and showed less emotional expression when compared to other mothers. However, the researchers did not label these women as having personality disorders, nor did they directly assess for personality disorders. Forty Dutch men confined in a forensic psychiatric hospital for committing serious crimes were all found to be insecurely attached, and separation from attachment figures in childhood was related both to attachment style, as assessed by the AAI, and personality disorders, as assessed by the SIDP-R (van Ijzendoorn, Feldbrugge, Derks, de Ruiter, Verhagen, Philispe, van der Staak, & Riksen-Walraven, 1997). These researchers suggested that individuals with schizoid personality disorder may have experienced lack of parental warmth in childhood, and individuals who are paranoid, avoidant, or antisocial, may have been more likely to experience parental rejection and hostility. They also reported that the number of diagnosed personality disorders were negatively correlated with patients' levels of secure attachment. Livesley, Schroeder, & Jackson (1990) found that insecure attachment is related to the development of dependent personality disorder, although insecure attachment and dependency behaviours are not the same thing. These authors note that insecure attachment is directed at a particular attachment figure or figures, whereas dependency behaviours are more generalized-56 that is, the target is not the issue, rather obtaining assistance, guidance, and approval is of uppermost importance. They also suggest that a categorical classification of dependent personality disorder should include insecure attachment and dependency as the two defining features. When a dimensional model is used, Livesely et al (1990) recommend the use of insecure attachment and dependence as separate dimensions. Research on borderline personality disorder has found that disregulation of interpersonal distance is a key factor in borderlines, as is a preoccupied attachment style (Melges & Swartz, 1989). These authors contend that borderline personality disorder involves an oscillation between attachment and detachment. In Alexander's (1992,1993) sample of 112 incest survivors, preoccupied attachment style was related to dependent, self-defeating and borderline personality disorders, and Rosenstein & Horowitz (1993; as reported in Diamond & Blatt, 1994) found that, among their sample of adolescent psychiatric inpatients, those classified as having a preoccupied attachment style, as assessed by the AAI, who also exaggerated affect and were overwhelmed by parental introjects5, tended to be diagnosed with histrionic, borderline, dependent, or schizotypal personality disorders, as assessed using DSM-II-R criteria. These relationships were not found when the patients were assessed using MCMI as a measure of personality dysfunction. Patrick, Hobson, Castle, Howard, & Maughan (1994) used the AAI with 12 women with borderline personality disorder, as assessed with DSM-III-R criteria, and found that these women tended to have a preoccupied attachment style and were also found to be unresolved with regards to loss or trauma. Similar results were found by Fonagy, Leigh, Steele, Steele, Kennedy, Mattoon, Target, & Gerber (1996) on a more diverse clinical sample, and by Stalker & Davies (1995) with a sample of women with histories of child sexual abuse. 57 Brennan & Shaver (1998) found that borderline personality disorder was associated with fearful and preoccupied attachment styles, as assessed with the Experiences in Close Relationships Inventory, and other researchers have found that fearfully avoidant and preoccupied adults have characteristics of borderline personality disorder (Alexander, 1993; Dutton, Saunders, Starzomski, & Bartholomew, 1994). Sack, Sperling, Fagen, & Foelsch (1996) found that borderline personality disorder was associated with all categories of insecure attachment styles, whereas van Ijzendoorn & Bakermans-Kranenburg (1996) found no association between AAI classifications and personality disorders in a sample of mentally disturbed criminal offenders. Avoidant personality disorder appears to be related more to both a desire for and a fear of close attachments than to lack of social skills or general discomfort (Sheldon & West, 1990). These researchers found that insecurity in attachment relationships involved an approach-avoidance conflict upon activation of that attachment system, which was behaviorally resolved with an avoidant response. They found that there was a high threshold for the activation of the attachment system, but a low threshold for the deactivation of the attachment system that severely restricted the range of attachment responses. Allen, Leadbeater, & Aber (1994) found a relationship between negative expectations about self-in-relationships and antisocial behavior. Allen, Hauser, & Borman-Spurrell (1996) found that adults with a dismissing attachment style were significantly more likely to report criminal behavior than secure adults. Luntz & Widom (1994) found a relationship between parental neglect and childhood abuse, both of which were associated with insecure attachment 5 Introjection is a psychoanalytic term that refers to the process of absorbing the superego from the parents; that is the child incorporates the attitudes of the parents as his own. 58 styles and antisocial personality disorder (ASPD). Rosenstein and Horowitz (1993; as reported in Diamond & Blatt, 1994) assessed the attachment styles of 60 adolescent psychiatric inpatients using the AAI and found that those classified as having a dismissing attachment style, who also denied and excluded negative affects or aspects of parental introjects, were more likely to be diagnosed as having narcissistic or antisocial (or conduct) personality disorders. Alexander (1992,1993) found that a fearful attachment style was related to interpersonal violence and self-defeating personality disorders. Mickelson, Kessler, & Shaver (1997) found that ASPD, as assessed using a structured interview, was associated with both anxious-ambivalent and avoidant attachment styles. Bartholomew et al (2001) suggest that ASPD may be linked to insecure attachment, but it may be that the extremity of insecure attachment is more predictive of ASPD than are the specific attachment styles. To summarize, some research has found that relationships exist between preoccupied attachment styles and borderline and dependent personality disorders, and between dismissing attachment styles and ASPD and conduct disorder. However, other research has not found this connection or has found more complex connections. It appears to be the case that there is some relationship between personality disorders and insecure attachment styles, but the precise nature of the relationship is complex and imprecise. Consequently, it may be more meaningful to simply use secure versus insecure attachment categories when making comparisons with personality disorder characteristics, at least with the current assessment tools. According to Greenberg (1999), psychological disorders that affect social functioning are not likely associated solely with attachment insecurity. Rather, a combination of risk factors will influence the likelihood of experiencing psychological problems and problems in social functioning. Additionally, there are several different pathways leading to a disorder, and insecure 59 attachment may be a risk factor in developing personality disorders and other pathologies, but only in combination with other contributing factors, such as difficult child temperament, poor parenting, and unfavourable family circumstances. As noted, the absence of insecure attachment does not guarantee the absence of personality pathology (Bartholomew et al, 2001). Further, Bartholomew et al (2001) propose that attachment insecurity alone is not likely to be associated with any particular disorder; rather the greater number of risk factors in a child's life, the more apt she is to develop a psychosocial disorder. However, when a parent has a personality disorder, this will likely affect the ability to give good and consistent caregiving to a child, thus increasing the possibility of the child developing an insecure working model of attachment (Colin, 1996). It is not the focus of this research to explain personality pathology, rather to investigate the role of attachment in all areas of dysfunction in a female offender population. Attachment theory lacks specificity a to which attachment styles relate to which personality disorders (definitively), but attachment theory is important in predicting social functioning. This female offender sample was assessed for personality disorder characteristics using a semi-structured clinical interview, the Structural Clinical Interview for DSM-IV Personality Disorders (First, Gibbon, Spitzer, Williams, & Benjamin; 1997). There is some prior evidence regarding possible relationships between particular personality disorder characteristics and attachment styles. Previous research has shown that female offenders are more likely to present with personality disorders than women in the general population (Tien et al, 1993), and that antisocial personality disorder is the dominant personality disorder found among female offenders. There was no reason to believe that the present sample would vary from this pattern. It was predicted that female offenders with an insecure attachment orientation would be more 60 likely to report characteristics of personality pathology than female offenders who reported a secure attachment orientation. Dysfunctional personality characteristics, when present, are likely to affect all of the individual's relationships, including relationships with their children. The presence of one or more personality disorders may hamper a mother's ability to bond with or provide adequate caregiving to her children. As adequate bonding and caregiving are important in the formation of secure attachments between mothers and their children, a review of this literature follows. Caregivine & Parental Bonding to Children As previously mentioned, Bowlby (1979) defined attachment behavior between and infant and caregiver as any form of proximity-seeking behavior towards a preferred, usually stronger and/or wiser, individual. He also described attachment between adult romantic partners as giving a sense a security and being a source of joy, adding that these emotions reflect the state of the individual's affectional bonds. However, there has been little research on parent-to-infant bonding and the research that has been conducted has focused on the mother's bonding to the infant within the first few hours or days after birth (see Chess & Thomas, 1982, for a review of this literature). Attachment orientations in adulthood may affect the manner in which parents act as attachment figures toward their own children (Main & Goldwyn, 1985). Bowlby (1969) claimed that the mother's contribution to the mother-infant relationships stemmed from her biological make-up, a long history of interpersonal relations with her family of origin, as well as cultural socialization. Other researchers believe that current working models of parenting reflect the mother's perceptions of her parent's roles, as well as her own role as a child. That is, a woman 61 enters into motherhood with preconceived ideas of her role as mother and is likely to replicate her childhood experience, or at least her perceived childhood experience, with her own child (Sroufe & Fleeson, 1986; Ricks, 1985; Main, Kaplan & Cassidy, 1985; Rholes, Simpson, Blakely, Lanigan, & Allen, 1997). This relationship between parents' and children's quality of attachment is a transgenerational process that is hypothesized by several researchers to be a three-step process that involves the parent's attachment in childhood, the effect that this attachment has on the development of parental caregiving behaviour with children, and the primary influence that this caregiving behaviour has on quality of attachment to the parent (Crittenden, Partridge, & Claussen, 1999; Ainsworth, 1085; Bretherton et al, 1989; Bowlby, 1973; Main & Goldwyn, 1990; Ricks, 1985, as cited in Crittenden et al, 1999). Further, Belsky (1984) suggested that separation from parents as a child is related to later difficulties in caring for young children. Main & Goldwyn (1984) studied thirty current transcripts of the Adult Attachment Interview of women in their 30's who had previously participated in research with their now almost 6-year-old children when they were infants. These thirty transcripts were coded for rejection by mothers, idealization of mothers, anger with mother, coherence of transcript, and inability to remember childhood. Videotapes of their children when they were infants (taken 5 years previously) were coded for avoidance of mother. When the mother's AAI transcripts was compared with the infant's videotapes it was found that the mother's apparent rejection by her mother was strongly related to the infant's avoidance of her following brief separations. However, if the mother expressed anger and resentment toward her own mother, but was rated as coherent in the interview, her infant was unlikely to avoid her. From these findings, the authors extrapolate that suppressed anger and resentment and an inability to coherently discuss attachment relationships of childhood experiences with parents 62 may indicate a distortion in the representation of an abusing parent and be a factor in the perpetration of child abuse. The authors also report that rejected and abused children tend to develop similar behavioral characteristics, that is, problems in aggression control, aversive unsympathetic responses to distress in others, and self-isolating tendencies, suggesting that psychologically rejection and abuse are experienced on a continuum. Of course, these theories need to be tested to see whether they have merit, and further, the coding criteria for rejection, inability to recall childhood, and incoherence need to be validated. For example, inability to recall childhood was coded according to the frequency with which the participant claimed to not be able to recall childhood, which is obviously problematic in a variety of ways, including the failure to take into account differences in individual reporting styles and differing levels of insight into quality and quantity of memory (i.e., metamemory). The earlier review of intergenerational transmission of attachment points to the importance of the mother's attachment style to the development of attachment orientation in children. Research on the correlation between attachment styles of mothers and their children assessed mothers using the AAI, and assessed infants using the Strange Situation. Main & Cassidy (1988) found strong ties between the mother's current attachment styles and the infants' attachment styles as assessed six years earlier at 12-18 months. Secure infants generally had secure or autonomous parents, avoidant infants were more likely to have dismissing parents, anxious-ambivalent infants usually had preoccupied parents, and disorganized infants were more likely to have parents who were categorized as unresolved. , It is apparent that the attachment style of the mother can affect her relationship with her child, the quality of her parenting, the attachment style developed by the child, the child's attachment style as an adult, and the parenting practices of the adult child. 63 Bowlby (1969) believed that the parent's bond to the child is related to attachment, but is a separate and independent behavioral system which he called "caregiving". However, in other writing, he also suggested that attachment figures could be spouses, parents, or children (Bowlby, 1979). Berman & Sperling (1994) agree with this latter conclusion, suggesting that the caregiving system develops directly out of the attachment system, and is integral to it. Parker (1994) distinguished between parent-child "attachment" and "bonding". He argued that instinctively determined behaviours are more appropriately termed "attachment", and how the parent and child each perceive, judge, and experience interactions with one another is more appropriately characterized as "bonding". Parker (1994) further commented that it is difficult to measure parental characteristics such as bonding and caregiving. Social desirability is likely to play a large part in the responses of parents. Therefore, much research has focused on obtaining the adult child's perception of the parents' bonding and caregiving toward him or her. However, this is problematic as it relies upon retrospective accounts, raising the issue of the quality of memory for these experiences. Further, using the memory of the child to assess the perceived bonding and caregiving of the parent only illuminates a portion of the story; that is, much information is lost about the parent's point of view, which one could expect may be very different than the retrospective point of view of a child. Also, retrospective depictions of childhood experiences with parents may also be subject to social desirability, either to protect the parent, or because the adult child blames the parent for current misfortunes. The present research takes a different approach; that is, the mother's (the female offender's) point of view is represented. However, it is recognized that this information is potentially biased and is therefore augmented with objective information gathered through interviews, questionnaires and prison records. 64 There are difficulties with assessing bonding and caregiving. There have been attempts to assess this bond from the grown child's perspective of his or her parents. For example, the Parental Bonding Instrument (PBI) was constructed by Parker, Tipling, & Brown, (1979), and consists of 25 items that assess parental care and overprotectiveness. This research was interested in investigating the women's perspectives on their relationships with their children, and it was unlikely that contact with children would be facilitated. The Maternal Separation Anxiety Questionnaire (Hock, Gnezda, & McBride, 1983; as described in Parker, Tilping, & Brown, 1997) focussed on issues of stress of separation from children. This measure would not be appropriate for the present research as it appears to be based on intact mother-child relationships; i.e., in cases where mothers have custody of their children, which previous research suggests is often not the case in female offender samples (MacLean, 1997). This measure appeared to be intended for investigating feelings of separation anxiety for temporary separations from children and could prove to be upsetting for mothers who no longer have custody of their children and in many cases have been forced to be separated from them. The Maternal Separation Anxiety Questionnaire (Hock et al, 1983) was modified by Cox, Owen, Henderson, & Lewis (1989; as found in Parker et al, 1997) into a broader scale that assessed attachment related behaviours and attitudes such as delight, acceptance, sensitivity, investment, time spent with infant, positive affect, animation, amount of reciprocal play, enjoyment of play, among other things. The addition of the behavioural observation component rendered this scale impractical for the purposes of the present research. . Bretherton, Birigin, Ridgeway, Maslin, & Sherman (1989) developed the Parental Attachment Interview (PAI), based on the Adult Attachment Interview, which focuses on 65 parental protectiveness and separation anxiety. Training for administering and coding the AAI measures is time intensive and costly, and also presumes that the parent-child relationship is intact. Parker et al (1997) developed a questionnaire to assess socioemotional investment of parents in their children, called the Parental Investment in Children (PIC). Parker et al (1997) astutely noted that detailed information regarding parental behaviour would assist in mapping the process that influence the intergenerational transmission of attachment. The PIC was designed to measure the following four constructs: delight, sensitivity, separation anxiety, and acceptance of the parenting role. Unfortunately, their analyses did not support the utility of the 'delight' construct, although the other constructs seemed valid. However, the PIC was harshly criticized by Hayes (1998), who described the PIC as "a series of taxing and exacting prescriptions for completely selfless, constant mother care" (p. 782). Hayes (1998) also took issue with the fact that the scale was validated on white middle class mothers, that it was entirely child-centered, and in her view, archaic. Bradley (1998) defended the PIC by stating that in the early stages of creating a theory, model, hypothesis, one uses the resources most readily at hand (i.e., middle class mothers) and that as one refines hypotheses and theories, there is ample opportunity to test measures on various, less accessible populations. Further, he argued that Hayes (1998) misunderstood how the scale was scored, and that even though it was a Likert scale from one to seven, the extreme score of an item was not necessarily the "best" answer to that item. In defense of Hayes (1998) criticisms, this should have been more apparent in the Parker et al (1997) article. Regardless, the PIC also appeared to be designed for intact parent-child relationships, and therefore was not appropriate for research with a female offender sample. 66 Crittenden & Ainsworth (1989) suggest that, even if there is no formal way of assessing a mother's "bond" to her infant, one could assume that if the mother provides enough nurturance for her infant to stay alive that a bond has been formed. This does not mean that all bonds are equal; that is, the bond of a sensitive and reliable mother toward her infant is likely different than the bond of a maltreating mother toward her infant (Crittenden & Ainsworth, 1989). For example, women who maltreat their infants may turn to their own mothers for assistance in raising the child in order to avoid forming a bond with their infant (Crittenden & Ainsworth, 1989). Further, neglecting mothers or mothers with difficulties caregiving may in fact have difficulty bonding to their children. As a result, these neglected children may themselves grow up with a desire to have better relationships, yet may regularly find themselves in difficult relationships: "Even an unpleasant but predictable world is preferable to a chaotic one" (p.213, Ricks, 1985, referring to a point made by Epstein, in press at the time of publication). These neglected children may grow up with an inability to trust others and may be wary of relationships. When they realize that a relationship is not working, they may jump to another on a never-ending quest for felt security, but their working models will continue to draw them into regular patterns of insecurity and possibly neediness, among other things. According to Crittenden & Ainsworth (1989), this type of individual may look to both the spouse and any children as sources of security rather than as individuals to whom they should provide nurturance. One could speculate that this need for felt security might actually be the motivating factor in becoming a mother for some insecurely (preoccupied) attached women. What if the preoccupied mother still thinks of herself as the child in the relationship? This flies in the face of attachment theory's emphasis on the primary goal of attachment being the protection of the child 67 (Lyons-Ruth, 1996). This goal relies upon the reciprocal behavioural systems of the parent and the child (George, 1996). Of course, other forms of insecure attachment would manifest in different ways. For example, an individual with an avoidant attachment style (i.e. dismissing according to Bartholomew's categories) might have less experience with caregiving, as well as less inclination to give care (Shaver & Hazan, 1988). It is precisely these suppositions that were investigated in this research. To summarize, research has found relationships between mother's attachment styles (presumably developed during childhood through relationships with primary caregivers) and the attachment styles of their infants. Additionally, research has identified differing parenting characteristics between securely and insecurely attached mothers. Finally, there is no appropriate instrument to measure maternal attachment towards children in a female offender sample. For this research, The Parental Stress Scale (PSS; Berry & Jones, 1995) was used as an indicator of parental satisfaction, parental stressors, parental rewards, and lack of control as a parent. This scale was chosen as it is somewhat less tied to current behaviour, and therefore custody of children, or lack thereof, should not be as much of an issue as it might be with the other parenting scales. In addition, the Parental Bonding Interview (PBI) was used characterize the finer points of the relationships between mothers and their children, such as how much time is spent together, as well as the mothers' subjective views of their parenting abilities. Although maternal "attachment" to children cannot be directly assessed, the results of the PSS and the PBI was used to characterize the mother-child relationship from the perspective of the mother, and the PSS characterizations were compared with the outcomes on the attachment measures being utilized. It was expected that female offenders would report higher levels of parental stress than has been found in both clinical and community samples of mothers (Berry & Jones, 1995), and 68 it was predicted that women who were categorized as insecurely attached would report higher levels of parental stress than women who reported a secure attachment orientation. As noted, attachment theory has been used with various populations. However, there has been little research using attachment theory with female offender populations. Reasons why attachment theory is appropriate for research with this population are outlined next, followed by a review of the research that has applied attachment theory to female offender populations. Usefulness of Attachment Theory in Research on Female Offenders Attachment theory provides a framework for characterizing aspects of interpersonal relationships and the potential origins of insecure attachment, adaptive and otherwise. This theory proposes the likely etiology of various attachment styles, the adaptive function of such styles, and the characteristics of the individual styles across the lifespan. It suggests how insecure attachment qualities may become so extreme as to manifest themselves, with the aid of other risk factors, as more severe problems, such as personality disorders. This theory also suggests that attachment style is dynamic, and therefore the possibility of change exists. Further, the mechanisms through which potential change may occur are proposed, and implications for therapy and treatment are clear: An alteration of dysfunctional working models of self and/or others, exposure to sensitive and consistent caregiving, romantic, or peer relationships, and/or a supportive therapeutic environment may provide the necessary support for developing a secure attachment style. Further, parenting programs that address attachment issues could be beneficial to female offenders in their relationships with their children, ultimately assisting both mother and child. Most research to date has included non-offender populations, however, there is a small literature on attachment in offender populations, the majority using male offender samples. 69 However, generalizing from a male offender to a female offender population cannot be justified (Tien et al, 1993), particularly when investigating matters of parenthood, as female offenders are more likely than male offenders to have been primary caregivers (Greenfield & Snell, 1999) and to have experienced sexual abuse at some point in their lives (Tien et al, 1993)- two of many factors that may affect an individual's attachment orientation. Female Offenders and Attachment It appears that research on attachment in adult female offenders has been limited to one investigation to date. Susan Turnbull (1996), from Simon Fraser University, investigated personality pathology and adult attachment with a sample of fifty female offenders at British Columbia Corrections for Women (BCCW) in Burnaby, B.C., for her doctoral dissertation. She found that her sample was characterized by low income, drug and alcohol abuse, relatively low levels of education and employment, and disrupted families of origin. She further found that most of the women she investigated were repeat offenders, and over half had five or more previous arrests. Also, Turnbull (1996) found high levels of problematic personality characteristics- high levels of emotional lability and reactivity, limited ability to regulate their affective experience, high levels of stimulus seeking and acting out against themselves and others, a distrustful and suspicious world view, somewhat limited communication patterns, and often rather rigid moral standards. Results of the attachment measures reflected very low levels of secure attachment and very high levels of insecure attachment, especially for the dismissing and fearful types, and lifestyle variables bore limited relationships to the personality and attachment measures collected. 70 Turnbull (1996) speculated that personality disorders that may be associated with various attachment styles include schizoid, dependent, avoidant, and borderline. Further, she noted that most women were raised in families in which there were early and frequent disruptions leading to non-continuous parenting. She also noted the importance of future research on the intergenerational transmission of attachment styles to children and the problems this may cause. Turnbull's (1996) research provided an important foundation upon which to build further research in the area of attachment in a female offender population. She illuminated the importance of including an attachment perspective when considering female offenders, and highlighted the unique qualities of the female offender - namely, the greater importance parenting plays in women's lives. The present research explored the various relationships of the female offender; relationships with children, romantic partners, peers, and parents. This research also added another measure of attachment, the Experiences in Close Relationships Scale, a multi-item scale to be compared with the Relationship Questionnaire. This research used a larger sample size, and included interviews with each participant, which adds a richness of understanding about the plights and individual concerns of these women. A sub-sample of these women were also interviewed using the Peer Attachment Interview in order to get a more detailed account of the peer and romantic relationship characteristics of the female offender population. Finally, this research interviewed female offenders about their current relationships with their children in order to investigate the possible implications of Turnbull's (1998) reported rates of insecure attachment among female offenders on the parenting experiences of these women. 71 Measurement & Methodological Issues There are several difficulties that are encountered when researching a female offender population. Firstly, because of the small numbers of female offenders relative to male offenders there has been little research conducted on the former, as funding and other resources are most often allocated to support research on the vastly larger male offender population. Further, the research that has been conducted has been primarily descriptive in nature- that is, is has addressed types of criminal behaviour, number of crimes, sentence length, mental health, offenders' reports of victimization in childhood, and drug and alcohol abuse. Additionally, the extant literature does not adequately outline the difficulties encountered when collecting data with a female offender population. Specifically, file information tends to be inconsistent and incomplete, and is often based on offender self-report, and is therefore not verifiable, or at least is not verified by institutional staff. As well, many of these women are in prison for charges such as theft or impersonation, and have a history of using one or more aliases. This speaks to the dishonest tendencies that are often found in offender samples, further attenuating the appeal of conducting research that is largely based on the self-reports of these women. The only reliable source of information is the local arrest record. Out-of-province or out-of-country charges are usually not included. This lack of verifiable file information renders it impossible to assess the veracity of the data gathered from the offenders, and therefore, the reliability of this information is called into question, and is likely another explanation for the lack of research conducted with this population. For these reasons and others, impression management must be evaluated and considered when analyzing and interpreting results of such research. Also relevant to the present research is the issue of measurement, and the availability, or lack thereof, of measures that have been normed and validated on a female offender population. 72 For example, the majority of researchers have used the various attachment measures on undergraduate samples or middle- to upper-middle-class samples who have been recruited through newspaper advertisements or university postings. Neither of these groups of individuals can be assumed to be representative of the general population, and in many ways are likely to be even less representative of a female offender sample. Therefore, it is not known whether these scales capture what they were originally designed to assess when they are administered to female offenders. In addition, disagreement exists between various adult attachment assessment measures regarding the classification of the various insecure attachment styles as well as their specific relationships to other variables (e.g., personality disorders). There also appears to be an issue as to whether current attachment tools measure state as opposed to trait characteristics, that is, whether or not what is being measured is a stable personality characteristic, or whether it is simply a representation of the current mood, perception, or life circumstances of the individual being assessed. Further, it is not clearly established whether attachment styles are continuous with, or simply analogous to, attachment styles in childhood. However, it does appear that there is a clear difference in the quality of functioning between individuals with a secure attachment style and those with an insecure attachment style. Mickelson, Kessler, & Shaver (1997) found that interpersonal traumas as well as maternal and paternal warmth (as assessed by the Parental Bonding Instrument) were related to insecure attachment, but that there were no patterns between types of insecure attachment. This higher level of classification (secure versus insecure) of attachment is used in the present research unless the lower-level distinctions provide more information. 73 A final issue involves the lack of appropriate control group for research with female offenders. The present research can be compared to previous research with university undergraduates or community samples, but these serve only to provide a base rate of various non-offender populations with which to compare the female offender sample. It cannot be assumed that these samples are representative of the general population. However, previous research on these samples does show that individuals who experience financial adversity in childhood (Mickelson et al, 1997) as well as low SES mothers tend to be classified with insecure attachment styles (van Ijzendoorn & Bakermans-Kranenburg, 1996) more often than is found in higher SES samples, who appear to be comparatively more stable in their interpersonal relationships (Egeland & Sroufe, 1981). It is difficult to find a control group that matches the particular combination of circumstances in that these women are confined to an institution and are therefore presently unable to function within their interpersonal relationships in the same manner that they would when not incarcerated. The majority of female offenders have serious drug and alcohol problems, few or no job skills, little or no economic security, histories of abuse both in childhood and adulthood, and friends, romantic partners, and associates who share similar sets of circumstances. It would be difficult to identify another group characterized by this level adversity in addition to being confined within an institution. Psychiatric patients may be similar in this regard (living arrangements), but are likely to have poorer functioning in many realms than female offenders due the requisite severity of Axis I disorders that would have led to psychiatric institutional confinement. Another option for a suitable control group might be drug treatment centres, or women who live in inner city neighbourhoods. However, for the purposes of this exploratory study, the inclusion of a control group was not deemed necessary, as this research is attempting to compare securely attached offenders with offenders who do not have a 74 secure attachment orientation. This within-group method of research has often been used in forensic research (e.g., psychopathy research; Rice, Harris, & Cormier, 1992; Serin, 1992; 1996; Hemphill & Hare, 1998). To summarize, conducting research with female offender samples is rife with obstacles, such as lack of corroborating information, there are inherent issues with the existing attachment measures, and none of the measures have been validated on female offender populations. Summary The previous literature review illuminates a number of points. The first is that adult attachment styles and personality disorders both appear to offer vital information regarding interpersonal relationships. The second is that the female offender population is characterized by a specific set of problems including drug and alcohol addiction as well as experiences of trauma and victimization. Thirdly, female offenders are often challenged by the parenting role, and many lose custody of their children, as they are unable or unwilling to adequately care for them. This has implications for the future interpersonal functioning of these children, as their working models are likely to have incorporated the patterns in their disruptive relationships with their mothers and subsequent primary caregivers. Finally, issues regarding measurement were addressed, as well as an account of the methodology used in this investigation. The purpose of this research was to explore the relationships between the attachment styles of female offenders, their early experiences with their caregivers, and their own parenting characteristics. Female offenders are extreme examples of what could go wrong in the development of attachment styles, and this is a particularly interesting population with which to explore attachment dysfunction. The interpersonal difficulties of female offenders with parents, 75 peers and romantic partners are likely to be reflected in their relationships with their children. These women have particular challenges to face in their roles as parents, and many decide not to attempt to play the role of mother. It is not possible to separate out all of the factors that contribute to female inmates' relationships with their children, particularly when taking into account the importance of other potential contributors such as previous abuse, addictions, and poverty. However, an attempt was made to garner a greater understanding of the nature of their mother-child relationships. Due to the fact that there is likely a high rate of victimization in this sample, and that their victimizers were likely males, the first hypothesis predicts that female offenders are more likely to report a secure attachment style in their platonic friendships than in their romantic relationships. The second hypothesis predicts that female offenders who report an insecure attachment style would be more likely to report characteristics of personality pathology than those women who report a secure attachment orientation. The third hypothesis addresses the relationships of female offenders with both their male and female caregivers. It was predicted that women who report an insecure attachment style would also report lower levels of parental warmth, higher levels of parental rejection, and higher levels of emotional, physical, abuse by both parents, and higher levels of sexual abuse by fathers (not mothers, as it is rare that mothers sexually abuse their children). The fourth hypothesis addresses the perceived amount of parenting stress reported by female offenders. It was predicted that female offenders would report high levels of parenting stress based on previous literature that suggests that female offender mothers encounter multiple difficulties in their roles as caregivers (e.g., battling drug addictions), and that many have lost custody of their children, suggesting that the parenting role 76 was challenging for them (MacLean, 1997). It was further predicted that women who report a secure attachment orientation would be less likely to report high levels of parental stress. In addition to the above hypotheses, it was predicted that women who were abused in childhood at the hands of primary caregivers and women who experienced high levels of parental rejection and low levels of parental warmth are more likely to experience high levels of personality pathology and high levels of parental stress than women who did not have such adverse childhood experiences. It was further expected that attachment orientation could mediate these relationships, particularly if insecure attachment orientation was found to be related to each of the above indices. Although the ability to find these mediational relationships may be limited by the measures and/or attachment construct problems, some of these variables are likely to be useful with respect to understanding the deficits in current knowledge about maternal behaviour and its association in attachment in a female offender sample, a group that seems most suitable for testing these relationships given the extreme nature of their interpersonal difficulties. It is possible that, if insecure attachment is indeed playing a part in the interpersonal difficulties being experienced by female offenders, addressing issues of attachment in therapy would assist mothers in need of finding more functional ways of interacting in her relationships, particularly with their children. It is also possible that parenting programs in the community that specifically address issues of attachment might benefit both the mother and the child currently and in perpetuity. 77 Method Participants Participants for this study were recruited on a voluntary basis from the Burnaby Correctional Facility for Women (BCCW), a British Columbia prison for federally sentenced, provincially sentenced, remanded, and immigration-hold women. There was an information notice posted at BCCW in February, 2001, and interested participants were asked to fill in a request form and send it to the project coordinator who forwarded this request to an interviewer. The interviewer then called the unit where the volunteer was housed and asked whether the volunteer wished to come down for the interview at that time. Often, when the participant was not available at the time of the call, the unit guard asked whether anyone else on the unit was interested in participating, and if so, another participant went to a closed interview room located in the visiting area, where many of the interviews were conducted. This was a frequent occurrence. Other participants were confined to their units, and thus were unable to come to the visiting area, due to risk to or from other inmates. In these cases, an interviewer went to the unit where the participant was housed. While in the unit, other women volunteered to participate, and they seemed to prefer staying 'close to home', as it were. Realizing this, some interviewers began to go to the units to recruit and interview participants. Participation was always voluntary, and participants' privacy was maintained while being interviewed, as each housing unit had an interview room within it. 78 Measures All participants were asked to complete a questionnaire package consisting of the following: The SCID-II Personality Questionnaire, Background information questionnaire (demographics), The Relationship Questionnaire (RQ), Experiences in Close Relationships Inventory (ECRI), Early Experiences in Childhood Questionnaire (EMBU), Balanced Inventory for Desirable Responding (BIDR), and the Parental Stress Scale (PSS) for those women who had raised children. Also, participants were interviewed using the SCID-II semi-structured interview for personality disorders, and women who had children were interviewed using a Parental Bonding Interview. SCID-II Personality Questionnaire This is a self-report personality questionnaire that was designed to be used as a screening tool to shorten the duration of the SCID-II Clinical Interview (First, Gibbon, Spitzer, Williams, & Benjamin, 1997). Participants circle "yes" or "no" for a number of statements and subsequently, the items that are circled yes on this questionnaire are then circled on the interview, and these are the only those questions are asked during the interview, unless there is reason to ask additional questions, i.e., additional probing is needed to score the item(s). The interviewer probes further into these items in order to determine whether a disorder is likely present or not. The rationale for the appropriateness of use of this screening tool is that a person who is not willing to acknowledge a symptom on a pencil and paper test is not likely to acknowledge it to an interviewer, and if the participant answers no to an item, follow-up questions are not germane, as the subject is unlikely to provide examples for a behavior that they claim is not present. Jacobsberg, Perry, & Francis (1995) found that the false-negative rate was very low for every diagnosis when they examined the validity of the SCID-II Personality 79 Questionnaire. These authors concluded that the SCID-II procedure of following up on positive responses was a valid method. Ekselius, Lidstrom, & Von Knorring, (1994) found good agreement between the questionnaire and the interview, establishing a correlation of .84 in the number of positive items and a kappa of .78. Background Information Questionnaire (Appendix 1) This questionnaire was adapted from a questionnaire created by Susan Turnbull for her dissertation research (1996). It explores the participant's offence history, occupation, education, socioeconomic status, drug and alcohol use, and family circumstances as a child. One question regarding ethnicity was added to Turnbull's questionnaire, and nine other questions regarding number and length of romantic and peer relationships, substance abuse and criminal activity by romantic partner, as well as satisfaction with current partner and intent on returning to partner, were omitted, but in retrospect should have been left in. It was thought that this information would be elicited during the interviews, however, the participants often did not disclose relationship information in sufficient detail. Relationship Questionnaire-RQ (Appendix 2) This is a four-item self-report questionnaire created by Kim Bartholomew (1990) that assesses attachment styles in adults. Each item is a description of each of Bartholomew's four attachment styles and is presented in paragraph form. The scale is answered four times- first, the participant chooses the description that best characterizes her friendship style, and second, she is asked to rate each of the four styles on a scale of one to seven according to how well each description characterizes her friendship style. Then, the participant chooses the description that best characterizes her romantic relationship style, and finally, she is asked to rate each of the four styles on a scale of one to seven according to how well each description characterizes her 80 romantic relationship style. Scharfe & Bartholomew (1994) found that, using the Relationships Questionnaire, 63% of females (k = .42) reported the same attachment patterns at two testing sessions, eight months apart. Experiences in Close Relationships Inventory-ECRI (Appendix 3) This is a 36-item self-report questionnaire created by Brennan, Clark, and Shaver (1998) that assesses attachment style, looking particularly at the dimensions of avoidance and anxiety. As noted previously, attachment-related anxiety involves feelings of discomfort stemming from fears of abandonment, rejection, and separation, and attachment-related avoidance or closeness involves the individual's emotional and/or behavioral reaction to attachment anxiety (Bartholomew et al, 2001). Each item is assessed on a 7 point scale (1= disagree strongly, 7= agree strongly), indicating how much the participant agrees or disagrees with statements describing how she generally experiences relationships. Test-retest reliability on both the anxiety and avoidance scales were found by the authors to be .70 with four weeks between tests. Egna Minnen Betraffande Uppfostran (EMBU; Appendix 4) This scale, in English, is entitled "My Memories of Upbringing". Gerlsma, Emmelkamp, & Arrindell (1990) performed a meta-analysis on the psychometric and validity qualities of measures of perception of parental rearing styles, and found the EMBU to be among three that best met the criteria and provided an adequate amount of psychometric research. Cronbach's alpha coefficient for the EMBU is .80, and reliability alpha is .82 (Ross, Campbell, & Clayter; 1982). Internal consistency was found to be very high, with the number of significant correlations between the items in the Rejection subscale for mothers being 9/10, for fathers 10/10 at the p <05 level (Perris, Jacobsson, Lindstrom, von Knorring, & Perris, 1980). These researchers also found the EMBU to be uncorrelated with social desirability. Four subscales of 81 the EMBU were used for the purposes of this research. These four subscales of the English version assess recollections of maternal warmth and rejection and paternal warmth and rejection, and have been used in ongoing research on domestic violence by Dutton (1994). The questionnaire asks a series of questions that are rated on a Likert scale that ranges from one (never occurred) to 4 (always occurred) regarding how often their female primary caregiver and their male primary caregiver displayed warm and/or rejecting behaviour towards them. Abuse Three questions were asked regarding emotional, physical, and sexual abuse by both male and female primary caregivers. The questions were appended to the EMBU, and therefore scored on the same four-point Likert scale. Participants were asked to rate from one (never occurred) to 4 (always occurred) how often their female primary caregiver and their male primary caregiver abused them in the various ways listed. It was suggested that a questionnaire method would be the least intrusive way of addressing these issues with this population, particularly because the institution had recently had negative feedback from the inmates as the result of a previous research project investigating sexual abuse (K. Bartholomew, personal communication, Jan. 2001). Parental Stress Scale (Appendix 5; only for women who have had children or who have played a parental role in the life of a child). This scale includes 18 items that are scored on a 5-point scale (1= totally disagree, 5 = strongly agree). Items describe levels of parental rewards, parental stressors, feelings of lack of control, and parental satisfaction (Berry & Jones, 1995). The PSS was found to have adequate reliability. Coefficient a was 0.83 for the total sample, and the mean inter-item correlation was .23. Test-retest correlation was .81 over a period of 6 weeks. To assess validity, Berry & Jones 82 (1995) compared scores on the PSS to a generic measure of stress, the Perceived Stress Scale. All correlations were significant and in the expected direction; total sample, r (233) = .50, p < .01. These comparisons were made for various groups, including a clinical sample, results showing r (51) = .41, p < .01. Several other tests showed this scale to be both valid and reliable. Further it is specific to the construct of interest and takes little time to administer and score. Balanced Inventory for Desirable Responding- (BIDR; Appendix 6) This scale assessed social desirability and includes 40 items that are scored on a Likert scale of 1 to 7 (1= not true, 7 = very true; Paulhus, 1991). This was used to assess whether or not the participant is likely to be responding truthfully (impression management). The Impression Management subscale is used to identify individuals who are intentionally attempting to present themselves in a favourable manner when using self-report questionnaires. Paulhus (1991) reports that values of coefficient alpha range from .75 to .86 for the Impression Management Scale, which is the scale being used for the purposes of this research. He further reported test-retest correlations over a 5-week period of .65 for the IM scale (Paulhus, 1988). In addition, the sum of all 40 BIDR items shows concurrent validity as a measure of socially desirable responding in correlating .71 with the Marlowe-Crowne scale (Paulhus, 1988). The IM scale of the BIDR has also been found to correlate highly with lie scales (e.g., MMPI Lie scale, Eysenck's Lie scale), and is responsive to demands for impression management, particularly from private to public conditions (Paulhus, 1988). Further, the IM subscale of the BIDR scale has been found to be valid for use with offender populations, although it has only been validated on male violent and sexual offenders to date (Kroner & Weekes, 1996). 83 Parental Bonding Interview (Appendix 7; only for women who have had children or have played a parental role in the life of a childJ This interview was created specifically for this research project and was modeled, with permission, after the Family Attachment Interview (Bartholomew & Horowitz, 1991). To date, there is no psychometrically sound scale nor established theoretical background for assessing parental attachment to children, however, for the purposes of this study it was important to understand how female offenders viewed their relationships with their children. The Family Attachment Interview asks participants about their family background - who their caregivers were, how much time spent with each caregiver, what sort of activities they participated in with their caregivers, how they were disciplined as children, whether parents were affectionate with them6. These and other related questions were included in the Parental Bonding Interview, but rather than assessing the participants' relationships with their parents, these questions were asked in regards to their relationships with their children (e.g., they were not being asked how much time their parents spent with them, rather how much time they spend with their children). Additional questions were included to determine overall satisfaction with the participant's parenting experience- i.e., whether they enjoyed being a parent, whether they felt they were good parents, whether they felt that they were there for their children in times of need, what they would do differently as a parent if given the chance, whether they are satisfied with the current care-taking arrangements for their children, and whether they planned on regaining custody of their children once released from prison. 6 Many other topics are covered in the Family Attachment Interview, such as parents' relationships with each other, abuse, suicide attempts, and parental substance abuse, but were not used in this interview, either because the information was being gathered elsewhere, or the information was too sensitive. 84 For women who have no children, an alternative interview was used to determine whether these women were interested in becoming mothers one day and whether they thought they would make good mothers (Appendix 8). SCID-II Clinical Interview for DSM-IVAxis II Personality Disorders This is a semi-structured diagnostic interview for assessing personality disorders, and can be used to make either categorical or dimensional characterizations of personality disorders (First, Gibbon, Spitzer, Williams, & Benjamin, 1997). Use of this interview protocol renders it possible to address select personality disorders, and for the purposes of this research (based on an analysis of previous research on attachment theory, as well as research on female offenders) avoidant, dependent, paranoid, schizotypal, schizoid, histrionic, narcissistic, borderline, and antisocial personality disorders were investigated (to the exclusion of obsessive-compulsive, passive-aggressive, and depressive personality disorders). However, as the interviewers were not clinicians, no diagnoses were made- rather, 'characteristics' of personality disorders were assessed. All SCID-II interviews were rated by two coders, - the interviewer, and a second, experienced coder. When disagreements arose between the two coders, a third coder was consulted and the discrepancy was discussed until all three raters agreed. If the three raters could not agree it was determined that an average of the three scores would be taken. However, this was not required, as agreement was reached in all cases. No data are available on the reliability or validity of the SCID-II using the DSM-IV, although several studies have investigated the reliability of the SCID-II that utilized the DSM-III-R. Test-retest reliability yielded an overall weighted kappa of .53 within in a two-week period for patient sites, and .38 for non-patient sites (First et al., 1995). Malow, West, & Williams 85 (1989; as found in First et al, 1995) administered the SCID-II sections for Borderline and Antisocial Personality Disorders on a sample of 29 selected from a larger sample of inpatients with either Cocaine or Opiod Dependence, and using a test-retest design within a 48-hour time span reported a kappa of .87 for Borderline Personality Disorder (base rate 16%) and .84 for Antisocial Personality Disorder (base rate 15%). Weiss, Najavits, & Muenz (1995; as found in First et al, 1995) studied 12-month test-retest reliability on 31 cocaine-dependent patients reported an overall kappa of .46. Concurrent validity of the SCID-II was demonstrated in a study that showed that in a group of primary care patients, having a personality disorder diagnosed by the SCID-II was associated with lower functional status, lower satisfaction with health care, and higher risk for depression and alcohol abuse (Hueston, Mainous, & Schilling, 1996; as found in First et al, 1995). In other research, the SCID-II was compared to the Personality Disorder Examination (Loranger, Susman, & Oldham, 1987; as found in First et al, 1995) and their agreement with each other was compared to the Longitudinal Expert evaluation using All Data (LEAD, a clinical standard; Spitzer, 1983). Results showed that the SCID-II had better agreement, as indicated by the kappa (exact kappa not reported), and better total predictive value for 8 of the 11 personality disorders (Skodol, Oldham, & Rosnick, 1991; as found in First et al, 1995). Peer Attachment Interview (Appendix 9; for sub-sample only) This in-depth semi-structured interview includes questions regarding attachment styles in friendships and romantic relationships (Bartholomew & Horowitz, 1991). Specifically, participants were asked questions such as how many close friends they had, what 'close friend' means to them, how they got along with their close friends, and how they resolved differences with close friends. They were also asked to describe all previous close romantic relationships-86 length, seriousness, and why relationship dissolved. As well, they are asked to describe current romantic relationship- length, seriousness, and how confident they are that this relationship would last. The PAI was primarily used to garner a more in-depth view of the relationship characteristics of these women. Interviewers Interviews were conducted by the author and six other female interviewers. Interviewers were trained to administer the SCID-II semi-structured interview using the SCID-II User's Guide (First et al, 1995) that included extensive explanations, descriptions, and examples of each personality disorder, as well as how these were to be scored. These were carefully studied, and then a training tape was viewed and scored by each participant. These were checked against a 'key' and the results were discussed until all discrepancies were analyzed and understood. Then each interviewer performed complete interviews (including the SCID-II as well as the Parental Bonding Interview) on two other interviewers, who 'acted' as female inmates. Some of these mock tapes included no personality disorders, although most included at least one. These interviews served as a data base to be used as a training tool for the interviewers-in-training. Each interviewer scored a total of 15 interviews, including the training tape, but these were only to familiarize them with the semi-structured interview and to encourage discussion and questions. Each interviewer participated in over 20 hours of group meetings and at least 40 hours of interviewing and coding practice. 87 Procedure Women who agreed to participate in the study were asked to read a consent form that described the research after which the researcher inquired whether the participant had any questions. The participant was then informed that she was only to sign the consent forms if she felt comfortable participating in the study. If she agreed, she signed two forms, one for her to take, and one for the researcher's records. The participant was advised that her signature indicated that she agreed to participate in the study, complete the interviews and questionnaires, and that the interviews would be audio-taped. She was also advised that she was agreeing to allow the researcher access to her criminal files for criminal history and background information. If she did not grant access her files, the interview was still completed (this only occurred in one case). Further, each participant was informed that she was free to stop participating at any time without penalty, and she was assured of the confidentiality of her participation and that her name would not be used on the research files. The participant was then given the SCID-II Personality Questionnaire, the Background Information Questionnaire, the Experiences in Close Relationships Inventory, the Relationship Questionnaire, and the EMBU (recollections of early childhood) consecutively. It was determined in the background questionnaire whether the participant had children or a parental role (if she reported no children on the background questionnaire, she was asked whether she has ever played a parental role). If so, she filled out the Parental Stress Scale and the BIDR. While the participant was filling out these questionnaires, the interviewer prepared the SCID-II interview based on the answers to the SCID-II personality questionnaire, but was also available for any questions that the participant had about the questionnaires. Participants who had children or played a parental role in a child's life were then interviewed using the Parental Bonding 88 Interview, followed by the SCID-II semi-structured interview. Women who did not have any children filled out the BIDR at this point and were given a brief interview followed by the SCID-II interview. If any participant became upset during the interviews, the researcher turned off the audio-tape and addressed the needs of the offender. A subset of participants (30) were asked whether they wished to participate in a more extensive interview about relationships- the Peer Attachment Interview. At the conclusion of the interview each participant filled out a feedback form (see Appendix 11), so she could provide her opinion about the experience. She was asked whether she felt comfortable about the questions she was asked, and whether she felt able to answer them honestly. She was also asked to provide further comments if she so desired. Ninety-four percent of the women who answered (N=102) claimed that they answered the questions honestly. However, interviewers and coders found inconsistencies within 30.6% of the interviews- i.e., contradictions within interviews or within file information or between interviews and file information. Some of these contradictions may have been unintentional, although some were likely deliberate. This is an inherent difficulty when conducting research with an offender population, and is always a consideration when interpreting results. Nearly 6% claimed that they answered the questions partly or mostly honestly, and no participant claimed to have not answered the questions honestly. Of the women who reported their comfort level during the interview (N=106), 76.4% claimed that they felt comfortable, 17.9% claimed feeling mostly or partly comfortable, and 5.7% reported feeling uncomfortable during the interview. Some women commented on their feedback form that they found the questions too personal. Interviewers were required to fill out an interview report within 24 hours of the interview in order to monitor problems or exceptional circumstances that took place in the interview room, 89 and in order to ensure that the interviewers are not adversely affected by the interviews (see Appendix 12), and they were subsequently provided with opportunities to discuss lingering issues with regards to interviews or the interviewing experience. The interviewers were resilient and none reported adverse affects from the interviews, although some did report feeling sad when hearing particularly tragic stories, and others reported liking the women they had interviewed. After meeting with the participant, the researcher obtained access to the inmate's file (for those who agreed) and completed the "Coding Checklist" (see Appendix 10). This checklist included information on crimes, number and types of offenses, history of violence, suicide attempts, family and relationship information (including criminality of partners or other family members). In order to maintain confidentiality, participants were numbered and no other identifying information was included in the interviews or on the questionnaires. The consent form was stapled to the coding checklist only until the file was reviewed, at which time the consent form was separated from the checklist and kept in a separate location in random order. Each participant was paid $5.00 for her participation. Those who participated in the follow-up interview were paid an additional $5.00. Participants were also given two small chocolate bars during the interview, and these were often more popular than the money. 90 Results Representativeness of the Sample. It was not possible to ascertain the exact number of women admitted to and discharged from BCCW during the period of data collection, as the institutional records do not distinguish between new and repeat admissions (there were at least three participants who were released and re-admitted while the study was in progress). However, available records show that there were 271 total admissions during the four-month data collection period. One hundred and twenty one women were interviewed, and one woman participated twice. This participant's interviews and questionnaires were checked for reliability and her second set of questionnaires was discarded. Ten more interviews were discarded because the level of completion was very low. The rates of completion for the remaining 111 participants varied, as will be reflected in the results reported below. Of these 111 participants, 13.5% (n = 15) were serving a federal sentence, 44.1% (n = 49) were serving a provincial sentence, 41.4% (n = 46) were being held in remand and awaiting trial or sentencing. One immigration-hold woman was included in the study (0.9%), as she was English-speaking and was being held in protective custody with other English-speaking women. Immigration- hold houses predominantly Asian-speaking women. On an average week between February and May, 2001, the months during which data collection took place, there were 137 female inmates within BCCW, not including those being held on immigration charges (an additional 10 women, on average). Of these 137 women, approximately 28% were sentenced federally, 46% were sentenced provincially, and 26% were on remand/hold7. Chi-square analyses show that there are significant differences in legal status 7 These are averages across the 4 month period of data collection. 91 between the sample from which data were collected and the population at B C C W during the study period (x2 = 18.742, N = 11 \,p = 0001). The B C C W population had a higher proportion of federal inmates and lower proportion of women being held on remand than did the present sample. Data were available on the charges for the B C C W population at the time of the study, and it was determined that some of these were appropriately represented in the study sample (see Table 1). Specifically, property offenses were similar, as were drug offenses. However, the present sample had more violent offenders, and fewer offenders in this sample who were charged with breach, failing to appear, unlawfully at large, and similar charges than in the B C C W populations as a whole during the study period. A chi-square analysis indicates that overall, the present sample was not representative of all admissions to B C C W over the study period with regard to offenses, f (N=270) = 7.45, p < .0001. 92 Table 1. Index Offenses of Present Sample Compared to Index Offenses of All Admissions to BCCW During Study Period Charge Category Present Sample All Admissions Property 30.4% 31.9% Drugs 14.8% 17.0% Motor Vehicle 1.7% 4.8% Violent 38.3% 19.3% Sex 2.6% .4% Breach/Obstruction 12.2% 25.9% Other Criminal Acts 0% .7% N 111 271 Demographics Most demographic data that were collected on the sample cannot be compared to the population at BCCW due to the inconsistencies and lack of record-keeping at that institution. However, it was possible to determine that the ethnic breakdown of this sample was representative of the ethnic breakdown of all admissions to BCCW during the study period. Sixty-three percent of women in this sample were Caucasian, almost 29% were Native Indian, and 8.3% were Asian, Black, Hispanic, or Indo-Asian. This compares to all admissions to BCCW during the study period that showed 60.5% as Caucasian, 31% as Native Indian, and 8.3% as Asian, Black, Hispanic, or Indo-Asian. Chi-square analyses indicates that there is no 93 significant difference in this distribution, x2 (N=l 11) = 2.71, p = .26. Ethnic breakdown was also compared to Turnbull's (1996) sample, and no significant differences were found (see Table 2). As previously reported, 63% of the women were Caucasian, and 28.7% were Aboriginal, accounting for over 90% of the sample. This is noted at this point as there were differences in rates of reporting between Caucasian and Native inmates in a variety of areas. Native inmates had a significantly lower level of reported educational attainment than Caucasians (9.13 years for Natives, 10.87 years for Caucasians, functionally a difference of almost two grade levels) , t (94) = 3.509, p =.001, greater number of drug offenses than Caucasians (55% of Natives had drug offences on record, 23% of Caucasians had drug offences on record), t' (50.4) = -3.068, p =.003, a higher rate of suicide attempts reported or on file than Caucasians (91% of Natives inmates, 57% of Caucasian inmates), t' (62.8) - -3.490, p =.001, and significantly more Caucasians reported that their parents stayed together throughout their childhood than were reported by Native inmates, t' (54.2) = 2.057, p =.045. Fifty-five percent of Caucasian inmates reported that their parents stayed together throughout childhood compared with 32% of Native inmates. Based on previous reports all mean differences found here were in the expected direction (e.g., Fournier & Crey, 1997; see introductory chapter). The finding that there were no significant differences in attachment styles reported between the Native and Caucasian women in the sample was interesting given their unique circumstances, such as the greater likelihood of being raised in foster home. However, no previous research has investigated attachment in this context. Data regarding educational level for all admissions to BCCW during the study period were collected by the institution in such a way as to prevent direct comparison with the present For some t-tests, the assumption of homogeneity of variance was untenable, therefore Welch's t' was used instead. 94 sample. However, it appears that the present sample claimed a higher level of education than was found for all admissions. Reported levels of education for this sample are presented in Table 2. Three-quarters of the sample were mothers; most had one or two children, but some had as many as seven. The remaining 25% of the sample reported having no children. Reports of income are presented in Table 2. These reports were inconsistent, as some inmates included income acquired through illegal means and others did not. Further, many were unsure of their annual income, and were asked to estimate how much money they earned per week or month and how many weeks or months per year they were earning, an annual income was calculated from these figures. Many of these women do not regularly pay bills, including rent, and may have found it difficult to estimate how much money they spent on living expenses; it is also unclear whether they were able to make accurate estimates of the cost of their drug/alcohol habits, although this may be one area in which they are very aware of their finances. Regardless, it must be noted that annual income estimates may be inaccurate in many cases. The women in the sample had a variety of family backgrounds. These are reported in Table 2 along with marital status, a reported numbers of previous arrests. Almost 70% of the sample had violent offenses on record. Also, 70% of all inmates reported that they were poly-substance abusers9 (i.e., reported abusing two or more substances), and 85.5% admitted to abusing at least one substance. This sample was very similar to Turnbull's (1996) BCCW sample of women in terms of age10, ethnic background, and education. The present sample tended to have higher annual 9 No diagnosis of drug/alcohol abuse/dependence was made. This was based on self-report responses written by inmates on the Background Information Questionnaire. 1 0 The mean age of all women admitted to BCCW over the study period was 32.62. 95 incomes, more stable family backgrounds, fewer heterosexuals, more single women, and more women with children. As most of the demographic data is based on self-report, differences may exist due to factors other than actual differences between the samples. For example, in the present study, women who expressed uncertainty about how to report their income, were told that they were to report all income, including that earned through illegal means'1 (see Table 2). ' 1 A problem with this is that not all women expressed uncertainty and many may have assumed that they were only to report income gained by legal means, including, and likely primarily, social assistance. Most adults have filled out forms asking for reports of income, and this always refers to income earned through legal means. Therefore, one can surmise that the report of annual income in both studies are under-reported. 96 TABLE 2 Summary of Demographic Information for this Sample versus Turnbull's (1996) Sample Present Turnbull N= 111 N=50 AGE MEAN 29.84 30.9 STANDARD DEVIATION 7.07 10.0 RANGE 18-56 18-66 RACE CAUCASIAN 63.0 54.0 NATIVE INDIAN 28.7 12.0 OTHER 8.3 12.0 EDUCATION K-GRADE 6 5.6 6.0 JR. HIGH 19.4 GRADE 10-12 51.9 78.0 GRADE 13+ 10.2 10.0 UNKNOWN 3.6 6.0 INCOME $0-$ 10,000 41.0 46.0 $10,001- $20,000 12.0 22 $20,001- $30,000 13.0 6.0 $30,001- $50,000 10.0 8.0 97 $50,001+* 24.0 18.0 WHO RAISED BY BOTH PARENTS 43.6 24.0 COMBINATIONS (e.g., step-parents, grandparents) 35.0 52.0 FOSTER CARE 21.8 24.0 SEXUAL ORIENTATION HETEROSEXUAL 73.0 84.0 LESBIAN 5.4 4.0 BISEXUAL 21.6 12.0 CURRENT MARITAL STATUS SINGLE 39.8 26.0 MARRIED OR COMMON-LAW 43.5 58.0 SEPARATED OR DIVORCED 16.7 14.0 NUMBER OF CHILDREN MEAN 1.77 1.5 STANDARD DEVIATION 1.48 1.6 RANGE 0-7 0-5 PERCENTAGE WITH NO CHILDREN 24.3 38.0 Note: Values shown are percentages unless otherwise indicated. *Women who earned over $50,000 per year were earning this through illegal means (as were many of the women who earned less than $50,000 per year). 98 Scoring The SCID-II Semi-Structured Interview, the Relationship Questionnaire (RQ; Kim Bartholomew, 1990), the Experiences in Close Relationships Inventory (ECRI; Brennan, Clark, & Shaver, 1998), the Early Experiences in Childhood Questionnaire (EMBU; Perris, Jacobsson, Lindstrom, von Knorring, & Perris, 1980 ), the Balanced Inventory for Desirable Responding (BIDR; Paulhus, 1991), and the Parental Stress Scale PSS; Berry & Jones, 1995) were all scored using standard procedures, as described by their respective authors. The Parental Bonding Interview and Background Information Questionnaire were coded as answered. Some answers were categorized into nominal groups and assigned arbitrary values to assist in data analysis. A trained coder according to standard coding procedures coded the Peer Attachment Interview (PAI), and the attachment category with the highest score was selected as the predominant attachment style for each participant. Reliability and Validity of Measures Relationship Questionnaire. For the present sample, the scores for the RQ when used categorically ("choose which of the four best describe your romantic relationships") were compared to the dimensional scores of the RQ for romantic relationships. Each category was significantly higher on its corresponding dimensional score,/? < .005. In other words, women who chose the "secure" description as most representative of their romantic relationships were significantly more likely to score the highest on the "secure" description when asked to rate each description dimensionally. The same held for the "fearful", "preoccupied", and "dismissing" categories. 99 Experiences in Close Relationships Inventory. A factor analysis of the ECRI items in the present sample found the expected underlying two factor structure. It appears that these two factors account for more than 42% of the variance in the present sample, demonstrating the validity of the use of this scale. Further, it was found that internal consistency of the ECRI with this sample was good, with an alpha coefficient for the avoidance factor of .88, and for the anxiety factor of .91. EMBU. A factor analysis of the EMBU found only two subscales, rather than four for the present sample. It appears that warmth and rejection are simply opposite ends of the same trait. That is, that high scores on warmth are correlated with low scores on rejection and vice versa, r = -.833 for mothers and r = -.738 fathers. This one factor, here entitled "rejection", accounts for more than 48% of the variance for fathers in the present sample and nearly 60% of the variance for mothers in the present sample. It was further found that internal consistency of the EMBU for this sample was adequate, with an alpha coefficient for fathers of .74, and for mothers of .74. Parental Stress Scale. A factor analysis of the PSS found that two factors account for more than 44% of the variance in this sample, which compares with 42% of variance accounted for by the highest two factors as reported by Berry & Jones (1995). Internal consistency of the PSS for this sample was adequate, with an alpha coefficient of .75. Balanced Inventory of Desirable Responding. A factor analysis of the BIDR showed that there was one factor accounting for 15% of the variance in this sample, but it was unclear whether this factor is Impression Management or Self-Deception, which is the other subscale of the measure. Internal consistency for the Impression Management scale was found to be adequate for this sample, with an alpha coefficient of .77. 100 Impression Management Due to the nature of the population being studied, it was deemed appropriate to administer a social desirability scale. The Balanced Inventory of Desirable Responding (Paulhus, 1984) was used, and indicated a rather low level of impression management on the part of the inmate sample studied. The mean score for Impression Management was 5.39, with a standard deviation of 3.20 (lower scores indicate lower levels of impression management). Mean norms for Impression Management on the BIDR include the following: in a sample of 884 religious adults, the mean for men was 7.3 (s.d. = 3.1), and for women was 8.9 (s.d. = 3.2) (Quinn, 1989). The women in the present research had a significantly lower mean impression management score, ;(110) = 6.3, p < .001, than the Quinn (1989) sample of women. In a sample of 100 college students, a mean of 11.9 (s.d. = 4.5) was reported for impression management in a public disclosure condition (Paulhus, 1984). Again, the mean for impression management for the women in the present sample (5.39) was significantly lower, /(110) = 21.44,/? < .001. A sample of 48 Alcoholics Anonymous members had a mean of 11.2 (s.d. = 4.9) for impression management using the BIDR (Mellor, Conroy, & Masteller, 1986). This is also significantly higher than the mean for impression management in the present sample of women, t(\ 10) = 19.14,/? < .001). However, in a sample of college students in a non-public disclosure condition, the mean impression management score for men was 4.3 (s.d. = 3.1) and for women was 4.9 (s.d.= 3.2) (Paulhus, 1988), and this was not significantly different than the mean in the present sample, t(\ 10) = 1.61, p = .111. Because most of the present sample had extensive criminal histories, and because their lifestyles were characterized by dishonest interactions with others (e.g., fraud, property crimes), it was thought that the impression management scores of this sample would be higher than those for the various comparison samples. 101 The Impression Management (IM) scale of the BIDR was compared with scores on the EMBU, RQ, ECRI, SCID-II, and PSS measures (see Table 3). No significant correlations were found between the BIDR and the EMBU scores for maternal or paternal warmth or rejection. However, it was found that higher Impression Management scores were significantly negatively correlated with the dimensional scores on the fearful category of the RQ, and positively correlated with dimensional scores on the dismissing category of the Relationship Questionnaire, although there was no significant difference between women categorized as secure and women categorized as not secure using the RQ on mean IM scores. Further, it was found that mean scores on the ECRI were highly significantly correlated with IM scores, both when compared as individual categories and when dichotomized into secure and not secure categories. Finally, it was found that there was a significant negative correlation between mean scores on the Parental Stress Scale and the IM scale. Women with high impression management scores tended to report lower levels of parental stress. It was also found that women with fewer reported personality disorder characteristics tended to have higher IM scores. 102 TABLE 3. Comparisons between Impression Management Scores and the EMBU, RQ, ECRI, & PSS. Scale IM score Sign. N Parental Stress Scale -.315** .003 87 Number of PDs -.252** .009 111 Maternal Rejection/EMBU -.115 .249 102 Maternal Warmth/EMBU .193 .051 102 Paternal Rejection/EMBU -.071 .511 88 Paternal Warmth/EMBU .130 .226 88 ECRI- Secure - 413*** .000 111 ECRI- Fearful 397*** .000 111 ECRI- Preoccupied . 4 1 9 * * * .000 111 ECRI- Dismissing -.357*** .000 111 RQ- Secure .122 .204 111 RQ- Fearful -.187* .050 110 RQ- Preoccupied -.179 .061 110 RQ- Dismissing .214** .024 111 Secure/Not Secure Categories Secure Not Secure Scale t df Sign. Mean s.d. Mean s.d. ECRI -4.018 109 .000*** 7.36 3.16 4.72 2.94 RQ -1.314 104 .192 6.22 3.54 5.29 3.05 *=p<. 05 **=/?<.01 *** =p< .001 (2-tailed significance). 103 New Admissions Women in this sample were categorized into "new" (14 days or less) or "not new" (more than 14 days) categories based on length of current incarceration. Their mean scores on the PSS, EMBU, ECRI, RQ, amount of personality pathology (mean number of personality disorders) and BIDR were compared in order to assess whether being newly admitted affected mean scores. T-tests found no significant differences between new admissions and women who had been incarcerated for at least 14 days on any measure. To summarize, all the scales show adequate validity for use with the sample, with the possible exception of the BIDR impression management scale, however the IM results should not be disregarded. Attachment Styles The Relationship Questionnaire The Relationship Questionnaire (RQ) was administered to each participant. Of four descriptive paragraphs, the women were asked to choose the category that most accurately characterized their romantic relationships, and again to describe their platonic friendships. Additionally, they were asked to characterize the degree to which each of these paragraphs described their romantic relationships, and separately, their platonic friendships. Using the best-fit (categorical) descriptors, 25.5% of the women characterized themselves as Securely attached, 42.5% characterized themselves as Fearfully attached, 12.3% characterized themselves as Preoccupied, and 19.8% characterized themselves as Dismissing in their romantic relationships. These results do not differ from Turnbull's (1996) previous findings using a sample of 50 women at the same institution. A chi-square analysis comparing 104 the present RQ results with the RQ results from Turnbull's (1996) sample RQ results was not significant, x2 (N= 111) = 4.602,/? = .203. However, these results do differ from previous findings reported by Scharfe and Bartholomew (1994), who used both self-report questionnaire (RQ) and the Peer Attachment Interview (PAI) in a sample of 80 university women (see Table 3). A chi-square analysis comparing the present RQ results with Scharfe and Bartholmew's (1994) RQ results from their university sample shows highly significant differences between the two samples, x2 (3, N = 111) = 75.26,/? <.001. Fewer women in the present sample characterized themselves as securely attached, and more women characterized themselves as fearful and dismissing, as compared to Scharfe and Bartholomew's (1994) university sample, which is not surprising, given that female offenders tend to have histories of abuse and unstable family constellations during childhood to a greater degree than women in the general population. The second categorical description characterized platonic friendships (see Table 4). It was expected that female offenders would be more likely to be able to form secure relationships with platonic (likely female) friends even if they were unable to form secure romantic relationships with men. However, a chi-square analysis comparing the platonic friendship RQ results to the romantic relationships RQ results in this sample yielded no significant difference, X2 (3, N=l 11) = 7.45,/? = .06. Nonetheless, it appears that there is a tendency for women in the sample to be slightly more secure, less fearful, less preoccupied, and more dismissing in their platonic friendships than in their romantic relationships. 105 TABLE 4. Relationship Questionnaire Across Various Samples RQ-ROM-GS RQ-ST RQ-SB PAI-SB RQ-FRND-GS SECURE 25.5 26 53 48 33 FEARFUL 42.5 36 21 18 36 PREOCCUPIED 12.3 20 20 31 7.3 DISMISSING 19.8 18 6 4 24 N 106 50 80 80 109 Note: Values reported in percentages. RQ-ROM-GS= Relationship Questionnaire best-fit (categorical) romantic relationships, present research RQ-ST= Relationship Questionnaire for Susan Turnbull's (1996) research. RQ-SB= Relationship Questionnaire for Scarfe & Bartholomew's (1994) research. PAI-SB= Peer Attachment Interview for Scarfe & Batholomew's (1994) research. RQ-FRND-GS= Relationship Questionnaire best-fit (categorical) friendship relationships, present research. The participants were then asked to rate the four descriptive RQ paragraphs on a scale from one to seven, to indicate the degree to which each description characterized their romantic relationships. This provided a dimensional score for each attachment style in addition to the categorical score previously given. T-tests indicated that the mean scores for the present sample on the secure, fearful, and dismissing categories were not significantly different from Turnbull's (1996) sample. There was a significant difference between the samples on the preoccupied score, t (109) = 2.98,/? = .004. Fewer women reported preoccupied attachment styles in the present sample. However, when compared to Scharfe & Bartholomew's (1994) university sample, the inverse was true: there were significant differences between the present sample and the 106 university sample for the secure, f (110) = 4.43,p < .001, fearful, t (109) = 5.34,/?<.001, and dismissing, t (110) = A.6\,p <.001, attachment styles, but for the preoccupied category, differences did not reach conventional levels of significance, t (109) = 1.78,/? = .08. Again, as expected based on what is known about the histories of female offenders, the present sample reported lower mean scores for the secure attachment style. They also had higher mean scores for fearful and dismissing attachment styles, although there was no difference between the present inmate sample and the university sample on the Preoccupied attachment score (see Table 5). The participants were then asked to rate the four descriptive RQ paragraphs on a scale from one to seven, to indicate the degree to which each description characterized their platonic friendships. These results were compared to the dimensional scores on the RQ for romantic relationships and, as with the categorical scores, t-tests indicated that the scores were not significantly different. 107 TABLE 5. Comparison Between the Current RQ Dimensional Scores, Those from Another Forensic Sample, and Those from a Non-Forensic Sample. Attachment Style GS RQ Results ST RQ Results SB RQ Results Mean s.d. Mean s.d. Mean s.d. SECURE 3.68 (2.15) 3.45 (2.24) 4.58***(1.83) FEARFUL 4.48 (2.15) 4.65 (1.94) 3.38***(2.03) PREOCCUPIED 2.87 (2.10) 3.47**(2.03) 3.23 (1.95) DISMISSING 3.66 (2.19) 3.67 (1.90) 2.70***(1.65) N 111 50 80 ***p = < 0001 **p = <.001 Significant findings indicate that the mean number indicated was significantly different from the present (GS RQ Results) for the designated attachment style. Note: Analyses from this point only used RQ-romantic scale, as there were no significant differences between the RQ romantic and RQ friendship results. RQ-GS= Relationship Questionnaire categorical romantic relationships, present research RQ-ST= Relationship Questionnaire for Susan Turnbull's (1996) research. RQ-SB= Relationship Questionnaire for Scarfe & Bartholomew's (1994) research. Experiences in Close Relationships Inventory The Experiences in Close Relationships Inventory (ECRI), asks participants to rate thirty-six questions on a scale from one to seven. Results are displayed in Table 6. A chi-square analysis comparing the present ECRI categorical results with the present RQ categorical results for romantic relationships revealed significant differences between the two methods of assessment, %2 (3, N=l 10) = 56.54, p < .0001. Results using the ECRI indicate fewer Fearful 108 attachment styles and far more Preoccupied attachment styles than results using the RQ. However, the ECRI and RQ were similar in terms of proportions of women who were classified as secure and not secure. These ECRI results differ, as expected, from previous research by Brennan, Clark, & Shaver (1998), who found the following proportions using the ECRI with a sample of 1,082 undergraduates: secure: 30.4%, fearful: 24.4%, preoccupied: 24.4%, dismissing: 20.8%. A chi-square analysis comparing the present results with the results from the Brennan et al. (1998) university sample was significant, x2 (N=l 11) = 15.73,p< .001. A smaller proportion of women in the present forensic sample identified themselves as secure, fearful, and dismissing, and a larger proportion described themselves as preoccupied than was found with the undergraduate sample (see Table 6). In summary, the results of both the RQ and the ECRI found that a higher proportion of female offenders in the present sample characterized themselves as one of the insecure attachment styles. However, there is clearly a problem with these measures with respect to classifying fearful and preoccupied attachment styles, therefore the remainder of the results will address secure and not secure attachment categories only. 109 TABLE 6. Experiences In Close Relationships Inventory Compared To Relationship Questionnaire for Present Sample and Non-forensic Norms ECRI- GS RQ-GS SECURE 25.2 25.5 FEARFUL 25.2 42.5 PREOCCUPIED 35.1 12.3 DISMISSING 14.4 19.8 N 111 111 Note: Values reported in percentages. ECRI-GS= Experiences in Close Relationships Inventory, present research. RQ-ROM-GS= Relationship Questionnaire categorical romantic relationships, present research. Peer Attachment Interview A sub-sample (approximately one quarter) of participants was interviewed using the Peer Attachment Interview in order to provide a richer understanding of their interpersonal relationships. It was not possible to interview all participants using the PAI due to time and financial constraints involved with training, interviewing, and coding, and as a result, too few women were interviewed using this tool to provide a reliable sample size. However, the PAI provided an excellent opportunity to speak with several of the inmates at length about their interpersonal relationships. Excerpts from some of these interviews are provided in the discussion section for illustrative purposes. 110 Personality Disorders Although personality disorders were not of primary interest in the present study, the presence of characteristics of personality disorders was evaluated using the SCID-II semi-structured interview, which yielded the following results with the present sample: 73% of the women interviewed (N=106) had characteristics consistent with one or more personality disorders; 32% had characteristics consistent with one personality disorder, a further 25% had characteristics consistent with two personality disorders, 11% had characteristics consistent with three personality disorders, and 4 % had characteristics consistent with four personality disorders. Percentages of women in this sample with characteristics related to each personality disorder representation are reported in Table 6. These results differ somewhat from previous findings in a Swedish community sample (N= 557) (Ekselius, Tillfors, Furmark, & Fredrikson, 2001). Significant differences were found between the present sample and this community sample for Narcissistic personality disorder %2 (N=106) = 30.91,p< .0001, borderline personality disorder, f (N=106) = 137.40,/? < .0001, and antisocial personality disorder %2 (N=106) = 2078.50,/? < .0001. Predictably, the community sample had lower rates of these personality disorders than were found in this female offender population (see Table 7). The present results also differ moderately from an average that was taken across four studies of clinical patients (overall N = 568) using DSM-III and DSM-III-R criteria (Widiger & Rogers, 1989). Significant differences were found between the clinical patients and the present sample for Schizotypal personality disorder x2 (N=106) = 22.70,/? < .0001, Histrionic personality disorder x2 (N=106) = 37.24,/? < .0001, and Antisocial personality disorder x2 (N=106) = 173.90,/? < .0001. More clinical patients were found to have Schizotypal and 111 Histrionic personality disorders than were characteristic of women in the present sample, and predictably, women in the present sample were more likely to present characteristics of Antisocial personality disorder than clinical patients (see Table 7). Inmates, both male and female, tend to have higher rates of antisocial personality disorder than do non-offenders. The present results were also compared to a Canadian sample of female offenders from the same institution (N=75), who were assessed using DSM-III-R criteria (Tien et al, 1993). Significant differences between the present sample and the other female forensic sample were restricted to Narcissistic personality disorder, x2 (N=106) = 11.78, p = .001, and Borderline personality disorder x2 (N=106) = 18.06,/? < .0001 (see Table 7), the present sample of women having higher rates of both. 112 TABLE 7 Comparison Between the Current Personality Disorder Scores and Those From Another Forensic Sample, A Study Of Four Clinical Samples, and a Non-clinical Swedish Sample. Female Inmates Male Inmates Clinical Sample Swedish Comm. GS Tien et al Widiger & Rogers Ekselius et al. AVOIDANT 12.3 15 22 6.6 Dependent 4.7 3 16 1.8 Paranoid 6.6 12 7 5.6 Schizotypal 2.8 3 22*** 5.2 Schizoid 1.9 3 5 0.9 Histrionic 2.8 1 2.2 Narcissistic 12.3 5** 10 2 Q * * * Borderline 31.3 i 38 Antisocial 61.0 49 1 g*** N 106 75 568 557 Note: Values reported in percentages. ***p<.0001 **/?<.001 GS = SCID-II results, present research. Tien et al = Tien et al (1993) research on a Canadian sample of female offenders, using a previous version of the SCID-II, the AutoScid. Widiger & Rogers = Widiger & Rogers (1989) research using 4 studies of clinical samples: Dahl (1986)- N= 108; Morey (1988)- N= 291; Pfohl, Coryell, Zimmerman et al (1986) N= 131; Zanarini, Frankenberg, Chauncey et al (1987)- N= 43. Ekselius et al= recent research (2001) on a community sample (postal survey), using the DIP-Q, a questionnaire. 113 Attachment styles and Personality Disorders When assessed using the Experiences in Close Relationships Inventory, it was found that there was a significantly greater number of personality disorders (or characteristics consistent with personality disorders) present in women who fell into one of the insecure attachment categories (mean = 1.53) than women who had a secure attachment style (mean = .73), as assessed using the ECRI, /' (57.486) = -3.869,/? < .0001. As noted in the introduction, female offenders generally have high levels of personality pathology, and they also have histories characterized by abuse and unstable relationships with primary caregivers, which may lead to personality pathology. Therefore, this finding was in accordance with expectations. However, when IM was controlled for, there remained a significant relationship between attachment styles (ECRI) and number of personality disorders, F= 6.62, p = .012. Recollections of Early Childrearing The EMBU questionnaire required participants were asked to rate a variety of childhood experiences with their parents on a Likert-type scale ranging from one to four (1= never occurred, 4 = always occurred). These answers produced means on two dimensions: parental warmth and parental rejection, with higher mean scores for maternal and paternal rejection indicating higher levels of perceived rejection, and higher mean scores for maternal and paternal warmth indicating higher levels of perceived warmth. Thirty-six women failed to complete all of the questions on the EMBU and scores for these participants were pro-rated accordingly. Some participants did not have a consistent primary caregiver (20 women did not complete the questionnaire for a male primary caregiver, and 7 women did not complete the questionnaire for a female primary caregiver) and therefore were unable to answer the questions. Some 114 participants refused to answer questions about their primary caregivers (usually the male) as it upset them to think about these past relationships. T-tests indicate that the mean scores for the current sample on maternal and paternal warmth and rejection are significantly different from previously reported results using a control community sample (Arrindell, W.A., Kwee, M.G.T., Methorst, G.J., Van Der Ende, J. , Pol, E., & Moritz, J.B.M., 1989; See Table 8). As expected, mean scores for paternal rejection, t (87) = 7.30, p < .0001, and maternal rejection, t (101) = 7.88,/? < .0001, were much higher in the present sample, and mean scores for paternal warmth, t (87) = -3.39,/? = .001, and maternal warmth, / (101) = -3.74,/? = .0001, were somewhat lower in the present sample than found in previous research using a non-forensic sample (see Table 8). No significant correlations were found between maternal or paternal warmth and attachment, however, mean scores for maternal rejection were significantly higher for women who fell into one of the insecure attachment categories (mean score = 56.62, s.d. = 23.58) than for women who fell into the secure attachment category (mean score = 44.48, s.d. = 19.88), as assessed using the ECRI, t' (50.92) = -2.56,p = .01. A significant relationship was also found between paternal rejection and attachment. Mean scores for paternal rejection were significantly higher for women who fell into one of the insecure attachment categories (mean score = 55.43, s.d. = 20.94) than for women who fell into the secure attachment category (mean score = 38.86, s.d. = 17.84), as assessed using the ECRI, / ' (51.47) = -3.73,/? < .0001. These relationships remained significant even after controlling for impression management scores. Additionally, scores for maternal and paternal warmth and rejection were compared and were found to be highly correlated (-.833 for mothers and -.738 fathers), suggesting that warmth and rejection may simply be opposite ends of the same dimension, rather than two separate dimensions. 115 In summary, women who fell intq one of the insecure attachment categories had higher mean scores for maternal and paternal rejection than women who fell into the secure attachment category when attachment was assessed using the ECRI. TABLE 8. Results of EMBU in present sample compared to Control sample REJECTION EMOTIONAL WARMTH Mean S.D. N Mean S.D. N Fathers Control 34.11 7.16 95 50.43 10.93 95 Present Sample 50.72 21.37 88 45.35 14.05 88 Mothers Control 35.39 7.75 95 52.88 10.01 95 Present Sample 53.52 23.22 102 45.78 14.38 102 Reports of Abuse by Primary Caregivers Participants were asked to rate on a Likert-type scale ranging from one to four (1 = never, 4 = always) whether their male and female primary caregivers abused them emotionally, physically, and/or sexually (see Table 9). 116 TABLE 9. Reports of Abuse by Primary Caregivers MOTHER n FATHER n ALMOST ALWAYS OR ALWAYS EMOTIONALLY ABUSED ME 41.1 % 42 38.9 % 35 Almost always or always physically abused me 31.3 % 31 31.4% 27 Almost always or always sexually abused me 0 % 0 12.3 % 11 It was not possible to compare these reports to national estimates of child abuse in the general Canadian population, as these estimates do not exist (Statistics Canada, 1998). However, reports of child sexual abuse in the United States have ranged from 9% to 54% of women (Finkelhor & Baron, 1986). A well-known study that included 1374 females interviewed in-depth over the telephone found that 27% of females surveyed reported childhood sexual victimization (Lewis, 1985; Timnick, 1985; as found in Finkelhor & Baron, 1986). According to Finkelhor (1993), the better studies- that is, the ones with more carefully designed questionnaires and better trained interviewers- generally found higher rates of sexual abuse. At first blush the present findings appear lower than the rates of abuse reported in the literature, however, it should be noted that previous studies usually included reports of abuse by anyone and, as a result, the telephone survey produced higher reported rates of abuse than the women in the present sample who were only asked about abuse by male and female primary caregivers. Therefore, these results are within the expected range. There were no relationships found between reports of abuse by mother and attachment styles, however there were relationships found between reports of emotional and physical abuse by father and attachment style when attachment styles were collapsed into secure and insecure 117 categories. When attachment was assessed using the Relationship Questionnaire, it was found that women who reported an insecure attachment style reported higher means of emotional abuse (mean score12 = 2.17, s.d. = 1.19) by the male primary caregiver than those women with secure attachment styles (mean score = 1.91, s.d. = 1.20), t' (48.664) = -2.261,/? = .028. When assessed using the ECRI, it was also found that women who reported an insecure attachment style reported higher means of emotional abuse by fathers (mean score = 2.26, s.d. = 1.20, on scale of one to four), than women who reported a secure attachment style (mean score = 1.72, s.d., 1.10, no scale of one to four), t' (47.381) = -2.037, p- .047. However, when IM scores were controlled for, only the attachment styles as assessed by the ECRI (not the RQ) were found to be significantly related to emotional abuse by fathers, F = 5.214,/? = .025. Further, when assessed using the ECRI, it was found that women who reported an insecure attachment style also reported higher means of physical abuse by fathers (mean score = 2.10, s.d. = 1.26), than women with a secure attachment style (mean score = 1.42, s.d. = .83), t' (63.538) = -2.915,/? = .005. When impression management was controlled for, attachment security was significantly related to physical abuse by father for both the RQ (F = 4.653,/? = .036) and the ECRI (F = 7.287,/? = .008) attachment measures. As previously noted, female offenders often report childhood abuse, and further, secure attachment of a child is related to sensitive and responsive caregiving by the parents. Therefore, the fact that women who were abused by their fathers tend to report insecure attachment styles is not surprising. In summary, it was found that women who reported an insecure attachment style had higher mean scores of both emotional (ECRI) and physical (ECRI & RQ) abuse by fathers than women who reported a secure attachment style. 1 2 Using a scale of one to four. 118 Inmates as Parents Approximately 24% of the sample (n = 26) reported not having any children. Twenty-five percent of non-mothers claimed that they did not want to have children in the future, 68% said that they did want to have children in the future, and the remaining 12% were undecided. The 26 non-mothers were also asked whether they thought they would make good mothers if they were to have children; only one said no (4%). The majority of women believed that they would make good mothers (85%), and three did not respond to this question. In this sample, women without children were significantly younger (mean age = 26.7 years) on average than women with children (mean age = 30.85 years), / (109) = -2.72,/? = .008. Child Custody Nearly 76% (n= 84) of the women in the present sample reported having given birth to at least one child; 88.4% (n= 74) of these women were not the primary caregivers for their children immediately prior to incarceration. Of the women who were not primary caregivers for their children prior to incarceration, 60.8% (n = 45) claimed that social services had taken at least one child away, 38.2% (n= 28) said that they voluntarily gave their children away, and 9.5% (n=7) reported that someone else had taken their child away13. Women were asked why they were no longer the primary caregivers for their children14. A summary of their responses can be found in Table 10. 1 3 This was usually the child's father, but may have also been another relative. 1 4 Totals add up to more than 100% because some mothers gave more than one reason for losing custody of their children. 119 TABLE 10. Reasons Given By Female Inmate Sample Regarding Loss Of Children REASON % N Drug or alcohol abuse 79.7 62 Abusive relationship with romantic partner 6.8 6 Child abuse/neglect 8.1 6 Divorce agreement 2.7 2 Prior incarceration or crimes 9.5 9 Mental illness 2.7 2 Too young when gave birth 1.4 1 No reason given 5.4 4 * Percentages calculated based on n = 74 women who had given birth and no longer has custody of children15. Inmate mothers were asked about the location of their children's fathers16. Results are displayed in Table 11. 1 5 Women were asked why they lost their children and these were the answers spontaneously given. No suggestions or checklist was provided. 16Based on the responses of 80 women rather than 84. Four women did not answer this question. 120 TABLE 11. Location Of Fathers Of Female Inmates' Offspring LOCATION % N STILL COUPLED 8.6 7 Know location and in contact 40.0 32 Know location and no contact 33.8 27 At least one father is in jail 22.5 18 Do not know location 21.3 17 At least one father is dead 11.1 9 Do not know identity of father 1.3 1 Of the women who had given birth, 69% (n=58) reported having more than one child, and these women were asked how many different fathers their children had (see Table 12). In total, 52.8% of participants who had more than one child had given birth to the children of more than one man. Of the 10 women who were still the primary caregivers for their children, seven of these had at least two children, and there was only one father for the children of five of these seven women. However, only one of the 10 women who were still the primary caregivers for their children prior to incarceration remained coupled with the father. As previously reported, 7 of the 81 mothers in the sample reported being still coupled with the father of their children. Only one of these was still the primary caregiver of her children. Therefore, 6 women were still coupled with the father of their children but were no longer the primary caregivers for their children. 121 TABLE 12. Number of Fathers of Female Inmates' Children (for those with more than one child) Number of fathers % n One father 47.2 25 Two fathers 37.7 20 Three fathers 11.3 6 Four fathers 3.8 2 No response 6.0 5 Women who were not the primary caregivers for their children prior to being incarcerated were asked who was currently playing the role of the primary caregiver17 (see Table 13). 1 7 Note that these add up to over 100% because some women have more than one child, and not all are being cared for by the same primary caregiver. 122 TABLE 13. Primary Caregivers for Inmates Who No Longer Have Custody of Children Primary Caregiver % n Grandparents 45.2 33 Father 21.9 16 Foster parents 37.0 27 Adopted parents 19.2 14 Other relative 13.7 10 Friend 1.4 1 Step-father 2.7 2 Child is dead 1.4 1 Inmates were also asked whom their children were closest to18. See Table 14 for their responses. 1 8 Note that these percentages add up to over 100% as some women had more than one child, and not all children were closest to the same individual. 123 TABLE 14. Inmates Reports of Who Child is Closest To % n Inmate 56.8% 42 Grandparents 30.1% 22 Father 23.3% 17 Adopted parents 13.7% 10 Foster parents 8.2% 6 Other relative 8.2% 6 Each other 2.7% 2 Don't know 1.3% 1 When inmates were asked whether or not they were affectionate with their children, 98.8% claimed they were, and 1.3% said they were not affectionate with their children. When inmates were asked whether or not they used physical discipline with their children, 83.3% claimed no use of physical discipline, 21.8% claimed they used mild physical discipline (swat on the bottom), and 2.6% admitted to using serious physical discipline with their children19. This finding was surprising, given that the majority of these women have had previous contact with Social Services, or had voluntarily relinquished custody of their children, presumably because they were overwhelmed by the responsibilities accompanying parenthood. Inmate mothers were also asked how often they saw their children when not in prison (see Table 15). 124 TABLE 15. How Often Inmate Sees Children When Not In Prison Everyday 25% Weekly 25% Monthly 10% More than once per year 17.5% Less than once per year 17.5% Never 23.8% Inmate mothers were additionally asked whether they were satisfied with the caregiving arrangement for their children while they are in prison and 13.6% replied no for all children, 4.9% were satisfied with the present caregiving arrangement for only some of their children, and 81.5% were satisfied with the caregiving arrangement for all of their children. When asked whether they planned to have their children living with them when released, 30% of inmates said they did not plan to have their children living with them, 45% said that they eventually planned to have their children living with them after release, and 25% (n=20) planned to have their children living with them upon their release. Note that only 10 (12.5%) of mothers in this sample had their children living with them prior to incarceration. When asked whether they enjoyed being parents, 1.2% of inmate mothers claimed that they did not, 4.9% replied that they "sort o f enjoyed parenthood, and 93.8% of those reporting 1 9 Again, these add up to over 100% as these women were reporting for each individual child, and some recognized 125 (N=81) said that they did enjoy being a parent. When asked whether they thought they were good parents, 5.1% said they were not good parents, 21.5% said that sometimes they were good parents, and 73.4% of those who reported (N=79) considered themselves to be good parents. When asked whether they felt that they "were there" for their children, 3.9% said they felt that they were not there for their children, 10.5% said that they felt that they were sometimes there for their children, and 85.5% of those who answered the question (N=76) said that they felt that they were there for their children. Parental Stress Scale Six of the participants who were assessed using the Parental Stress Scale (PSS) did not complete all of the items. When these six respondents were removed from the analyses, the mean PSS score was found to be 33.38 (s.d. = 9.28, n= 82). Five of the six respondents who did not complete all items missed only one item, and the completed items were pro-rated to take this into account. The PSS mean for this larger, more inclusive group was only slightly different, with a mean of 33.43 (s.d. 9.27, n = 87). This mean is compared to a mean score of 43.2 for a clinical group (parents of children with behaviour problems) and 37.1 for a non-clinical control group (parents of children with typical development; Berry & Jones, 1995). Higher scores are indicative of higher levels of parental stress; thus, the results indicate that the inmate mothers in the current sample reported experiencing lower levels of parental stress than Berry & Jones' (1995) clinical, t (86) = -9.83,p < .0001, and control samples, t (86) = -3.69,p< .0001. Because of the fact that most mothers in the sample were no longer playing the role of primary caregiver for their children, it was expected that these women would report much higher rates of parental stress than the control sample, if not the clinical sample. that did not treat all children in an identical manner. 126 When attachment styles were assessed using the Relationship Questionnaire and were collapsed into secure and insecure categories, it was found that women with an insecure attachment style reported higher means of parental stress (mean score = 34.07, s.d. = 9.82) than women with an secure attachment style (mean score = 30.03, s.d. = 6.39), t' (49.61) = -2.117,/? = .039. The same relationship was also found when attachment styles were assessed using the ECRI. Women with an insecure attachment style reported higher means of parental stress (mean score = 34.45, s.d. = 9.87) than women with a secure attachment style (mean score = 30.59, s.d. = 6.77) t' (58.84) = -2.062,/? = .044. However, when impression management was controlled for, the relationship between attachment security (as assessed by RQ & ECRI) and PSS scores were no longer significant. It had been predicted that women with an insecure attachment orientation would report more parenting stress than women who reported secure attachment orientations, simply because an insecure attachment orientation is characterized by interpersonal difficulties, difficulties that often stem from problems with the individual's own upbringing, and often characteristic of the parenting they themselves received. 127 It should be noted that there was a significant relationship between low scores on the PSS and high scores on the Impression Management scale of the BIDR, and when IM scores were co-varied with attachment security and PSS scores, it was found that impression management was a better predictor of parental stress scores than was attachment security (as measured by both the RQ and the ECRI). Impression management is a multi-dimensional construct and this complex relationship between measures in not surprising given the potential differences in motivations of female offenders from the college student samples that were used for the development of this scale. Mediation In order to test for a mediating relationship, two critical assumptions must be met. The first is that there is some basis for hypothesizing temporal precedence. In the present case this is not an issue, as childhood experiences of abuse and parental warmth and rejection clearly precede adult personality pathology and parental stress. The second assumption that must be met is that there be a significant relationship between the independent variables (childhood variables) and dependent variables (adult variables) that the other variable (attachment) is purportedly mediating. Because no significant relationship was found between abuse experienced in childhood or parental warmth and rejection and levels of personality pathology or parental stress scores, it was not possible to pursue this hypothesis. However, it should be noted that there were significant relationships within the independent variables (i.e., between child abuse and parental warmth and rejection) and within the dependent variables (i.e., between # of personality disorders and PSS scores), as would be expected. These latter findings speak to the validity of these measures with this sample. 128 The possibility also had to be considered that there was a factor moderating the relationships between the childhood and adulthood variables of interest, and this was investigated. A series of univariate ANOVAs were conducting for both the PSS and personality pathology variables in relationship to the measures of childhood adversity (parental warmth and rejection, abuse by parents) and attachment styles (as assessed by ECRI and RQ), controlling for the effects of impression management. Results showed that there were no noteworthy interactions involving attachment orientations, ruling attachment style out as a moderating factor. However, as noted previously, complex relationships were found between impression management in that it was significantly related to a number of variables when controlled for in the various analyses, although these relationships did not alter the findings with the exception that relationship between the PSS and attachment styles was found to be no longer significant when controlling for IM scores. In addition, the strength of some of the relationships tended to vary somewhat when IM was controlled for. The relationships between attachment security, as assessed by the ECRI, and emotional and physical abuse by fathers, personality pathology, and parental rejection, were particularly robust in light of the IM issues, which is interesting given that each subscale of the ECRI was significantly negatively correlated with the IM scale. However, this appears to reflect a tendency by the participants to not endorse the high end of the scale, and this tendency was consistent across attachment styles. 129 Discussion This discussion begins with a recapitulation of the main findings, a commentary on the representativeness of the sample, the veracity of reports as well as impression management issues, followed by select demographic variables of the sample. Next the childhood experiences as well as the personality pathology of the women in this sample are examined. The discussion then turns to the general implications and concerns of the attachment variables and the relationships between attachment and childhood experiences of warmth and rejection as well as experiences of abuse at the hands of caregivers. Further, the characteristics of these women as parents are explored. Finally, a discussion of the limitations of this investigation and suggestions for future research is presented, followed by concluding remarks. Recapitulation of Main Findings It was found that three-quarters of the women in the sample had given birth to children, although 88.4% of those were no longer playing the role of primary caregiver upon incarceration. Nearly 61% of respondents claimed that Social Services had seized custody of at least one of their children, and 38.2% claimed that they had voluntarily relinquished custody of at least one of their children. Reasons for no longer having custody varied, although the majority of women (80%) reported that drug and alcohol abuse were key factors. It was found that women in the sample were characterized by insecure attachment styles. Characteristics of personality disorders were present in over 73% of the sample, and results showed high rates of abuse in childhood by primary caregivers, high levels of parental rejection, and low levels of parental warmth. In addition, participants reported lower than expected levels of parental stress and unstable relationships with the fathers of their children. Further, the 130 presence of an insecure attachment style was related to higher mean number of personality disorder characteristics, higher reported means of maternal and paternal rejection, and higher reported means of emotional and physical abuse by fathers, and impression management was found to be an issue with most of the measured utilized in this research. Representativeness of Sample The BCCW population had more federal inmates and fewer women held on remand than the present sample. The recruiting procedure may provide one possible explanation for this difference. Rather then making appointments with inmates, a plan that was quickly deemed to be an inefficient means of procuring participants for a variety of reasons, interviewers telephoned individual units from the interview rooms to ask for interested participants. A high rate of response was consistently received from the unit that primarily housed remand-hold women, as most had only recently arrived, and were not yet allowed to work. Consequently, the majority women on remand hold did not have money in their prison accounts for cigarettes or other "canteen" items. Also, many of these women did not appear to feel settled in their environment, and often did not have a routine or an established group of friends with whom to spend time, as did the more permanent residents in the institution. This may also account for the lower rate of participation by the Federal inmates. The preponderance of these women had regular jobs at the prison, and federally and provincially convicted inmates are permitted to have money deposited in their canteen accounts by outside persons. A little money goes a long way in prison, where items are relatively inexpensive (except for cigarettes, which are not tax exempt), and the federal inmates may have had less need for the $5.00 being offered for participation in this project. As well, women serving longer sentences have established schedules for both work and recreation 131 and may find comfort in their routines. They are likely to have more social support among the other inmates than women who are new to the prison, and the social activities provide vital opportunities to foster and maintain social support, as well as for maintaining and enhancing social status among the prison populace. Further, there were reports of discontent due to a recent research project that left many of the inmate participants upset. The longer-term inmates would have more knowledge of this than remand-hold inmates who had just arrived to the prison and had not heard about the adverse experiences of others, or who had not yet been accepted into social circles and hierarchies that existed in the prison.20 Veracity of reports The vast majority of this research relies upon self-report of the inmate. This population is by nature prone to dissembling. For example, various studies have found rates of psychopathy in female offenders ranging from 9%-31% (Salekin, Rogers, and Sewell, 1997; Loucks and Zamble, in press; Vitale, 2000; Strachan, 1995), and of Antisocial personality disorder ranging from 49% (Tien et al, 1993) to 61% (present study21), and one of the key symptoms of both psychopathy and Antisocial personality disorder is deceitful and manipulative behaviour (Hare, 1993; APA, 1994). From this, the inference is made that researchers should be aware of the likelihood that at least some participants in a female offender population will intentionally attempt to be deceptive. In some cases, the offenders may deliberately prevaricate in order to 2 0 There were even greater differences in legal status between the present study and Turnbull's (1996) study, x 2 = 24.509, df= 2,p = .0001. The percentage of federal participants were the same, but Turnbull had many more provincial than remand participants, reportedly due to the fact that she scheduled interviews, and the majority of the remand-hold population spend a week, on average, at BCCW, thus not leaving enough time for recruiting, scheduling, and conducting the interview. 2 1 Note that for the present study only "characteristics" of personality disorders are assessed. No diagnoses for personality disorders were made. 2 2 Psychopathy was not assessed in this sample due institutional policy. 132 spare themselves punishment or embarrassment, and in other cases, they may simply lack the insight to be able to accurately answer questions about their feelings and motives. Further, severe and consistent alcohol abuse may confuse the mind's ability to provide a factual account of past occurrences. However, some research has found that substance abusers can be quite accurate in their self-reports about their substance use and related behaviours, even when these events occurred several years prior (Sobell, L.C., Sobell, M.B., Riley, D.M., Schuller, R., Pava, D.S., Cancilla, A., Klajner, F., & Leo, G., 1988). This implies the possibility that these female offenders gave accurate answers to their substance use and related problems, although this does not speak to whether they were accurate in other areas of self-report. The women in this sample were particularly reluctant to talk about their children and their relationships with their children, and many became very defensive when doing so. Interestingly, several of these women easily admitted being addicted to drugs, to participating in prostitution and other illegal activities, but appeared very concerned that they be regarded as good mothers. The irony of this inconsistency did not seem to occur to them. More specifically, many women had lost custody of their children to a child welfare agency, citing drug addiction as the primary reason, yet when asked whether they felt they were good mothers the majority said that yes, indeed, they did think that they were good mothers and that they were there for their children when needed. Further, it was found that women who scored lower on the Parental Stress Scale, indicating that they experience lower levels of stress as parents, were more likely to score higher on the Impression Management scale. For these reasons, the answers given on the Parental Bonding Interview and Parental Stress Scale are considered suspect, and caution must be used when interpreting the findings that utilize this information. 133 Impression Management It appears that the women in the sample scored lower, on average, than previous research using a sample of religious women, a sample of college women in a public disclosure condition, and a sample of alcoholics, but had similar scores to another group of college women not in a public disclosure condition. In other words, their impression management scores seem to suggest that there was no attempt to present themselves in a socially desirable manner. This finding is interesting, as this population is generally considered to lack honesty, insight, and the sophistication to malinger on scales. However, these theories are untested, both in previous studies and the present investigation. Women in this sample do not appear to be managing their impressions on others, at least when answering the BIDR scale. Further, it is unlikely that they knew that malingering their replies on this scale would reduce their impression management scores, nor what the outcome of this would be. In many ways, this finding corresponds with interviewers' experiences with participants. As mentioned, they were often candid about their drug addictions, their crimes, and their participation in prostitution. On the other hand, they appeared to be defensive and managing impressions when discussing their current relationships, particularly with their romantic partners and with their children. Higher mean impression management scores were positively correlated with mean dimensional scores on the dismissing attachment category, as assessed by the RQ, suggesting that women who scored high on dismissing attachment styles were managing their impressions. However, IM scores were negatively correlated with mean dimensional scores on the fearful attachment category, as assessed by the RQ, suggesting that women who scored low on the fearful category were managing their impressions. It seems reasonable to suggest that portraying oneself as independent and not afraid of romantic relationships may be an adaptive characteristic for these 134 women. Women who did endorse a fearful attachment orientation did not appear to be managing their impressions. Interestingly, when the RQ scores were collapsed into secure and insecure categories, there were no significant differences found on mean impression management scores between the two groups, although women endorsing a secure attachment orientation still tended to have higher mean IM scores, indicating that they may have been attempting to portray themselves and autonomous and comfortable depending on others. These two findings contradict each other and speak to the complexity of the measurement issues and the interpretational difficulties involved. To further confuse matters, a highly significant negative correlation was found between each of the attachment styles, as assessed by the ECRI and mean IM scores. That is, low scores on each style were related to high IM scores, and conversely, high scores on each style were related to low IM scores. When attachment styles were dichotomized into secure and not secure, the secure group showed higher mean IM scores than found with the insecure group. Overall, these findings indicate that the women in this sample were managing impressions, at least on their responses to the ECRI, and therefore results utilizing the ECRI should be viewed with caution. The ECRI appears to be more susceptible to impression management, which may explain the measurement error differences between the two scales. This finding suggests that the RQ is a better measure of attachment, at least with a female offender population. Parenting and impression management is another area of concern. It appears that this sample's reports of their parent-child relationships may be less reliable than their reports regarding the other areas of their lives. This seems more likely given that a significant negative relationship was found between PSS scores and Impression Management scores. As noted, women who reported lower levels of parental stress tended to have higher scores on the Impression Management scale. Further, women who reported fewer characteristics consistent 135 with personality pathology had higher IM scores. It should be noted that, in the present sample, the factor analysis for the BIDR did not produce the two main factors that the scale purports to represent, namely, Impression Management and Self-Deception. Only one factor accounting for 15% of the variance was found. Therefore, the findings that there were significant relationships between the IM scale and the PSS, the RQ, the ECRI, the SCID-II should be viewed with caution. It is important to recognize that this impression management surrounding questions about female offenders' relationships with their children may be due to the fact that many of these women have had contact with Social Services (now called the Ministry for Children and Families23) or others regarding the care of their children. Further, by virtue of being in prison, most of the women who still had custody of their children needed to make childcare arrangements for their children. It would be difficult for an inmate not to interpret this, however fleetingly, as a failure on her part to be a consistent caregiver to her children. Unfortunately, in this sample most mothers had already lost custody of their children prior to incarceration, so presumably they have had to in some way acknowledge that they were unable to perform the role of mother, at least on a full-time, consistent basis. This may account for the observation that many of the inmate mothers who were interviewed gave contradictory and defensive answers when questioned about their children. It is interesting that they understood and accepted what was socially desirable in their roles as mothers, and wanted to maintain an appropriate image in others eyes in this regard, however, they did not seem to place as much value on the other aspects of their image that may also be perceived as not 2 3 This is the government agency that deals with child protection, adoption, foster care, and youth corrections, among other things. Social Services was the name used by the female offenders in this sample, so it will be called such for the remainder of the discussion. 136 socially desirable. Is the role of mother so much more central to a woman's self-image- inmate or non-inmate? Or is this the one area of an inmate's life that she feels she should be able to manage adequately? Alternatively, this may simply be the way the inmate had learned to deal with the "system"- not admitting fault to anyone, knowing that to do so could change their life and the lives of their children. Several women told stories of approaching Social Services for temporary help because they knew that they were not capable of adequately caring for their children, and wanted a little assistance and some time to get back on their feet- to get off drugs, to get out of an abusive relationship, to make some money- and this request for assistance would result in a permanent loss of their children. They did not trust Social Services because of this, and many felt that it would have been better to just try and manage on their own. They believed that they were going to get the kind of help that they asked for, for the length of time that they asked for it, and discovered that once asking for this help, they were no longer in control of the outcome unless they sufficiently convinced Social Services that they were again able to care for their children. For many this was not possible because in most cases serious drug addiction was the origin of the problem, and most of these women were unable to remain off drugs for a sufficient period of time to regain custody of their children. Of course, parenthood, personality pathology, and attachment issues were not the only areas in which there were impression management issues. For example, one of the case studies included in this discussion (Offender 1) involves a 19-year-old drug dealer who reported to the interviewer that she was angry with her father for leaving her mother to take care of the family, a task her mother apparently did not do to either the offender's or Social Services' satisfaction. This teenager stated that her father was in prison for murder but would be released in a few 137 years. Interestingly, file information for this offender indicated that her father had died from a heroin overdose. Further, when asked whether she had ever attempted suicide, this offender claimed she had not, whereas file information indicated that she had at least one previous suicide attempt. Clearly, impression management was an issue with at least some participants from this population, and this should be taken into consideration when interpreting findings. Regardless, it can be assumed the some of the relationships investigated in this research may have been stronger if impression management had not been an issue and/or that impression management added too much error to find significant relationships that might actually exist. New Admissions It was thought that there might be reporting biases among women who were newly admitted to the institution at the time of assessment. Specifically, it was thought that women might report higher levels of pathology or lower levels of attachment security due to the stress of their new surroundings and their recent dealings with the criminal justice system. It was also thought that many may have been withdrawing from drugs and/or alcohol and that this might have increased stress and affected reporting patterns. However, analyses showed that there were no differences between women who had been at the institution for less than two weeks and women who had been at the institution for more than two weeks on any measure. This was an arbitrary cut-off point, but two weeks was thought to be a reasonable amount of time for them to settle into their new surroundings, although the majority were repeat offenders and had previous incarceration experience. 138 The Women The women in this sample were, on average, 30 years of age and had a variety of family backgrounds. Nearly 44% of participants were raised by both parents, and the remaining 56% were raised by some other combination of caregivers. A large minority of the women claimed to be married or in common-law relationships (43.5%). This number is likely inflated, as several women did not seem to realize that the legal definition of common law in British Columbia includes the requirement that the couple be cohabitating for at least two years duration. Many of these women had been in relationships for a substantially shorter period of time. Most of the women in the sample had many previous arrests. One-quarter of the sample had 30 or more previous arrests, and only 6.3% were at BCCW for a first offence. Nearly 70% of the sample had violent offences on record. Also, 70% of these inmates were poly-substance abusers, that is, reported abusing two or more substances, and 85.5% reported abusing at least one substance. Not surprisingly, the results of the EMBU showed that, regarding women's relationships with their primary caregivers in childhood, mean scores for paternal and maternal rejection were much higher and mean scores for paternal and maternal warmth were somewhat lower in the present sample than found in previous research using a non-forensic sample. Case Study Offender 1 Interviewer: So you've never given birth. Why? Offender 1: Because when I was little my brother pushed me down the stairs and I fell on the baseball bat and I broke my ovaries or something or whatever it is I don't know... I screwed up my insides, so.... Interviewer: Any other reasons why you haven't had children? Offender 1: Because I always use condoms (laughing). Interviewer: Do you like children? 139 Offender 1: Depends on the kid. Interviewer: Would you like to have kids one day? Offender 1: Maybe when I'm older. Ummm, if they turn out anything like me (derogatory). Interviewer: Do you think you would make a good parent? Offender 1:1 guess... because I wouldn't beat them, and I'd find other forms of punishing them besides beating them like no TV, in the corner, no outside. Interviewer: What else? Offender 1: I'd listen to my kids... sometimes kids have something important to say. Interviewer: Anything else? Offender 1: I wouldn't criticize them or something. This is the young woman whose mother introduced her to crack and whose father overdosed on heroin, thus is it not surprising that her responses indicate the likelihood that her childhood was characterized by low levels of parental warmth and high levels of parental rejection. This Native woman was classified with a non-secure attachment style on both attachment measures. Rates of reported abuse by primary caregivers was expected to be high in this sample, and this was borne out. There were particularly high rates of emotional and physical abuse. As mentioned previously, it is not possible to compare inmates' reports of abuse by caregivers to national estimates of child abuse in the general Canadian population (Statistics Canada, 1998). However, nearly 1% of the population of children in Canada (N=61,000) under 15 years of age during 1998 were involved in substantiated maltreatment investigations (Health Canada, 2001). This is most certainly an underestimation of actual occurrence of child maltreatment, as an unsubstantiated claim of maltreatment is not verification that abuse did not happen, only that it could not be proven with the available evidence. Further, there are undoubtedly many more incidences of child maltreatment that are never reported, and there is no accurate way of 140 estimating these rates. Yet, as reported, over 31% of women in this sample reported physical abuse by their female primary caregiver, and over 31% of women in this sample reported physical abuse by their male primary caregiver. Greenfield & Snell (1999) found that many female offenders were neglected or abused in childhood, although it is not clear how these researchers assessed abuse. However, it is assumed that rates of maltreatment at the hands of primary caregivers is a great deal higher in this sample than in the general population, presumably playing a factor in their overrepresentation of insecure attachment styles, personality pathology, and parenting difficulties (Main & Goldwyn, 1984; Alexander, 1992; 1993). In addition, more than 12% of the sample reported that their male primary caregiver sexually abused them24, and given that 75% of the present sample report being mothers, it is meaningful to note that evidence suggests that daughters of sexually abused women are more likely to become abused themselves (Goodwin , McCarthy, & DiVasto, 1981). In other words, the female offspring of these women may be at greater risk for sexual abuse than other female children. At the request of BCCW, it was not possible to probe further into the sexual victimization experiences of this sample, although presumably further probing would uncover a great deal more abuse, given that research had found that much sexual abuse is perpetrated by individuals other than primary caregivers (Finkelhor, 1993). Further, female offenders are at greater risk of experiencing sexual abuse as adults, particularly those who engage in prostitution as a means of earning income. Research has shown that more in-depth interviews result in higher reported rates of abuse, and these reports are assumed to be more accurate (Finkelhor, 1993). As noted earlier, a previous research project conducted at the institution was mishandled, and some participants reported feeling re-victimized as a result. Therefore, we had to strictly limit our questions to lNo women reported that their female primary caregiver sexually abused them. 141 those mentioned above, and it was deemed most appropriate to approach this topic in questionnaire form (Bartholomew, personal communication, Jan. 2001). The following is an excerpt from an interview with an inmate who reported experiencing sexual abuse at the hands of a trusted other. Case Study Offender 2 Interviewer: ... you have to keep an eye out to stop people from hurting you... Offender 2: Well, yeah. And it's because someone who I thought really loved me when I was a little girl hurt me, right... and when my mother wasn't looking after me and stuff because she neglected me a lot and this man, I thought he loved me, so I did what he wanted just just because I wanted the attention... and I didn't know it was wrong, right? And that's why I'm like-1 don't trust very many people. Interviewer: Um, you've said that it's best not to let people know much about you because they'll use it against you.... When has this happened- can you tell me about that? Offender 2: Ugh. Ha! All the time! Um, I don't know, what I meant by that was that you shouldn't- people that are you're friends you know- until you've known someone for like 10 or 15 years, you don't tell anyone your secrets... because if you make them mad or they get angry at you or something, then they could turn around and use that against you, right?... and then it gets to the extreme, you know?... and then you really don't trust anybody. I just don't give them the ammunition, right? Interviewer: All right, um, you've said that they are many people that you can't forgive because they did or said something to you a long time ago. Can you tell me about that? Offender 2: Well, my ex for one thing. I lost my kids and he abused me for 4 years and I was in the hospital and I almost had a nervous breakdown. I was only 17 years old, you know. Ahh, my mother, al, I don't think I'll ever forgive her the hell she put me through when I was growin' up and the things that happened to me that she was just in oblivion to or too drunk or high to notice, and you know there's a lot of things.... My dad for leavin' and just forgettin' all about me and starting a whole new life without me. My mom for giving my brother up, my uncle, my grandfather... There's-1 could go on and on, you know. There's a lot of people. Interviewer: You said that before you were 15, you ran away from home and stayed away overnight. Was that more than once? Offender 2:1 used to run away all the time... yeah, all the time. Interviewer: Who were you with? Offender 2: I-I started workin' the street when I was eleven, so I was really young and I was being abused and stuff at home.... I got put in about 4 or 5 foster homes and they couldn't- man, I was unmanageable. I was-1 was such a messed up kid, right? So then they just ended up putting me back with my mother. They didn't know what else to do with me. Interviewer: Okay, um, before you were 13 you would often stay out very late, long after the time you were supposed to be home? 142 Offender 2: Well, my mother never gave me a curfew... I just- she just let me run the roads and do what I wanted. Interviewer: Okay, so did you ever- did it ever happen that you went out despite what your mom had said- your mom, maybe one time or something? Offender 2: No, she never ever- she never said anything. She was always drunk all the time. She never said "You be back at this time or that time." She probably didn't even know I was gone. This offender experienced sexual abuse at the hands of someone that she cared about and trusted. She felt abandoned by her father, and neglected, and was likely emotionally, if not physically, abused by her mother. Not surprisingly, she did not report a secure attachment style, however, there are clearly many other factors that could contribute to interpersonal difficulties that this woman may be experiencing. One such factor could be the presence of personality disorder characteristics. The SCID-II results indicate high levels of personality pathology within the present sample relative to that found in another sample of female offenders at the same institution (Tien et al, 1993) using the same structured measure, although the Tien et al (1993) research utilized a computerized program in order to arrive at specific diagnoses. They may account, at least in part, for the differences in the overall rates of personality pathology between the two samples, and for the elevated rates for antisocial and narcissistic personality disorders in the present sample. Livesley et al (1990) note that there is substantial overlap between the personality disorder diagnoses, calling into the question the validity of distinguishing between personality disorders. However, the comparisons between the present sample and community and clinical samples that were significant showed differences in the expected direction. The majority of women in the sample were also found to be drug abusers and violent, repeat offenders. Many had experienced low levels of parental warmth, high levels of parental rejection, and abuse by their primary caregivers. Many were also characterized by personality pathology, adding an additional component to their difficulties in relating with others. Therefore, 143 it is not surprising that the present as well as previous research (Turnbull, 1996) has found that female offenders are more likely than non-offenders to have insecure attachment orientations. Attachment Styles As attachment theory formed the basis for this research, the following excerpts were taken from interviews with four women in the sample as a demonstration of the four attachment styles. Each interview represents a woman with a different attachment style, and preceding each transcript is the Bartholomew (1990) description her theoretical model of that attachment orientation: Secure Secure: Positive model of self/positive model of other Low level of dependence/low level of avoidance These individuals are comfortable with intimacy and autonomy. (This woman is married and has known her husband for 14 years.) Interviewer: When did you start dating? Secure Offender: Actually dating... it's quite a different relationship... a month after I met him he got posted over in (a foreign country) for two years. So a lot of our courtship... was by letter and phone calls... But I would say that we didn't start getting serious until a year after I met him. Interviewer: How quickly did you become serious? Secure Offender: It was fairly quick because a lot of the courtship that people would do face-to-face we had done through letters.... There's times when we have lived in certain places where he's had to go away [for work]. Interviewer: How did it feel being away from him? Secure Offender: On the one hand I was used to it, but on the other hand I really missed him, and there were times that I wished that he didn't have to go away. But I was accepting because I knew what his job was. Interviewer: What do you like about the relationship? Secure Offender: I like the fact the we compliment each other. Areas that he's not strong in I am and vice versa... we suit each other really well. Interviewer: What do you like most about him? 144 Secure Offender: That he- even though we are not together, that he still thinks about me and... he includes me in major decisions to do with the family and tries to keep things as normal as possible for us right now. Fearful Fearful: Negative model of self/negative model of other High level of dependence/high level of avoidance These individuals display fear of attachment and are socially avoidant. Interviewer: If you feel unhappy or upset about something, what are you likely to do? Fearful Offender: Attempt suicide... I have had a very interesting time out... .1 got out in February, [out for] two months... I tried to kill myself three times. First time I slashed my arms, and second time I overdosed on pills, and the third time was actually just last week. I tried to strangle myself.... I've never really had a serious relationship. Because I'm afraid.... My self confidence isn't that great... just waking up in the morning and looking crappy, or you know, him touching my flab... I've always been afraid. Plus I was sexually abused as a child so that has an effect on it. By my [biological] father every Christmas Eve because that is the only time that I would see him. Interviewer: Have you been looking for a relationship? Fearful Offender: When I was younger I was too messed up as a teenager to really- it was more one-night-stand dating... but now that I'm older I really crave a relationship, but I'm just so afraid of it. Preoccupied Preoccupied: Negative model of self/positive model of other High level of dependence, low level of avoidance These individual are overly dependent. At the beginning of this interview, this participant stated that she does not ever cry anymore. However, when asked about her previous romantic relationships she begins to cry when talking about her ex-boyfriend. Much of the interview was incoherent, so is not transcribed here. However, some of the contents are reported. This offender had a relationship with a man for eight years which was serious to her but not to him. He broke up with her because of another woman and she was very upset talking about this. She claims that she did not date after this because she was a 'working girl' and wasn't looking for anyone, she was just looking for him to come back. She is now seeing a man that she says she has known 'all her life' because he is her dream man and she has been thinking of him (her dream man) since she was a little girl (figuratively, not literally). He has also recently had his heart broken. It appears from her description of their relationship that they are consoling one another. However, there is conflict in this relationship because he does not think that she should be a prostitute anymore, but she thinks that he should just let her be herself, particularly because this is how they met. She would not quantify anything with regards to this relationship (e.g., duration), nor will she compare it to past relationships. She is very confident that this relationship will last in the long term and believes that only God himself could break them up. Dismissing Dismissing: Positive model of self/negative model of other Low level of dependence/high level of avoidance These individuals deny attachment and are counter-dependent. Interviewer: In what circumstances are you less emotional than other people? 145 Dismissing Offender: When it comes to being really sad and stuff I just turn it off- basically choke it down and build it up [comes out as anger]. Interviewer: How do you express your emotions if you are sad? Dismissing Offender: I just kind of choke it down. Some situations are harder like with my daughter. Social Services had her- they had her close to a month before I came in here... because my mom called and said I was dealing drugs and I had guns. She knew that I was selling dope- my daughter wasn't at my home she was at the sitter's and when you're packin' dope you gotta carry something. I'm not gonna take the chance of getting robbed... Interviewer: So your child's in Foster care? Dismissing Offender: [She] first went to my mom's but they had to take her out of my mom's because... she houses him [a known sex offender]. She gets paid $3500 a month to house this guy. And she had told me for my daughter's safety and then she told Social Services. And I wasn't supposed to say anything because... so Social Services put it on me but I'm like I'm not supposed to say anything because you know if I would have said anything she would have lost her job [for telling her daughter that he was a sex offender- it was confidential information] and then there's something else to blame on me. So instead of giving up the client she gave up my kid. And so my daughter went to one of [my mother's] friends who was an ex coke head so that right there just built up more animosity in me and I was just like okay, screw all of you. You can all go to hell.... My mom beat me when I was growing up... my mom's an ex junkie, alcoholic [with an extensive criminal history]. She's clean and sober for almost ten years now, but it doesn't change the past.... Her fourth [boyfriend] beat on my brother and molested me. My mom was always in abusive relationships. Interviewer: How often do you cry? Dismissing Offender: When I'm pissed off or frustrated. Not too often. I don't want to cry in front of these people [other inmates]. Interviewer: What about when you're not in here? Dismissing Offender: Not that often. Enough to be human (laughing). Interviewer: Do your friends tend to be male or female? Dismissing Offender: Male, because females gossip and bitch too much, and females tend to strike you in the back too quick. Interviewer: Do you find that there are differences between your male and female friendships? Dismissing Offender: [With females it's] ... grocery shopping, shopping, doing the mommy thing, activities... where my buds [male friends] we can go and sit down and watch sports or go do fun stuff. They're my buddies and my female friends I find are more protective because like my male buds look at me like one of the guys, like I can take care of myself, but the females are just like mother hens. Interviewer: When you meet people do you think they will like you? Dismissing Offender: I don't care. Interviewer: But do you think they would? Dismissing Offender: No, I think their first impression of me is a bitch. Interviewer: How confident are you about making new friends? Dismissing Offender: Pretty damn confident. I get what I want. 146 Interviewer: Are you shy? Dismissing Offender: I'm not shy. I'm standoffish. Interviewer: The six years you were single, why were you single? Dismissing Offender: I hated men. Interviewer: Were you dating? Dismissing Offender: I guess you could call it dating, yeah. Find somebody I like, I'd see 'em for a couple of weeks, get what I want- bye. Interviewer: Were you looking for a relationship? Dismissing Offender: Nope. Didn't want one. Too heavy. The preceding interviews are examples that illuminate how the various attachment styles manifested themselves in this sample. It was expected that there would be significant differences between friendship and romantic relationships, as assessed by the Relationship Questionnaire, as it was thought that inmates would be more likely to be able to form secure attachments with their platonic (particularly female) friends. However, these expectations were not borne out for either the categorical or dimensional ratings. As was the case with the Dismissing Offender, many of these women found other women, female friends and acquaintances, even less trustworthy than their romantic partners. Many commented that they did not have female friends, explaining that they were usually too "catty" and could not be trusted with someone else's secrets, boyfriends, money, or drugs, among other things. Because no differences were found between the friendship and romantic relationship scores, only the results from the romantic portion of Relationship Questionnaire are included for purposes of the remainder of this discussion. However, it is important to note that the fact that the majority of these women have an insecure attachment orientation towards both their romantic and peer relationship speaks to the general difficulties 147 that they are experiencing across their interpersonal relationships, as well as the problems that they face when attempting to develop a positive social support network post-release, an important factor in preventing recidivism. When the RQ and ECRI four-category classifications were compared, there were significant differences, particularly for the fearful and preoccupied categories. However, when the RQ and ECRI were divided into secure and non-secure categories, they were nearly identical. This demonstrates a high degree of convergent validity between the two attachment measures used with this sample. This is in contrast with previous findings that showed the ECRI as more likely to classify individuals as insecure than the Relationship Questionnaire, apparently because the ECRI is more precise in discriminating among people with different levels of insecurity (Brennan et al, 1998). The following is an example of a woman who has a preoccupied attachment style, but reported a secure attachment style on the RQ categorical measure. Her score on the ECRI reflected her preoccupied attachment orientation. Case Study Offender 4: This inmate refused to answer any questions about her child. She started to cry immediately when asked, so the interviewer turned off the tape recorder while Offender 4 composed herself. They spent a few minutes making small talk, and then Offender 4 informed the interviewer that she would be willing to talk about other things on tape. Interviewer: Okay, so do you want to tell me a little bit about your childhood? Offender 4: Ah, okay. First of all, I didn't have a good childhood. I had a really bad childhood. I was abused in every way possible- emotionally, physically. Um... as far back as I can remember being, um three years old. I was sexually molested up until I was 13 years old when I left home. Started working the street when I was 12 uh behind everybody's back, but um they found out and introduced me to my first sugar daddy. Um I was with him until I was 16, ummm... [I am the] Second youngest, I started using-1 started smoking marijuana when I was 12, drinking when I was 10, smoking when I was 9,1 started-1 met my son's father when I was 14, un I haven't seen him [her son] since he was 5 'A weeks old and the father started using needles behind my back. I lost him to welfare. Uh, that was it. I went downhill. I became, I became a user real badly. Um, I started fixing when I was 16 and I'm 29 now and I did crank 2 Vi weeks ago I did a fix before I came here.... I been coming in and out of jail since I was 19-18. It sucks. I'm tired of this life. I've been here at least 10 times [for possession for purposes of trafficking]. 148 [Interviewer asks who she was referring to at the beginning when she reported abuse]. Offender 4: Everybody. All my uncles, all my guy cousins, um, a couple of brothers, there were quite a few. I forgave 'em... there was a lot of incest in my family... and 1 hate men. I don't trust men, and I have a hard time trusting my step-father [who she previously reported being a person who is there for her- the one person in the family that she can count on]. In a follow-up interview: Interviewer: Have you had any serious previous relationships? Offender 4: Yeah, I've had maybe four. Interviewer: I'm going to ask you some questions about those okay? Go back to the first of those four... Offender 4: Okay, I was 16- she was20.... Not that serious.... We met through family... it took me a year to get her to be my girlfriend. Interviewer: Why did you guys break up? Offender 4: Um, ... and then because she was working the streets... Interviewer: So you were alone a week before the next relationship... okay so you were 18? Offender 4: Yes, until I was 22... she was 29... she got pregnant on me and I went through with it, with the baby, picked up the baby, held the baby, and the baby was mine. I'm the baby's godmother. That was the most serious relationship I've ever been in because I had responsibilities. She has 2 other kids as well. I had a built-in family there... Interviewer: Why did you guys break up? Offender 4: She started working the streets behind my back... she got pregnant again by a trick... I left her... it was hard. Interviewer: How long were you along before your next relationship? Offender 4: In order for me to get over her I had to be with somebody else right away and if I can remember correctly... I was with a women who was 20 years older than me... in a about a month.... For two years... Interviewer: Why did you guys break up? Offender 4: Because she married my mark... back then 1 wasn't working the streets, I had marks... was selling drugs... [The offender then reports that her next relationship, which started 6 years later, was with a women who was at BCCW at the time, one of the other participants in this study]. Interviewer: Do you see any patterns across your relationships? Offender 4: Yes, I need to date different kinds of women... they use me for my dope. Interviewer: [Regarding most recent relationship with the BCCW inmate] Any major separations since you've been together? Offender 4: Yes, last month. [Offender then reports that she was out of jail and purposely got caught selling drugs the day she heard that her girlfriend was seeing someone else at BCCW]. Interviewer: So how long do you think you'll be here for? 149 Offender 4: Six months. [She and this woman were no longer in a relationship at this point, and this offender reports that this woman has been physically abusive towards her particularly since her return to prison. They live in the same unit and the woman is indeed in a relationship with someone new.] This woman clearly did not have a secure attachment style, even though she chose the secure description of the categorical portion of the Relationship Questionnaire for romantic relationships. It is possible that the description indicated how she would like to feel and behave in a romantic relationship, and it is also possible that she did not understand the instructions. Another explanation may be that she strongly identified with one or more portions of the secure paragraph description and therefore adopted the entire description as most representative of her romantic attachment style. The ECRI was more successful in tapping into the attachment orientation, at least for this participant. The interview with Offender # 4 illustrates the sorts difficult life circumstances faced by many female offenders, and illuminates how experiences in childhood shape the lives over time as well as the kinds of difficulties the can encounter in their relationships with others: their children, their families, and their romantic partners. For example, it was also found that women with insecure attachment styles had higher means of maternal and paternal rejection than women with secure attachment styles. Attachment theory espouses that sensitive and responsive caregiving is the foundation of the development of a secure attachment style, and that the lack thereof (for example, rejection by parents) can lead the affected child to develop one of the insecure attachment styles (Ainsworth & Bowlby, 1991). As noted, scores for warmth and rejection were compared to one another and were found to be highly correlated, suggesting that warmth and rejection may simply be opposite ends of the same dimension, and this may account for the fact that there were no differences between attachment styles in reported levels of maternal or paternal warmth. 150 Regarding abuse by primary caregivers, it was found that women with insecure attachment styles had higher reported mean scores of emotional (as assessed by the ECRI) and physical (as assessed by the RQ & ECRI) abuse by fathers than women with secure attachment styles, when controlling for the effects of impression management. There were no relationships between attachment security and sexual abuse by fathers, nor between attachment security and any form of abuse by mothers. It is possible that for many of these women sexual abuse by fathers was only one source of sexual abuse they had encountered, and it is possible that experiences of sexual abuse by men other than the father, or a lack of such experiences, may have been more relevant to their current attachment functioning. Regarding the lack of relationship between attachment security and abuse by mothers, it may have been the case that these women had resolved some of their childhood issues with their mothers in adulthood. It may be more difficult for some women to resolve poor childhood relationships with fathers, particularly if the parents are divorced, as the mother most often retains custody of children upon marital dissolution, and women may not have much contact with their fathers. It was further found that women with insecure attachment styles had a higher mean number of personality disorder characteristics than women who had a secure attachment style. This is not surprising, as both insecure attachment and personality pathology are characterized by difficulties in interpersonal relationships. Clearly, these women had many different issues than what might be found in a non-offender sample of women, and therefore, attachment styles, and how they should be addressed in the context of research, is likely to be different between female offender and non-offender populations. Factors such as drug abuse, victimization, and criminal lifestyles have a large impact on the interpersonal relationships of female offenders, whereas, these factors are less likely to be present in the general population, and if they are 151 present, may be to a lesser degree than for female inmates. Perhaps these differences between female offenders and the samples usually utilized for attachment research render it more difficult to attribute interpersonal dysfunction to attachment insecurity alone, although this does not render futile attempts to do so. As noted in the introduction, attachment theory was initially developed to understand and explain interpersonal difficulties in at-risk families (i.e., Bowlby's work at the school for maladjusted children who had experienced maternal rejection/separation). It would almost certainly be more beneficial to address drug abuse, victimization, and criminal lifestyles of the offender population in tandem with attachment orientation in addressing their interpersonal relationships problems than it would be to focus on attachment orientation alone. As can be seen from the preceding commentary, attachment orientation was related to early experiences with parents, as well as personality pathology. However, for numerous female offenders, problems such as drug addiction, abusive relationships, lack of education, and poverty were largely responsible for the interpersonal difficulties that these women faced, and these issues also affected their experiences as parents as well as their relationships with the fathers of their children. Inmates as Parents More than three-quarters of women in this sample had borne children; the remaining quarter had not. The only apparent difference between these two groups of women was that the women without children were, not surprisingly, younger than the women who had children. The small sample size of the non-mothers may have prevented finding further differences. Of the women who did have children, the vast majority of them did not have custody of their children prior to their present incarceration. Many of them reported that Social Services had apprehended 152 at least one of their children (n = 45), and several reported that they had voluntarily relinquished custody of at least one of their children (n = 28). When asked why they were no longer the primary caregivers for their children, the most common reason given (70%, n=59) was drug and alcohol use. Other reasons included abusive relationships (with a romantic partner), child abuse and/or neglect, losing their children in a divorce agreement, prior incarcerations or crimes, mental illness (on the part of the mother), and being too young when they gave birth. No participant reported that they lost custody of their children due to financial difficulties, although it is likely that this was a factor in several cases. Regardless, it appears that among this sample drug and alcohol abuse was the superseding reason for loss of child custody. Further, there were no systematic relationships between attachment security and whether the mother relinquished custody or had their children apprehended. Case Study Offender 3: She has two children. The father of the first child was granted custody when the child was 5, because the offender was in an abusive relationship with another man. The youngest child was taken away by Social Services and put in Foster Care because of drug use. Interviewer: Do you think you were a good parent? Offender 3:1 was at first. Interviewer: What changed? Offender 3: Drugs. Well, somebody came over with drugs and then I'd throw the kids in their room and start doing drugs, then I wasn't with them as much as I was in the beginning. Then it started affecting my kids. Didn't neglect them, though. Like... food and stuff. I'd never make them go without, but I'm glad somebody stepped in before it came to that point. For this offender, parental responsibility was overridden by drug addiction. One of her children was taken away by the father, and the other was taken away by the government. This is not an atypical story among this population. What was atypical was that this woman reported being relieved that someone made the decision to relieve her of her parental responsibilities because she was aware that things were spiralling downwards for her and she did not want her children to suffer for it. Interestingly, she was one of few women interviewed who spontaneously admitted being glad that her children were taken away from her. When inmate mothers were asked where the father of their children was, it was discovered that a small proportion of them were still coupled with the at least one of their children's fathers (8.6%), 22.5% reported that at least one of the children's fathers were in jail; and 11.1% reported that at least one of the children's fathers are dead. Of the 10 women who were still the primary caregivers for their children, seven of these had at least two children, and there was just one father for the children of five of these seven women. In other words, regardless of number of children (one or more), eight of the 10 women who were still the primary caregivers for their children had the same father for all their children. However, only one of the 10 women who remained the primary caregivers for their children prior to incarceration stayed together with the father. That is, of these 10 women, only one is still with the father of her children, regardless of the fact that there may have only been one father for all of her children. Paradoxically, as previously reported, 7 women were still together with the father of their children. Only one of these is still the primary caregiver. Therefore, 6 women are still together with the father of their child(ren) but are no longer the primary caregivers for those children. In sum, even women in the sample who still had custody of their children upon incarceration were unlikely to be in a relationship with the father of their children, regardless of whether he was the only father, and conversely, of the relatively few women who were still together with the father of their children upon incarceration, only one remains primary caregiver to her children- that is, the others were able to maintain their romantic relationships with their children's fathers, but were not capable of playing the role of primary caregiver to the couples' 154 children. If these romantic relationships were further investigated, it would not be surprising to find high levels of drug and alcohol abuse by both partners, as the presence of such addictions appeared, at least in this sample, to take precedence over child-rearing. Also illuminating are the above-mentioned percentages of women who claim that at least one of the children's fathers are in prison or dead speaks to the risky lifestyles and criminal associates with whom they were involved, again not conducive to a secure and stable base from which to form child-mother attachments. Recall Offender 2, the woman who claimed that someone she trusted abused her as a young teenager. This woman has given birth to 7 children. All of her children were given up for adoption between the ages of two weeks to two years. There are four different fathers for these children, and none of them have helped with any of the children, and none have contact with any of the children. One of her children's fathers went to prison for abusing her, the second25 is also in prison, she has occasional contact with the third father, and the fourth father was a 'date'- a regular customer ("John")- who she reports did not care that she was pregnant with his child. Although this is an extreme example, many female offenders had unstable relationships with the fathers of their children. Whether their pregnancies were planned or accidental, these women on the whole did not seem to consider their ability to provide stable home environments for their children in their decisions to become parents, and more often than not, custody was either relinquished or taken away from them. When women in the sample were asked who was currently playing the role of primary caregiver for their children, the vast majority (more than 80%) reported that either grandparents or foster parents has custody of their children. 155 It is likely that women who had voluntarily given their children away had given them to the person most likely to accept such a responsibility, and many may have relied on their parents or their children's fathers to perform this task. However, it is also the case that when Social Services takes a child from a mother, they try to place the child with the nearest dependable relative, rather than placing the child into a foster home of strangers. Therefore, the striking finding here is that 37% of women reported that foster parents were the primary caregivers for at least one of their children- that is, neither the mother herself nor Social Services could find a suitable alternative within the children's family for anything other than foster placement. This speaks to the breakdown of the mothers' relationships with her parents and other relatives, as well as with the father of the children. It also suggests that some of these women may come from families in which parents were not dependable throughout their childhood, nor were their parents a present source of support to their daughters or their grandchildren. Women in the sample were asked whom their children were currently closest to: Just over half of the mothers said that at least one of their children was closest to them. However, it should be noted that during the interview when this question was asked, it appeared that some women claimed that the children were closest to her when, in fact, the balance of the interview did not indicate that this could possibly be the case- in other words, it can be assumed that this proportion is actually an over-representation. Therefore, it is assumed that fewer children are closest to their mothers than was reported here. When asked whether they were affectionate with their children, only one woman claimed that she was not affectionate. When women described what they meant by affectionate, they usually reported that they hugged or kissed their children. However, this was not quantified in These numbers are randomly assigned. This does not necessarily correspond with the order in which the children 156 any way so there is no way of determining the frequency of or circumstances surrounding these reported affectionate behaviours. Inmate mothers were also asked whether they physically discipline their children. The majority (83.3%) said that they never physically disciplined their children when the children were in their care. This is also assumed to be an under-representation in reporting, and possible reasons for this will be addressed shortly. However, it is implausible that many of these women, with their interpersonal difficulties, challenging backgrounds, and drug abusing lifestyles, would have that degree of control during frustrating times of children not obeying, fighting amongst themselves, breaking things, being stubborn, among the many other challenges to patience that children pose. Only half of the women in this sample reported seeing their children daily or weekly, and a quarter of them reported that they never saw their children when not in prison. Further, most inmate mothers indicated that they were satisfied with the caretaking arrangement of their children while they are in prison, and almost three-quarters said they planned to have their children living with them eventually or immediately after release from prison. As previously noted, only 10 (12.5%) of mothers in this sample had their children living with them prior to incarceration. In addition, the vast majority of inmate mothers related that they enjoyed parenthood, and nearly three-quarters of those who reported said that they considered themselves to be good parents, and that they felt that they were there for their children. These findings are yet another indication of the considerable discrepancies between views of parenthood and one's parenting ability and styles in contrast to the actual state of affairs of the samples' relationships with their children. The preponderance of mothers maintained that they did not physically punish their were born. 157 children, however, recent research has found that nearly half of parents in the United States use physical punishment for child discipline (Wissow, 2001). Other recent research found even higher rates- the percentage of parents using physical punishment reached a peak of 94% when the children were ages 3 and 4 yrs (Straus & Stewart, 1999). It is untenable that 83.3% of this female offender sample did not use any physical discipline with their children, although this number might have been inflated by those women who gave their children up as infants and had little opportunity to encounter the frustration that toddlerhood can bring to a parent. However, many of the mothers in sample did raise their children for a period of time past infancy. Further probing into the inmate mothers' experiences of parenting found that the majority of inmate mothers claim that they enjoyed parenthood, that they planned to become the primary caregiver again at some point after release from prison, that they were good parents, and that they were there for their children. However, as noted, very few of inmate mothers in this sample were primary caregivers for their children prior to incarceration, and most of them had either voluntarily relinquished custody, or had at least one child seized by Social Services. This is an indication that parenthood was not as easy or enjoyable as was reported, and the majority of mothers in the sample experienced difficulties with the responsibilities of child-rearing. Interestingly, the results of the Parental Stress Scale indicate that the inmate mothers in this sample report experiencing lower levels of parental stress than clinical and control samples (Berry & Jones, 1995). There are a variety of potential explanations for this: first, these women may indeed be experiencing less parental stress due to the fact that nearly all of them no longer had custody of their children. Likely, the foremost reason why these women no longer have custody may be the fact that they were overwhelmed by the experience, and were not able to adequately care for their children. In other words, they were experiencing intolerable levels of 158 parental stress. Now that they are no longer burdened with the responsibility of being primary caregiver, their interactions with their children might be comparatively stress-free. A second explanation is that these mothers do not know what constitutes proper parenting, and therefore do not know that they are not doing it well. However, a third explanation deserves consideration. As previously discussed, women who reported lower levels of parental stress tended to have higher scores on the Impression Management scale of the BIDR, suggesting that may have been purposely trying to present themselves in a favourable light. Further, it was found throughout interviews that the participants were particularly defensive about their relationships with their children. Many refused to answer questions, and some became very upset or angry when questions pertained in any way to their parenting experiences. Additionally, many of these women have had unfavourable contact with Social Services, and have learned through experience that being honest about such matters may lead to great heartbreak and malcontent. For instance, as previously mentioned, some female offenders related that they had approached Social Service in the past, of their own free will, in order to obtain some assistance with their children, as the mother was well aware that they were doing a substandard job of taking care of their children (usually due to drug and alcohol abuse). These women were hoping for a reprieve while they could get a handle on their addictions, their finances, and their general living situations. However, several of these requests for help resulted in circumstances in which these mothers had no control over their children, often had no access to them, and in some cases, were reportedly coerced26 into agreeing to give their children up permanently. 2 6 This coercion might have been in the form of repeated denials of requests to see children, repeated assertions that giving the child up for adoption would be best for the child, and most importantly, settlements were reportedly 159 Whether or not these reports are accurate, what is relevant here is the fact that this might have been the way these inmates perceived their experiences with regards to reporting parent-child circumstances, and therefore they would understandably be reluctant to discuss such matters with a virtual stranger, regardless of assurances of confidentiality. In addition, these women were clearly warned at the beginning of the interview that any information that they may give involving harm to a child (although this is intended to prevent future or impending harm to a child) must be reported by the interviewer. This may have sent up a 'warning flag' of sorts to the participants that they should not speak freely of their relationships with their children, or more accurately, they should only speak of these relationships in positive terms. Regardless of the reasons for the lower-than-expected scores on the PSS, it is possible that the scores of this sample are unreliable, and in this light, any findings relating to the PSS, or lack thereof, should be viewed with caution. Nevertheless, it was found that attachment styles, when collapsed into secure and insecure categories, were related to scores on the Parental Stress Scale. Specifically, women with insecure attachment styles reported higher levels of parental stress than women with secure attachment styles. So while the sample as a whole seems to have under-reported levels of parental stress, the differences between securely and non-securely attached women on levels of parental stress were in the expected direction. It must be noted that previous research has found that there are many variables that can be responsible for higher levels of parental stress, for example, financial difficulties, number of children, child characteristics, and being a single parent (Bendell, Stone, Field, & usually made while the mother was at a low point in her recovery, i.e., was seriously abusing drugs and alcohol, and 160 Goldstein, 1989; Conger, McCarty, Yans, Lahey, & Kropp, 1984; Belle, 1982; Beckman, 1983). It is these issues, as well as issues of substance abuse and personality pathology that should be addressed in programs that could help the female offender cope with the parenting role once released from prison. In addition, tailoring these programs to address attachment issues may be more effective, in that women with varying attachment styles are likely to respond to different types of treatment. For example, a woman with a dismissing attachment style might benefit more from group therapy because she could learn alternative ways of relating to people by hearing how non-dismissive women describe their feelings and relationships (Horowitz, Rosenberg, & Bartholomew, 1993). Further, dismissing individuals are more likely to benefit from extended rather than brief treatment (Horowitz et al, 1993). The benefits of treatment could be maximized if the treatment providers take these sorts of individual differences in interpersonal styles into account during the development and administration of programs. Limitations and Recommendations for Future Research Several difficulties are inherent when investigating a female offender population. Firstly, there is an inconsistency in data collection by the criminal justice system from the time of arrest to release from prison, resulting in a lack of accurate information about victims, offenders, and the crimes themselves (Johnson and Rodgers, 1993). Further, there is an inconsistency in data collection within the correctional institutions, resulting in a dearth of information on offenders, with the exception of Federal offenders who constitute a small minority of all female offenders. In addition, data that is available is based largely on self-report, and no attempt is made to verify or controvert these reports. In fact, some files had contradictory information within themselves, was likely unable to foresee favourable alternatives to relinquishing custody. 161 and these obvious discrepancies were not investigated by the institutional staff. Also, these findings cannot be assumed to be generalizable to other female offender populations, as this sample is comprised of women who have federal or provincial sentences, or are awaiting sentencing or trial. Most facilities do not house all three groups of women together. There is another fundamental difficulty in working with this population. There is no means through which to verify their responses, and many of these women are seasoned liars- it is one of their survival tools, although some may do it just because they can (i.e., "duping delight"; Ekman, 1992). As noted earlier, some topics lend themselves to misrepresentation more than others. In this project, women were reluctant to speak candidly about their relationships with their children, although contradictions within interviews or between interview and file information occurred across a variety of topics as serious as suicide attempts or as comparatively minor as educational level attained. Impression management was an issue, not only in women's responses regarding their children, but also in their responses regarding their attachment orientations, particularly on the Experiences in Close Relationships Inventory, the PSS, and the SCID-II. Therefore, results from these measures should be interpreted with caution. Additionally, time and financial constraints as well as institutional policy prevented the depth of interviewing necessary to gather the quality and quantity of information necessary to explore the intricacies of these women's interpersonal experiences. Because many of these women were not comfortable discussing (or acknowledging to themselves) private issues with respect to their relationships, and because they were only being paid $5.00 for their time (future research should offer at least $10.00), there was a sense of urgency to complete the 2-hour interview and questionnaire session and get back to their units. It would have been desirable to have further time to ask more questions. For example, abuse by primary caregivers does not even scratch the 162 surface of the possible abuse experiences to which this population is vulnerable. However, as mentioned previously, BCCW was opposed to further questioning in this area. As well, it would have been beneficial to delve more deeply into whether or not abuse has been reported, and if so, what was done about it, and if nothing, why? Also, although women were reluctant to discuss motherhood, questions that would have been informative include: What are the things that you found the hardest about being a mother? What sorts of things frustrated you? Further, as can be seen by the preceding transcripts, which were excerpts from the Peer Attachment Interview, this interview delved more deeply into the interpersonal relationships of the participant- how easily she got into and out of romantic involvements and the difficulties she encountered therein. Unfortunately, only a portion of the present sample was assessed using the Peer Attachment Interview, due to time and financial constraints involved in training interviewers, interviewing participants, and coding interviews. Without such constraints, attachment styles as assessed by the PAI would be a logical choice when performing analyses in this research, because the PAI is an interview rather than a self-report questionnaire, it permits a much more comprehensive look into the interpersonal lives of these offenders. It is recommended that future research utilize the PAI or the Adult Attachment Interview when feasible. A meta-analysis of the AAI has indicated that this measure can predict the quality of the infant-parent attachment relationship as well as the parents' responsiveness to their infants' attachment signals (van Ijzendoorn, 1995). Unfortunately, this interview is time-consuming, and the training involved is expensive and lengthy, reducing its practicality for many researchers. As well, due to time constraints with the current sample, several questions were omitted from Susan Turnbull's (1996) background information questionnaire adapted for this project. In retrospect, questions that should have been retained include: How many serious romantic 163 relationships have you had? How long did your longest romantic relationship last? How many close personal friendships do you have? How long has your longest friendship lasted? Has a previous partner's use of drugs and/or alcohol had a negative effect on any of your romantic relationships? Has your current partner's use of drugs and/or alcohol had a negative effect on your current romantic relationship? Has your current partner ever been arrested? If yes, how many times and is he/[she] currently in jail? Do you plan on returning to your partner when you leave BCCW? How would you rate your level of satisfaction with your current romantic relationship? Methodologically, the use of interviews is recommended with this population. Reasons for this include a potential literacy deficit- inability to read or a limited vocabulary-, or possibly lack of insight on the part of the offender. If an offender fills out a self-report questionnaire and does not understand a word or a question, she may guess or skip it and move on to another item. If she is being interviewed and hesitates or gives an answer that does not appear to follow the line of questioning, the interviewer can clarify, thus increasing the probability of greater accuracy in reporting. Further, many women in this sample were defensive when answering questions regarding their relationships with their children, and this will need to be addressed in future research. It is recommended that a questionnaire format be introduced in order to avoid the level of impression management that was found with this sample. As well, it is recommended that existing scales be validated with a female offender population, and that future scales attempt to address the issue of reliability of the measures, particularly the stability of test-retest outcomes and the accuracy of categorization within the insecure attachment styles. This research found discrepancies between the RQ and the ECRI on the categorization of women into the three insecure attachment styles. The most salient discrepancy was found in 164 analyses of the fearful and preoccupied attachment styles {fearful 25.2%, ECRI; 42.5%, RQ; preoccupied 35.1% ECRI; 12.3% RQ), whereas there was a much smaller discrepancy between the two measure for those classified as dismissing (14.4% ECRI; 19.8% RQ), and almost no difference between the two measures for those categorized as securely attached (25.2% ECRI; 25.5% RQ). It appears that these measures cannot discriminate between fearful and preoccupied attachment styles in a reliable way. Perhaps individuals with fearful and preoccupied orientations are more similar than different in that both want to have close relationships and both are fearful of getting hurt, they simply manifest these needs and fears differently; the preoccupied individual clings to her love object whereas the fearful individual pushes her love object away. Attachment theory purports that it is the underlying working models or unconscious representations/expectations of others that determines how we experience relationships, so perhaps the underlying needs and fears of the preoccupied and fearful individuals are more salient or are being tapped into more effectively by these measures that their reactions to their needs and fears (clinging versus pushing away behaviours). Maybe fearful and preoccupied attachment orientations have the same underlying construct, and that the behavioural outcome is irrelevant to the underlying construct and its measurement. Perhaps one or both of the measures used in this research are not tapping into the outcomes or behavioural manifestations that distinguish these two styles. For instance, it may be that the paragraph description of the preoccupied attachment orientation is too extreme, in that a person has to endorse that they want to be completely emotionally intimate with others. It also presumes that the individual has the insight to recognize that others are reluctant to get as close as they would like, and finally, these female offenders may not identify with the term "value" as it is used in the final statement of the preoccupied description: "... I sometimes worry that others don't value me as much as I value 165 them", or perhaps this is not the way that they characterize their interpersonal difficulties with their romantic partners. Impression management scores were not significantly related to the preoccupied attachment style, as assessed by the RQ, although there was a trend (r = -A79,p=. 061) that suggests that women who scored high on impression management were less likely to endorse the preoccupied attachment style, and there was a very similar, but significant relationship between IM and fearful attachment, as assessed by the RQ, (r = -.187,/) =. 050). However, there was a stronger significant relationship between IM and the dismissing attachment style, as assessed by the RQ (r = .214, p =. 024), suggesting that individuals who scored high on IM were likely to endorse the dismissing attachment style. It is possible that women in this sample who may in fact have a fearful or preoccupied attachment style, may endorse a dismissing attachment style, in that this style is characterized by independence and self-sufficiency, and does not indicate a need for close relationships. Therefore, if a woman is without a close relationship, it is her choice, according to this attachment orientation. However, there was not a large over-representation of dismissing women, as assessed by the RQ, in this sample. In contrast to the fearful and preoccupied orientations, there was less discrepancy between the two measures for those with a dismissing attachment style, indicating that this may be a more distinctive insecure attachment style with its own underlying construct. The dismissing individual does not wish to be close to others or to have others depend on her. Fear is not a salient factor of the dismissing construct as it is with the two other insecure attachment styles, and although those with a dismissing style may be behaviourally manifested as pushing others away as do those with a fearful style, there is (purportedly) no underlying desire to be close. This supports the suggestion that the behavioural manifestation of the insecure attachment 166 is less salient to the underlying construct and perhaps the measurement of attachment than is the underlying motivations for those behaviours. Hazan & Shaver's (1987) adult attachment scale did not discriminate between fearful and preoccupied attachment styles, nor did Main (1985). The former entitled the corresponding attachment style "preoccupied" and the latter entitled it "anxious-ambivalent", indicating that there may be a lack of a valid distinction between the fearful and preoccupied categories. Regardless, these attachment styles have repeatedly been found to be less stable than either the secure or dismissing (avoidant) attachment styles (see "Continuity and Stability of Attachment Styles Across the Lifespan" on page 38 of this document for extensive review of these findings). However, it does appear that both the RQ and the ECRI can consistently discriminate between those with secure attachment styles versus those with insecure attachment styles, however, the finer distinctions of the insecurely attached women are problematic. Further recommendations include scale development and validation research that attempts to account for state versus trait responding in order to discriminate between temporary and stable personality characteristics as they relate to attachment style, taking such things into consideration such as drug/alcohol withdrawal and stress surrounding their prison admission, their recent relationship experiences with family members, romantic partners, and children, as well as any recent experiences as victims of crime. The ability to focus on trait rather than state characteristics would increase test-retest stability, and may produce stronger relationships between attachment styles and other interpersonal variables, therefore potentially increasing the validity of future attachment research. In addition to dealing with the measurement issues surrounding stability, it would be interesting to see whether or how attachment styles alter when childhood living circumstances alter; for example, does a child's attachment style change after 167 the child moves into a foster home? If so, does this depend on the age of the child, the number of times the child moves, the attachment styles of the foster parents, the attachment style of the initial primary caregivers, the attachment style of the child before moving into the foster home, or any or all of the above? Further, is a child more likely to have a secure attachment orientation if the child is placed with a family member, and does this depend upon the attachment style of the family member? And importantly, what characteristics or set of circumstances make it possible or likely that a child's attachment style can change, particularly from insecure to secure, and is there a sensitive period in which this transition is most likely to occur? Future research should also continue to address how to more accurately measure the relationship between personality disorders and insecure attachment styles, as well as the extent to which insecure attachment is related to drug addiction, the effects of stress on reports of personality pathology and on reports of insecure attachment orientation. Also, it would be interesting to investigate potential reasons for the finding that there was no relationship between attachment security and emotional or physical abuse by mothers, or between attachment security and sexual abuse by fathers. In particular, the latter question may be addressed by delving further into the offenders' extra-familial victimization experiences. Additionally, future research should investigate further the contributions of fathers to attachment orientation as well as personality disorders, parental stress, and childhood experiences with primary caregivers. Finally, future research on attachment styles in female offenders should include the use of control groups, either from drug treatment groups or using an inner city sample, and some attempt should be made to corroborate offenders' (and control participants') reports, possibly through the use of peer or family reports. However, this method also has its limitations. 168 Conclusion This research has attempted to illuminate the circumstances surrounding the interpersonal difficulties of the female offender. As can be seen, the majority of these women are imprisoned by, and at least indirectly because of, their addictions to substances such as crack, heroin, and cocaine. The predominant lifestyle is characterized by crime, drug use, and relationships with like-minded individuals. Three-quarters of these women have given birth to children, and nearly 90% no longer had custody of their children prior to their current incarceration. More than half did not have a cohesive family upbringing, and several had terribly fragmented upbringings that involved neglect, abuse, foster homes, group homes, parental criminality, and/or parental substance abuse. In addition, participants reported unstable relationships with the fathers of their children, as well as low levels of parental warmth and high levels of parental rejection. Further, the presence of an insecure attachment style was related to higher mean number of personality disorder characteristics, higher reported means of maternal and paternal rejection, and higher reported means of emotional and physical abuse by fathers. Attachment theory was helpful in understanding the interpersonal difficulties that these women face and although it was not a mediating factor that tied all the variables together in a direct way, this may have been due to the impression management issues found within most of the measures. Alternatively, attachment orientation may be more proximal to the outcome variables and there may be less of a direct relationship between adverse childhood experiences with parents and later parental and personality functioning. Bowlby (1988) ventured that there is a strongly interactive effect of adverse experiences, and that someone who has had one adverse experience has an increased likelihood of having another. Bowlby (1988) further suggested that adverse childhood experiences make an 169 individual more vulnerable to future adverse experiences (this being out of the individual's control) and also make it more likely that such adverse experiences will actually occur (this being the consequences of the individual's actions that occur as a result of the personality problems that adverse experiences contributed to). This indirect cycle of adversity certainly appears to be manifesting itself among the female offenders in this sample. As Sirkia (2000) suggests: Perhaps we should not be querying why it is that they have arrived in prison, but rather marvel at those with similar experiences who have not. What attachment [theory] teaches us is that their behaviour may be considered an adaptive response to the conditions they found themselves in; the child who learns to avoid [her] caregiver reduces [her] anxiety, as does the inmate with a Dismissing attachment style who does not consider the impact of [her] behaviour on others (p.89). Survival is inherently essential to us all, and whereas a secure attachment orientation might be adaptive to an educated middle class woman who has no history of abuse and has a relationship with a dependable and responsive partner, the open and trusting nature of a securely attached individual would be foolhardy in a woman who is caught up in a cycle of adversity. Future research should address whether programs and treatment should incorporate information about individual offender's attachment styles as a method of increasing the likelihood of successful long-term outcomes after they are released from prison. This is even more vital when the number of female offenders who have lost custody of their children is taken into account. These children are now embarking down their own paths of unstable relationships with their parents, and this instability may be increasing these children's risks of victimization, future personality pathology, and attachment insecurity, as was the case with many of their mothers. The needs of these children as well as the needs of female offenders as parents must be taken into consideration both in prison and post-release, as their parenting difficulties are adversely affecting a new generation, and timely and consistent intervention may ameliorate some of the challenges both 170 mothers and children will face. Attachment theory provides a rich basis for research in this area and has much to offer towards this end. 171 References Adelberg, E. , & Currie, C. (1987). Too few to count: Canadian women in conflict with the law. Vancouver: Press Gang Publishers. Ainsworth, M.D.S. (1967). Infancy in Uganda: Infant care and the growth of love. Baltimore: Johns Hopkins University Press. Ainsworth, M.D.S. (1978). Patterns of attachment. Hillsdale, NJ: Erlbaum. Ainsworth, M.D.S., Blehar, M.C. , Waters, E. , & Wall, S. (1978). Patterns of attachment: A study of the strange situation. Hillsdale, NJ: Erlbaum Ainsworth, M.D.S., & Bowlby, J. (1991). An ethological approach to personality development. American Psychologist. 46 (4), 333-341. Ainsworth, M.D.S., & Eichberg, C. (1991). Effects on infant-mother attachment on mother's unresolved loss of an attachment figure or some other traumatic experience. In P. Marris, J. Stevenson-Hinde, & C. Parkes (Eds.), Attachment across the life cycle (pp. 160-183). New York: Routledge. Alexander, P.C. (1990). Interventions with incestuous families. In S.W. Henggeler & C M . Borduin (Eds.), Family therapy and beyond (pp.324-344). Pacific Grove, Ca: Brooks/Cole. Alexander, P.C. (1992). Application of attachment theory to the study of sexual abuse. Journal of Consulting & Clinical Psychology, 60 (2), 185-195. Alexander, P.C. (1993). The differential effects of abuse characteristics and attachment in the prediction of long-term effects of sexual abuse. Journal of Interpersonal Violence.8 (3). 346-362. Allen, C , & Lee, C. (1992). Family of origin structure and intra/extrafamilial childhood sexual victimization of male and female offenders. Journal of Child Sexual Abuse. 1 (3). 31-45. Allen, J.P., Leadbeater, B.J., & Aber, J.L. (1994). The development of problem behavior syndromes in at-risk adolescents. Development and Psychopathology. 6. 323-342. American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders. Fourth Edition. Washington, D.C. American Psychiatric Association. Arrindell, W.A., Kwee, M.G.T. , Methorst, G.J., Van Der Ende, J., Pol, E. , & Moritz, B.J.M. (1989). Perceived parental rearing styles of Agoraphobic and Socially Phobic In-patients. British Journal of Psychiatry. 155. 526-535. Baldwin, M.W., & Fehr, B. (1995). On the instability of attachment ratings. Personal Relationships. 2. 247-261. Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social and Personal Relationships. 7 (2). 147-178. 172 Bartholomew, K. (1997). Adult attachment processes: individual and couple perspectives. British Journal of Medical Psychology. 70 (3), 249-263. Bartholomew, K. & Horowitz, L . M . (1991). Attachment styles among young adults: A test of the four-category model. Journal of Personality & Social Psychology, 61, 226-244. Bartholomew, K., Kwong, M.J., & Hart, S.D. (2001). An attachment perspective on personality disorders. In J. Livesly (Ed.), Handbook of personality disorders: Theory, research. & treatment (pp. 196-230). New York: Guilford. Bartholomew, K., & Shaver, P.R. (1998). Methods of assessing adult attachment: Do they converge? In J.A.Simpson, & W.S. Rholes (Eds.), Attachment theory and close relationships (pp. 25-45). New York: Guilford Press. Bauserman, R., & Rind, B. (1997). Psychological correlates of male child and adolescent sexual experiences with adults: a review of the nonclinical literature. Archives of Sexual Behavior, 26 (2), 1 OS-M I . Beckman, P.J. (1983). Influence of selected child characteristics on stress in families of handicapped infants. American Journal of Mental Deficiency, 88, 2, 150-156. Belle, D. (1982). Social ties and social support. In D. Belle, (Ed.), Lives in Stress: Women and Depression. Beverly Hills, Sage. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55. 83-96. Bendell, R.D., Stone, W., Field, T, & Goldstein, S. (1989). Children's effects on parenting stress in a low income, minority population. Topics in Early Childhood. 8. 4, 58-71. Benoit, D., & Parker, K. (1994). Stability and transmission of attachment across three generations. Child Development, 65 (5), 1444-1456. Benward, J. & Densen-Gerber, J. (1975). Incest as a causative factor in antisocial behaviour: An exploratory study. Contemporary Drug Problems. 4. 323-340. Berman, W.H., & Sperling, M.B. (1994). The structure and function of adult attachment. In M.B. Sperling & W.H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 3-28). New York: Guilford Press. Berry, J.O, & Jones, W.H. (1995). The Parental Stress Scale: Initial psychometric evidence. Journal of Social and Personal Relationships, 12 (3). 463-472. Blanchette, K., & Motiuk, L . L . (1996). Female offenders with and without major mental health problems: A comparative investigation. Correctional Services of Canada, May 1996, R-46. Blatz, W.E. (1966). Human Security: Some reflections. Toronto, Canada: University of Toronto Press. 173 Bowlby, J. (1944). Forty-four juvenile thieves: Their characters and home lives. International Journal of Psychoanalysis. 25. 19-52. Bowlby, J. (1969). Attachment and loss. Vol. I: Attachment. New York: Basic Books. Bowlby, J. (1973). Attachment and loss. Vol. II: Separation. New York: Basic Books. Bowlby, J. (1979). The making and breaking of affectional bonds. London: Tavistock. Bowlby, J. (1980). Attachment and loss. Vol. Ill: Loss, sadness. & depression. New York: Basic Books. Bowlby, J. (1988). A Secure Base: Parent-Child Attachment and Healthy Human Development. New York: Basic Books, Inc. Bradley, R.H. (1998). In defense of parental investment. Journal of Marriage & the Family, 60. 3, 791-795. Bradley, R.H., Whiteside-Mansell, L . , Brisby, J.A., & Caldwell, B .M. (1997). Parents' socioemotional investment in children. Journal of Marriage and the Family. 59, 77-90. Brennan, K.A. , Clark, C.L. , & Shaver, P.R. (1998). Self report measures of adult attachment: An integrative overview. In J.A. Simpson & W.S. Rholes (Eds.), Attachment theory and close relationships (pp. 46-76). New York: Guilford. Brennan, K.A. , Shaver, P.R., & Tobey, A .E . (1991). Attachment styles, gender, and parental problem drinking. Journal of Social and Personal Relationships, 8, 451-466. Brennan, K.A. , & Shaver, P.R. (1995). Dimensions of adult attachment, affect regulation, and romantic relationship functioning. Personality and Social Psychology Bulletin, 21, 267-283. Bretherton, I. (1985). Attachment theory: Retrospect and prospect. Monographs of the Society for Research in Child Development. 50 (1-2), 3-38. Bretherton, I. (1987). New perspectives on attachment relations: Security, communications and internal working models. In J.D. Osofsky (Ed.), Handbook of infant development (2nd Ed.) (pp. 1061-1100). New York: Wiley & Sons. Bretherton, I. (1990). Communication patterns, internal working models, and the intergenerational transmission of attachment relationships. Infant Mental Health Journal. 11(3). 237-252. Bretherton, I. (1992). The origins of attachment theory: John Bowlby and Mary Ainsworth. In S. Goldberg, R. Muir, & J. Kerr (Eds.), Attachment theory: Social, developmental and clinical perspectives (pp. 45-84). Hillsdale NJ: Analytic Press. Bretherton, I., Birigin, Z., Ridgeway, D., Maslin, C , & Sherman, M . (1989). Attachment: The parental perspective. Infant Mental Health Journal, 10, 203-222. Brownstone, D.Y., & Swaminath, R.S. (1989). Violent behavior and psychiatric diagnosis in female offenders. Canadian Journal of Psychiatry. 34, 190-194. 174 Campos, J.J., Barrett, K . C . , Lamb, M.E. , Goldsmith, H.H., & Sternberg, C. (1983). Socioemotional development. In M . M . Haith & J.J. Campos (Eds.), Handbook of child psychology: Vol.2. Infancy and psychobiology (pp. 783-915). New York: Wiley. Carnelly, K.B. , Pietromonaco, P.R., & Jaffe, K. (1994). Depression, working models of others, and relationship functioning. Journal of Personality and Social Psychology, 66, 127-140. Cheek, J.M. & Buss, A . H . (1981). Shyness and sociability. Journal of Personality & Social Psychology. 41. 330-339. Chess, S., & Thomas, A. (1982). Infant bonding: mystery and reality. American Journal of Orthopsychiatry, 52, 213-222. Cohn, D.A., Silver, D.H., Cowan, C P . , Cowan, P.A., & Pearson, J. (1992). Working models of childhood attachment and couple relationships. Journal of Family Issues. 13 (4), 432-449. Cole, P., & Woolger, C. (1989). Incest survivors: The relation of their perceptions of their parents and their own parenting attitudes. Child Abuse and Neglect. 13 (3), 409-416. Cole, P., Woolger, C , Power, T., & Smith, D. (1992). Parenting difficulties among adult survivors of father-daughter incest. Child Abuse and Neglect. 16 (2), 239-249. Colin, V. (1996). Human attachment. Philadelphia: Temple University Press. Coletta, N. (1979). Support systems after divorce: Incidence and impact. Journal of Marriage & the Family. 41. 837-846. Collins, N.L. , & Read, S.J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology. 58 (4"). 644-663. Collins, N. L. , & Read, S. J. (1994). Cognitive representations of adult attachment: The structure and function of working models. In K. Bartholomew & D. Perlman (Eds.) Advances in personal relationships, Vol. 5: Attachment processes in adulthood (pp. 53-90). London: Jessica Kingsley, Inc. Conger, R., McCarty, J., Yans, R., Lahey, B., & Kropp, J. (1984). Perception of child, child-rearing values, and emotional distresses mediating links between environmental stressors and observed maternal behaviour. Child Development. 55, 6, 2234-2247. Crittenden, P.M., (1997). Patterns of attachment and sexual behavior: risk of dysfunctional versus opportunity for creative integration. In L. Atkinson and K.J. Zucker (Eds.), Attachment and psychopathology (pp. 47-93). New York: Guilford Press. Crittenden, P.M., & Ainsworth, M.D.S. (1989). Child maltreatment and attachment theory. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and research on the causes and consequences of child abuse and neglect (pp. 432-463). New York, NY, USA: Cambridge University Press. Crittenden, P.M., Partridge, M.F. , & Claussen, A . H . (1999). Family patterns of relationship in normative and dysfunctional families. Development & Psychopathology, 3, 491-512. 175 Crowell, J. A. , Fraley, R.C., & Shaver, P.R. (1999). Measurement of individual differences in adolescent and adult attachment. In J. Cassidy & P.R. Shaver (Eds). Handbook of attachment: Theory, research. & clinical applications (pp. 434-465). New York: Guilford Press. Crowell, J. A., & Treboux, D. (1995). A review if adult attachment measures: Implications for theory and research. Social Development. 4 (3), 294-327. Dahl, A. (1986). Some aspects of the DSM-III personality disorders illustrated by a consecutive sample of hospitalized patients. Acta Psychiatra Scandanavia. 73. 61-66. Daniel, A . E . & Harris, P.W. (1981). Female offenders referred for pre-trial psychiatric evaluation. Bulletin of the American Academy of Psychiatry and Law. 9. 40-47. Daniel, A . E . , & Harris, P.W. (1982). Female homicide offenders referred for pre-trial psychiatric examination: A descriptive study. Bulletin of the American Academy of Psychiatry and the Law. 10 (4), 261-269. Daniel, A . E . , Robins, A.J. , Reid, J.C., & Wilfley, D.E. (1988). Lifetime and six-month prevalence of psychiatric disorders among sentenced female offenders. Bulletin of the American Academy of Psychiatry and the Law. 16 (4), 333-342. Diamond, D., & Blatt, S.J. (1994). Internal working models and the representational world in attachment and psychoanalytic theories. In M.B. Sperling & W.H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 72-97). New York: Guilford Press. D'Orban, P.J. (1979). Women who kill their children. British Journal of Psychiatry. 134. 560-571. Dutton, D.G., Saunders, K., Starzomski, A., & Bartholomew, K. (1994). Intimacy-anger and insecure attachment as precursors of abuse in intimate relationships. Journal of Applied Social Psychology. 24. 1367-1386. Edwards, S. (1981). Female sexuality and the law. Oxford: Martin Robertson. Egeland, B., Jacobovitz, D., & Sroufe, L.A. (1988). Breaking the cycle of abuse. Child Development. 59. 1080-1088. Egeland, B., & Sroufe, L.A. (1981). Attachment and early maltreatment. Child Development. 52. 44-52. Ekman, P. (1992). Telling lies. (2nd ed.). New York: Norton & Co. Ekselius, L . , Bodlund, O., von-Knorring, L. , Lindstroem, E. , & Kullgren, G. (1996). Sex differences in DSM-III-R, Axis II~Personality disorders. Personality and Individual Differences. 20 (4), 457-461. Ekselius, L . , Lindstrom, E. , & Von Knorring, L. (1994). SCID-II interviews and the SCID Screen Questionnaire as diagnostic tools for personality disorders in DSM-III-R. Acta Psychiatra Scandanavia. 90, 120-123. 176 Ekselius, L . , Tillfors, M . , Furmark, T., & Fredrikson, M . (2001). Personality disorders in the general population: DSM-IV and ICD-10 defined prevalence as related to sociodemographic profile. Personality and Individual Differences, 30. 311-320. Feeney, J.A., & Noller, P. (1990). Attachment style as a predictor of adult romantic relationships. Journal of Personality and Social Psychology. 58. 281 -291. Feeney, J.A., & Noller, P. (1992). Attachment style and romantic love: Relationship dissolution. Australian Journal of Psychology, 44 (2), 69-74. Feeney, J.A., & Noller, P. (1996). Adult Attachment. Thousand Oaks: Sage Publications. Feeney, J.A., Noller, P., & Callan, V.J. (1994). Attachment style, communication and satisfaction in the early years of marriage. In K. Bartholomew & D. Perlman (Eds.), Advances in personal relationships (Vol. 5, pp. 269-308). London: Jessica Kingsley. Feeney, J.A., Noller, P., & Hanrahan, M . (1994). Assessing adult attachment: Developments in the conceptualization of security and insecurity. In M.B. Sperling & W.H. Berman (Eds.), Attachment in adults: Theory, assessment, and treatment (pp. 128-152). New York: Guilford. Feeney, J.A., Noller, P., & Patty, J. (1993). Adolescents' interactions with the opposite sex: Influence of attachment style and gender. Journal of Adolescence, 16, 169-186. Fey, W.F. (1955). Acceptance by others and its relation to acceptance of self and others: A revaluation. Journal of Abnormal Social Psychology. 50. 275-276. Finkelhor, D. (1986). A sourcebook of child sexual abuse. Beverly Hills: Sage. First, M.B., Gibbon, M . , Spitzer, R.L., Williams, B.W., & Benjamin, L.S. (1997). Structured Clinical Interview for DSM-IV Axis-II Personality Disorders. Washington, DC: American Psychiatric Press. Fonagy, P., Leigh, T. Steele, M . , Steele, H. , Kennedy, R., Matoon, G. , Target, M . , & Gerber, A. (1996). The relation of attachment status, psychiatric classification, and response to psychotherapy. Journal of Consulting & Clinical Psychology, 64. 22-31. Fonagy, P., Steele, H. & Steele, M . (1991). Maternal representations of attachment during pregnancy predict the organization of infant-mother attachment at one year of age. Child Development, 62, 891-905. Founder, S., & Crey, E. (1997). Stolen From Our Embrace: The Abduction of First Nations Children and the Restoration of Aboriginal Communities. Vancouver: Douglas & Mclntyre. Fraiberg, S., Adelson, E. , & Shapiro, V. (1975). Ghosts in the nursery: A psychoanalytic approach to the problems of impaired mother-infant relationships. Journal of the American Academy of Child Psychiatry. 14. 387-421. Fraley, R.C., & Shaver, P.R. (2000). Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology. 4 (2), 132-154. 177 Fraley, R.C., & Waller, N. (1998). Adult attachment patterns: a test of the typological model. In J.A. Simpson & W.S. Rholes (Eds.), Attachment theory and close relationships (pp. 77-114). New York: The Guilford Press. Frommer, E. & O'Shea, G. (1973). The importance of childhood experiences in relation to problems in managing their infants. British Journal of Psychiatry, 123. 157-160. George, C. (1996). The representational perspective of child abuse and prevention: Internal working models of attachment and caregiving. Child Abuse & Neglect, 20 (5), 411-424. George, C , Kaplan, N., & Main, M . (1985). The Berkeley Adult Attachment Interview. Unpublished protocol, University of California at Berkeley. Gerlsma, C , Emmelkamp, P.M.G., & Arrindell, W.A. (1990). Anxiety, depression, and perception of early parenting: A meta-analysis. Clinical Psychology Review, 10, 251-277. Gold, E.R. (1986). Long-term effects of sexual victimization in childhood: An attributional approach. Journal of Consulting & Clinical Psychology, 54, 471-475. Goldberg, S. (1991). Recent developments in attachment theory and research. Canadian Journal of Psychiatry. 36. 393-400. Golomb, M . , Fava, M . , Abraham.M., & Rosenbaum, J.F. (1995). Gender differences in personality disorders. American Journal of Psychiatry, 152 (4), 579-582. Goodwin, J., McCarthy, T , & DiVasto, P. (1981). Prior incest in mothers of sexually abused children. Child Abuse & Neglect, 5. 87-96. Greenberg, M.T. (1999). Attachment and psychopathology in childhood. In J. Cassidy & P.R. Shaver (Eds.), Handbook of Attachment: Theory, Research. & Clinical Applications (pp. 469-496). New York: Guilford Press. Greenfield, L.A. , & Snell, T.L. (1999). Women Offenders (Rep. No. NCJ 175688). Washington, DC: U.S. Department of Justice Griffin, D.W., & Bartholomew, K. (1994). The metaphysics of measurement: The case of adult attachment. Advances in Personal Relationships. 5. 17-52. Grilo, C M . , Becker, D.F., Fehon, D . C , & Walker, M.L. (1996). Gender differences in personality disorders in psychiatrically hospitalized adolescents. American Journal of Psychiatry, 153 (8), 1089-1091. Grossman, K., Fremmer-Bombik, E. , Rudolf, J., & Grossman, K. (1988). Maternal attachment representations as related to patterns of infant-mother attachment and maternal care during the first year. In R.A. Hinde & J. Stevenson-Hinde (Eds.), Relationships within families: Mutual influences (pp.241-260). Oxford, England: Clarendon Press. Gwinn, B. (1992). Linking intimate families together. Federal Prisons Journal. 3, 7-40. Haft, M . G . (1980). Women in prison: Discriminatory practices and some legal solutions. In S.K. Datesman & F.R. Scarpitti (Eds.), Women, crime. & justice (pp. 320-338). New York: Oxford University Press. 178 Haft, W.L., & Slade, A. (1989). Affect attunement and maternal attachment: A pilot study. Infant Mental Health Journal. 10 (3). 157-172. Hare, R.D. (1993) Without Conscience. New York: Simon & Schuster, Inc. Harlow, H. (1976). Effects of maternal and peer separations on young monkeys. Journal of Child Psychology & Psychiatry & Allied Disciplines. 17, 2, 101-112. Harmer, A . L . M . , Sanderson, J., & Mertin, P. (1999). Influence of negative childhood experiences on psychological functioning, social support, and parenting for mothers recovering from addiction. Child Abuse & Neglect. 23 (5), 421-433. Havens, J.F., Mellins, C.A., Pilowski, D. (1997). Mental health issues in HIV-affected women and children. International Review of Psychiatry, 8, 2-3, 217-225. Hays, S. (1998). The fallacious assumptions and unrealistic prescriptions of attachment theory: A comment on "Parents' Socioemotional Investment in Children". Journal of Marriage & the Family, 60. 782-795. Health Canada. (2001). Final Report, Ottawa, Ontario: Canada. Hemphill, J.F. (1998). Psychopathy and recidivism: A review. Legal & Criminological Psychology. 3. 139-170. Herjanic, M . , Henn, F.A., & Vanderpearl, R.H. (1977). Forensic psychiatry: Female offenders. American Journal of Psychiatry, 134 (5), 556-558. Hock, E. , Gnezda, M.T., & McBride, S. (1983, April). The Measurement of Maternal Separation Anxiety. Paper presented at the biennial meeting of the Society for Research in Child Development. Detroit, MI. Horowitz, L . M . , Rosenberg, S.E., Baer, B.A., Ureno, G. & Villasenor, V.S. (1988). Inventory of Interpersonal Problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology. 56. 885-892. Horowitz, L . M . , Rosenberg, S.E., & Bartholomew, K. (1993). Interpersonal problems, attachment styles, and outcome in brief dynamic psychotherapy. Journal of Consulting and Clinical Psychology, 61,4, 549-560. Horowitz, L . M . , Rosenberg, S.E., Ureno, G., Kalehzan, B .M. , & O'Halloran, P. (1989). Psychodynamic formulation, consensual response method, and interpersonal problems. Journal of Consulting and Clinical Psychology, 57, 599-606. Hueston, W.J., Mainous, A .G. , & Schilling, R. (1996). Patients with personality disorders: functional status, health care utilization, and satisfaction with care. Journal of Family Practice. 42, 54-60. Ingram-Fogel, C. (1991). Health problems and needs for incarcerated women. Journal of Prison and Jail Health. 10 (1), 43-57. 179 Jacobsberg, L. , Perry, S., & Francis, A. (1995). Diagnostic agreement between the SCID-II screening questionnaire and the Personality Disorder Examination. Journal of Personality Disorder Assessment, 65, 428-433. Johnson, H. (1987). Getting the facts straight: A statistical overview. In E. Adelberg & C. Currie (Eds.), Too Few To Count: Canadian Women in Conflict with the Law (pp. 23-46). Vancouver: Press Gang Publishers. Johnson, H. , & Rodgers, K. (1993). A statistical overview of women and crime in Canada. In E. Adelberg & C. Currie (Eds.), In conflict with the law: Women and the Canadian justice system (pp.95-116). Vancouver: Press Gang Publishers. Keeney, B., & Heide, K. (1994). Gender differences in serial murderers: A preliminary analysis. Journal of Interpersonal Violence. 9 (3), 383-398. Kirkpatrick, L A . & Hazan, C. (1994). Attachment styles and close relationships: A four-year prospective study. Personal Relationships. 1, 123-142. Kobak, R.R., & Hazan, C. (1991). Attachment in marriage: effects of security and accuracy of working models. Journal of Personality and Social Psychology, 60 (6), 861-869. Kobak, R.R., & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation,and representations of self and others. Child Development, 59.135-146. Kroner, D.G., & Weekes, J.R. (1996). Balanced Inventory of Desirable Responding: Factor structure, reliability, and validity with an offender sample. Personality & Individual Differences, 21, 3, 323-333. Lake, E.S. (1993). An exploration of the violent victim experiences of female offenders. Violence & Victims. 8(1). 41-51. Levy, M.B. , & Davis, K . E . (1988). Lovestyles and attachment styles compared: Their relations to each other and to various relationship characteristics. Journal of Social & Personal Relationships, 5, 439-471. Livesley, W.J., Schroeder, M.L. , & Jackson, D.N. (1990). Dependent personality disorder and attachment problems. Journal of Personality Disorders. 4 (2), 131 -140. Loranger, A.W., Susman, V . L . , & Oldham, J.M. (1987). The Personality Disorder Examination: a preliminary report. Journal of Personality Disorders, 1, 1-13. Loucks, A.D. , & Zamble (in press). Predictors of criminal behavior and prison misconduct in serious female offenders. Empirical and Applied Criminal Justice Review. Lukianowicz, H. (1972). Incests: 1. Paternal incest. British Journal of Psychiatry, 120, 301-313. Lyons-Ruth, K. (1996). Attachment relationships among children with aggressive behavior problems: The role of disorganized early attachment patterns. Journal of Consulting & Clinical Psychology. 64 (1), 64-73. 180 MacLean, D. (1997). Mother and child live-in programs in Canadian prisons: A policy analysis of program development and implementation. Unpublished Master's thesis, Simon Fraser University. Maden, T., Swonton, M . , & Gunn, J. (1994). Psychiatric disorder in women serving a prison sentence. British Journal of Psychiatry. 164. 44-54. Main, M . & Cassidy, J. (1988). Categories of response to reunion with the parent at age 6: Predictable from infant attachment classifications and stable over a 1-month period. Developmental Psychology, 24. 415-426. Main, M . , & Goldwyn, R. (1984). Predicting rejection of her infant from mother's representation of her own experience: Implications for the abused-abusing intergenerational cycle. Child Abuse & Neglect. 8. 203-217. Main, M . & Goldwyn, R. (1985). Adult attachment rating and classification systems. In M . Main (Ed.), A typology of human attachment organization assessed in discourse, drawings and interviews. New York: Cambridge University Press. Main, M . & Hesse, E. (1990). Lack of resolution of mourning in adulthood and its relationship to infant disorganization: Some speculations regarding causal mechanisms. In M.T. Greenberg, D. Cicchetti, & E . M . Cummings (Eds.), Attachment in the Preschool Years: Theory, Research, and Intervention (pp. 160-183). Chicago: University of Chicago Press. Main, M . , Kaplan, N., & Cassidy, J. (1985). Security of infancy, childhood, and adulthood: A move to the level of representation. Monographs of the Society for Research in Child Development. 50 (1-2) 66-104. Main, M . , & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In M. Greenberg, D. Cicchetti, & M . Cummings (Eds.), Attachment in the preschool years: Theory, research, & intervention (pp. 121-160). Chicago: University of Chicago Press. Main, M . & Weston, D. (1981). Quality of attachment to mother and to father: Related to conflict behavior and the readiness for establishing new relationships. Child Development. 52, 932-940. Malow, R.M., West, J.A., & Williams, J.L. (1989). Personality disorders classifications and symptoms in cocaine and opioid addicts. Journal of Consulting and Clinical Psychology, 57, 765-767. Marshall, W.L. (1989). Invited essay: Intimacy, loneliness & sexual offenders. Behavior Research & Therapy. 27. 491-503. Matheson, P.B., Schoenbaum, E. , Greenberg, B., & Pliner, V. (1997). Association of maternal drug use during pregnancy with mother-to-child HIV transmission. AIDS. 11, 7, 941-942. Mathews, R., Matthews, J., & Speltz, K. (1990). Female sexual offenders. In M . Hunter (Ed.), The sexually abused male (pp. 275-293). Lexington, MA: Lexington Books/ D.C. Heath and Company. McClellan, D.S., Farabee, D., & Crouch, B .M. (1997). Early victimization, drug use, and criminality: A comparison of male and female prisoners. Criminal Justice & Behavior. 24 (4), 455-476. 181 McLean, D. (1997). Mother and child live-in programs in Canadian prisons: A policy analysis of program development and implementation. Unpublished master's thesis, Simon Fraser University, Burnaby, British Columbia. Melges, F.T., & Swartz, M.S. (1989). Oscillations of attachment in Borderline Personality Disorder. American Journal of Psychiatry. 146 (9). 1115-1120. Mickelson, K.D., Kessler, R.C., & Shaver, P.R. (1997). Adult attachment in a nationally representative sample. Journal of Personality & Social Psychology. 73. 1092-1106. Mikulincer, M . , Florian, V. , & Tolmacz, R. (1990). Attachment styles and fear of personal death. Journal of Personality and Social Psychology. 58 (2), 273-280. Morey, L. (1988). Personality disorders in DSM-III and DSM-III-R: Convergence, coverage, and internal consistency. American Journal of Psychiatry. 145, 573-577. Morrissette, P.J. (1994). The holocaust of First Nation People: Residual effects on parenting and treatment implications. Contemporary Family Therapy: An International Journal, 16 (5), 381-392. Pagelow, M.D. (1982). Child abuse and delinquency: Are there connections between childhood violence and later deviant behaviour? Paper presented at the Tenth World Congress of the International Sociological Association, Mexico City. Parker, G. (1994). Parental bonding and depressive disorders. In M.B. Sperling & W.H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 299-312). New York: Guilford Press. Parker, G., Tipling, H., & Brown, L.B. (1979). A parental bonding instrument. British Journal of Medical Psychology, 52, 1-10. Patrick, M . Hobson, R.P., Castle, D., Howard, R., & Maughan, B. (1994). Personality disorder and the mental representation of early social experience. Development & Psychopathology, 6. 375-388. Paulhus, D.L. (1984). Two-component models of socially-desirable responding. Journal of Personality and Social Psychology, 46, 598-609. Paulhus, D.L. (1988). Assessing self-deception and impression management in self-reports: The Balanced Inventory of Desirable Responding. Unpublished manual, University of British Columbia, Vancouver, Canada. Paulhus, D.L. (1991). Measurement and control of response bias. In J.P. Robinson, P.R. Shaver, & L.S., Wrightsman, (Eds.), Measures of Personality and Social Psychological Attitudes (pp. 17-59). San Diego, C A : Academic Press, Inc. Perris, C , Jacobsson, L. , Lindstrom, H., von Knorring, L . , & Perris, H. (1980). Development of a new inventory for assessing memories of parental rearing behaviour. Acta Psychiatra Scandanavia, 61. 265-274. 182 Peters, J.J. (1976). Children who are victims of sexual assault and the psychology of offenders. American Journal of Psychotherapy, 30, 398-421. Pfohl, B., Coryell, W., Zimmerman, M . , & Stangl, D. (1986). DSM-III personality disorders: Diagnostic overlap and internal consistency of individual DSM-III criteria. Comprehensive Psychiatry. 27 (1), 21-34. Pistole, M . C . (1989). Attachment in adult romantic relationships: Styles of conflict resolution and relationship satisfaction. Journal of Social & Personal Relationships, 6, 505-510. Rholes, W.S., Simpson, J.A., Blakely, B.S., Lanigan, L. , & Allen, E.A. (1997). Adult attachment styles, the desire to have children, and working models of parenthood. Journal of Personality, 65, 2, 357-385. Rice, M.E . , Harris, G.T., & Cormier, C A . (1992). An evaluation of a maximum security therapeutic community for psychopaths and other mentally disordered offenders. Law & Human Behaviour. 16. 4, 399-412. Ricks, M.H. (1985). The social transmission of parental behavior: Attachment across generations. Monographs of the Society for Research in Child Development, 50 (1-5), 211-227. Rivera, B., & Widom, C. (1990). Childhood victimization and violent offending. Violence and Victims. 5 d) . 19-35. Robertson, R.G., Bankier, R.G., & Schwartz, L. (1987). The female offender: A Canadian study. Canadian Journal of Psychiatry, 32 (9), 749-755. Rosenberg, M . (1965). Society and the Adolescent Self-image. Princeton, NJ: Princeton University Press. Rosenblatt, E. , & Greenland, C. (1974). Female crimes of violence. Canadian Journal of Criminology & Corrections, 16 (2), 173-180. Ross, M.W., Campbell, R.L., & Clayter, J.R. (1982). New inventory of measurement of parental rearing patterns: An English form of the E M B U . Acta Psychiatrica Scandinavica. 66, 499-507. Rothbard, J C , & Shaver, P.R. (1994). Continuity of attachment across the lifespan. In M.B. Sperling & W.H. Berman (Eds.), Attachment in Adults: Clinical and Developmental Perspectives (pp. 31-71). New York: Guilford Press. Rutter, M . , Quinton, D., & Hill, J. (1990). Adult outcome of institution-reared children: Males and females compared. In L.N. Robins & M . Rutter (Eds.), Straight and Devious Pathways from Childhood to Adulthood (pp. 135-157). Cambridge, England: Cambridge University Press. Sack, A., Sperling, M.B., Fagen, G. , & Foelsch, P. (1996). Attachment style, history, and behavioral contrasts for a borderline and normal sample. Journal of Personality Disorders, 10, 88-102. Salekin, R.T., Rogers, R., & Sewell, K.W. (1997). Construct validity of psychopathy in a female offender sample: A multitrait- multimethod evaluation. Journal of Abnormal Psychology, 106, 576-585. 183 Salekin, R. T., Rogers, R., Ustad, K . L . , & Sewell, K.W. (1998). Psychopathy and recidivism among female inmates. Law & Human Behavior. 22 (1), 109-128. Salter, M.D. (1940). An evaluation of adjustment based on the concept of security. University of Toronto Studies, Child Development Series, No. 18. Toronto, Canada: University of Toronto Press. Scharfe, E. , & Bartholomew, K. (1994). Reliability and stability of attachment patterns. Personal Relationships. 1. 23-43. Scher, A. , & Mayseless, O. (1997). Changes in negative emotionality in infancy: The role of mother's attachment concerns. British Journal of Developmental Psychology. 15. 311-321. Senchak, M . & Leonard, K . E . (1992). Stability of spouses' attachment styles over the first year of marriage. Unpublished raw data. Serin, R.C. (1992). The clinical application of the PCL-R in a prison population. Journal of Clinical Psychology. 48. 5, 637-642. Serin, R.C. (1996). Violent recidivism in criminal psychopaths. Law & Human Behavior. 20. 2, 207-217. Shaver, P.R., & Hazan, C. (1993). Adult romantic attachment: Theory and evidence. In D. Perlman & W. Jones (Eds.), Advances in personal relationships (Vol. 4, pp.29-70). London: Jessica Kingsley Publishers. Shaw, M . , Rodgers, K , Blanchette, J., Hattem, T., Thomas, L.S., & Tamarack, L . (1991). Survey of Federally Sentenced Women: Report to the Task Force on Federally Sentenced Women (User Report No. 1991-4). Ottawa: Ministry of the Solicitor General. Sheldon, A.E.R., & West, M . (1990). Attachment pathology and low social skills in avoidant personality disorder: An exploratory study. Canadian Journal of Psychiatry. 35. 596-599. Short, R.H., & Hess, G.C. (1995). Fetal Alcohol Syndrome: Characteristics and remedial implications. Developmental Disabilities Bulletin. 23. 1, 12-29. Silbert, M.H. & Pines, A . M . (1981). Sexual child abuse as an antecedent to prostitution. Child Abuse and Neglect. 5. 407-411. Simmons, D. (1992). Gender issues and borderline personality disorder: Why do females dominate the diagnosis? Archives of Psychiatric Nursing. 6 (4), 219-223. Simpson, J., & Rholes, W., & Nelligan, J. (1992) Support seeking and support giving within couples in an anxiety provoking situation: The role of attachment styles. Journal of Personality and Social Psychology. 62 (3). 434-446. Simpson, S.S. (1989). Feminist theory, crime, and justice. Criminology, 27, 605-631. Sirkia, T.D. (2000). Childhood maltreatment experiences and adult attachment styles in male offenders. Unpublished Master's thesis, University of British Columbia. 184 Skodol, A . E . , Oldham, J.M., & Rosnick, L. (1991). Diagnosis of DSM-II-R personality disorders: a comparison of two structured interviews. International Journal of Methods in Psychiatric Research, 1, 13-26. Snell, T.L. , & Morton, D.C. (1991). Women in Prison (Rep. No. NCJ- 145321). Washington, DC: U.S. Department of Justice. Sobell, L . C . , Sobell, M.B. , Riley, D.M. , Schuller, R., Pavan, D.S., Cancilla, A. , Klajner, F., & Leo, G.I. (1988). The reliability of alcohol abusers' self-reports of drinking and life events that occurred in the distant past. Journal of Studies on Alcohol, 49, 3, 225-232. Sommers, I., & Baskin, D. (1993) Situational context of violent female offending. Journal of Research in Crime & Delinquency, 30 (2), 136-162. Sommer, R., Barnes, G.E., & Murray, R.P. (1994) Alcohol consumption, alcohol abuse, personality and female perpetrated spousal abuse. Personality and Individual Differences, 13 (12), 1315-1323. Sperling, M.B. , & Berman, W.H. (1991). An attachment classification of desperate love. Journal of Personality Assessment, 56 (1), 45-55. Spitzer, R. (1983). Psychiatric diagnosis: are clinicians still necessary? Compr Psychiatry, 24, 399-411. Sroufe, L .A. (1988). The role of infant-caregiver attachment in development. In J. Belsky & T. Nezworski (Eds.), Clinical Implications of Attachment, (pp. 18-38). Hillsdale, NJ: Erlbaum. Sroufe, L .A. , & Fleeson, J. (1986). Attachment and the construction of relationships. In W. Hartup & Z. Rubin (Eds.), Relationships and development (pp. 239-252). Hillsdale, NJ: Erlbaum. Stalker, C , & Davies, F. (1995). Attachment organization and adaptation in sexually abused women. Canadian Journal of Psychiatry, 40, 234-240. Statistics Canada. (1998). A statistical profile. (Catalogue No. 85-224-XPE, pp. 21-24). Ottawa, Ontario: Canada. Steele, M . , Steele, H., & Fonagy, P. (1993). Associations among attachment classifications of mothers, fathers, and their infants, Child Development, 67, 541-555. Strachan, C.E. (1995). The assessment of psychopathy in female offenders. Unpublished Doctoral Dissertation. University of British Columbia, Vancouver, B.C., Canada. Straus, M . (1978). Wife beating: How common and why? Victimology, 2, (3-suppl-4), 443-458. Straus, M.A. & Stewart, J.H. (1999). Corporal punishment by American parents: National data on prevalence, chronicity, severity, and duration, in relation to child and family characteristics. Clinical Child and Family Psychology Review, 2 (2), 55-70. Sultan, F., & Long, G. (1988). Treatment of the sexually/physically abuse female inmate: Evaluation of an intensive short-term intervention program. Journal of Offender Counselling, Services & Rehabilitation. 12 (2). 131-143. 185 Sultan, F.E. , Long, G.T., Kiefer, S.A., Schrum, D., Selby, J.W., & Calhoun, L . G . (1984). The female offender's adjustment to prison life: A comparison of psychodidactic and traditional supportive approaches to treatment. Journal of Offender Counseling. Services. & Rehabilitation, 9 (1 & 2), 49-56. Tien, G., Lamb, D., Bond, L., Gillstrom, B., & Paris, F. (1993). Report on the needs assessment of women at Burnabv Correctional Centre for Women. B.C. Institute on Family Violence. Trevethan, S. (1999). Women in federal and provincial-territorial correctional facilities. CSC Forum. 11 (3). 9-12 Turnbull, S.D. (1997). Personality pathology and adult attachment in female offenders (Doctoral Dissertation, Simon Fraser University, 1997). Dissertation Abstracts International, 54 (4-B), 2170. Van Buskirk, S.S. & Cole, C F . (1983). Characteristics of eight women seeking therapy for the effects of incest. Psychotherapy: Theory, Research, & Practice, 20 (4), 503-514. van Ijzendoorn, M.H. (1995). Adult attachment representations, parental responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin. 117 (3), 387-403. van Ijzendoorn, M.H. , & Bakermans-Kranenburg, M.J. (1996). Attachment representations in mothers, fathers, adolescents, and clinical groups: A meta-analytic search for normative data. Journal of Consulting & Clinical Psychology, 64, 8-21. van Ijzendoorn, M . H . , Feldbrugge, J .T.T.M., Derks, F.C.H. , de Ruiter, C , Verhagen, M.F .M. , Philipse, M.W.G. , van der Staak, C.P.F., & Riksen, W.J.M.A. (1997). Attachment representations of personality-disordered criminal offenders. American Journal of Orthopsychiatry. 6 (3), 449-459. Vaughn, B.E., & Bost, K.K. (1999). Attachment and temperament: Redundant, independent, or interacting influences on interpersonal adaptation and personality development? In J. Cassidy & P.R. Shaver (Eds). Handbook of attachment: Theory, research, & clinical applications (pp. 198-225). New York: Guilford Press. Vitale, J.E., & Newman, J.P. (2001). Using the Psychopathy Checklist-Revised with female samples: Reliability, validity, and implications for clinical utility. Clinical Psychology: Science and Practice. 6(1). 117-132. Waganaar, W.A. (1986). My memory. A study of autobiographical memory over six years. Cognitive Psychology, 18, 225-252 Waters, E. (1978). The reliability and stability of individual differences in infant-mother attachment. Child Development, 49, 483-494. West, M . , & Sheldon-Keller, A. (1994). Psychotherapy strategies for insecure attachment and personality disorder. In M.B. Sperling & W.H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives (pp. 313-330). New York: Guilford Press. West, M . , Sheldon, A. , & Reiffer, L. (1987). An approach to the delineation of adult attachment: Scale development and reliability. The Journal of Nervous and Mental Disease, 175 (12), 738-747. 186 Widiger, T.A., Rogers, J.H. (1989). Prevalence and comorbidity of personality disorders. Psychiatric Annals. 19, 132-136. Widom, C S . (1989). Does violence beget violence? A critical examination of the literature. Psychological Bulletin, 106, 3-28. Wilbanks, W. (1983). Female homicide offenders in the U.S. International Journal of Women's Studies. 6 (4). 302-310. Wissow, L.S. (2001). Ethnicity, income, and parenting contexts of physical punishment in a national sample of families with young children. Child Maltreatment: Journal of the American Professional Society on the Abuse of Children, 6 (2), 118-129. Wyatt, G.E. , & Mickey, M.R. (1987). Ameliorating the effects of child sexual abuse: An exploratory study of support by parents and others. Journal of Interpersonal Violence, 2, 403-414. Zanarini, M . , Frankenburg, F., Chauncey, D. & Gunderson, J. (1987). The Diagnostic Interview for Personality Disorders: Interrater and test-retest reliability. Comprehensive Psychiatry, 28 (6), 467-480. Zeanah, C , Benoit, D., Barton, M , Regan, C , Hirshberg, L. , & Lipsett, L. (1993). Representations of attachment in mothers and their one-year-old infants. Journal of the American Academy of Child & Adolescent Psychiatry, 32, 278-286. Zimmerman, M . , & Coryell, W. (1989). DSM-II personality disorder diagnoses in a non-patient sample: Demographic correlates and comorbidity. Archives of General Psychiatry, 46. 682-689. 187 Appendix 1 Background Information Questionnaire Title: Female Offenders: Attachment Styles, Personality Characteristics, and Perceived Parental Bonding Instructions: Please answer the following background information questions: 1. Age: 2. Marital Status: Single Separated Divorced Common-Law Married Widowed 3. Are you working at BCCW? Yes No 4. If Yes: What is your occupation here? 5. Are you going to school at BCCW? Yes No 6. What grade did you get to in regular school? Through Upgrading? 7. What is your most common occupation outside of BCCW? 8. What is the longest period of time you have spent in the same job? 9. When did you last work outside of BCCW? 10. What was your approximate income in the year before your current arrest? $0 - $ 10, 000 $ 10, 001 - $20,000 $20,001 - $30,000 $30,001 - 40,000_ $40,001 - $50,000 $50,001 - $ 100,000 Over $ 100,000 11. How many siblings (brothers and sisters) do you have? 12. How many of these siblings were raised in the same home as you were? 13. Were y O U : Raised by your birth mother and father Raised by your birth mother Raised by your birth father Raised by your birth mother and step-father Raised by your birth father and step-mother Raised by grandparents Raised by other relatives Raised by adoptive parents Raised in foster care (Place a check beside one or more that best describe your situation) 14. Are you currently in a romantic relationship? Yes No 15. Do you have romantic relationships: Only with men Only with women With both men and women 16. How many children do you have? 17. If you have children: When you are at home, do your children: Live with you Live with their father Live with a relative Live in foster care Live with adoptive parents (If your children are not all living in the same place, please circle more than one) 18. Current legal StatUS: Remand Provincial Sentence Federal Sentence Immigration Hold 19. Most serious current charges: 1) 2) 3) 20. How long have you been at BCCW for these charges? 21. If sentenced: a) How long is your sentence? b) How much of it have you served? 22. Number of previous arrests: 1 2 3 4 5 or more 23. Number of previous convictions: 1 2 3 4 5 or more 24. How old were you at your first arrest? 188 25. What is the longest previous sentence you have received? 26. What is yOUr ethnic background? Caucasian Black First Nations Asian Hispanic Indo-Asian 27. Which of the following have you used? Please circle the letter beside each substance you have tried. a) Alcohol b) Marajauna (Pot) or Hashish c) Speed, amphetamines, or other stimulants d) Heroin, morphine, Percodan, or other opioid e) Methadone f) Ritalin and Talwin (Rs & Ts) g) Cocaine or crack h) Valium, Quaaludes, or other sedatives i) MDA, Extasy, or Poppers j) Other Please List: Now please go back to the list and UNDERLINE those substances you feel you have been addicted to or dependant on at some point in time. Finally, please place a CHECKMARK beside those substances that you used in the six months prior to your last arrest. 189 Appendix 2 Participant #: BCCW-2001-RELATIONSHIP QUESTIONNAIRE P L E A S E R E A D THE DIRECTIONS! 1. Following are descriptions of four general relationship styles that people often report. Please read each description and C I R C L E the letter corresponding to the style that best describes you or is closest to the way you generally are in your close f r iendships. A. It is easy for me to become emotionally close to others. I am comfortable depending on them and having them depend on me. I don't worry about being alone or having others not accept me. B. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others. C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. D. I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. 190 Page 2 2. Please rate each of the following relationship styles according to the extent to which you think each description corresponds to your general f r iendship style. A. It is easy for me to become emotionally close to others. I am comfortable depending on them and having them depend on me. I don't worry about being alone or having others not accept me. B. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others. C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. D. I am comfortable without close emotional relationships, It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. Not at all like me Somewhat like me Very much like me Style A . Style B. Style C. Style D. 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 191 RELATIONSHIP QUESTIONNAIRE P L E A S E R E A D THE DIRECTIONS! The following are descript ions of four general relationship styles that people often report. P l e a s e read each descript ion and C I R C L E the letter corresponding to the style that best descr ibes you or is closest to the way you general ly are in your romantic relat ionships. A . It is easy for me to become emotionally c lose to others. I a m comfortable depending on them and having them depend on me. I don't worry about being a lone or having others not accept me. B. I a m uncomfortable getting c lose to others. I want emotionally c lose relat ionships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I al low myself to become too c lose to others. C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as c lose as I would like. I am uncomfortable being without c lose relat ionships, but I somet imes worry that others don't value me as much as I value them. D. I am comfortable without c lose emotional relat ionships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. 192 Page 2 2. P l e a s e rate each of the following relationship styles according to the extent to which you think each descript ion corresponds to your romantic relationship style. A . It is easy for me to become emotionally c lose to others. I a m comfortable depending on them and having them depend on me. I don't worry about being a lone or having others not accept me. B. I a m uncomfortable getting c lose to others. I want emotional ly c lose relat ionships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I al low myself to become too c lose to others. C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as c lose as I would like. I am uncomfortable being without c lose relat ionships, but I somet imes worry that others don't va lue me a s much as I value them. D. I a m comfortable without c lose emotional relat ionships, It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. Not at all like me Somewhat like me Very much like me Style A . Style B. Style C. Style D. 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 193 Appendix 3 ECRI The following statements concern how you feel in romantic relationships. We are interested in how you generally experience relationships, not just in what is happening in a current relationship. Respond to each statement by indicating how much you agree or disagree with it. 1 2 3 4 5 6 7 Disagree Strongly Neutral/ Mixed Agree Strongly I. 1 prefer not to show a partner how I feel deep down. 2.1 worry about being abandoned. 3.1 am very comfortable being close to romantic partners. 4.1 worry a lot about my relationships. 5. Just when my partner starts to get close to me I find myself pulling away. 6.1 worry that romantic partners won't care about me as much as I care about them. 7.1 get uncomfortable when a romantic partner wants to be very close. 8.1 worry a fair amount about losing my partner. 9.1 don't feel comfortable opening up to romantic partners. 10.1 often wish that my partner's feelings for me were as strong as my feelings for him/her. II. 1 want to get close to my partner, but I keep pulling back. 12.1 often want to merge completely with romantic partners, and this sometimes scares them away. 13.1 am nervous when partners get too close to me. 14.1 worry about being alone. 15.1 feel comfortable sharing my private thoughts and feelings with my partner. 16. My desire to be very close sometimes scares people away. 17.1 try to avoid getting too close to my partner. 18.1 need a lot of reassurance that I am loved by my partner. 19.1 find it relatively easy to get close to my partner. 194 20. Sometimes I feel that I force my partners to show more feeling, more commitment. 21.1 find it difficult to allow myself to depend on romantic partners. 22.1 do not often worry about being abandoned. 23.1 prefer not to be too close to romantic partners. 24. If I can't get my partner to show interest in me, I get upset or angry. 25.1 tell my partner just about everything. 26.1 find that my partner(s) don't want to get as close as I would like. 27.1 usually discuss my problems and concerns with my partner. 28. When I'm not involved in a relationship, I feel somewhat anxious and insecure. 29.1 feel comfortable depending on romantic partners. 30.1 get frustrated when my partner is not around as much as I would like. 31.1 don't mind asking romantic partners for comfort, advice, or help. 32.1 get frustrated if romantic partners are not available when I need them. 33. It helps to turn to my romantic partner in times of need. 34. When romantic partners disapprove of me, I feel really bad about myself. 35.1 turn to my partner for many things, including comfort and reassurance. 36.1 resent it when my partner spends time away from me. Brennan, Clark, & Shaver (1998) 195 Appendix 4 Recollections of Early Childrearing (EMBU) Did your parents remain together during your childhood? Yes No . If "no," please indicate your age at the time of separation: years old. Who did you then live with? Mother Father Other (specify) . Beside each statement, please write in the number of the response listed below (1 - 4) that best describes how often the experience happened to you with your mother (or female guardian) and father (or male guardian) when you were growing up. If you had more than one mother/father figure, please answer for the persons who you feel played the most important role in your upbringing. I 1 2 3 I 4 I | never occurred occasionally occurred often occurred I always occurred | Father or Guardian Mother or Guardian 1. My parent showed with words and gestures that he/she liked me. 1 2 3 4 1 2 3 4 2. My parent refused to speak to me for a long time if I had done anything silly (stupid). 1 2 3 4 1 2 3 4 3. My parent punished me even for small offenses. 1 2 3 4 1 2 3 4 4. I think that my parent wished I had been different in some way. 1 2 3 4 1 2 3 4 5. If I had done something foolish, I could go to my parent and make everything right by asking for his/her forgiveness (apologize). 1 2 3 4 1 2 3 4 6. I felt that my parents liked my brother(s) and/or sister(s) more than he/she liked me. 1 2 3 4 1 2 3 4 7. My parent treated me unjustly (badly) and compared with how he/she treated my sister(s) and/or brother(s). 1 2 3 4 1 2 3 4 8. As a child I was physically punished or scolded in the presence of others. 1 2 3 4 1 2 3 4 9. If things went badly for me, I felt my parent tried to comfort and encourage me. 1 2 3 4 1 2 3 4 1— 1 I 2 3 I | never occurred | occasionally occurred often occurred | always occurred | Father or Guardian Mother or Guardian 10. My parent gave me more corporal (physical) punishment than I deserved. 1 2 3 4 1 2 3 4 11. My parent would get angry if I didn't help at home when I was asked to. 1 2 3 4 1 2 3 4 12. I felt that it was difficult to approach my parent. 1 2 3 4 1 2 3 4 13. My parent would narrate or say something about what I had said or done in front of others so that I felt ashamed. 1 2 3 4 1 2 3 4 14. My parent showed he/she was interested in my getting good marks. 1 2 3 4 1 2 3 4 15. If I had a difficult task in front of me, I felt support from my parent. 1 2 3 4 1 2 3 4 16. I was treated as a the "black sheep" or "scapegoat" of the family. 1 2 3 4 1 2 3 4 17. My parent wished I had been like somebody else. 1 2 3 4 1 2 3 4 18. I felt my parent thought it was my fault when he/she was unhappy. 1 2 3 4 1 2 3 4 19. My parent showed me that he/she was fond of me. 1 2 3 4 1 2 3 4 20. I think my parent respected my opinions. 1 2 3 4 1 2 3 4 21. I felt that my parent wanted to be with me. 1 2 3 4 1 2 3 4 22. I think my parent was mean and grudging toward me. 1 2 3 4 1 2 3 4 23. I think my parent tried to make my adolescence stimulating, interesting, and instructive (for instance, by giving me good books, arranging for me to go to camp, taking me to clubs). 1 2 3 4 1 2 3 4 24. My parent praised me. 1 2 3 4 1 2 3 4 197 1 1 I 2 3 I | never occurred I occasionally occurred often occurred | always occurred | Father or Guardian Mother or Guardian 25. I could seek comfort from my parent if I was sad. 1 2 3 4 1 2 3 4 26. I was punished by my parent without having done anything. 1 2 3 4 1 2 3 4 27. My parent allowed me to do the same things my friends did. 1 2 3 4 1 2 3 4 28. My parent said he/she did not approve of my behaviour at home. 1 2 3 4 1 2 3 4 29. My parent criticized me and told me how lazy and useless I was in front of others. 1 2 3 4 1 2 3 4 30. Of my sister(s) and brother(s), I was the one my parent blamed if anything happened. 1 2 3 4 1 2 3 4 31. My parent was abrupt with me. 1 2 3 4 1 2 3 4 32. My parent would punish me hard, even for trifles (little things). 1 2 3 4 1 2 3 4 33. My parent beat me for no reason. 1 2 3 4 1 2 3 4 34. My parent showed an interest in my own interests and hobbies. 1 2 3 4 1 2 3 4 35. My parent treated me in such a way that I felt ashamed. 1 2 3 4 1 2 3 4 36. My parent let my sister(s) and brother(s) have things that I was not allowed to have. 1 2 3 4 1 2 3 4 37. I was beaten by my parent. 1 2 3 4 1 2 3 4 38. I felt that warmth and tenderness existed between me and my parent. 1 2 3 4 1 2 3 4 39. My parent respected the fact that I had other opinions than had he/she. 1 2 3 4 1 2 3 4 198 never occurred | occasionally occurred | often occurred | always occurred Father or Guardian Mother or Guardian 40. My parent would be angry with me without letting me know why. 1 2 3 4 1 2 3 4 41. My parent let me go to bed without food. 1 2 3 4 1 2 3 4 42. I felt that my parent was proud when I succeeded in something I had undertaken. 1 2 3 4 1 2 3 4 43. My parent hugged me. 1 2 3 4 1 2 3 4 44. My parent emotionally abuse me. 1 2 3 4 1 2 3 4 45. My parent physically abused me. 1 2 3 4 1 2 3 4 46. My parent sexually abused me. 1 2 3 4 1 2 3 4 199 Appendix 5 Participant #: BCCW- 2001-Do you have or have you ever had children? If no, have you ever had a parental role- for example, with younger siblings or someone else's children? If no, please DO NOT fill out this questionnaire. Please answer these questions about your relationship(s) with your child(ren). If you no longer have contact with your children, please answer according to your perceptions and feelings at the time when you were in contact with your children. Every statement is this quest ionnaire is fol lowed by five numbers, as is shown in the example below. P l e a s e give a response (answer) to each statement indicating how much you agree or d isagree with each statement. There is an example of a statement given below: Example: 1. I like my family. 1 2 3 4 5 Meaning of the numbers is as fo l lows: 1= Totally D isagree 2= Disagree 3= Neutral 4= Agree 5= Strongly Agree P l e a s e put a circle around the one number after each statement that best descr ibes your situation. It is very important that you answer every statement in the quest ionnaire. 1. I am happy in my role as a parent. 1 2. There is little or nothing I wouldn't do for my child(ren) if it was necessary. 1 3. Caring for my child(ren) sometimes takes more time and energy than I have to give. 1 4. I sometimes worry whether I am doing enough for my children. 1 5. I feel close to my child(ren). 1 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 200 6. I enjoy spending time with my child(ren). 7. My child(ren) is (are) an important source of affection for me. 8. Having children gives me a more certain and optimistic view for the future. 9. The major source of stress in my life is my child(ren). 10. Having children leaves little time and flexibility in my life. 11. Having children has been a financial burden. 12. It is difficult to balance the different responsibilities because of my child(ren). 13. The behavior of my child(ren) is often embarrassing or stressful to me. 14. If I had to do it over again, I might decide not to have children. 15.1 feel overwhelmed by the responsibility of being a parent. 16. Having child(ren) has meant having too few choices and too little control over my life. 1 2 17.1 am satisfied as a parent. 18.1 find my child(ren) enjoyable. Are you currently in contact with your child(ren), or did you answe your past relationship(s)? Current Past 2 3 4 2 3 4 these questions based Appendix 6 BIDR Participant #: BCCW-2001-INSTRUCTIONS: Using the scale below as a guide, please circle the number beside each statement to indicate how true it is. 1 Not True I Somewhat True 1 Very True 1 1 1 2 3 4 5 6 7 I 1. My first impressions of people usually turn out to 1 be right. 2 3 4 5 6 7 2. It would be hard for me to break any of my bad 1 habits. 2 3 4 5 6 7 3. I don't care to know what other people really think of me. 1 2 3 4 5 6 7 4. I have not always been honest with myself. I 2 3 4 5 6 7 5. I always know why I like things. 1 2 3 4 5 6 7 6. When my emotions are aroused, it biases my thinking. 1 2 3 4 5 6 7 7. Once I've made up my mind, other people can seldom change my opinion. 1 2 3 4 5 6 7 8. I am not a safe driver when I exceed the speed limit. 1 2 3 4 5 6 7 9. I am fully in control of my own fate. 1 2 3 4 5 6 7 10. It's hard for me to shut off a disturbing thought. 1 2 3 4 5 6 7 11. I never regret my decisions. 1 2 3 4 5 6 7 12. I sometimes lose out on things because I can't make up my mind soon enough. 1 2 3 4 5 6 7 13. The reason I vote is because my vote can make a difference. 1 2 3 4 5 6 7 14. My parents were not always fair when they 1 2 3 4 5 6 7 202 punished me. | Not True | Somewhat True 1 Very True 1 1 1 2 3 4 5 6 7 | 15. I am a completely rational person. 16. I rarely appreciate criticism. 17. I am very confident of my judgments. 18. I have sometimes doubted my ability as a lover. 19. It's all right with me if some people happen to dislike me. 20. I don't always know the reasons why I do the things I do. 21. I sometimes tell lies if I have to. 22. I never cover up my mistakes. 23. There have been occasions when I have taken advantage of someone. 24. I never swear. 25. I sometimes try to get even rather than forgive and forget. 26. I always obey laws, even if I'm unlikely to get caught. 27. I have said something bad about a friend behind his/her back. 28. When I hear people talking privately, I avoid listening. 29. I have received too much change from a salesperson without telling him or her. 30. I always declare everything at customs. 31. When I was young I sometimes stole things. 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 203 1 Not True | Somewhat True 1 Very True 1 1 1 2 3 4 5 6 7 I 32. I have never dropped litter on the streets. 1 2 3 4 5 6 7 33. I sometimes drive faster than the speed limit. 1 2 3 4 5 6 7 34. I never read sexy books or magazines. 1 2 3 4 5 6 7 35. I have done things that I don't tell other people 1 about. 2 3 4 5 6 7 36. I have never taken things that don't belong to me. 2 3 4 5 6 7 37. I have taken sick-leave from work or school even though I wasn't really sick. [ 2 3 4 5 6 7 38. I have never damaged a library book or store merchandise without reporting it. I 2 3 4 5 6 7 39. I have some pretty awful habits. 1 2 3 4 5 6 7 40. I don't gossip about other people's business. 1 2 3 4 5 6 7 204 Appendix 7 Script for: Parental Bonding Interview (IF previously determined that there are no children and that the participant has never had a parenting role with a child, skip to the Relationship Characteristics Questionnaire). How many children do you have? How old are they? What are their first names? (if she doesn't want to answer, find another way to differentiate between them- maybe by numbers: 1st, 2nd, 3rd.) Background of Child(ren) - ** get this information for each child, as they may not have all been living in the same home at the same time with the same people- in your notes, identify which child is being referred to if they need to be discriminated between** Until you were incarcerated, were you the primary caregiver for your child? ** There may be times when they were, and times when they weren't. Try to establish a time-line for this. If it was on and off, at what ages were they the primary caregiver. And if they were not the primary caregiver, who was? And are they in contact with their child, or were they before being incarcerated? What did you do for a living? How much time did you spend away from home? (Day-to-day at work, and any work related travel). How much time did you spend with your child(ren) when you were home? (Evenings? Weekends? Holidays?) Did you have any 'special' things that you did with them/him/her? (Bedtime rituals, reading, etc.). When you did spend time with your child(ren) what did you do together? (Get several examples). Were there any other times when you were away from your child(ren) overnight? Other Caregiver(s): Did your children have other primary caregivers besides yourself? If no, skip the following questions. If yes, who was/were the other primary caregiver(s)? Did the other primary caregiver(s) work outside the home? If yes: What was the occupation(s) of the other primary caregiver(s)? How much time did the other primary caregiver(s) spend away from home (Day-to-day at work, and any work related travel). How much time did the other primary caregiver(s) spend with your child(ren) when home? (Evenings? Weekends? Holidays?) Do you know whether the other primary caregiver(s) have/had any 'special' things that they did/do with your child(ren)? (Bedtime rituals, reading together, etc.). When the other primary caregiver(s) did spend time with your child(ren) what did they do together? (Get several examples). Were there any instances when your child(ren) were separated from other primary caregiver(s) overnight? If yes - How old were they and how long were they away from their other primary caregiver(s)? How did you celebrate special occasions with your child(ren) (e.g., birthdays, etc.). Have you and your child(ren) moved a lot? (Get frequency of moves, when, distance, why move, etc.) If there were frequent moves ask - How did the child(ren) take to moving so often? If good experience, get example of why. If bad experience, get example of why. If you have school-aged children, have they changed schools a lot? (Get frequency of moves, when, why, etc.). Where is the father(s) of your child(ren)? Who is/are your child(ren) closest to? (get examples to substantiate claims) Were you affectionate with your child(ren)? (Get examples). 206 If your child(ren) was/were unhappy or upset what would they do? Did they come to you or another adult? What did you do to discipline your child(ren)? (Get ages, frequency, reason discipline needed, get examples to illustrate: pay special attention if physical discipline was involved). How often do you talk to your child(ren)? How often do you see your child(ren)? Do you have children that are eligible to participate in the mother/child program with you? Are they here? If not, why? And where are they now? Why? (taken away, given away, temporarily being watched by relatives or friends). Are you satisfied with the caretaking arrangement for your children while you are at BCCW? When you get out of BCCW, do you plan to have your children live with you? Why or why not? If no, where will they live? Do you presently have any contact with your children? If yes, describe. Do you/did you enjoy being a parent? Expand on this. Do you think that you are/were a good parent? Expand on this. During the time that you were taking care of your children, do you feel that you were there for them when they needed you- e.g., when they were sick, when they has important moments or milestones to share? What, if anything, would you do differently as a parent if you were given the chance? 207 Participant #: BCCW-2001-Appendix 8 For women who do not have children: (*NOTE: if woman does normally take care of boyfriend's children or other child on a regular basis- i.e., plays the role of primary caregiver, please have her still fill out the PSS and administer the Parental Bonding Interview. However, STILL ask these questions.) Have you ever given birth to a child? If yes, how many? How old were you? Where are they now? (i.e, adopted?) If no, is there any reason why you have not had any children? Do you like children? Would you like to have children one day? Why or why not? DO you have a relationship with any child(ren)? E.g., nieces, nephews, romantic partner's children. And/or have you ever worked or volunteered with children? Do you think that you would make a good parent? Why or why not? 208 Appendix 9 Script for: Peer Attachment Interview. Participant #: BCCW-2001-Part 1: Personal Information General: What is your age? How long have you been incarcerated? Do you work at a job at BCCW? What? Are you married? If no, are you involved in a romantic relationship? Compared to other people you know, how emotional are you? Why do you say that? How do you express your emotions? If you feel unhappy or upset about something, what are you likely to do? If necessary: Are you more likely to go to other people or do you tend to deal with it on own? What kinds of things do you tend to get most upset about these days? (If necessary give examples, e.g., school, relationships, work, etc.). How often do you cry? (If necessary give examples - once a day, every few days, etc.) What do you cry about? Do you cry more often alone of with others? If you cry with others, with whom do you cry? Part 2 : Friendships. General: About how many friends do you have? Of those, how many do you consider close friends? What does it mean to you to say someone is a close friend? Do your friends tend to be more male or female? If so, do you have a sense of why that might be? Has this changed over time? If so, why the change? Are there any differences between your male and female friendships? People often report that in opposite sex friendships, one person or the other becomes romantically interested. Has that ever been a problem for you? If so, how do you handle it? Specific Friend: Now I would like you to choose one of your close friends as a reference so I can ask you some questions about a particular friendship. (Get name). How long have you known each other? How much time do you spend together? What kinds of things do you do? Who organizes your get-togethers? How do you feel about that? Why do you think you've become good friends? Could you give me an example of things you've done together or experiences you've had together that have brought you to be good friends? What do you like about (name of friend)? What don't you like about (name of friend) ? 21 Do you ever talk to them about it? (what they dislike). Do you and ever have conflicts? Explore - what are conflicts about; how are they resolved, i f they are resolved? Do you ever feel angry with ? What do you do when angry? Have you ever had your feelings hurt by ? (Example). Have there been times when you and haven't talked to each other? Who is more involved in the relationship? Do you discuss personal matters with ? Are there things that you wouldn't talk about or that would be difficult to talk about? (Get examples). Why? What changes would you like to see in your friendship over time? 211 General Friends: How does your friendship with compare to your other close friends? (Probe especially for: closeness, time spent together, different groups of friends.) How often are your feelings hurt by friends? Examples. Have you ever had conflicts between friends and your romantic partners? (Explore - drop friends when involved; intimate others jealous offriends, etc). If applicable: If currently in a relationship, how often do you socialize with and without your partner? How much time do you spend alone (estimate per day)? What changes would you like to see in your friendships in general? (Probes: number offriends, amount of time with friends, quality offriendships?) When you meet new people, do you think they will like you? How confident are you about making new friends? Are you shy? What impression do you think you make on other people? What impression would you like to make? 212 Script for: Romantic Attachment Interview Part 3: Relationship History What is your sexual preference? How long have you been involved with your current partner? (If applicable). Have you had any (previous) serious relationships? If yes: I'd like you to give me a brief history of your serious romantic/sexual involvement's (how old each party was at the start, length of relationship). For all previous relationships, briefly describe major issues: How serious the relationship was; How long it took to get to that point; Reason for breakup; who initiated it, how did it feel, how long did it take to breakup, how long did it take to get over it, and time alone before next relationship During periods of non-involvement: Why not? Are/were you dating? Are/were you looking for a relationship? Was there any physical conflict in your past relationships? Explore. Do you see any patterns across your relationships? If no, to question regarding relationships (i.e., has never had a serious relationship) ask: Were you dating before becoming incarcerated? If yes, ask: Were you looking for a relationship? If no, why not? Have there been times when a relationship looked like it might get serious? 213 If yes, what happened? Is having a long term relationship at some time in the future important to you? Why or why not? Do you worry about not finding the right person? Why or why not? Part 4: Current Relationship How long have your know each other? When did you start dating? How quickly did you become serious? Have there been any major separations since you've been together? How did it feel when you were apart? Have there been any other involvement's since you've been together? How serious is your relationship? Are you sexually involved? Have you considered future plans? How much time do you spend together? What do you like about your relationship? What don't you like? Describe your partner? What do you like most about her/him? What don't you like about her/him? Do you talk to your partner about it? What do you think your partner likes most about you? What do you think your partner likes least about you? How does your current relationship compare with past ones? 214 Part 5: Communication and Support in Current Relationship How comfortable are you discussing personal matters with your partner? What are some of the topics of conversation you avoid with your partner because they're awkward to talk about or they lead to disagreements? (e.g., money, sex, family, etc.) How does your partner respond when you would like help or support? (what do they say/do?) How does your partner respond when you're hurt or sick? (what do they say/do?) What about emotional upset? What would your partner do? Examples. Do you ever feel your partner is not responsive enough or too responsive? Do you feel comfortable crying in front of your partner? If not, why? How does your partner respond? How well does your partner understand you? (Get example) Part 6: Love-worthiness and Trust in Current Relationship Have you ever felt rejected by your partner? Describe. Have you ever had your feelings hurt by your partner? Have you ever doubted that your partner loves or cares about you? How does your partner show they care about you? Have you ever felt neglected by your partner? Have there been situations when you felt your partner was not honest with you? Are you always honest with your partner? Do you say "I love you" to each other? How often? Does one say it more than the other? 215 Part 7: Conflict Resolution in Current Relationship How often do you have disagreements or arguments? What are they about? What happens when you disagree? Does it get resolved? How is it resolved? As necessary (if they're denying ever having conflicts): Do you ever have differences of opinion? About movies, music, etc.? Do you ever wonder if your partner disagrees with you, but doesn't say anything? How long do you stay angry? Who initiates the arguments and the resolution? If necessary: Could you give me an example of a typical conflict and describe how it tends to go? Have you ever felt afraid of your partner? Have your conflicts ever become physical, such as breaking or throwing things, pushing, or slapping? How often? Can you describe what happens? How does this compare with past relationships? How often are you and your partner mean to each other or critical? Has your partner ever done anything that you consider abusive? Have you done anything that they would consider abusive? Have the two of you had any issues about the amount of time you spend together? How do the two of you go about making decisions? Is it mutual? 216 Part 8: Physical Relationship with Current Partner How affectionate are the two of you within the relationship? Is one of you more so than the other? Is this ever an issue, in private or in public? Do you ever feel that your partner is not warm or affectionate enough? How do you feel about the sexual aspect of your relationship? How do you think you partner feels about it? Is sex more important to you or your partner? Explore. Has that changed over time? What do you do when your partner initiates sex and you don't feel like it? Vise Versa? How often does this happen? If necessary: Is your partner aware that you sometimes don't feel like having sex? What changes would you like to see in your sexual relationship? Part 9: Separations from Current Partner Have you and your partner ever been apart for any length of time? (e.g., for holidays, business trips, etc.) Explore. If not, how would you feel if it happened? Part 10: Mutuality in Current Relationship People in relationships commonly report that one partner seems more invested or committed to the relationship. Has this ever been the case in your relationship? If so, describe. Some people feel concerned about being too dependent in a relationship. Is this a concern for you? For your partner? 217 Have you or your partner ever felt jealous in the relationship? Describe. If your partner is jealous, how does it make you feel? If necessary: Are you ever possessive of your partner? Is your partner ever possessive of you? If yes, how do you feel about it? Part 11: Regrets and Breakups in the Current Relationship Have you ever thought about separating? When? Why? How difficult would it be for you to end this relationship? If necessary: Have you ever had any regrets or doubts about becoming involved with your partner? If appropriate: How much faith do you have that your relationship with last in the long term? If it doesn't, who would be most likely to initiate the breakup? Why? If necessary: If you and your partner ever did breakup, how difficult would it be for you? How difficult would it be for your partner? Part 12: General Evaluation of Current Relationship How would you like to see your relationship change over time? If necessary: Any changes in the way you relate to your partner? In the way your partner relates to you? If you could have an ideal relationship, how would it differ from your present relationship? How does this relationship compare with past relationships? 218 Part 13: General Social Relations What kinds of changes would you like to see in the way you relate to others? What kinds of changes would you like to see in the way others relate to you? Is there anything else about your social relationships that we haven't hit upon that important? How did you feel about this interview? Are these things that you've talked about with other people? Appendix 10 Title: Female Offenders: Attachment Styles, Personality Characteristics, and Perceived Parental Bonding Ethical Approva l Number: Part icipant Number: BCCW-2001-File Coder : Date: Cod ing Checkl is t Item Source: CSC Files 1. Index Offence(s): 2. Sen tence 3. Total # of Of fences: 4. Type(s) of Of fences: • 5. Instrumental or React ive V io lence 6. D iagnos is of Adult Psychopatho logy 7. Psychopa thy rating 8. Admiss ion of Responsibi l i ty for Offence 9. Admiss ion of Guilt for Of fence 10. Institutional Adjustment 11. Institutional Status (Protective Cus tody or Genera l Populat ion) 12. Subs tance A b u s e Prob lems Drugs? A lcoho l? 13. Suic ida l Ideation 13. Su ic ide Attempts 220 14. Homic idal Ideation (inside) 15. Sexua l Deviat ion 16. Treatment Type(s) 17. Treatment R e s p o n s e 18. Fami ly Info Corroborated 19. Relat ionship Info Corroborated 20. A g e of First Cr iminal Offence 21 . Parental Criminality 22. Spousa l Criminality Appendix 11 Participant Number: UBC ethics approval: B00-0485 Participant Feedback Form Dr. John C. Yuille and Gayla Swihart (Ph.D. Candidate), of the Department of Psychology at the University of British Columbia, are interested in the Family, Peer (friends), and Romantic Attachment Patterns and personal interaction styles of adult female inmates BCCW. Attachment style can be described as the usual way a person interacts with people that are close to him. The people that we are close to are usually our family members, friends and romantic partners. We are using Attachment Style interviews developed by Dr. Kim Bartholomew of Simon Fraser University. She thinks that there are four attachment patterns. These are: Secure, Fearful, Preoccupied and Dismissing. She thinks that a person can be one of these patterns, or can be a combination of several different patterns. She also thinks that people may be one pattern in one type of relationship (for example, family relationships) and another in a different type of relationship (for example, with friends or romantic partners). We are hoping that a better understanding of what an individual's personality and relationship style with their parents, peers, and romantic partners relates to their relationships with their children. Sometimes our attachment patterns can prevent us from functioning well in day-to-day life, and how we relate to others. It may also be a factor in how well inmates respond to treatment. If you feel the need to talk to someone about the things that have been discussed during this research please let the researcher know, and she will make sure that BCCW will provide you with assistance. We thank you for your valuable participation. If you are interested in the results of this study, please ask your researcher to make a note on your file, and a copy of the results will be forwarded to you when available. The interviewer may ask you if you would be willing to participate in a more in-depth interview on your relationships with family and friends. You would be given another $5.00 for your time, and the interview would take place in the near future. If you are interested, please print your name here, and someone may contact you. Please note that we are only re-interviewing a small number of women, so you may not be contacted. Yes, I am interested in participating in another interview: No, Please do not contact me: just circle if this applies to you- no need to write name. Did you answer the questions honestly? Did you feel comfortable answering the questions you were asked? If you have any comments about this interview, the interviewer, or any other comments, please write them here or on another sheet of paper if you wish to remain anonymous. We appreciate any feedback you may have. 222 Appendix 12 Interview Report Participant number: B C C W -2001 -Interviewer: Date & Time of Interview: Please characterize the interview. Were there any difficulties? Was the inmate upset at all? D id she answer questions easily? And anything else that is salient to you about the interview. Please include a quick synopsis of her circumstances, but please do not write her name. A lso, please write how you felt about the interview- did you feel upset, tired, drained, or otherwise? Please write this report within 24 hours of the interview. 223 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0091293/manifest

Comment

Related Items