’--_- I .I- I 1 111‘, 111,1 1‘1 .~. .. , )1 1L VI?" Y‘II, III [III-‘3 11‘“ 31111.43”); "H, I) ”III”. ‘v‘I‘ “‘1‘|I1“‘kul*“”““'l“ I I... ‘ 11‘“. .1 -. 1.311;. ‘H‘-‘1ol‘ , 1 .‘k 11""".11"‘111"‘|1“‘I‘:‘r" ""1'1‘1‘3‘ .N“. ‘ "II‘I"'-,‘ 11‘ “' 11"“: 51“]: 11111.11 7151'. '1""‘1‘1"“‘1“ ,1 1‘ 11111111151111.1-“131i‘11 {1&1 ‘1‘1‘1‘1‘1‘1‘““I“““‘I‘"“ “I. :‘ ,I‘A‘? ”1‘1"“ “2‘1; “M“ W“ ‘ "‘ ‘ 1‘1‘ I“1"“"‘a$‘. 11 1M1,“““:‘I‘.‘“‘..‘ ‘ 11‘1‘ . ‘.‘1‘:-1““.“1‘:"1‘.“4.."‘I 9‘1 1111:11111.1"“.‘1 - . 5 ' )1“ "‘1’" 1111! 1‘5‘111 1'1 .131 “‘11"“"‘ 1'11“ 111131111111111‘1“ 1 ‘1‘??? 1111.11 211111.111 111:1: ‘ ‘ 1‘ 1“ 111111.111 1 1.1111" 1: 11111511111911111 .1 5:1. ‘15 1- .1111: 1111111115.,” ‘11? 1.11111 1"1 "1‘11: 1 “‘1'1‘1'11‘11‘11‘11‘1 111111 ‘- 11133511‘11‘11’11‘1’111‘11111‘111111111113111111111.11111.1 I‘LLI‘ ‘5“."1'3‘ $511] E}.‘J;;1.,1‘I"‘1‘1“1“‘-"II“ '11'1‘I11‘11' i“1“‘$‘l‘1‘1‘1‘.“1."~“1'“ If} Tflf' «““1‘5‘1‘25{‘I““"“.II‘E‘1“ ‘ .1- 13 ’ i". 1111111“‘11“'~“‘1"1"11111111 111.1 ‘11:. TM 11 11231113151 1315111."? 1W" 1.113111% iIk-t I11. I i‘11“;1II‘.‘1‘,1'11:=;{1LI1 1531' ‘JI‘III‘E’II 131' ‘ ' ’ 1.1.‘1'1'11‘2111‘: 1131111‘11‘11111'111111111111111 1 1‘11‘1‘1“ .‘J 7 h " “‘ "‘2 3h" 15:33.... 1'11“. :13 ’ 55-2“ 5 ‘I .; ..“ 4:. ‘ .1511“ 4.1191113? 1““‘1‘ I1 $19311: 1111111; .‘115111 111111“ -1 ' . I {‘11 E3,‘ 3: ‘3‘11'4-1“. “ ' ' " “11111.1’1111111‘11111:‘1‘1‘1-‘11111111.111“ 1‘1“ I .1"11-J=1‘;".1'-1:1.‘I 11111251121151? 131-3111112114 11 I l's IIS‘.01;.II;‘I}III!' I 'EIoU J l' II'II'.'1;‘,}1;I',I '1 3.34151... ‘1151'5‘1V31313‘41?‘§;‘g11;,'1fi?7,13fi:. . 1’"1' 11.1‘.‘ "‘- ‘ ”14 ' 1“ x -:‘~“i .. 1 1135111 111‘“? .- : ' 3‘39: “ a“ if . ““3' 11 ~ . 11111. . '1 7 -‘ “‘1“ ('5 ‘ tl'n. 1. ‘ 2. ‘ 1-1211 112“»? 1:" g. l ' xi. "2; .7 .' “ ,_, ...... .. .557: :.un.> . .. ...-..'..-v .. w-.. .. .. ... - . .. 4"». “33:! "f C15..- .f- ' 4.". i1). “‘3— '4 3'3? .-t v 1': ”-1. .VA u. . .— ._v¢ -'. '1; ‘ ' ‘ 1I - -' ”HH‘I I‘ ‘- :7 - {at "'13“. I‘- 'Istlm“ - I -‘ 7 ' 15“" {I “ ”111m 1‘" l.'. “‘I 11 23.... ~.,’ '“fw ”-WA .11 .mp4- “-3. _ _.._-. . . .453 ow “-“JT _- THESlS I IGAN Z TTTTTLT 1 T TE UNIVERSITY Ll 71THTTTTTTT‘TTTT 9 1707 4208 This is to certify that the thesis entitled ADOLESCENT SURVIVORS OF CHILDHOOD SEXUAL ABUSE: THE MEDIATING ROLE OF ATTACHMENT STYLE AND COPING IN PSYCHOLOGICAL AND INTERPERSONAL FUNCTIONING presented by DEBORAH LYNN SHAPIRO has been accepted towards fulfillment of the requirements for MA- degree in ESXLHQLQEL or professor Date .0ecetnber_l,_199]_ 0.7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY . Michigan State University PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return. on or before date due. I DATE DUE DATE DUE DATE DUE AW 9&3923512004 1/98 ClClFCJDIIIDuthS—p.“ ADOLESCENT SURVIVORS OF CHILDHOOD SEXUAL ABUSE: THE MEDIATING ROLE OF ATTACHMENT STYLE AND COPING IN PSYCHOLOGICAL AND INTERPERSONAL FUNCTIONING By Deborah Lynn Shapiro A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1998 ABSTRACT ADOLESCENT SURVIVORS OF CHILDHOOD SEXUAL ABUSE: THE MEDIATING ROLE OF ATTACHMENT STYLE AND COPING IN PSYCHOLOGICAL AND INTERPERSONAL FUNCTIONING By Deborah Lynn Shapiro Childhood Sexual Abuse (C SA) can affect the psychological and interpersonal fimctioning of its victims, with the extent of disorder and level of fianctioning varying greatly within a population of survivors. In an attempt to explain the variability in functioning seen among survivors of CSA, this study examined attachment style and coping strategies as potential mediating variables. Eighty adolescent females, aged 14-16 years, answered questions regarding abuse history, attachment style, coping with an interpersonal stressor, depression and trauma symptomatology, and conflict with a best fiiend. Structural equation modeling analyses indicated that a secure attachment Style mediates the effects of family violence, leading to lower psychological distress and interpersonal conflict. Avoidant and cognitive coping strategies increase following CSA, increasing distress and conflict. In combination, a secure attachment style influences coping, with attachment style accounting for greater variance in psychological distress, and avoidant and cognitive coping accounting for more variance in interpersonal conflict. In addition, a multivariate analysis of variance indicated that at high levels of distress, survivors utilize more avoidant coping techniques than nonvictims, but there were no differences in coping strategies between survivors and nonvictims at low levels of distress. For my fiiends who are survivors, and who have inspired me iii ACKNOWLEDGMENTS The assistance and support of many people contributed to my ability to successfirlly finish this project. First, I would like to thank my committee members, Drs. Bogat and Reischl, for their guidance and contributions along the way. A Special thanks also goes to Dr. Neal Schmitt, for his patience and assistance with the structural equation modeling analyses. I am deeply indebted to my advisor, Dr. Alytia Levendosky, for her enormous support, caring, warmth, and for contributing to my grth as a researcher and clinician. My graduate school experience has been enriched by working with her, and I look forward to continuing to learn from her. The support I received from my family and friends throughout the progress of this project was a vital part of its completion. My parents and brother have continued to provide me with never-ending support and encouragement, and have never stopped believing in me. To the friends who provided me with a perspective beyond graduate school, and to those who were experiencing it alongside me - thank you all. Deb, without you I wouldn’t have made it this far, and you deserve a special thank you. I additionally owe gratitude to everyone in my life for tolerating me through this process, understanding the emotional toll it occasionally took, and for not taking it personally. There are a number of people working in the community that assisted me in making contact with their clients and supported my research efforts. I will not name them iv individually, to protect the confidentiality of the participants, but I do thank them for their help. I am especially gratefiil to the adolescent girls who participated in this project, particularly to those who shared the personal and painful information Of their abuse. I admire their strength and honesty, and learned something from each of them. When my strength to continue wavered, I was inspired by their continued strength and amazing ability to persevere. TABLE OF CONTENTS LIST OF TABLES ........................................................................................................ viii LIST OF FIGURES ........................................................................................................ ix INTRODUCTION .......................................................................................................... l CSA and Psychopathology: Theoretical Explanations .......................................... 4 Developmental Psychopathology ............................................................... 5 Attachment Theory ................................................................................... 7 CSA and Psychopathology: Empirical Evidence .................................................. 12 Coping ............................................................................................................... 17 Coping Theory ........................................................................................ 17 Coping in Adolescence ............................................................................ 20 Coping with CSA .................................................................................... 24 Summary ............................................................................................................ 29 HYPOTHESES ............................................................................................................. 34 METHOD ...................................................................................................................... 36 Participants ......................................................................................................... 36 Measures ............................................................................................................ 39 Abuse History ......................................................................................... 39 Coping Strategies .................................................................................... 40 Internalized Model of Relationships ......................................................... 41 Interpersonal Functioning ........................................................................ 42 Psychological Functioning ....................................................................... 43 Demographics ......................................................................................... 44 Procedures ......................................................................................................... 44 RESULTS ..................................................................................................................... 46 Attachment as a Mediator ................................................................................... S3 Coping as 3 Mediator .......................................................................................... 53 Attachment and Coping as Mediators .................................................................. 55 Coping Strategies of CSA Survivors ................................................................... 57 DISCUSSION ............................................................................................................... 61 Attachment as a Mediator ................................................................................... 61 Coping as a Mediator ......................................................................................... 64 Attachment and Coping as Mediators ................................................................. 67 Coping Strategies of CSA Survivors ................................................................... 71 Limitations ......................................................................................................... 74 Future Directions ................................................................................................ 78 FIGURES ...................................................................................................................... 82 APPENDIX A ............................................................................................................... 87 APPENDIX B ................................................................................................................ 90 APPENDIX C ................................................................................................................ 93 APPENDIX D ............................................................................................................... 95 APPENDIX E .............................................................................................................. 100 REFERENCES ............................................................................................................ 103 vii LIST OF TABLES Table 1 - COPE Principal Component Analysis ............................................................... 47 Table 2 - Correlations of Observed Variables .................................................................. 52 Table 3 - Differences Between CSA Survivors and Non-Victims .................................... 58 Table 4 - Interaction of CSA, CD1, and TSC-40 on Avoidant Coping ............................. 6O viii LIST OF FIGURES Figure 1 - Structural Model Incorporating all Hypothesized Pathways ............................ 30 Figure 2 - Attachment Style as a Mediator ...................................................................... 82 Figure 3 - Avoidant Coping as a Mediator ...................................................................... 83 Figure 4 - Cognitive Coping as a Mediator ..................................................................... 84 Figure 5 - Attachment and Avoidant Coping as Mediators .............................................. 85 Figure 6 - Attachment and Cognitive Coping as Mediators ............................................. 86 Figure 7- Rate Of Reported CSA Experiences ................................................................ 76 ix INTRODUCTION Estimates of the prevalence of childhood sexual abuse (CSA) range fi'om 15%- 33% of females and 13%-16% of males in the general population (Polusny & Follette, 1995). Not only does the incidence vary between males and females, but many factors surrounding the abuse experience, individual reactions to abuse, and subsequent firnctioning often do as well. Therefore, most research studies in this area focus on one gender, primarily females. The first national survey of adults conceming CSA found that 27% of the women polled reported experiencing some form of sexual abuse in childhood (Finkelhor, Hotaling, Lewis, & Smith, 1990). Furthermore, for girls, the median age at onset was 9.6 years; 98% of the abuse was perpetrated by a male; they were more likely to be abused by a family member; and the majority of respondents reported experiencing only one incident of abuse. There are a variety of psychological, interpersonal, and social dysfiinctions associated with CSA, indicating its global impact on individual life functioning. The extent of disorder and level of functioning vary greatly within a population of CSA survivors, with consequences ranging from mild emotional or behavioral problems, such as low self-esteem and poor school performance, to severe psychopathology, such as debilitating depression and anxiety (Browne & Finkelhor, 1986). The search for a causal relationship between CSA and psychopathology has produced inconclusive findings, with a variety of factors (such as family dysfirnction, abuse characteristics, and coping) implicated as contributing to a survivor’s prognosis. The fact that many survivors are capable of functioning adaptively provides evidence that internal mechanisms do exist for dealing with severe childhood trauma and avoiding severe consequences, and learning how these mechanisms firnction may assist CSA survivors in their recovery process. Until recently, CSA research lacked a strong theoretical base, instead focusing on investigating correlations between abuse history and mental disorders with no theoretical explanation as to how or why these sequelae develop. A number of theories have now been proposed to explain this relation, yet many questions still exist concerning the underlying mechanisms involved. This study utilized theories of attachment and coping, within a developmental psychopathology fiamework, to study the influence of CSA on subsequent firnctioning in adolescents. One of the primary methodological flaws in much of the CSA literature is the retrospective nature of the studies. Most research focuses on adult survivors, and is thus dependent on the questionable accuracy of self-reports and recollection of childhood trauma. By focusing on adolescents, this study eliminated the potential for inaccurate memories due to time lapse, as participants were in the early stages of the recovery process, or possibly even experiencing ongoing abuse. The present coping strategies utilized by the individuals, as well as their current psychological and interpersonal functioning, were therefore a more accurate reflection of the immediate or short-term impact of the abuse. While research findings continue to support the idea that psychological and interpersonal functioning are affected by childhood sexual abuse, the wide variability of functioning seen in survivors cannot yet be firlly explained. This study focused on mediating models to explain this relation, incorporating the roles of attachment style and COping strategies in the psychological and interpersonal firnctioning of adolescents who have experienced sexual abuse. Because of the many unique developmental changes occurring during adolescence, a time in which one faces many challenges related to physical, social, and interpersonal development, this age group is of particular interest in this area of research. How one adjusts to the stress of this time is influenced by earlier experiences, and will affect subsequent development. CSA can produce a range of devastating immediate effects, including depression, anxiety, self-destructive behavior, and many other related symptoms, which may be aggravated by the already stressful time of adolescence. By examining the abuse history, attachment style, coping strategies, psychological functioning, and interpersonal functioning of adolescents, this study attempted to increase the understanding of the initial effects and potential mediating factors of childhood sexual abuse. Through self-report, specific details of the abuse experience and attachment style were assessed. Coping was measured in the context of a recent interpersonal conflict reported by the adolescent. Psychological functioning was measured through self-reports of psychological symptomatology, and interpersonal firnctioning was based on a measure of the conflict resolution strategies used in a best fiiend relationship. C SILand Psychopathology: Theoretical Explanations An abundance of research has been conducted to study the connection between childhood sexual abuse and psychopathology, but the product of this effort is primarily a collection of mental disorders that have been found to correlate with previous abuse. Until recently, the majority of this research was lacking a strong theoretical foundation that explained how the trauma of CSA led to the development of these disorders. However, researchers have now recognized the need for theory-based models in the study of CSA, producing a diverse collection of theories to explain the connection between sexual abuse and the development of psychopathology. Among the more prominent theories that have emerged are the traumagenic model (F inkelhor & Browne, 1985), post- traumatic stress disorder (e.g. Herman, 1992; Rowan & Foy, 1993 ), psychoanalytic theory (e. g. Davies & Frawley, 1994; Price, 1993), cognitive theories (e. g. Janoff-Bulman, 1992), biological theories (e. g. van der Kolk, 1987) and attachment theory within a developmental psychopathology framework (e. g. Alexander, 1992; Main, 1996; Cole & Putnam, 1992). Certain models provide valuable information on symptomatology related to CSA, but do not incorporate developmental factors into their models. Alternatively, developmental theories, such as psychoanalytic and attachment theories, encompass the processes involved in psychological and social development for both survivors and nonvictims, and can therefore examine how development differs for someone experiencing childhood trauma. Developmental Psychopathology. Developmental psychopathology “examines the evolution of psychological disturbance in the context of development” (Cole & Putnam, 1992; p. 174). By applying a developmental psychopathology perspective to the study of childhood sexual abuse, one can understand how certain sequelae appear at difi‘erent points in development, how developmental factors afi‘ect outcomes, and how events during childhood affect later adjustment (Cole & Putnam, 1992). This model can be used to account for the potential developmental differences found in CSA survivors, while other models and theories Often do not. The development of self is a central factor in understanding the relation between CSA and subsequent psychopathology. The sense of self develops fi'om the selfs differentiation fi'om, and interactions with, significant others. The influence on self- development can occur in the development of physical and psychological self-integrity and the development of self-regulatory processes (Cole & Putnam, 1992). While self- development may potentially be disturbed in all instances of CSA, incest may be viewed as more traumatic than extrafamilial abuse to the process of self-development, in that it not only involves the sexual violation, but a great amount of guilt, fear, and a breach of trust by the perpetrator. Many of the disorders reported by incest survivors share the common aspect of disturbed development of self as well as disrupted social relationships. Dysfunction of self inevitably leads to related social dysfunction, due to the distorted perceptions and expectations of social relationships that result from a distorted sense of self. In order to fully understand the impact of self-development dysfirnction on the individual, the greater social context must be examined. Cole and Putnam (1992) focused on the trauma of incest to explain how disruptions in self-development at specific stages of child development lead to difl‘ering effects. Although it has not been examined as extensively, it is likely that similar developmental processes occur in situations of extrafamilial abuse. During infancy and toddlerhood, early attachment experiences influence the initial development of a sense of self, while self-regulatory functions and social relations begin to develop as well. Sexual abuse during this period may destroy the infant’s sense of integrity of self, trust in the parent, and sense of control. During the preschool years, the child must learn to integrate his/her self with the realities of the social world. Denial and dissociation become common methods of coping with stress during this developmental period, particularly when other methods of coping are prevented, as in the case of incest. The ongoing self-development is disturbed in cases of abuse, and accomplishments of infancy and toddlerhood may be compromised. During childhood, prior to puberty, self-development centers on an increasing sense of cognitive and social competence and control. The initiation of sexual contact in cases of incest is most common during this period of development, disrupting the child’s ability to expand social experience and self-competence outside of the home, and to integrate positive and negative aspects of self. Adolescence is marked by the onset of puberty and emerging sexuality. Puberty is often considered to be a critical period that is particularly affected by previous sexual abuse (Putnam & Trickett, 1993). Adolescents become capable of introspection and gain heightened self-awareness, as they begin to form their true identities. Abuse may continue into this period, impeding the process of self integration. In addition, sexual abuse victims often utilize immature coping strategies, exhibited in common conduct disturbances such as substance abuse, running away, and other self-destructive behaviors. The tasks of adulthood, in sexual relationships, marriage, and parenthood, are subsequently affected by childhood incest as well. As the framework of developmental psychopathology is rather broad, applying a more specific theoretical model, such as attachment theory, within the overall model, allows for a strong theoretical base with specific, testable hypotheses. Furthermore, attachment theory allows for the inclusion of the family context in examining the effects of C SA, which is necessary when examining the overall childhood experience (Alexander, 1992). Whether attachment style develops as a result of abuse by a caregiver, or is disrupted due to trauma inflicted by another individual, it can potentially have a great influence on subsequent emotional and psychological development. A_tt_achment Theog. A child’s attachment to the primary caregiver appears to be relatively stable over time, and although infant-mother attachment predicts firture social interactions better than infant-father attachment, secure attachment to both parents is predictive of the best outcome (Alexander, 1992). Through years of work with the Strange Situation procedure, four categories of infant attachment style have been identified. These categories include secure attachment and three categories of insecure attachment: avoidant, resistant-ambivalent, and disorganized-disoriented (Main, 1996). The fourth category was added in more recent research, to attempt to explain behaviors seen in some infants that did not fit into the other classifications. Children classified as securely attached use their mothers as a base for exploring their environment, appear to miss their mothers during separation, and greet their mothers warmly upon return. The resistant-ambivalently attached child appears preoccupied with his/her mother and exhibits passive-aggressive behavior toward her, as well as exhibiting more anxiety in general. The insecure-avoidantly attached child, in contrast, ignores the mother in the Strange Situation procedure, as well as displaying anger and anxiety in the home. The mothers of these insecure-avoidantly attached children are often rejecting of tactual contact with their children. All forms of insecure attachment are commonly found in children who have experienced abuse or neglect, with the disorganized attachment style relating to unresolved trauma from the mother’s or father’s past (Alexander, 1992). In this category, there is no coherent strategy for dealing with interactions with the attachment figure, because of the dual role of the caregiver in both producing and relieving the child’s anxiety (Alexander, 1992). The disorganized attachment style overlaps with the other insecure types in this way, but often implies more severe abuse or trauma, as well as unavailability of the caregiver. These children are believed to be at the greatest risk for later psychopathology in adolescence and adulthood, although the other insecure attachment styles are also risk factors for psychopathology. According to Bowlby’s (1973) concept of the internal working model, early experiences with the attachment figure lead a child to develop expectations about his/her role in relationships as well as others’ roles in relationships, as the internal working model serves as a basis for interpretation of later experiences. Within this model, attachment is seen as a stable process that endures into adulthood. Relationships with peers in adolescence and adulthood should then exhibit similar attachment characteristics as those with the early caregiver. Within an attachment theory framework, psychopathology is seen as a deviation from normal development in an attempt to adapt to developmental change (Rosenstein & Horowitz, 1996). The quality of attachment plays a vital role in an individual’s development, as it affects the degree to which an individual can adapt to deviation without the development of psychopathology. Alexander (1992) proposed that the range of psychological and interpersonal symptomatology experienced by abuse survivors is mediated by their attachment experiences, helping explain some of the variability. In addition, early attachment can impact later social relations by reducing social support, with insecure attachment styles creating “enduring vulnerabilities to psychopathology by impairing an individual’s ability both to participate in satisfying social relationships and to appropriately understand and evaluate social interactions” (Allen, Hauser, & Borman- Spurrell, 1996; p. 254). Insecure attachment is often seen as a risk factor for the development of childhood psychopathology (Alexander, 1992; Rosenstein & Horowitz, 10 1996), and each type of insecure attachment is associated with different types of family dysfiinction, such as neglect, emotional abuse, and physical abuse. Insecure attachment in the parent may precede onset of abuse of the child, as either the abusive father or the unavailable mother may have an avoidant attachment style as a result of their own histories of abuse or rejection (Alexander, 1992). The child then develops an avoidant attachment style due to this parental rejection, and is less likely to defend herself against abuse or to seek outside support. Related to the resistant- ambivalent attachment category is role-reversal or parentification, i.e. the expectation that a child will firlfill the parental role. Parents’ own childhood experiences may cause them to expect the same behavior fiom a child that was expected from them growing up, or they may simply be unable to meet the needs of the abused child. The disorganized attachment pattern is often associated with unresolved trauma in the parent’s past. This trauma is often a repressed part of the perpetrator’s history, or may be in the nonabusive parent’s history, causing the current abusive family situation to be unconsciously ignored. The child who develops a disorganized attachment style often has no effective method of coping, and may possess distorted perceptions of social relationships, aggravating the effects of the abuse and increasing risk for psychopathology. The child’s insecure attachment does not always precede abuse, as in situations where a previously secure attachment is destroyed by an act of abuse or the consequences of disclosure. The mediating role of attachment can also be viewed as a positive influence, such that support 11 from the non-abusive parent may decrease subsequent psychological distress (Alexander, 1992) Rosenstein and Horowitz (1996) examined the role of attachment in the development of adolescent psychopathology and found a relation between type of insecure attachment and psychiatric diagnosis. Adolescents with an avoidant attachment style were more likely to exhibit symptoms of conduct disorder and substance abuse, while adolescents with a resistant-ambivalent attachment style often sufl’ered from affective disorders. The authors explained that the early development of insecure attachment does not produce the pathology, but rather increases one’s vulnerability to disorder, particularly when combined with other risk factors, such as family dysfunction and trauma. The specific developmental issues of adolescents, such as the search for autonomy, need to be considered in this process as well. Alexander (1992) proposed that specific symptoms commonly experienced by CSA survivors can be matched to the specific insecure attachment subtypes, as evidenced by adults. For example, an adult with a resistant-ambivalent attachment style will most likely experience depression and anxiety, while the adult with a disorganized attachment style will be more likely to suffer from the more severe disorders, such as PTSD and dissociation. Disturbances of self, which can involve low self-esteem and loss of identity, can also be explained within an attachment fi'amework. In a situation of ongoing sexual abuse, an insecure attachment style often develops, which in turn will produce a negative sense of self in the development of the internal working model. The symptoms of 12 borderline personality disorder, often found in survivors of sexual abuse, are also common to individuals with a resistant-ambivalent attachment style. Working within a developmental framework (specifically attachment theory) in the study of CSA provides the advantage of utilizing an established and widely accepted theory in a new area, rather than creating a new theory applicable only within the limited scope of interest. Rather than focusing on the common sequelae of CSA to propose ways in which they developed, the family context of the abuse victim and development following abuse can be examined to determine the relation between early attachment, abuse, and psychological firnctioning. One limitation of attachment theory is the question of whether it is applicable to extrafamilial abuse, or if it is restricted to cases of incest. In the present study, both intrafamilial and extrafamilial childhood sexual abuse were examined within the attachment theory framework. CSA_ and Psychopathology: Empirical Evidence There are a variety of Diagnostic and Statistical Manual of Mental Disorders, (DSM-IV; American Psychiatric Association, 1994) diagnoses associated with CSA, pervading all axes, thus illustrating the global impact that it has on life functioning. However, the extent of disorder and level of functioning vary greatly within a population of survivors, making it difficult to draw any concrete conclusions as to a causal relation between CSA and specific psychopathology. To study this relation, a distinction should be made between short-term, or initial, effects, and long-tenn effects of the abuse, although many of the symptoms seen in children and adolescents continue into adulthood. 13 In general, research in this area focuses on female victims, due to a greater incidence of abuse in females, and the confounding influence of including both genders in a single study. The empirical evidence discussed below is that pertaining to studies of female survivors of CSA. The short-term, or initial, effects of childhood sexual abuse are often forgotten, as researchers focus on the lifelong impact of trauma. However, the influence of such a traumatic experience can greatly affect the psychological well-being of children and adolescents. Browne and F inkelhor (1986) categorized initial efi‘ects as those that occur within 2 years of termination of the abuse, but not all researchers agree on this time period. Therefore, some studies of short-term effects may include those occuning in conjunction with the abuse, or those appearing more than 2 years after the abuse has ceased. Until recently, an additional problem in this area of research was that many studies of short-term effects were actually based on adult survivors’ retrospective accounts. More recent research has focused on child victims, and illustrates the differential reaction seen in children and adolescents as compared to adults (Kendall- Tackett, Williams, & F inkelhor, 1993). Illustrating the difficulty in developing one diagnosis for the diverse psychopathology often experienced by CSA survivors, Kendall-Tackett et al. (1993) found that for each symptom studied, the percentage of victims reporting that symptom ranged from 20% to 30%, but no single symptom was found in a majority of victims. Thus, no “survivor’s syndrome” exists in which to group individuals experiencing abuse- 14 related symptomatology. However, this variation in symptoms was lessened when children were divided according to age, indicating the importance of considering age of the victim in further study. Often, developmental differences are not considered in research with children, a potential explanation for the inconclusive findings thus far. Relatedly, Beitchman, Zucker, Hood, daCosta, and Akman (1991) suggested that it is not symptomatology that changes across age levels, but the fact that certain behaviors (e. g., suicide attempts) are age-related and will therefore be reported more often at certain age levels. Changes in externalizing versus internalizing behaviors across age levels also need to be examined firrther to help clarify the relation between age and reported symptomatology (Browne & Finkelhor, 1986). Among the reported symptoms in children who have been sexually abused are anxiety, depression, PTSD, somatic complaints, withdrawn behavior, delinquent behavior, self-injurious behavior, aggression, school problems, guilt, shame, and sexualized behavior (Browne & F inkelhor, 1986; Kendall-Tackett et al., 1993). Most studies have found that children who have been sexually abused are significantly more symptomatic in these areas than nonabused children of the same age, with sexualized behavior to be one of, if not the most, commonly reported symptoms in children and adolescents. While some adolescents may not have experienced the actual abuse in a number of years, the difficulties and symptoms they display may be better conceptualized as short- terrn efl’ects of the abuse, or at least should be studied as age-specific, due to the developmental changes and challenges of adolescence. Often, the “problem behavior” of 15 an adolescent, such as promiscuity, running away, substance abuse, criminal behavior, self- injurious and suicidal behavior, and depression, is actually a manifestation of the effects of ongoing or previous sexual abuse (Beitchman et al., 1991; Van Gijseghem & Gauthier, 1994). Of the most common behaviors reported by adolescent survivors in studies reviewed by Kendall-Tackett et al. (1993 ), self-injurious behavior was particularly high, with 71% of the subjects exhibiting such symptoms. Research has less consistently found that externalizing behaviors, as compared to internalizing behaviors, correlate with abuse history. This finding may be partly a function of the developmentally specific nature of these behaviors, and partly a result of outside variables. For instance, alcohol abuse and suicidal ideation in adolescents have both been found to correlate with a history of sexual abuse, but the extent to which these problems exist may be moderated by outside variables, such as family support and recency of abuse (Luster & Small, 1995). Adolescents are often thought to be particularly vulnerable to psychopathology following abuse, particularly for adjustment problems. Orr and Downes (1985) examined the impact of sexual abuse on adolescents’ self-concept, which has been correlated with overall mental health and adjustment. In this study, adolescents who had recently been sexually victimized had more problems than nonabused adolescents in three main areas: 1) negative attitudes toward sexuality and sexual behavior, 2) more conflicted relationships with parents, and 3) inability to master or control the environment, with related depression. Experiencing these difficulties in adolescence is often associated with more severe problems in adulthood, particularly depression. 16 Obviously, the short-term effects of CSA are Significant and may be severe. One should not be deceived by the designation “short-term” however, as many of these symptoms do not simply disappear and may continue into adulthood without appropriate intervention. The consequences of not providing intervention may range fi'om mild emotional or behavioral problems to severe psychopathology (Browne & Finkelhor, 1986; Polusny & Follette, 1995). Adolescents, in particular, may be more vulnerable to the impact of CSA, due to the developmental challenges and changes they face regardless of abuse history. The range of problems experienced by survivors of CSA extends beyond any one diagnostic category and fi‘ustrates those researchers attempting to develop a syndrome that contains a specific constellation of symptoms. Often, survivors are given a collection of diagnostic labels to describe the effects of abuse, which in itselfcan be stigrnatizing and isolating, aggravating the existing distress. Briere and Runtz (1988) suggested an alternative category of “post-sexual abuse trauma,” a broader description than PTSD, to describe symptoms that were adaptive responses to childhood abuse, but have become harmfirl or inappropriate. What causes this broad range of symptomatology among survivors is unknown, but coping has been a recent focus of research, as certain methods of coping may mediate the relation between sexual abuse and psychological functioning, to affect either positive or negative outcomes. 17 Coping Coping Theory. The concept of coping has gained a great deal of attention in research over the years, but lacks a common definition and accepted model. Two contrasting views of coping are coping as a style versus coping as a process (Lazarus, 1993). In the first model, coping is seen as a stable aspect of personality. Alternatively, the process perspective views coping as changing over time and across situations, and is the more accepted model of coping today. Although many researchers discount the concept of coping as a trait or style, choosing instead to accept the view of coping as a process, individual differences may play a role in coping, in that people may have certain preferred coping strategies, which then interact with the environment (Carver, Scheier, & Weintraub, 1989). Research has found that certain individual coping strategies are relatively consistent across situations while others are more dependent on the context of the situation (Lazarus, 1993). People utilize a number of cOping strategies in each stressful situation they encounter, due to the complexity of these experiences, and in any given stressfirl situation, coping changes as the experience progresses through stages (Lazarus, 1993). All of these findings illustrate the complexity of defining and conceptualizing coping, and tend to support the View of coping as a process, but further research needs to study coping across time and situations to determine the degree to which coping is a trait vs. process. Lazarus (1993) set forth some basic principles to the consideration of coping as a process. The first is that coping methods under stress must be assessed separately fiom 18 their outcome to examine their adaptiveness or maladaptiveness. This principle is based on the idea that the adaptiveness of a coping mechanism depends on a host of factors, and each strategy cannot simply be designated as adaptive or maladaptive. He also suggested that rather than viewing coping in the context of the overall stressor, it should be examined in terms of the specific threats that are of immediate concern to the individual. A further principle defines coping as “ongoing cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person” (p.237), once again separating the efforts fiom the outcome and taking individual differences into account. Finally, COping can be analyzed in terms of its function, as problem-focused or emotion-focused. Problem-focused coping functions to change the stressfirl situation by acting on the environment or oneself. Emotion-focused coping attempts to change either how the situation is attended to, or the meaning of what is happening. Problem-focused coping is often perceived to be more desirable and adaptive, however, in situations where one lacks control, emotion-focused coping may result in less distress. What is adaptive versus non-adaptive appears to depend on the context of the stressor, and generalizations cannot be made about coping strategies that are consistently one or the Other. Similar to the problem- vs. emotion-focused model of coping is the approach- avoidance model, in which cognitive or emotional thoughts and/or actions are directed either toward or away from a threat (see Roth & Cohen, 1986, for a review). This model emphasizes the focus of coping, rather than the firnction, in that approach strategies are 19 oriented towards the threat, whereas avoidance strategies are oriented away from the threat (Ebata & Moos, 1991). The appropriateness or adaptiveness of each type of strategy once again depends on the context of the stressor, with both types of coping having potential benefits and costs. Within the approach-avoidance model, there are a number of conceptualizations of coping that differ slightly fiom each other, although each is based on the focus of coping. The coping strategies of approach and avoidance are also compatible with either the style or process theories discussed previously. One benefit of the approach-avoidance model is that rather than categorizing specific c0ping methods as either problem- or emotion-focused, which often presents emotion-focused COping as a maladaptive or less desirable response, methods that may center around affect or other internal processes can be seen as effective ways of dealing with a situation. For example, venting emotions would be considered emotion-focused coping, but may also be an approach strategy, depending on the context. As the situation of child sexual abuse is virtually uncontrollable for the victim, many available methods of coping are emotion-based, but may be approach-oriented given the circumstances, such as venting emotions or seeking emotional support. In this uncontrollable situation, many strategies that would be considered problem-focused, such as disclosure, may not be available options for the victim, and many of the problem-focused strategies that are employed are actually avoidant coping methods, such as running away, or avoiding the situation or the perpetrator. 20 Based on the premise that the problem-focused vs. emotion-focused coping model is too simplistic, Carver et al. (1989) created a measure to assess a number of factors that represent a more diverse array of coping strategies. Their model is based on a combination of the Lazarus model and their own model of behavioral self-regulation. Although their model and measure began with a large number of distinct categories of coping, when testing adolescents with the measure, four factors emerged: active coping strategies, avoidant strategies, emotion-focused strategies, and acceptance strategies (Phelps & Jarvis, 1994). Active strategies include planning, suppression of competing activities, and seeking social support for instrumental reasons. Avoidant strategies include denial, behavioral disengagement, and alcohol-drug disengagement. Emotion-focused strategies involve seeking social support for emotional reasons and focusing on and venting emotions. Finally, the acceptance strategies consist of restraint, positive reinterpretation, growth, acceptance, and mental disengagement. Thus, these factors integrate the problem/emotion-focused and the approach/avoidance-oriented theoretical models of coping, without oversimplifying the coping process. Coping in Adolescence. Adolescence includes many challenges and stressors that are Unique to this developmental stage, thus it is important to examine adolescent coping separately fi'om adult or child coping. Children and adolescents are dependent on their parents or other adults, necessitating the inclusion of the family context in examining their ability to cope with stress. Adolescents differ from young children in that while they are still dependent on their parents, they are going through a process of individuation and 21 breaking away from them. Such challenges are unique to this developmental stage. An important part of development is learning to cope successfirlly with stress in order to decrease psychological distress (Compas, 1987). Early attachment and separation from the mother can be viewed as the first coping experience of an individual, and circumstances surrounding this early experience with coping and stress may affect coping later in development (Compas, 1987), particularly during the separation of adolescence. The distinction between coping resources (individual and social), coping styles, and coping strategies is Often not made in research, though each is a separate process in overall coping (Compas, 1987). Social COping resources include social support from significant others, while personal coping resources may include self-esteem, individual skills, and perceptions of coping ability. Coping style may refer to some form of personality trait, but can also imply the consistency of coping strategies in similar situations across time. Strategies are simply those methods of coping used in specific situations. This distinction is often not made in research with children and adolescents, blurring the lines between these factors and making it difficult to draw conclusions as to the relative importance and stability of each. In connection to the concept of coping resources, the transactional model of stress asserts that it is not so much the demands placed on a person that produce stress, as his/her perceived ability to cope with the situation (Allen & Hiebert, 1991). Thus, individuals with fewer personal coping resources are more vulnerable to stress, while others may not be adversely affected by similar circumstances. This theory holds 22 particular importance in studying adolescent COping, as adolescents have been found to possess fewer coping resources than adults (Allen & Hiebert, 1991). Those adolescents that possess greater coping resources report lower levels of stress, anxiety, and stress- related symptoms than those with fewer available resources. The greater stress levels of adolescents, as compared to adults, can be attributed to a variety of factors, including perception of coping ability, the demands of adolescence, or any number of other variables. Because adolescents are known to face many developmentally-specific challenges and demands, such as emerging sexuality, greater importance of their peer group, and increased independence fiom their parents, their coping skills and perception of those skills are vital to their psychological well-being. As discussed previously, the adaptiveness of a coping strategy depends on the type Of stressor involved, including whether or not the stressfirl situation is controllable or not. The connection between controllability of the stressor and type of coping utilized appears in adolescents as well as adults. In fact, adolescents who fail to match these factors, either by employing problem-focused strategies when they perceive little control or by turning to emotion-focused strategies when they do feel a sense of control, report more behavior problems overall (Compas, Malcame, & Fondacaro, 1988). The context of the stressor, such as academic versus social stress, does affect the relation between type of coping and behavior problems, indicating that these findings may not be generalizable across situations, and more research needs to be conducted in this area. 23. Ebata & Moos (1994) have found that personal factors, situational factors, and contextual factors interact to influence the coping strategies utilized in stressful circumstances. This model extends beyond a person-environment interaction in that it not only considers individual differences and the stressor, but also the ecological context, including school, family, and community. Therefore, in this model, current coping strategies may be affected by previous life stressors and coping with past events. Ebata and Moos (1994) utilized the approach-avoidance model to explain adolescent coping behavior, emphasizing the focus of the coping response, rather than its firnction. Each type of strategy includes cognitive and behavioral efforts to either deal directly with the stressor or to somehow avoid it. They created a hierarchical structure based on focus (approach vs. avoidance) and method (cognitive vs. behavioral) of coping, which incorporates the various conceptual models of coping that exist. Overall, approach methods correlated with greater well-being, and avoidance methods related to high levels of distress. However, no individual coping strategy can predict adjustment. Rather, certain combinations of approach and avoidance methods are correlated with adolescent adjustment vs. distress, depending on the situation, the specific stressor, and other factors. It is not known whether distress leads to the use of avoidant coping, or if use of avoidant coping methods somehow causes psychological distress. It appears that a cycle develops, in which already distressed adolescents employ more avoidant coping Strategies, and adolescents who persistently engage in avoidant coping do not adjust as well when faced with life stressors, thus increasing distress. Results of one study indicate that previous use of each type of coping (approach and avoidance) predicts later COping, indicating some 24 amount of stability in the coping process (Ebata & Moos, 1994). Together, these results demonstrate the importance of the COping process in psychological adjustment, and suggest that coping may be a point of intervention for those adolescents who are already experiencing distress. Coping with CSA. Coping may play a dual role in psychological functioning, both in how one deals with a specific stressor and how that situation and the coping utilized affect subsequent coping and firnctioning. The types of coping strategies available to and utilized by adolescents in cases of severe trauma differ from the coping methods used for everyday stress, due to the overwhelming nature of the trauma, and its uncontrollability. In situations of childhood sexual abuse, a victim’s available coping resources may be limited or perceived as useless. For adolescents who have experienced sexual victimization during development, particularly severe, ongoing abuse, the dual role of coping may be especially salient. First, coping with this traumatic situation can be viewed in the specific context of the abuse itself, and how this coping mediates later psychological functioning. Furthermore, coping with previous abuse may afl’ect the types of coping strategies employed for the common developmental stressors of adolescence, and the subsequent adaptation, or maladaptation, that follows. This second role of coping was the focus of the present study, in examining the mediating role of coping with an interpersonal stressor following CSA. This conceptualization is a means of indirectly examining the effects of CSA on current coping abilities. The review which follows focuses on coping with CSA itself, fi'om which subsequent effects on coping strategies can be inferred. 25 Children’s methods of coping with neglectful or abusive families can be examined in the framework of attachment theory (Crittenden, 1992). The developmental stage at which these neglect or abuse experiences initially occur influences which coping strategies are available to the child, and these strategies will likely change as skills increase throughout development. Whereas many researchers argue for the lack of stability in coping over time and across situations, attachment theory proposes the construct of the internal representational model, which provides a basis for stability. Experiences with the attachment figure lead to expectations about firture interpersonal situations, based on the internal working model, which subsequently affects coping (Crittenden, 1992). Through experiences with neglectful or abusive parents, children may develop maladaptive coping strategies. Certain strategies may be adaptive in the short-term, in protecting the child in an adverse environment, but a consequence may be that the internal working model cannot be revised, so that coping strategies based on the insecure attachment become maladaptive as the individual carries the internal representations, often accompanied by a detached and avoidant coping style, into later interpersonal relationships. In the present study then, it was expected that adolescents who had been sexually abused would differ from nonvictims in the coping strategies utilized for everyday stressors, especially when associated with interpersonal conflicts. In addition, attachment style was expected to mediate this connection between CSA and coping. Research has discovered that an individual’s cognitive interpretation of a traumatic event determines the response as much as the event itself does, and in some cases the 26 survivor’s appraisal and coping methods are better predictors of post-disclosure psychological adjustment than are specific characteristics of the abuse (Johnson & Kenkel, 1991; Newberger & DeVos, 1988). The child’s initial interpretation of the abuse will have long-term consequences in the attributions made and the coping mechanisms employed as a function of that cognition. In considering cognitive appraisals and coping, one must also consider the influence of development. During adolescence, cognitive abilities increase, including a greater understanding of sexuality which may change appraisals of sexual abuse. Adolescents may also be more likely to consider the impact of disclosure on family relationships, and to blame themselves for the abuse as they develop secondary sex characteristics (Johnson & Kenkel, 1991). According to the transactional model, negative cognitive appraisals and maladaptive coping strategies lead to increased symptomatology (Spaccarelli, 1994). The highest risk for psychological dysfunction occurs with the use of an avoidant coping style, especially cognitive avoidance. More active coping strategies, such as resistance, disclosure, and support seeking, may serve beneficial functions in increasing a sense of control and decreasing self-blame (Spaccarelli, 1994). Feelings of control may produce positive effects, or in contrast, may lead to poorer outcome due to self—blame (Newberger & DeVos, 1988). Alternatively, Johnson and Kenkel (1991) found that survivors who perceived the abuse as more threatening and themselves as having less control experienced greater distress. The extent of distress resulting fiom the abuse experience would 27 naturally have long-lasting effects on development, and the ability to cope with stress in other areas of victims’ lives. Many self-destructive behaviors common to adolescents, including substance abuse, attempted suicide, self-isolation, promiscuity, running away, and self-harming behavior, can be viewed as adaptive COping responses to childhood sexual abuse, particularly incest (Lindberg & Distad, 1985). Sexual abuse often causes an overwhelming sense of helplessness, accompanied by anger, guilt, fear, and stress, against which the child must develop some type of protective response. The “acting out” behaviors often exhibited in adolescence may be adaptive methods of survival for abuse survivors, as they serve as attempts at exerting some control over themselves to compensate for the lack of control they had over their own bodies during the abuse, or simply to escape the abusive situation. Unfortunately, while these methods of coping may serve adaptive functions during the abuse, and even in the immediate time following its cessation, they become maladaptive and potentially harmfirl. The types of coping mechanisms used by survivors may also differ depending on characteristics of the abuse, specifically whether the abuse was intrafarnilial or extrafamilial. Intrafamilial abuse, particularly when the perpetrator is a close family member, is believed to be more traumatic for the victim, suggesting that available coping resources and strategies might be affected more negatively than by extrafamilial abuse (DiLillo, Long, & Russell, 1994). For instance, victims who are abused by a close relative Often perceive less control over the situation and fewer available coping resources, and 28 might therefore employ more emotional and avoidant strategies to lessen internal distress. Intrafamilial abuse is also associated with longer duration, more severe sexual acts, greater frequency, and other factors that may be considered to be more traumatic. DiLillo, Long, and Russell (1994) found that victims of intrafamilial abuse utilized more emotion-focused coping strategies than extrafamilial abuse victims, specifically wishful thinking, self-blame, self-isolation, and detachment. They hypothesized that the use of more emotion-focused coping may be a result of perceived lack of control resulting from the qualitative differences between intrafamilial and extrafamilial abuse. However, the intrafamilial abuse victims also employed more problem-focused coping strategies, potentially due to the fact that over a longer period of time, there are more opportunities for difi’erent types of coping. The greater use of all types of coping strategies may reflect the distress that these victims felt, and their attempts to try any means of ending the abuse. One particular coping mechanism that is commonly associated with childhood sexual abuse is dissociation. According to McElroy (1992), the onset of dissociation occurs “when child sexual abuse victims find they must submit, inhibit, or dissociate the strong, assertive part of themselves during the actual abuse in order to survive. Hence, dissociation becomes a fixed way to ‘avoid’ the pain of psychological surrender” (p.839). While dissociation may be viewed as pathological, in the context of overwhelming trauma, such as CSA, it is often considered to be an adaptive defensive response (Putnam, 1991). Often, the greater the duration or severity of trauma associated with the abuse, the more the victim relies upon the dissociative mechanism, leading to the development of a pattern 29 of dissociation (Classen, Koopman, & Spiegel, 1993). Symptoms become “abnormal,” or maladaptive, when they persist and interfere with social and occupational firnctioning, and while dissociation may begin as a defense mechanism against overwhelming anxiety, it may subsequently be utilized in maladaptive as well as adaptive circumstances. Dissociation is an example of an extreme method of coping with severe, uncontrollable trauma, such as CSA. Though an extreme example, it illustrates the important role of coping, both during abuse, and in the long-term effects that employing certain coping mechanisms may have on subsequent functioning. Certain coping mechanisms may be associated with greater psychological distress in the context of other stressors, but they may have initially begun as adaptive responses to the incredible stress of sexual abuse. Summgy The relation between CSA and subsequent psychological and interpersonal functioning potentially involves a very complex path involving a number of factors with differential influence. Figure 1 represents the part of this complex path model examined in the present study. The full model was not tested, due to the excessive number of pathways represented, but each hypothesized model is contained within the firll model. The primary focus was to examine the mediating roles of attachment and coping in the psychological and interpersonal functioning of adolescent females who reported a range of sexual abuse and family violence experiences. 30 The negative efi‘ects of CSA on psychological and interpersonal fimctioning have been well-established through empirical research, but not as well-explained. Symptoms reported by sexually abused children and adolescents include anxiety, depression, PT SD, somatic complaints, withdrawn behavior, delinquent behavior, self-injurious behavior, aggression, academic problems, guilt, shame, and sexualized behavior (Browne & Finkelhor, 1986; Kendall-Tackett et al., 1993). In adolescence, behavioral symptoms such as promiscuity, rumring away, substance abuse, criminal behavior, self-injurious and suicidal behavior, as well as depression, are common manifestations of sexual abuse (Beitchman et al., 1991; Van Gijseghem & Gauthier, 1994). In addition to studying the mediators of CSA, it was also expected that some direct efi’ects would also exist, as suggested by the empirical literature. o -- , 4 Distrou § ' -‘v Attachment - ‘\ A. Figure 1 Structural Model Incorporating All Hypotheajzed Pathways 31 The effects of family violence on adolescents’ firnctioning was not the primary focus of the study, yet experience of family violence was measured due to the expected rate at which an adolescent who had been sexually abused would have also been exposed to some form of family violence. The effects of witnessing domestic violence and/or experiencing neglect or abuse have been studied extensively, particularly in relation to psychological firnctioning. Symptoms reported in children and adolescents include depression, psychosomatic complaints, anger, anxiety, fears and phobias, dissociation, posttraumatic stress, insomnia, and nightmares (Fantuzzo & Lindquist, 1989; Singer, Anglin, Song, & Lunghofer, 1995). Thus, it was expected that family violence would have a direct effect on psychological, and possibly interpersonal, functioning. Once again, the process by which these effects occur has not been studied as extensively as the direct effects. Thus, examining attachment and cOping as mediators following family violence was also included in the present study. The recent inclusion of theory-driven models has failed to produce a single theory to explain CSA effects, yet attachment theory has been utilized and accepted in so many areas of human development that it provides a solid basis for firrther research. Utilizing Bowlby’s (1973) internal working model, attachment can be seen as a stable process in which relationships with peers and dating partners in adolescence resemble attachment characteristics of the caregiver relationship. Insecure attachment is often viewed as a risk factor for the development of childhood psychopathology, and it has been suggested that 32 attachment experiences may mediate the psychological and interpersonal symptomatology of abuse survivors (Alexander, 1992; Rosenstein & Horowitz, 1996). It was expected that attachment style would mediate the experience of childhood trauma, in such a way that a secure attachment style would be associated with less psychological distress and interpersonal conflict following childhood trauma. Coping mechanisms have recently become a research focus in trying to explain the differential outcomes in a population of survivors. Retrospective accounts of coping with CSA are difficult to measure, so this study examined the role of current coping strategies in response to an interpersonal stressor. Coping in an interpersonal situation in adolescence was expected to reflect coping with CSA, as coping with the sexual trauma would influence the types of strategies that a survivor would choose to employ later. A similar effect should also occur following exposure to family violence. The adaptiveness of any particular coping strategy depends on situational factors (Lazarus, 1993), so while coping was expected to play a mediating role, the specific relations between different strategies and distress and conflict were not predicted. The literature also suggests some connection between attachment style and coping, leading to the expectation that attachment style mediates the relation between childhood trauma and coping (Crittenden, 1992). The developmental process of separation and individuation in adolescence reflects the experience of early attachment and separation from the parent, which can be viewed as the first coping experience of an individual (Compas, 1987). Therefore, coping in the context of early attachment experiences may affect coping later in development, particularly during adolescence. Children may develop 33 particular coping patterns in response to interactions with neglectfirl or abusive parents, which become maladaptive in subsequent interpersonal situations that are no longer threatening. It was expected that a secure attachment style would be related to more adaptive coping strategies, while an insecure attachment style would be associated with more maladaptive coping, in the context of an interpersonal conflict. Additionally, coping was expected to mediate the paths to psychological and interpersonal functioning, with adaptiveness detemrined by the extent of psychological distress and interpersonal conflict. HYPOTHESES 1) Attachment style mediates the relation from CSA and family violence to psychological and interpersonal firnctioning, with a secure attachment associated with decreased distress and conflict following childhood trauma. The direct effects of CSA and family violence will also be examined. 2) The strategies utilized to cope with interpersonal stressors during adolescence influence psychological and interpersonal functioning, mediating the effects of negative or traumatic childhood experiences on present adjustment. An exploratory analysis will identify how coping strategies utilized in interpersonal conflicts during adolescence affect the relation between childhood abuse experiences and subsequent firnctioning, in conjunction with the direct effects of these traumatic experiences. 3) Attachment style serves a mediating role between the experience of C SA and coping, with a secure attachment style being associated with the development of adaptive coping mechanisms and their related firnctioning, and an insecure attachment relating to maladaptive coping strategies and the relative distress that develops. Both indirect and direct effects of C SA and family violence will be examined. 4) Adolescent survivors of CSA will be compared to non-victimized adolescents at comparable levels of fiinctioning, with a prediction that successful everyday coping strategies of non-victims will differ from the adaptive daily coping strategies of victims. Healthy, non-victimized adolescents are expected to respond to a recent interpersonal stressor with active, approach-oriented coping strategies. However, CSA survivors, at 34 35 levels of psychological and interpersonal functioning equal to their non-victim counterparts, are expected to utilize more avoidant and accepting coping strategies, focusing on affect and internal relief. METHOD Participants The sample consisted of 80 female adolescents, ages 14-16 years, 26 of whom reported a history of childhood sexual abuse, and 54 of whom did not. Mothers of the adolescents were asked to participate in order to obtain demographic data on the families. The adolescents were recruited through several sources throughout the Mid-Michigan area, including domestic violence shelters and programs, at-risk teen programs, the Family Independence Agency, and the general community. Fifty-four of the 80 girls who were included in the present study were recruited for a larger Family Relationships Study, of which the present study was a small part. Most of these participants were recruited through flyers posted in the community. Of these girls, 13 reported a history of CSA, and 41 did not. Additional adolescent girls were recruited solely for the present study, including active recruitment of CSA survivors. Service providers were asked to specifically contact clients with a history of CSA, and potential participants were screened on the phone to determine if they had been sexually abused. Twenty-six adolescents participated solely in this study. Six of these girls were ineligible for the larger study because they had siblings participating (only one child from each family could participate), so they were referred for this study. Another 9 girls were recruited through flyers posted in the community, and 11 girls were specifically recruited through agencies that provide services to adolescent CSA survivors. A total of 13 CSA 36 37 survivors came from this group: 10 from service providers and 3 fiom the community. One interesting methodological note is that although a number of agencies across the state Offered to help recruit participants (one agency sent out over 300 letters to eligible clients), only 10 CSA survivors came through this channel. Through the community, although not a truly random sample, CSA was reported by 23% of the participants. In the total sample, the mean age of the adolescents was 14.96 years (sd=0.83), and the mean grade in school was 9.36 (sd=1. 12). The adolescents’ ethnicity was diverse: 51% Caucasian, 21.3% Afiican American, 15% Biracial, 10% Hispanic/Latina, 1.3% Asian American, and 1.3% other. The mothers’ mean age (N=75) was 39.37 years (sd=6.47). Many of the women (41.3%) were married at the time of the study, and 88.8% either had firll or joint custody of the participating adolescent. The mean monthly household income was $1866.46 (sd=l644.32), ranging from $0 to $9000/month. Sixty percent of the mothers participating reported that they had been in an abusive relationship, but it is not know if the participating adolescents were witnesses to the violence. There were no significant differences between the non-abused group and the group of CSA survivors on any demographic variables examined. Specific information about the abuse is only known for 20 girls, due to methodological problems at the beginning of the project. Originally, those girls that reported sexual abuse were asked to return for a follow-up interview, however, many of these initial participants could not be reached. As a result, the procedure was altered so that the abuse interview occurred at the end of the regular interview, had CSA been 38 reported. The mean age at onset of abuse was 8.10 years (sd=4.24), and 55% were abused for less than 1 year, although 15% reported the abuse lasted 3 or more years. The frequency of abuse varied from once (3 5%) to daily (15%), with another 35% reporting weekly abuse, 10% several times/year, and 5% monthly. Adolescents reported up to four different perpetrators, the majority of whom were men, although three participants were abused by women. Overall, 21.9% of the abuse was perpetrated by the victim’s father or a father figure, 25% by a “secondary” relative (uncle, aunt, grandfather, or cousin), 18.8% by a stranger, 9.4% by a peer acquaintance, 6.2% by the mother’s boyfi'iend, 6.2% by a neighbor, and the remainder by a brother, boyfriend, or unidentified other. Age differences between victims and perpetrators ranged from 1 year (5%) to 50 years (5%), with a mean of 19.15 years (sd=13.20). When asked to report types of abuse they experienced, 73.7% had their genitalia fondled, 68.4% reported kissing/fondling, 57.9% were forced to touch the perpetrator’s genitalia, 52.6% reported vaginal intercourse, 26.3% had no physical contact, 21.1% experienced oral-genital contact, 10.5% reported simulated intercourse, and 5.3% (1 participant) suffered anal intercourse. The majority (89.5%) of the adolescents had experienced some form of coercion to engage in sexual activity, of which 42.1% reported force, 47.4% a threat of force, 21.1% were forced to use drugs and/or alcohol, 26.3% were given bribes or presents, 52.6% experienced verbal pressure, 10.5% heard threats to other family members, 10.5% were threatened with loss of love, and another 10.5% reported other forms of coercion. Of the 60% that disclosed the abuse at the time it was occurring, 66.7% of them reported it stopped as a result of the 39 disclosure. Eighty—five percent of the girls had attended therapy at least once since the sexual abuse had ceased. In comparison to published data, this sample reported a greater rate of sexual intercourse, more abuse perpetrated by a father figure, and a longer duration of abuse. However, the age at onset, use of force, and rate of disclosure were similar to national survey data (see Finkelhor et al., 1990). These differences are probably a result of the small number of participants included in the present statistics, but they indicate that the survivors in this study may have experienced more severe abuse than would be expected from national survey reports. Those adolescents who had been recruited through service providers specifically because of their abuse history tended to report more severe abuse (as measured by the CTQ), with some exceptions. Those survivors who were not actively recruited reported a range of abuse severity, including the more severe. This finding is not surprising, as one would expect that more severe abuse would create a greater need for professional services. Measures Abuse History. Each adolescent rated the amount of abuse she had experienced using the Childhood Trauma Questionnaire (CTQ), a 70-item self-report retrospective measure of child abuse and neglect (Bernstein, Fink, Handelsman, F oote, Lovejoy, Wenzel, Sapareto, & Ruggiero, 1994). Items on the scale begin with the phrase “When I was growing up,” and are rated on a 5-point Likert-type scale, ranging from “never true” 40 to “very often true.” For example, questions include, “When I was growing up, someone )3 6‘ in my family hit me or beat me, When I was growing up, I felt that I was loved,” and “When I was growing up, I believe that I was emotionally abused.” When tested by the authors, Cronbach’s alpha for the four factors of physical and emotional abuse, emotional neglect, sexual abuse, and physical neglect ranged from .79 to .94, indicating high internal consistency. In addition, the authors found good test-retest reliability over a 2- to 6- month interval (intraclass correlation=.88), and convergent validity with the Childhood Trauma Interview, an instrument developed by the authors. For the purpose of the present study, all non-sexual abuse questions were combined to form the “family violence” variable. Adolescents reporting sexual abuse on the CTQ (assessed by 6 items) also completed a Characteristics of Abuse questionnaire (adapted from Westerlund, 1992), to assess specific details about the abuse. The information obtained from this questionnaire was used to describe the sample, but was not used in quantitative data analyses, due to the low number of participants completing this questionnaire. Questions included those concerning the identity of the perpetrator, type of abuse experiences, and age at onset, among others. Subjects were asked to indicate which situations they had experienced, and to explain in more detail. No validity or reliability data are available for this measure, as it was created for the purpose of this study. See Appendix A for a copy of this measure. Coping strategies. The adolescents’ current coping strategies were measured in relation to a recent interpersonal stressor using the COPE (Carver, Scheier, & Weintraub, 41 1989). The measure was revised somewhat, with the original 60 items and an additional section in which adolescents were first asked to describe a recent fight with a best fiiend or dating partner, and to rate how upsetting the situation was, according to a 4-point Likert-type scale ranging from “Not at all upsetting” to “Very upsetting.” They then completed the 60-item scale of coping responses in reference to that Situation. For each item, subjects rated the extent to which they utilized each coping method on a scale of 1-4, from “I didn’t do this at all” to “I did this a lot.” Sample items include “I looked for something good in what was happening,” “I learned to live with it,” and “I got upset and let my emotions out.” When tested with middle adolescents (ages 14-18), Cronbach’s alpha reliability coefficients for the 15 subscales ranged from .51 - .87 (Phelps & Jarvis, 1994). Appendix B contains the version of the COPE used in this study. Interaajged model of relationships. Adolescents completed the Adult Attachment Scale (AAS, Modified version; Collins & Read, 1990). The AAS is an 18-item questionnaire in which respondents rate how characteristic each statement is of their feelings on a 5-point Likert-type scale ranging from “Not at all characteristic of me” to “Very much characteristic of me.” Statements include “I am comfortable depending on others,” “I am nervous when anyone gets too close to me,” and “I find it difficult to trust others completely.” Participants receive scale scores for the three attachment styles measured by this questionnaire: Secure, Anxious-Avoidant, and Anxious-Resistant. Factor analysis of the AAS reveals three dimensions underlying attachment style: comfort with closeness, degree to which one can depend on others, and anxiety about being 42 unloved or abandoned (Collins & Read, 1990). When analyzed by Collins and Read, Cronbach’s alpha for the three factors (Close, Depend, Anxiety) were .69, .75, and .72, respectively. Appendix C contains the version of the AAS used in this study. Intenfirsonal functioning. The quality of participants’ peer relationships was assessed using the Conflict in Relationships Scale (CIR; Wolfe, Reitzel-Jafi‘e, Gough, & Wekerle, 1994). The CIR is an 35-item measure of the frequency of physically, sexually, and emotionally abusive and non-abusive behaviors committed and/or experienced by the 9’ ‘6 respondent. Participants indicate if each behavior has occurred “never, once,” “2-3 times,” or “more than 3 times” in the past year during a conflict or argument. Items are repeated in 4 sections: Parts 1 and 2 reflect behaviors toward the respondent’s same-sex best fiiend and received from the best fiiend, respectively. Items include “Threatened to end the relationship,” “Gave in, but brought it up again later,” and “Gave reasons for my/her side of the argument.” For the purpose of this study, only behavior directed towards the best fiiend was assessed, as interest was in the participants’ interpersonal behavior. While behavior towards dating partner had originally been planned as part of the analyses, due to a low number of participants who had ever had a dating partner for at least 1 month (N=64), these data were collected but not analyzed. A revised version of the Network of Relationships Inventory (NR1; Furman & Buhnnester, 1985) measured adolescents’ perception of the conflict in, and quality of, ‘ their romantic and best fiiend relationships. The NR1 is a 30-item questionnaire assessing 10 relationship qualities: reliable alliance, enhancement of worth, instrumental help, 43 companionship, affection, intimacy, relative power of the child and other, conflict, satisfaction, and importance of the relationship. Participants rated how much a relationship quality occurred in each relationship, on a 5-point Likert-type scale ranging from “little or none” to “very much,” except in the case of relative power, which ranges from “I almost always do” to “they almost always do.” Items include “How satisfied are you with your relationship with this person?,” “How much does this person treat you like you’re admired and respected?,” and “How much do you and this person argue with each other?” In testing with children 11-13 years of age, Furman and Buhrmester found acceptable internal consistency (Mean Cronbach’s Alpha=.80). Due to poor measurement quality, the data from this measured were not analyzed. See appendix D for the modified version of the NR1. Psycholpgigcal functioni_ng. Adolescents completed the Children’s Depression Inventory (CD1; Kovacs, 1985) and the Trauma Symptom Checklist (TSC-33/40; Briere & Runtz, 1989; Elliott & Briere, 1992), as measures of psychological functioning. The CDI is a 27—item self-report scale measuring a variety of depressive symptoms. Each item consists of three choices, scored from 0 to 2 in the direction of increasing severity, and for each item, a participant selects the sentence that best describes her feelings for the past 2 weeks. For example, one item includes the statements “I hate myself,” “I do not like myself,” and “I like myself,” of which the respondent chooses one. A sum score is obtained which indicates severity of depressive symptoms. When completed by a sample of 8- to 13-year—old teens with psychiatric problems, the internal consistency, measured by 44 coefficient alpha, was .86 (Kovacs, 1985). A pediatric-medical outpatient group, tested 1 year after diagnosis with insulin-dependent diabetes mellitus, produced an alpha=.82. The TSC-40 is a 40-item scale assessing the traumatic impact of childhood sexual abuse, as it relates to psychological functioning. The measure yields six subscale scores (Anxiety, Depression, Dissociation, Sexual Abuse Trauma Index, Sexual Problems, and Sleep Disturbance), as well as a total score. Participants rate, on a scale fiom 0 (never) to 3 (often), how often each symptom has been experienced in the last two months. Items include headaches, stomach problems, sadness, and fear of men. In a national survey of 2,963 professional women, the measure was found to be reliable (alpha=.90), and to exhibit predictive validity, as each subscale score and the total TSC-40 score were significantly greater for abused than nonabused subjects. For the purpose of the present study, only the total score was utilized, indicating overall severity of traumatic symptoms. Demoggaphics. Family income, education, occupation, and marital status were compiled using a brief demographic questionnaire completed by the mother. See Appendix E. Procedures Adolescents were recruited both through a larger study and specifically for this study. For the larger Family Relationships Study, adolescents and their mothers each met with a trained interviewer for approximately 2 hours to complete the series of questionnaires. Both mothers and daughters were asked to sign an informed consent form, and questionnaires were administered separately to the adolescent and mother. 45 Interviewers were advanced undergraduates in psychology and clinical psychology graduate students. Adolescents and mothers were each paid $30 for their participation. Additional adolescents girls were recruited solely for participation in the present study. The initial procedures to arrange appointments and obtain consent were the same as the larger study. However, the mothers were only asked to complete a demographic questionnaire and wait in a separate room while the adolescents completed their questionnaires, which lasted approximately 1 1/2 hours. Adolescents were paid $20, and their mothers received $5 for their participation. Due to the sensitive nature of the questionnaires, particularly for adolescents who had been sexually abused, interviewers were prepared to provide support and resources to participants in need. The interviews did not appear to upset any of the participants, and many reported that they liked having the opportunity to talk about their abuse. Referral information was provided to all participants, regardless of their reports of abuse. RESULTS All scales included in statistical analyses had adequate reliability coefficients. Measures of coping and interpersonal conflict were factor analyzed prior to conducting structural equation modeling (SEM) analyses. For the COPE, previously published factors did not appear to fit the current data. Although the 15 subscales did arise fi'om an initial factor analysis of the data, the four factors established in the literature on adolescents were not supported when these subscales were then analyzed. Because the coping scales appeared to be highly correlated, an oblique principal component analysis was conducted with the 15 subscales from the COPE, producing 4 factors: active, avoidant, cognitive, and humor. The “humor” subscale loaded as its own factor, and was subsequently eliminated from further analyses. A second principal component analysis without this subscale produced the other three factors. The active scale consisted of strategies such as planning, seeking social support, focus on and venting of emotions, suppression of competing activities, and the subscale of active coping. The avoidant strategies included alcohol/drug use, behavioral disengagement, and denial. The cognitive factor consisted of strategies involving internal processes, such as restraint, positive reinterpretation and growth, religion, acceptance, and mental disengagement. Cronbach’s alpha reliability coefficients for the three coping factors were .84, .66, and .72, respectively. Table 1 lists the factor loadings for the 14 subscales on the 3 factors. 46 Table l COPE Principal Component Analysis 47 Active Avoidant Cognitive Planning .885 .005 .055 Active Coping .858 -.113 .055 Seeking Social Support: .790 .033 -.052 Instrumental Seeking Social Support: .691 .091 -.084 Emotional Focus on and Venting of .614 .091 .003 Emotions Suppression of Competing .410 -.029 -.409 Activities Alcohol/Drug Disengagement .047 .851 .312 Behavioral Disengagement -.O79 .696 -.339 Denial .165 .678 -.217 Mental Disengagement -. 135 .204 -.773 Restraint .084 .1 14 -.688 Positive Reinterpretation and .168 -. 139 -.658 Growth Acceptance .093 -. 102 -.63 8 Religion .33 l -.042 -.350 48 It is possible that different factors were obtained in this study because of the specificity of the stressor the adolescents were asked to consider. They were first asked to remember a recent conflict with a best fiiend or dating partner, or someone else who was close to them, and to respond in terms of how they had coped with this situation. While previous studies have looked at interpersonal stressors, this methodology was slightly different. Furthermore, the Phelps and Jarvis (1994) study was the only study using the COPE with an adolescent population, and it is possible that their results were simply not generalizable to all adolescents. Analyses of the measures of interpersonal fimctioning led to the elimination of one of the measures, the NR1, due to its poor measurement quality in this study. Specifically, it was found to not adequately measure the construct of “Interpersonal Conflict” when utilized in the structural equation analyses. The Conflicts in Relationships scale provided scores for both best fiiend and dating partner, but only 64 of the participants had ever been in a dating relationship, so this scale was dropped from the analyses due to the complexity of the model and the need for a greater sample size. Only the scores for behavior towards the best fiiend, versus behavior received from the best fiiend, were included, as this measure was more relevant to the hypotheses. The measure had previously been factor analyzed by Wolfe et a1. (1994) for the dating partner scales, but not for the best fiiend scales. A principal axis analysis of the best fiiend CIR scale produced three factors: abuse, positive communication, and negative communication. These scales were the same as those that had previously been found for the dating partner 49 scale. Alpha reliability coefficients for these factors were .91, .87, and .89, respectively. Interestingly, they were all significantly positively correlated (p<.0001). Correlation coefficients ranged from .40 (positive communication and abuse) to .62 (positive and negative communication). In the SEM analyses, negative communication was found to load highest on the factor and was therefore used to define the interpersonal conflict variable. Both abuse and positive communication also had positive loadings on the variable, which was expected due to the positive correlations. One potential explanation for the positive relation between the two types of communication is that, for adolescents, someone who communicates in one fashion tends to communicate in the other as well, rather than there being a strong distinction between positive and negative communication. The alpha reliability coefficients for the CTQ were sexual abuse=.93, physical abuse=.96, physical neglect=.95, and emotional neglect=.86. As the sexual abuse factor on this scale was the only measure of CSA, it was defined in the SEM analyses as having no error variance, thus having a loading value of 1.00. The other three CTQ scales were combined to form the family violence variable, with standardized loadings between .89 - .94. These variables were measured on continuous scales, versus categorical. The CSA variable was severely positively skewed, due to the majority of the participants reporting no history of sexual abuse. While this skew limits generalizability of the findings and imposes limitations on interpretation of the data, it is also somewhat representative of the population. A larger, random sample of females would produce a less skewed distribution, but there would still be a high anchor due to the number of girls and women 50 reporting no sexual abuse history. The effects of this skew are discussed further in the Limitations section below. In all of the models examined, attachment is seen as a positive variable, as it was defined by secure attachment, with the insecure attachment styles loading negatively on the factor. The scale reliability coefficients were .65 for secure, .64 for anxious-avoidant, and .77 for anxious-resistant. The two measures of psychological distress, the CD1 and the TSC-40, were highly correlated (r=.78) and had high standardized loadings on the distress variable, with values ranging from .78 to .99. Cronbach’s alpha for the CDI was .83, and .91 for the TSC-40. Table 2 contains the correlation matrix for the 15 observed variables that were analyzed, with significant alpha levels indicated. These correlations served as the basis for the SEM analyses, after being converted into covariances. Of particular interest are the correlations between CSA and the endogenous variables. Significant correlations were found with all three attachment styles, avoidant coping, and both measures of psychological distress (p<.01). Furthermore, avoidant coping was correlated with each attachment style (p<.05), and both measures of psychological distress (p<.01), and all attachment style and psychological distress measures were significantly correlated (p<.05). Only avoidant coping was significantly correlated with the measures of interpersonal conflict (p<.05). All SEM analyses were conducted using Lisrel 8 (JOreskog & SOrbom, 1993). A variety of Goodness of Fit indices can be examined in determining model fit. For the 51 purpose of this study, the Chi-Square statistic (x2), Goodness of Fit Index (GFI), Non- Norrned Fit Index (NNFI), and Comparative Fit Index (CF I) were used, as recommended by Hoyle and Panter (1995). The 12 is a direct derivation of the fitting firnction which indicates how much the model deviates from a perfect fit. The GFI is analogous to R2 used in multiple regression analyses, indicating the amount of variance accounted for by the model. The NNFI (or Tucker-Lewis index) estimates improvement of the target model over a baseline model, and the CF 1 indicates the reduction in lack of fit in a target model compared to a baseline model, as estimated by the noncentral xz. Overall, a small x2 is desired, the GP] should be greater than .90, and the NNFI and CFI are expected to near 1.00 when model fit is good. It is common in research utilizing SEM to estimate the number of subjects required based on either the number of variables or number of parameter estimates in the structural model. One common estimate is 8-10 subjects per parameter (N. Schmitt, personal communication, 1997). Based on this estimate, 80 participants was deemed adequate for analyses in this study. 52 .835: .Svfi. «ova... daemon—:58 o>EmoaanAEm ”connoasaaoo gnawoaucmznfim ”83m 25.5 muffin—m ”oarmfioonu 82qu «Enachuom... 59:02: commoaon Bog—EUUEU ”quoo oEEwoOUOOU ”wince E36>mHBO>< ”wince o>uoauhu< 358588 238n~50m~m 35:30.33 Efimfifiéaogmuhmummm 35:2083 EmEo>$§o§u>< oooe ooo.- mom. ow _.. mm... moo. 9:. BF. m>_.—o< ooo._. .mRr groovy .ommr mom: .mmm.- foam: maomw ooo. _. :Lom. Imam. Immm. Imwv. :Nvm. 56mm ooo.? t Em. :vmm. :me. zoom. >< Fo< mnoww ~2me > 82030 «O 20:38.80 N 033. 53 Attachment as a Mediator The hypothesis that attachment style serves to mediate the effects of childhood trauma on psychological distress and interpersonal conflict was partially supported by the data. For this model, the overall fit was x2 (48, N=80)=53.56, p>.05, GFI==0.91, NNFI=0.99, CFI=0.99 (See Figure 2). The squared multiple correlation (r2) for attachment was .29, and for psychological distress, r2=.45. For these variables, the model explained a significant part of the variance. A model generating approach was utilized in this and all other SEM analyses, in which the target model was respecified after initial estimation (Hoyle & Panter, 1995). For example, the model was tested including direct effects of the exogenous variables on psychological distress and interpersonal conflict, but no additional variance was accounted for and model fit was not improved, so these paths were eliminated. As indicated in Figure 2, the only paths that were significant (p<.05) were family violence to attachment (-0.44; t=-2.52), and attachment to psychological distress (-0.67; r—--3.49). Attachment was defined by secure attachment, thus family violence was negatively related to secure attachment. Also, secure attachment was negatively related to psychological distress. The other variables and paths were retained in the model, although not significantly related, as they were integral parts of the initial hypotheses. Coping as a Mediator Subsequently, coping strategies were examined as mediating variables. The initial model that was examined included all three coping factors as mediating variables; 54 however, the results of that analysis can not be interpreted as valid, as the number of paths exceeded what would be considered acceptable with a sample size of 80. Therefore, each coping strategy was tested in a separate model. For the avoidant coping model, the overall fit was x2(29, N=80)=3 9.20, p>.05, GFI=0.92, NNFI=0.97, CFI=0.98. (See Figure 3). Once again, not all paths were significant, but they were all retained to help with interpretation or for theoretical significance. Other non-significant paths that were eliminated following respecification were considered less important for interpretation, or their removal greatly improved model fit. The paths from avoidant COping to psychological distress (0.25; t=2.33, p<.05), and to interpersonal conflict (0.32; F277, p<.05) were both significant, indicating a positive relation. In addition, family violence was significantly related to psychological distress (0.43; F290, p<.05). It is evident that the model does not account for all factors in this process, with only psychological distress being explained adequately (12:27). Although other factors are involved, avoidant coping does increase psychological distress and interpersonal conflict, in addition to the increased psychological distress following family violence experiences. The path from CSA to avoidant coping was not significant, but it was positive. The model with cognitive coping also contained some significant findings (see Figure 4). The overall fit was x2(28, N=80)=32.76, p>.05, GFI=0.93, NNFI=0.98, CFI=0.99. Specifically, at the p<.05 level, the path from cognitive COping to psychological distress was significant (0.24; t=2.33), as was the path from family violence 55 to psychological distress (0.44; t=2.92). In addition to cognitive coping being positively related to psychological distress, there was also a similar trend with cognitive coping being positively related to interpersonal conflict, although it was not significant. In addition, the path from C SA to cognitive coping was not significant, but it was positive. R2 values were very low for all variables except for psychological distress (r2=.26). Active coping did not play any significant mediating role in the model, with no paths either to or from active coping achieving significance (Figure was thus not included). Specifically, neither the path from active coping to psychological distress (0.14, t=1.28), nor to interpersonal conflict (0.04, r-=0.33) was significant. The paths fiom CSA to active coping (0.17, t=1. 17) and from family violence to active coping (0.02, t=0.11) were also not significant. Furthermore, for active coping r2=.03, and r2 was equal to zero for interpersonal conflict. The Goodness of Fit statistics were high: x2(28, N=80)=29.29, p>.05, GFI=0.93, NNFI=1.00, CFI=1.00. Attachment and Copiagas Mediators Finally, models were run including both attachment and each type Of coping method. (They were also run separately for each coping strategy due to the limitations of the sample size). First, the model with attachment and avoidant coping was tested, resulting in Figure 5. The model fit was x2(57, N=80)=66.77, p>.05, GFI=0.90, NNFI=0.98, CFI=0.98. The high r2 values were attachment=.29 and psychological distress=.46. This model accounts for more of the avoidant coping variable than the previous model, with r2=. 12. The significant paths (p<.05) in this model were family 56 violence to attachment (-0.43, t=-2.50), attachment style to psychological distress(-O.64, t=-3.44), and avoidant coping to interpersonal conflict (0.32, t=2.77). Family violence is negatively related to secure attachment, which is negatively related to psychological distress, and avoidant coping is positively related to interpersonal conflict. Furthermore, although the paths were not significant, there was some indication of a negative relation between secure attachment and avoidant coping, and a positive relation between CSA and avoidant coping. The attachment and cognitive coping model can be seen in Figure 6. This model also had adequate fit indices: x2(5 7, N=80)=67.97, p>.05, GFI=0.90, NNFI=0.97, CFI=0.98. The significant paths, at the p<.05 level, were family Violence to attachment (-0.42, t=-2.45), attachment to cognitive coping (-0.30, t=-2.02), and attachment to psychological distress (-0.67, t=-3.43). The only substantial r2 values were for attachment (.27) and psychological distress (.44). As before, family violence was negatively related to secure attachment, which was negatively related to cognitive coping and psychological distress. The final attachment and coping model focused on active coping, but once again active COping did not appear to serve any significant role in relation to the other variables (Thus, the figure was not included). The path from attachment to active coping (-0.15, t=-1.03) indicates a negative relation between secure attachment and active coping, but it was not statistically significant. None of the other variables were significantly related to 57 active coping. However, the model did achieve good fit indices: x2 (57, N=80)=57.62, p>.05, GFI=0.91, NNFI=1.00, CFI=1.00. Coping Strategies of C SA Survivors A secondary focus of the study was the examination of coping in CSA survivors versus non-victims, according to level of psychological functioning. For these analyses, other types of abuse were not taken into account, due to the lack of significant correlations between the family violence factor and COping factors (only physical neglect and avoidant coping were significantly correlated; r=.22, p<.05), and the absence of a significant relation between these variables in the SEM analyses. First, a t—test was conducted to examine differences in CD1 and TSC-40 scores between the two groups. Significant difi’erences were found for both CDI and TSC-40 total scores (p<.001), with sexually abused girls reporting significantly more distress than nonabused girls (See Table 3). Next, a one-way ANOVA was used to compare the survivors and non-victims on the three types of coping strategies. A significant difference between groups was only found on the avoidant coping factor (p<.05), with sexually abused girls utilizing more avoidant coping strategies than nonabused girls (See Table 3). CSA survivors also reported more active and cognitive coping, but the differences were not Statistically significant. 58 Table 3 Differences Between C SA Survivors and Non-Victims (N=80) CSA No CSA CDI 14.31" 8.17 TSC-40 37.04“ 20.39 Active Coping 10.50 9.81 Avoidant Coping 7.09* 5.88 Cognitive Coping 10.12 9.51 *p<.05. **p<.OOI. To examine the differences in COping between groups incorporating both abuse history and psychological distress, multivariate analyses were conducted, using the MANOVA technique. The hypothesis was that CSA survivors would utilize different coping strategies than non-victims at comparable levels of firnctioning. In order to test this idea, psychological firnctioning was dichotomized by dividing all participants into low/high CDI groups, and low/high TSC-40 groups, based on median scores on each measure. The median score for the CDI fell in the range of mean scores for school-aged children exhibiting symptoms of depressed mood and affective disorders (Kovacs, 1985), and therefore appears to be a reasonable cut-Off point. The median TSC-40 score is 59 indicative of some distress, and fell between reported mean scores for nonabused and sexually abused women (Elliott & Briere, 1992). First, there was a main effect of CD1 scores, with those girls scoring high on the CDI reporting Significantly more avoidant and cognitive coping (p<.05). In addition, there was an interaction between C SA and CD1 score, with sexually abused girls with high CDI scores utilizing more avoidant coping than nonabused girls with high CDI scores, and than both groups with low CDI scores (p<.01). A second interaction was found between CD1 and TSC-40 scores, on both active and cognitive coping. For both types of coping, the Low CDI/Low TSC group scored significantly lower than the other three groups (p<.01). The final significant interaction, which was the test of the final hypothesis, occurred for avoidant coping, from the CSA*CDI*TSC interaction (p<.05; See Table 4). Post-hoc :- tests were conducted to determine significance between the CSA group and the Non-C SA group at each level of psychological distress. The greatest use of avoidant coping was reported by the sexually abused girls with High CD1 and Low TSC-40 scores, at a much higher rate than the comparative nonabused group, with high depression and low trauma symptoms. However, Levene’s Test for Equality of Variances was significant (p<.0001), and the subsequent significance test not assuming equal variances indicated that there was not a significant difference between the CSA and non-C SA groups (p=.338). Table 4 InteraIction of CSA, CD1, and TSC.40 on Avoidant Coping (N=79) CSA No CSA Low CDI/ Mean 5.00 5.75 Low TSC-40 SD .67 1.87 N 3 25 Low CDI/ Mean 6.17 5.92 High TSC-40 SD 1.82 1.73 N 4 8 High CDI/ Mean 10.33 5.58 Low TSC-40 SD 4.71 .90 N 2 8 High CDI/ Mean 7.46 6.31 High TSC-40 SD 2.31 1.38 N 16 13 DISCUSSION Attachment as a Mediator The analyses indicate that attachment style does serve as a mediating variable in the relation between family Violence and psychological distress, with a secure attachment style relating negatively to both factors, and apparently aiding in psychological adjustment following childhood abuse or neglect. The large amount of variance in psychological distress that is accounted for by attachment firrther emphasizes the importance of attachment in psychological development, and the difference that a secure attachment style can make in psychological firnctioning. Furthermore, the direct effects of childhood trauma were accounted for by the indirect effects through attachment, suggesting that attachment plays a vital role in adolescent psychological and interpersonal functioning. The mediating role of attachment can be interpreted as either a positive or a negative influence, depending on the quality of attachment style, as the anxious-avoidant and anxious-resistant styles relate positively to increased psychological distress. The quality of initial attachment can affect many aspects of an individual’s life. For instance, a secure attachment to the primary caregiver will potentially alleviate the distress that subsequent trauma will inflict, while an insecure attachment will not provide the necessary support and security to do so. Thus, a secure attachment allows one to adapt to change or deviation from “normal” development without developing psychopathology, 61 62 while insecure attachment serves as a risk factor for the development of childhood psychopathology (Rosenstein & Horowitz, 1996; Alexander, 1992). Previous research with adolescents has consistently provided evidence for a relation between insecure attachment and psychological symptomatology (Rosenstein & Horowitz, 1996). Much of this previous research has measured attachment style with the Adult Attachment Interview (AAI; George, Kaplan, & Main, 1984), while the results of this study provide firrther empirical evidence to support these hypotheses, using an alternative measure of attachment, the AAS (Collins & Read, 1990). Although family violence was not the primary focus of this study, it was measured due to the expected co-occurrence with CSA. The relation of family violence to attachment was not unexpected, as previous research has found a correlation between family violence and attachment style (Crittenden, 1992). All types of insecure attachment have been found in children who have experienced abuse or neglect, be it physical, sexual, or emotional abuse (Alexander, 1992). Recent research finding stability of attachment across the lifespan (see Main, 1996) suggests that an adolescent’s attachment style also reflects the influence of such negative childhood experiences, as was indicated in the present study. Although the relation between CSA and attachment style was not significant, it was negative, as expected. An insecure attachment style often develops prior to sexual abuse, particularly in cases of incest, although it may form as a result. Previous research on the effects of CSA on attachment has focused primarily on incest, and has established a 63 fairly stable negative relation between abuse and secure attachment, as well as other aspects of self firnctioning (Alexander, 1992; Cole & Putnam, 1992). Early sexual abuse can disrupt development of a sense of self, as well as subsequent self-competence and self integration (Cole & Putnam, 1992). This self development is an integral part of the attachment process, in terms of the internal working model and views of self and other. In a situation of ongoing sexual abuse, an insecure attachment style often develops, which involves the development of a negative sense of self in the internal working model. While much of the evidence for these disruptions has come from work with survivors of incest, this study included both intrafamilial and extrafamilial abuse survivors, with the hypothesis that extrafamilial CSA also has a negative effect on the attachment system. The sample size is insufficient to separate these two groups in the analyses, but it is expected that a significant relation would have been found had only incest survivors been included in the sample, as was true in previous research. Future research might focus specifically on extrafamilial abuse, to determine if this relation is significant. Even though the path is not statistically significant, a secure attachment style appears to have the potential to compensate for the negative effects of CSA on psychological functioning, and the mediating effects of attachment style should be considered in future research. Specific problems with the measurement of CSA that may have contributed to the lack of significance are discussed below. 64 Coping as a Mediator The influence of various coping strategies indicates some potential mediating effects, for both psychological and interpersonal functioning. Avoidant coping strategies appear to increase with the experience of CSA, and are then positively related to psychological distress and interpersonal conflict. The positive relation with CSA is not significant; however, it is fairly strong. Interestingly, family violence had very little effect on avoidant coping, and the small amount of variance in avoidant coping explained by this model indicates that there are other factors contributing to its use. The positive relations between avoidant coping and psychological distress and interpersonal conflict were expected, as avoidant coping methods are usually viewed as maladaptive, particularly in the context of an interpersonal conflict. Thus, the increased use of avoidant coping strategies by CSA survivors can help account for their increased psychological distress and interpersonal conflict. Obtaining a greater number of CSA survivors might produce a significant relation between CSA and avoidant coping, at which point its role as a mediator would be determined. Avoidant coping did account for more of the variance in psychological functioning in comparison to the direct effects of CSA, although not compared to those of family violence. Previous use of avoidant coping strategies has been found to predict subsequent avoidant coping (Ebata & Moos, 1994). Avoidant coping methods are at times the only means of escaping sexual victimization, so the positive relation between CSA and avoidant coping supports the idea that coping with previous abuse affects how one copes with 65 interpersonal stressors. Furthermore, although avoidant coping may be the most adaptive Option in dealing with the sexual abuse itself, it becomes maladaptive as it is used to cope with other stressors, such as peer relationships, as is indicated by the greater distress and conflict exhibited by the adolescents in this study. Cognitive coping showed effects similar to those of avoidant coping, which was somewhat unexpected. Some of the strategies that contributed to this factor were restraint, positive reinterpretation and growth, and religion, which one might expect would be adaptive when dealing with an interpersonal stressor. However, other strategies included acceptance and mental disengagement, which may be maladaptive at times, explaining the contribution to increased distress. The fact that these strategies loaded on one factor was interesting in itself, and difficult to interpret. However, they can all be viewed as internal processes, utilizing cognitive skills. Although in some situations these strategies might be useful, one can see that when dealing with an interpersonal conflict they could potentially increase distress by internalizing the conflict. In previous research, the highest risk for psychological dysfirnction occurred with the use of an avoidant coping style, especially cognitive avoidance (Johnson & Kenkel, 1991; Spaccarelli, 1994). More active coping strategies are often considered to be beneficial because they provide the individual with more of a sense of control (Spaccarelli, 1994). Thus, one reason for the deleterious effects of both avoidant and cognitive coping may be the lack of a sense of control that accompanies the use of these methods. 66 Similar to avoidant coping, cognitive COping was positively related to CSA (but not significantly), with practically no relation to family violence. Overall, it appears that there are more complex processes, and effects beyond childhood trauma, involved in the development of coping methods. Family violence is strongly related to an increase in psychological distress, regardless of type of coping strategies utilized. This finding supports the literature on the effects of experiencing violence on children and adolescents. The family violence factor was a composite of the physical abuse, physical neglect, and emotional neglect scales on the CTQ. Thus, this factor includes anything from witnessing domestic violence to being physically abused oneself. A substantial amount of research has examined the effects of various aspects of family violence on a child, particularly in the realm of psychological functioning. In a meta-analytic study, Weaver and Clum (1995) found a significant efi'ect of interpersonal violence, including family violence, on psychological distress in participants ranging from 6 to 41 years of age. In terms of witnessing family violence, a variety of both internalizing and externalizing problems have been reported in children. Some of these reported problems that were also measured in this study include depression, psychosomatic complaints, fears and phobias, insomnia, and nightmares (F antuzzo & Lindquist, 1989). Gauthier, Stollak, Messe, and Aronoff (1996) found experience of neglect to be a strong predictor of psychological distress in young adults, even in the absence of any physical abuse. Additional research, using a measure similar to the TSC-40 used in the present study, found that the variance in trauma symptoms in adolescents, such 67 as depression, anger, anxiety, dissociation, posttraumatic stress, and total trauma were explained by exposure to violence (Singer, et al., 1995). Specifically, having been a witness or victim of family violence was strongly related to total trauma, and to anxiety, dissociation, stress, and depression. The measurement of family violence in this study encompassed all of these aspects that have been studied previously, so the strong effect of family violence on psychological distress concurs with these findings. Attachment and Cop'mgas Mediators In combination, attachment style and coping both influence psychological and interpersonal firnctioning, and attachment affects the relation to coping from CSA and family violence, although not all paths were significant. First, the model examining attachment and avoidant COping illustrates some interesting relations between these two factors and the others. The negative path from attachment to avoidant coping neared significance, indicating that insecure attachment styles are positively associated with avoidant coping, while a secure attachment style is negatively related to avoidant coping strategies. Due to the maladaptive influence of avoidant coping, interpersonal conflict then increases, although it appears that the negative relation of secure attachment to psychological distress is more significant than that which is mediated by avoidant coping. Again, the paths from CSA to attachment and coping did not reach significance, but they were in the expected directions, with CSA negatively related to secure attachment, and positively related to avoidant coping. Combining CSA and attachment accounted for more of the avoidant coping variance than CSA alone, providing evidence that attachment 68 does serve a role in the process of developing coping strategies. Once again, attachment mediated the effects of family violence on psychological distress, with a secure attachment relating negatively to both family violence and distress. The model including attachment and cognitive coping provides slightly different information than that of avoidant coping. This model did not predict variance in the constructs as well, with the influences on cognitive coping and interpersonal conflict not adequately explained. There is a significant negative relation between secure attachment and cognitive coping, which was somewhat unexpected. However, for the same reasons that were discussed above in examining the increased distress associated with cognitive coping, one can see how an insecure attachment might lead one to utilize more internal coping mechanisms, especially in the context of interpersonal conflict. In particular, many of the cognitive strategies may serve to avoid seeking outside support or interaction, which could result fiom an insecure attachment style. Similar to the previous model, there is a stronger association between secure attachment and psychological distress than between cognitive coping strategies and distress. However, increased interpersonal conflict is positively related to cognitive coping, although the relation was not statistically significant. The integration of attachment style and coping has been previously examined by Crittenden (1992), in the context of child abuse and maltreatment. She proposed that children’s coping strategies vary as a function of their internal representational models, which develop following maltreatment or abuse. Her model explains how experiences 69 with the attachment figure lead to expectations about firture interpersonal situations, affecting coping, and also how coping with these early experiences can then influence subsequent coping, especially in similar conflict situations. Experiences with neglectfirl or abusive parents both lead to the development of an insecure attachment style and maladaptive coping strategies. As a result of the internal working model, the COping strategies that were adaptive in dealing with an adverse environment may be continued in later interpersonal relationships, but they are often no longer adaptive. Rather, they are Often detached or avoidant. One assumption of the present study was that coping with early abuse experiences would affect coping strategies employed during adolescence, which can be theoretically explained in terms of the internal working model and the influence of attachment style. Thus, the positive relation between insecure attachment styles and maladaptive coping strategies was as expected. Overall, it appears that attachment style is strongly associated with amount of psychological distress, and weakly associated with interpersonal conflict, as measured in this study. Alternatively, avoidant and cognitive coping strategies are more strongly associated with interpersonal conflict than psychological distress. These findings were somewhat surprising, as it was expected that attachment would influence interpersonal firnctioning, and that coping would have more of an effect on psychological firnctioning than results indicate it did. One explanation involves how the constructs were measured; psychological functioning was measured primarily in terms of internal distress, while interpersonal functioning was measured as behavior towards the best fiiend. Attachment 70 style is an internal process, and may therefore have been more strongly associated with the internal processes of psychological distress. Alternatively, coping was measured in the context of an interpersonal conflict that had occurred in the adolescent’s life, potentially leading to the stronger correlation with the measure of behavioral conflict. Furthermore, adolescents’ attachment styles and internalized models of relationships may not have been outwardly expressed in the self-report of behavior in a conflict situation. The fact that coping was limited to this interpersonal conflict may have also afi'ected the results, as the context of a stressor influences the relation between type of cOping and subsequent behavior problems (Compas, Malcame, & F ondacaro, 1988). Thus, the measurement may have limited reports of psychological distress in relation to coping, due to the focus on behavioral problems. In addition, family violence has a significant negative influence on the development and/or maintenance of a secure attachment style, but little effect on the types of coping strategies utilized to deal with an interpersonal stressor. Although the influence of CSA on the other variables is inconclusive, there are interesting trends. First, CSA appears negatively related to secure attachment, implying that the experience of CSA might lead to the development of an insecure attachment style. Second, CSA is positively related to the more “maladaptive” coping strategies, avoidant and cognitive. This finding can be interpreted in a number of ways, and will be discussed below in reference to the final hypothesis. 71 Copirg Strategies of CSA Survivors It was hypothesized that C SA survivors utilize different coping strategies than nonvictims, although they may be firnctioning at the same levels. Interest in this facet of the study resulted from the labeling of common strategies used in the context of sexual abuse as maladaptive, and the possibility that these strategies may not be maladaptive given such a traumatic situation. Adolescent girls who had been sexually abused appeared more distressed in general, and reported greater use of avoidant coping strategies, both findings being expected. In addition, girls experiencing more depressive symptomatology also utilized more avoidant and cognitive cOping strategies than those girls with low depression scores. When analyzing the groups at comparative levels of functioning, it was found that sexually abused girls with higher scores on the depression index (CDI) reported more avoidant coping strategies than nonabused girls at the same level of depression. However, abused and nonabused girls with low CDI scores (no or little depressive symptomatology) reported similar levels of avoidant cOping. For active and cognitive coping, there was an interaction between CD1 and TSC-40 scores, indicating a greater use of both types of coping at higher levels of distress, for all participants. The examination of the three-way interaction (CSA*CDI*TSC-40) affecting avoidant coping indicated that CSA survivors utilize more avoidant coping at higher levels of distress. However, the results were inconclusive, due to the small sample sizes and unequal variances in certain cells. Potentially, with larger cell sizes this finding might reach significance. Overall, it appears 72 that greater psychological distress is associated with greater use of a variety of coping strategies. Combined with the structural models, these findings imply that adolescent girls who have experienced childhood sexual trauma employ more coping strategies to deal with an interpersonal conflict than those girls who have not been victimized, but the fact that they are more distressed indicates that they are not using these coping methods effectively. It is not possible based on the information obtained in this study to determine if the ineffectiveness is a result of their using inappropriate strategies for dealing with interpersonal conflict, thus making them maladaptive, or if other factors are contributing to this result. Whatever the reason, these girls appear to be trying to cope in any way possible, but they are not succeeding in dealing effectively with these stressors. Survivors of incest have been found to utilize more coping strategies than extrafamilial CSA survivors, possibly due to the fact that it often continues over a longer period of time, or potentially reflecting the greater distress these victims feel, and their attempts to end the abuse by any means (DiLillo, Long, & Russell, 1994). The victimized participants in this study, although approximately half intrafamilial and half extrafamilial abuse survivors, may have been exhibiting similar attempts to cope with their distress. It is also interesting to note the similar types of coping Used by the two groups at low levels of distress. It may be that these survivors learn that the types of coping strategies available during the time of victimization are different from what is available in “normal” times of stress, and adapt accordingly. For the more distressed girls, the pattern 73 of ineffective coping may simply continue following the cessation of the sexual abuse. Another possibility is that fewer coping strategies are required to deal with stress in the lives of girls with little psychological distress. Alternatively, distressed adolescents who have been sexually abused tend to employ more coping strategies, and those adolescents persistently utilizing avoidant or other maladaptive strategies do not adapt as well, increasing their distress (Ebata & Moos, 1994). The number of participants from the non-C SA group falling in the high CDI/high TSC cell is also of interest. As discussed, family violence did not influence amount or type of coping reported. However, a history of family violence did have a strong effect on increasing psychological distress. Therefore, it is likely that this group is made up of girls who have experienced other forms of childhood trauma, causing high levels of depression and trauma symptoms. Interestingly, they still do not utilize avoidant coping strategies to the same extent as the C SA survivors, possibly because a different process is involved in dealing with physical abuse and/or neglect. An additional possibility is that the family violence was still occurring in the households of these participants, while there was no ongoing sexual abuse reported, and different coping strategies are required to deal with an ongoing trauma or stressor. Overall, the hypothesis was partially supported by the data. Specifically, at high levels of psychological distress, CSA survivors and nonvictims utilized different coping strategies when faced with an interpersonal conflict. Causation can not be determined by the available data, that is, whether maladaptive coping causes distress or vice versa. In 74 other words, is avoidant coping causing greater distress, or does someone who is more distressed use more avoidant coping strategies? They likely affect each other, causing a maladaptive cycle that may be complicated even more by a history of sexual abuse. Limitations Attachment style and preferred coping strategies did serve some mediating role between abuse history and current psychological distress and interpersonal conflict, but it is not possible to draw any definitive conclusions based on the results of this study alone. A variety of problems, primarily related to measurement error, render the data difficult to interpret. The two most prominent issues were skewness and the measurement of coping, which will be discussed briefly. In structural equation modeling, there is an assumption of normality, or that the distribution of each observed variable is normal. Violation of this assumption imposes certain limitations and causes problems in interpretation of analyses. The CSA construct was measured by only one observed variable, the CSA score from the CTQ. This variable was severely positively skewed, with a skewness statistic of 1.442 (SE=0.269). The mean score was 8.7 (sd=6.18), with a range of 5-25. (See Figure 7). The strong positive skew is due to the number of participants reporting no experience of sexual abuse, which earned a score of 5. As is evident in Figure 6, such a large number of sexually abused adolescents would have been necessary to achieve a normal distribution (over 500), that it was not possible given time, recruitment, and other resource constraints. Furthermore, a random sample of adolescents would never produce a normal distribution, due to the greater 75 number of girls who have not been sexually abused. One effect of the skewed distribution is that some of the scores at the high positive end became outliers, so that they might have had more influence on analyses than normal, and they had the potential to affect model fit, parameter estimates, and standard errors (West, Finch, & Curran, 1995). Combined with a small sample size, nonnormality Often leads to higher Type I error rates, and analyses that fail to converge or result in an improper solution. In the present study, these problems did occur, leading to model modifications and respecification until a viable solution was obtained. Fit indices, such as the Tucker and Lewis Index and the Comparative Fit Index, are often underestimated in cases of nonnormality, as are standard errors of parameter estimates (West, Finch, & Curran, 1995). As a result, the tests of parameter estimates should not be considered trustworthy, and should be viewed as tentative. This potential also explains the failure of particular fit indices to reach 1.00, as would be expected. 76 Mean=8.7 SD=6.18 6 7.6 10 12.6 16 17.6 20 22.6 26 Sexual Abuse Experienced Figure 7 Rate of Reported CSA Experiences As CSA was a central focus of this study, the fact that this variable possessed so many potential problems is of great concern. Although correlations are also affected by skewness, they are not as sensitive to nonnorrnality as SEM. Thus, some trends in the models can be supported by significant correlations, assisting in interpretation of the data. Many of the expected relations that were not significant in the models were supported by significant correlations, particularly those between CSA and other variables (See Table 2). Although measurement error appears to have affected many aspects of the study, the measurement of coping in particular was a difficult task that was not completely resolved. As discussed, the COPE measure has been used, and has established and acceptable reliability and validity in both adult and adolescent samples. However, it does 77 not seem appropriate for the current population of interest. In this study, the adolescents were asked to specifically consider a recent interpersonal conflict, so one explanation is that the coping strategies offered by the COPE were not completely applicable. It is also possible that in the situation of an interpersonal conflict, adolescent girls employ such a variety of coping strategies that factors could not be easily distinguished. One additional factor is that coping was assessed through self-report, as is usually the case in this type of research. Participants’ reporting of their coping strategies might have been affected by a variety of influences, including the type of COping they may have employed during the process of discussing personal information with a stranger. For example, it seems logical that an individual who utilizes denial to cope with interpersonal conflict is not likely to report so when asked, whether due to unconscious or conscious processes. It is clear fiom examining the available literature on adolescent coping that more work needs to be conducted to try to formulate a model of coping that incorporates both theoretical and empirical evidence. Adolescence is a unique developmental period, with its own stressors and conflicts, particularly surrounding social and interpersonal issues. In addition, the coping strategies available to an adolescent are different from those that were available during childhood, and different from those of an adult, which contributes to the measurement problems in this area of research. The lessons learned from this empirical work can be applied within a developmental framework to improve upon the conceptualization of adolescent coping. For instance, this study provides evidence of a connection between coping and attachment, 78 one theoretically-based model that could be examined further. The merging of theory and empirical evidence is a receiving more emphasis in psychological research, and is particularly important in this area of adolescent development. The issue Of self-report arises again in reference to abuse history. Although the procedures were designed to make it as comfortable as possible for the adolescents to report a history of sexual (and other) abuse, it is suspected that several participants had been sexually abused and did not report so. The CTQ was given to the adolescents to complete themselves, rather than having the interviewer record their responses, in order to limit feelings of intrusiveness and to try to make them more comfortable in answering honestly. However, it was the experience of the interviewers that some of the adolescents may not have endorsed any items reflecting negatively on their family or any aspect of their childhood, when other responses implied that they had some negative life experiences (such as high trauma symptoms). Obviously, there is no means of confirming this suspicion, although the situation of mandated reporting of child abuse often leads to underreporting in research, making it likely that it also occurred in this study. Utilizing reported cases of abuse through Child Protective Services is one means of ensuring reporting; however, one is then limited to reported cases, which can cause other problems in generalizing results. Future Directions The models derived in this study should be considered tentative until firrther research can be conducted in this area, as the results were largely based on the 79 idiosyncrasies of the sample, and all of the final models were obtained through modification and respecification of the originally tested models. Not only is replication recommended, but research which includes a greater number of C SA survivors is necessary to eliminate the skewness effects, and to enable more conclusive analysis of the data. In order to achieve a normal distribution of this variable, one has a number of options to consider. One possibility is to only include CSA survivors, with the expectation that the range of abuse experiences will be well-distributed, as was the case with the 26 survivors in the present study. Or, if there is interest in including girls who have not been sexually abused, the sample size must be increased to accommodate the number of girls that will anchor the low end of the distribution. It is often difficult to obtain large numbers of participants, particularly in this age group, and to obtain those who are willing to discuss their abuse experiences, but it would be useful to do so. The strength of many of the correlations indicate that the proposed mediating processes are worth examining further, particularly the role of attachment style. It is believed that coping is also a valuable area of study, although its measurement remains difficult. This study was conducted not only to identify mediating variables that explain variability in functioning among CSA-survivors, but to also identify potential points of intervention for adolescent girls who have been victimized. The important firnction of a secure attachment style is one area to consider; for instance, interventions can be based on a theoretical framework such as attachment. Allen, Hauser, and Borman-Spurrell (1996) found that severe psychOpathology in adolescence is predictive of an insecure attachment 80 style 11 years later, Often accompanied by a lack of resolution of trauma. This information emphasizes the need for early intervention, to aid in the resolution process. Although attachment is somewhat stable, it is apparent that it can be disrupted by traumatic events, and should therefore also be amenable to positive change following intervention. Furthermore, intervention could focus on coping strategies, and teaching adolescents to reduce their use of avoidant coping methods when it has proven to be maladaptive. Thus, not only can future research expand on this integration of theory into empirical testing, but results can be interpreted so that they may be utilized in applied settings, to assist victims and survivors in the recovery process. This study added to the small research base using adolescent participants, providing valuable information during this developmental period. The information obtained from the girls about their sexual abuse was particularly valuable, in an area of research based largely on retrospective reporting. While reports were still retrospective, there was less of a chance of forgetting due to time lapse, as Often occurs when conducting studies with adult women. Also evident is the lack of research on adolescent psychological and interpersonal functioning, and tying these processes into a developmental fiamework, especially in the context of attachment. Adolescence appears to be a time when the effects of CSA and other childhood trauma may become more salient, and may exert a strong influence on social development. Understanding the processes involved is one way to improve survivors’ chances of adapting to these developmental changes without negative consequences. While the results of this study 81 must be considered preliminary, they do provide evidence for areas that should be included in future research. FIGURES 82 a .HEO as BREE .movs. 6300 + EBOOI 0.3 _ AZ..— mmEo .2320 .2320 2. mm. mm. «confined Hango>< Insomxc< -uaoaxut. 835:0 U «cannon; a — u m fini \vw; no; 3. 2e :5 «no 8:0“: < 3. mo. _ .5... on. on. 3053 D ~fl0mfl0~0£0~hmm ocsoom nv. 83:62 a me 2 m Engage. lawn... a. goo—moz “wowwoz ousn< Jon—m . 5m , .nme mm. mm. mm. 00:23 > Ea: evvf < “om—mac U mmoo _e meagre“: co; .mv. o_. . {NM \ HOD winoo aaaomo>< oo. mo. 83:82 a me 5 .00 EwEo>< goo—moz on “09:92 ..»Ea m o :mwm ausn< cone—3" vo.- on. a: a: mm on. 333.1 mmmHU \. mm. OVN. , 33qu unan< Hoodmuz .manm .nxgm Scam no. «a. I... 02::on mmOO ooze—om> a: a: co.— wE a c U 03::on Ho; 85 ooUEO 82882 8 8 co 8an>< can 2 m 86883. as 88-95 wove... is; o a «co—moz Sou—moz 5?: ”MM“. .3 8m no. 8800 + 8800- on=p< mo. mmMHU hmdno hmmuo / o O G 0— Om > 2:33;... Eo>< ”WNW—”MW -naomxcx‘ ouaoom %—~ Edna 32.23 U 3:022:35 /~..- ”3.- Z. .2..- “no 8no~3< 3:qu wouBo>< mm cc. 03.2300 mmOU ooé 884802 «430300 26.300 2m 23m 808:... 0. 0.3. -<.._ goo—moz J08”..— “038.01% 30082;.2 unnum0>< -uaomxat. ocaoom ”emf on..- wcaou 0>£0w00 Li .59. en. Io 0.8mm ace—woz . 0252‘. .AEE ..aaa .7 0N0... 00:23 > 3:5; 5. 00.0 co.— APPENDIX A APPENDIX A Please answer the following questions to the best of your knowledge. If there is anything you cannot recall or anything that does not apply to you, please indicate this. Please take your time and ask the interviewer if you do not understand a question. 1. How old were you when the sexual abuse began? 2. What was the relationship of the offender(s) to you (please specify gender if necessary)? 3. What was the age difference between you and the offender(s)? 4. Did the offender(s) use particular forms of coercion either to engage in sexual activity with you or to keep you fi'om telling anyone? If yes, please describe under the appropriate choice(s): a. Force: b. Threat of force: . Gave you drugs and/or alcohol: 0 d. Bribes/Presents: Verbal Pressure: .0 f") Threat of hurting other family members: . Threat of loss of love: 00 h. Other: 87 88 5. What kinds of sexual acts did the offender(s) engage in with you? Please describe fully in your own words under the appropriate choice(s). a. Sexual activity not involving direct physical contact: b. Kissing and/or fondling: c. Touching of male and/or female genitalia: d. Oral-genital contact: e. Simulated intercourse (no penetration): f. Vaginal intercourse: g. Anal intercourse: h. Other: 6. How often did the abuse occur? Please describe under as many choices as necessary to include variation based on differing sexual acts. a. Once: b. Several times a year: c. Monthly: (1. Weekly: e. Daily: f. Other: 7. How long did the sexual abuse continue? (Please circle the appropriate choice, and explain if necessary.) a. less than 1 year b. 1-3 years c. 3-5 years d. more than 5 years 89 8. At the time of the sexual abuse, did you tell anyone about it? If yes, a. Who? b. Did the abuse stop? What was the outcome? 9. Did you feel like you had any control over the abuse? Please explain: 10. What were some things you did to try to cope with the abuse? (For example, pretended it wasn’t happening, ran away, daydreamed, used drugs, hurt yourself, etc.) 11. After the abuse stopped were you ever raped or assaulted by adifferent offender? Please explain. 12. Have you ever been in therapy or counseling? Ifyes, a. Individual, group, or both? b. How often (weekly, etc)? 0. For how long (months, years, etc)? (1. Are you presently in therapy? APPENDIX B APPENDIX B We are interested in how people respond when they confront difficult or stressful events in their lives. There are lots of ways to try to deal with stress. Try to remember a time in the last 2 months when you fought with your best friend or boyfriend! girlfriend. Ifyou can not remember a recent incident, think of what it generally feels like when you are fighting with an important person in your life. Briefly describe the incident (who you were fighting with, what the fight was about, etc): On the following scale, rate how upsetting this situation was for you: 1 2 3 4 Not at all A little Somewhat Very upsetting upsetting upsetting upsetting Indicate what you did, or generally do, when you experienced this situation. Please answer every item. There are no “right” or “wrong” answers, so choose the most accurate answer for YOU—not what you think most people would say or do. l=l didn’t do this at all 2=I did this a little bit 3=I did this a medium amount 4=I did this a lot 1. I tried to grow as a person as a result of the experience. 2. I turned to work or other substitute activities to take my mind off things. 3. I got upset and let my emotions out. 4. I tried to get advice from someone about what to do. 5. I concentrated my efforts on doing something about it. 6. I said to myself “this isn’t real.” 7. I put my trust in God. 8. I laughed about the situation. 9. I admitted to myself that I couldn’t deal with it, and quit trying. 10. I restrained myself from doing anything too quickly. 11. I discussed my feelings with someone. 12. I used alcohol or drugs to make myself feel better. 13. I got used to the idea that it happened. 90 14. 15. l6. l7. l8. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 91 1=I didn’t do this at all 2=I did this a little bit 3=I did this a medium amount 4=I did this a lot I talked to someone to find out more about the situation. I kept myself from getting distracted by other thoughts or activities. I daydreamed about firings other than this. I got upset, and was really aware of it. I sought God’s help. I made a plan of action. I made jokes about it. I accepted that this had happened and that it couldn’t be changed. I held off doing anything about it until the situation permitted. I tried to get emotional support from friends or relatives. I just gave up trying to reach my goal. I took additional action to try to get rid of the problem. I tried to lose myself for a while by drinking alcohol or taking drugs. I refused to believe that it had happened. I let my feelings out. I tried to see it in a different light, to make it seem more positive. I talked to someone who could do something concrete about the problem. I slept more than usual. I tried to come up with a strategy about what to do. I focused on dealing with this problem, and if necessary let other things slide a little. I got sympathy and understanding from someone. I drank alcohol or took drugs, in order to think about it less. I kidded around about it. I gave up the attempt to get what I wanted. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 92 1:1 didn’t do this at all 2=1 did this a little bit 3=l did this a medium amount 4=I did this a lot I looked for something good in what was happening. I thought about how I might best handle the problem. I pretended that it hadn’t really happened. I made sure not to make matters worse by acting too soon. I tried hard to prevent other things from interfering with my efforts at dealing with this. I went to the movies or watched TV, to think about it less. I accepted the reality of the fact that it happened. I asked peOple who had similar experiences what they did. I felt a lot of emotional distress and I found myself expressing those feelings a lot. I took direct action to get around the problem. I tried to find comfort in my religion. I forced myself to wait for the right time to do something. I made fun of the situation. I reduced the amount of effort I put into solving the problem. I talked to someone about how I felt. I used alcohol or drugs to help me get through it. I learned to live with it. I put aside other activities in order to concentrate on this. I thought hard about what steps to take. I acted as though it hadn’t even happened. I did what had to be done, one step at a time. I learned something from the experience. I prayed more than usual. APPENDIX C APPENDIX C PART 2 Questions #6 to 23 Please read each of the following statements and rate the extent to which it describes your feelings about romantic relationships. If you have never been involved in a romantic relationship, answer in terms of how you think you would feel. Answer questions #6 to 23 using the scale below. 1 2 3 4 5 Not at all Somewhat Fairly Mostly Very Characteristic of Me 6) I find it relatively easy to get close to people. 7) I find it difficult to allow myself to depend on others. 8) In relationships, I often worry that my partner does not really love me. 9) I find that others are reluctant to get as close as I would like. 10) I am comfortable depending on others. 11) I do no_t worry about someone getting too close to me. 12) I find that people are never there when you need them. 13) I am uncomfortable being close to people. 14) In relationships, I often worry that my partner will not want to stay with me. 15) When I show my feelings for people, I’m afraid they will not feel the same about me. 16) In relationships, I ofien wonder whether my partner really cares about me. 17) I am comfortable developing close relationships with others. 18) I am nervous when anyone gets too close to me. 19) I know that people will be there when I nwd them. 20) I want to get close to people but I worry about being hurt by them. 21) I find it difficult to trust others completely. 22) Often, people want me to be closer than I feel comfortable being. 23) I am not sure that I can always depend on people to be there when I need them. 93 94 Note: There were changes made from the original Collins and Read (1990) version of the AAS in the wording of some items, as well as adding or eliminating items from the original Aas for the modified AAS used in the current study. The changes were not made by the present investigator, but had already been made on the version of the AAS used in the study. Below is an explanation of the differences between the two version. Secure 6) I find it relatively easy to get close to people. (“Others” in original version changed to “people”). 11) I do not worry about someone getting too close to me. (“Ofien” removed from original). 17) I am comfortable developing close relationships with others. (“Having others depend on me” changed to “developing close relationships with others”). 19) I know that people will be there when I need them. (“Others changed to “people”). Items from the original AAS Secure scale not used in the modified Secure scale: I do not often worry about being abandoned. I am comfortable having others depend on me. Avoidant 7) I find it difficult to allow myself to depend on others. (Added on “to allow myself”). 12) I find that people are never there when you need them. (Original: “People are never there when you nwd them”). 13) I am uncomfortable being close to people. (Original: “1 am somewhat uncomfortable being close to others”). 18) I am nervous when anyone gets close to me. (Added “to me”). 22) Often, people want me to be closer than I feel comfortable being. (Original: “Often, love partners want me to be more intimate than I feel comfortable being”). Resistant 8) In relationships, I often worry that my partner does not really love me. (Added “In relationships”). 14) In relationships, I often worry that my partner will not want to stay with me. (Added “In relationships”). 16) In relationships, I often wonder whether my partner really cares about me. (Item added). 23) I am not sure that I can always depend on people to be there when I need them. (Changed “others” to “people”). APPENDIX D APPENDIX D Please answer each of the following questions for your relationship with your dating partner and your best friend of the same sex. For example. the first question asks: “How much free time do you spend with this person?” Ifyou spend very much free time with your best friend, circle 5 on that line. If you spend little or no free time with your best friend, circle I on that line. Circle a number for each person asked about. If you do not have a current dating partner or best friend, please answer the questions about your most recent dating partner or best friend of the same sex. Sometimes the answers for difference people might be the same, but often they are different. 1. How much free time do you spend with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 2. How much do you and this person get upset with or mad at each other? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 3. How much does this person teach you how to do things that you don’t already know how to do? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 4. How satisfied are you with your relationship with this person? Best Friend I 2 3 4 5 Dating Partner 1 2 3 4 5 5. How much do you tell this person everything? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 6. How much do you help this person with things he/she can‘t do by him/herself? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 95 96 7. How much does this person like or love you? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 8. How much does this person punish you? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 9. How much does this person treat you like you’re admired and respected? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 10. Who tells the other person what to do more ofien, you or this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 11. How sure are you that this relationship will last no matter what? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 12. How much do you play around and have fun with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 13. How much do you and this person disagree and argue? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 14. How much does this person help you figure out or fix things? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 97 15. How happy are you with the way things are between you and this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 16. How much do you share your secrets and private feelings with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 17. How much do you protect and look out for this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 18. How much does this person really care about you? Best Friend I 2 3 4 5 Dating Partner 1 2 3 4 5 19. How much does this person discipline you for disobeying him or her? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 20. How much does this person treat you like you’re good at many things? Best Friend I 2 3 4 5 Dating Partner 1 2 3 4 5 21. Between you and this person, who tends to be the boss in this relationship? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 22. How sure are you that your relationship will last in spite of fights you have with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 98 23. How often do you go places and do enjoyable things with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 24. How much do you and this person argue with each other? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 25. How often does this person help you when you need to get something done? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 26. How good is your relationship with this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 27 . How much do you talk to this person about things that you don’t want others to know? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 28. How much do you take care of this person? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 29. How much does this person have a strong feeling of affection (love or liking) toward you? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 30. How much does this person scold you for doing something you’re not supposed to do? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 99 31. How much does this person like or approve of the things you do? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 32. In your relationship with this person, who tends to take charge and decides what should be done? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 33. How sure are you that your relationship will continue in the years to come? Best Friend 1 2 3 4 5 Dating Partner 1 2 3 4 5 APPENDIX E APPENDIX E THE ADOLESCENT SURVEY PROJECT 1) First name of the ADOLESCENT participating in the survey 2) Please list the members of your family (include yourself, the adolescent participating in the survey, and all member of your household): First Name Relationship .338 ABE Livingwith you? ”SQMPP’NT‘ 3) Marital status [ ] Single [ ] Living with partner [ ] Married [ ] Separated (how long? ) [ ] Divorced [ ] Widowed 4) Who has custody of the children? 5) Race or ethnic group? you adolescent [ ] [ ] Native American I I [ ] Asian [ ] [ ] Black, African-American [ ] [ ] Latino, Hispanic-American [] [ ] Biracial (mixed): specify [ I [ I White I I [ ] Other IOO 101 6) What is the highest level of education you have completed? [ ] Grade school or less [ ] Some high school [ ] High schol degree/GED [ ] Some college [ ] College degree [ ] Some graduate school [ ] Graduate degree 7) What grade is your adolescent currently in? 8) Which one of the following best describes your occupation? Artist, writer, designer, crafisperson Farmer, agricultural worker Homemaker Manager, administrator Professional: specify type Technician, skilled worker Student Serrriskilled or unskilled worker White-collar (sales, clerical, secretary) Retired Unemployed Other HHHF‘HHHHl—HHHHF—l HHHHHHHHHHHH 9) Are you working at this time? [ ] Yes Hours per week? [ ] No 10) What was your total income last month? 1 1) Which of the following best describes your religious affiliation? [ ] Protestant (what type? ) [ ] Catholic [ ] Jewish [ ] Muslim [ ] Atheism/Agnosticism [ ] No religious affiliation 102 12) How long have you been living in your current housing? less than 1 month 1 to 3 months 4 to 6 months 7 to 12 months 1 to 3 years more than 3 years r—Hr-Ht—Hr—ir—tr—i 13) How long have you lived in this state? 14) How many times have you moved since your adolescent was born? [ ] once [ ] 2 to 3 times [ ] 4 to 5 times [ ] more than 5 times 15) Have you ever stayed in a shelter before? [ ] Yes, how many times? I ] No 16) If Yes, what kind of a shelter did you stay in? [ ] For battered women [ ] For homeless women [ ] Other 17) Have you ever been in an abusive relationship? 18) If yes to 17, what are the dates of that relationship(s). Indicate the number of monts for each relationship (both dates and months). REFERENCES REFERENCES Alexander, PC. (1992). Application of attachment theory to the study of sexual abuse. Journal of Consulting and Clinical Psychology, 60(2), 185-195. Allen, J .P., Hauser, S.T., & Borman-Spurrell, E. (1996). Attachment theory as a framework for understanding sequelae of severe adolescent psychopathology: An ll-year follow-up study. Joumflf Consulting and Clinical Psychology, 64(2), 254-263. Allen, S. & Hiebert, B. (1991). Stress and coping in adolescents. Canadian Journal of Counseling, 25(1), 19-32. Ammican Psychiatric Association. (1994). Diagnostic and gatistical manual of mental disorders (4th ed). Washington, DC: Author. Beitchman, J.H., Zucker, K.J., Hood, J.E., daCosta, G.A., & Akman, D. (1991). A review of the short-term effects of child sexual abuse. Child Abuse & Neglect. 15, 537- 556. Bernstein, D.P., Fink, L., Handelsman, L., Foote, J., Lovejoy, M., Wenzel, K., Sapareto, E., & Ruggiero, J. (1994). Initial reliability and validity of a new retrospective measure of child abuse and neglect. American Journal ofPsychiatry, 151(8), 1132-1136. Bowlby, J. (1973). Michment and loss: Vol. 2. Separatiom anxiety and anger. New York: Basic Books. Briere, J ., & Runtz, M. (1988). Post sexual abuse trauma. In G. E. Wyatt & G. J. Powell (Eds), Lasting Effects of Child Sexgrl Abuse (pp. 85-99). Newbury Park, CA: Sage Publications. Briere, J ., & Runtz, M. (1989). The Trauma Symptom Checklist (TSC-33): Early data on a new scale. Journal otlnterpersong Violence, 4, 151-163. Browne, A., & F inkelhor, D. (1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99(1), 66-77. 103 104 Carver, C.S., Scheier, M.F., & Weintraub, J .K. (1989). Assessing coping strategies: A theoretically based approach. Jorml of Personality and Social Psychology, 5_6_(2), 267-283. Classen, C., Koopman, C., & Spiegel, D. (1993). Trauma and dissociation. Bulletin of the Menninger Clinic, 57(2), 178-194. Cole, P.M., & Putnam, F.W. (1992). Efl‘ect of incest on self and social fiinctioning: A developmental psychology perspective. Journal of Consulting and Clinicaljsychology, 69(2), 174-184. Collins, N.L., & Read, 8.]. (1990). Adult attachment, working models, and relationship qualities in dating couples. Jourg of Personality and Social Psychology, 3(4), 644-663. Compas, BE. (1987). Coping with stress during childhood and adolescence. Psycholgg’cal Bulletin, 101(3), 393-403. Compas, B.E., Malcame, V.L., & F ondacaro, KM. (1988). Coping with stressful events in older children and young adolescents. Journal of Consultingand Clinifl Psychology, 56(3), 405-41 1. Crittenden, PM. (1992). Children's strategies for coping with adverse home environments: An interpretation using attachment theory. Child Abuse & NgLect, 16, 329-343. Davies, J .M., & Frawley, MG. (1994). Treating the adult survivor of childhood sexualfiabuse: A psychoa_nalytic perspective. New York: Basic Books. DiLillo, D.K., Long, P.J., & Russell, L.M. (1994). Childhood coping strategies of intrafamilial and extrafamilial female sexual abuse victims. Journal of Child Sexual Abuse, 3(2), 45-65. Ebata, A.T., & Moos, RH. (1991). Coping and adjustment in distressed and healthy adolescents. J cm of Applied Developmental Psychology, 12, 33-54. Ebata, A.T., & Moos, RH. (1994). Personal, situational, and contextual correlates of coping in adolescence. Journal of Research on Adolescence, 4(1), 99-125. Elliott, D.M., & Briere, J. (1992). Sexual abuse trauma among professional women: Validating the Trauma Symptom Checklist-40 (TSC-40). Child Abuse & Neglect, 16, 391-398. 105 Fantuzzo, J .W., & Lindquist, C.U. (1989). The effects of observing conjugal violence on children: A review and analysis of research methodology. Journal of Family Violence 4(1), 7 7-94. F inkelhor, D., & Browne, A. (198 5). The traumatic impact of child sexual abuse: A conceptualization. American Jougl of Orthopsychiatry, 55(4), 530-541. Finkelhor, D., Hotaling, G., Lewis, I.A., & Smith, C. (1990). Sexual abuse in a national survey of adult men and women: Prevalence, characteristics, and risk factors. Child Abuse & Neglect, 14, 19-28. Furman, W., & Buhrmester, D. (1985). Children’s perceptions of the personal relationships in their social networks. Developmental Psychology, 21(6), 1016-1024. Gauthier, L., Stollak, G., Messe, L., & Aronoff, J. (1996). Recall of childhood neglect and physical abuse as differential predictors of current psychological functioning. Child Abuse and Neglect, 20, 549-559. George, C., Kaplan, N., & Main, M. (198). Att_achment interview for adults. Unpublished manuscript, Department of Psychology, University of California, Berkeley. Herman, J .L. (1992). Complex PTSD: A syndrome in survivors of prolonged and repeated trauma. Journal of Traumatic Stress 5(3), 377-391. Hoyle, R.H., & Panter, AT. (1995). Writing about structural equation models. In RH. Hoyle (Ed), Structural equation modeling: Concepts, issues, and applications (pp. 158-176). Thousand Oaks, CA: Sage Publications. JanofllBulman, R (1992). Shattered assumptions: Towards a new psyfichology of traunra. Johnson, B.K., & Kenkel, MB. (1991). Stress, coping, and adjustment in female adolescent incest victims. Child Abuse & Neglect, 15, 293-305. Joreskog, K.G., & Sorbom, D. (1993). Lisrel 8 user’s reference gtride. Chicago: Scientific Software International, Inc. Kendall-Tackett, K.A, Williams, L.M., & Finkelhor, D. (1993). Impact of sexual abuse on children: A review and synthesis of recent empirical studies. Psychological Bulletin, 113(1), 164-180. Kovacs, M. (1985). The Children’s Depression, Inventory (CDI). Psychopharmacology Bulletin, 21(4), 995-998. 106 Lazarus, RS. (1993). Coping theory and research: Past, present, and future. Psychosomatic Medicine, 55, 234-247. Lindberg, F.H., & Distad, L.J. (1985). Survival responses to incest: adolescents in crisis. Child Abuse & Neglect, 9, 521-526. Luster, T., & Small, SA. (1995). Sexuaflruse history and problemfl adolescence: Explonngahe effects of moderating variables. Paper presented at the National Council on Family Relations, Portland, OR Main, M. (1996). Introduction to the special section on attachment and psychopathology: 2. Overview of the field of attachment. Journal of Conaulting and Clinical Psychology, 64(2), 237-243. McElroy, LP. (1992). Early indicators of pathological dissociation of sexually abused children. Child Abusaand Neglect, 16, 833-846. Newberger, C.M., & DeVos, E. (1988). Abuse and victimization: A life-span developmental perspective. American Journfi of Orthopsychiatry, 58(4), 505-511. Orr, D.P., & Downes, MC. (1985). Self-Concept of adolescent sexual abuse victims. Journal of Youth and Adolescence, 14(5), 401-410. Phelps, S.B., & Jarvis, PA. (1994). Coping in adolescence: Empirical evidence for a theoretically based approach to assessing coping. Journal of Youtmd Adolescence, Q0), 359-371. Polusny, M.A., & Follette, V.M. (1995). Long-term correlates of child sexual abuse: Theory and review of the empirical literature. Applied & Preventive Psychology, 4, 143-166. Price, M. (1993). The impact of incest on identity formation in women. Journal of The Amman Academof Psychoanalysis, 21(2), 213-228. Putnam, F.W. (1991). Dissociative disorders in children and adolescents: A developmental perspective. The Psychiatric Clinics of North America, 14(3 ), 519-531. Putnam, F.W., & Trickett, PK. (1993). Child sexual abuse: A model of chronic trauma. Psychiatry, 56, 82-95. Rosenstein, D.S., & Horowitz, HA. (1996). Adolescent attachment and psychopathology. Journal of Conaultmgfld Clinical Psychology, 64(2), 244-253. 107 Roth, 8, & Cohen, L.J. (1986). Approach, avoidance, and coping with stress. American Psychologist, 41(7), 813-819. Rowan, A.B., & Foy, D.W. (1993). Post-traumatic stress disorder in child sexual abuse survivors: A literature review. Journal of Traumatic Stress 6(1), 3-20. Singer, M.I., Anglin, T.M., Song, L.Y., & Lunghofer, L. (1995). Adolescents’ exposure to violence and associated symptoms of psychological trauma. The Journal of the American MedigAssociation, 273(6), 477-482. Spaccarelli, S. (1994). Stress, appraisal, and coping in child sexual abuse: A theoretical and empirical review. Psychological Bulletira116(2), 340-362. van der Kolk, B. (1987). Psychological trauma. Van Gijseghem, H., & Gauthier, MC. (1994). Links between sexual abuse in childhood and behavioural disorders in adolescent girls: A multivariate approach. Canadian Journal of Behavioural Science, 26(3), 339-352. Weaver, T.L., & Clum, GA. (1995) Psychological distress associated with interpersonal violence: A meta-analysis. Clinical Psychology Review, 15(2), 115-140. West, S.G., Finch, J .F., & Curran, P]. (1995). Structural equation models with nonnorrnal variables. In RH. Hoyle (Ed), Structural equation modeling: Concepts, issues, and applications (pp. 56-75). Thousand Oaks, CA: Sage Publications. Westerlund, E. (1992). Women’s sexuath after childhood incest. New York: W.W. Norton. Wolfe, D.A., Reitzel-Jafl'e, D., Gough, R, & Wekerle, C. (1994). Conflicts in relationships: Measuring physical and sexuafloercion among youth. Available from the Youth Relationships Project, Department of Psychology, The University of Western Ontario, London, Canada, N6A 5C2.