UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The Rosebush picture sort : a diagnostic technique to differentiate sexually abused children from other… Carter, Mary Ann Sheller 1994

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

Item Metadata

Download

Media
831-ubc_1994-893354.pdf [ 9.78MB ]
Metadata
JSON: 831-1.0053989.json
JSON-LD: 831-1.0053989-ld.json
RDF/XML (Pretty): 831-1.0053989-rdf.xml
RDF/JSON: 831-1.0053989-rdf.json
Turtle: 831-1.0053989-turtle.txt
N-Triples: 831-1.0053989-rdf-ntriples.txt
Original Record: 831-1.0053989-source.json
Full Text
831-1.0053989-fulltext.txt
Citation
831-1.0053989.ris

Full Text

THE ROSEBUSH PICTURE SORT: A DIAGNOSTIC TECHNIQUE TO DIFFERENTIATE SEXUALLY ABUSED CHILDREN FROM OTHER CHILDREN by MARY ANN SHELLER CARTER M.A., University of Minnesota, 1971  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE  REQUIREMENTS FOR THE DEGREE OF DOCTOR OF EDUCATION in  THE FACULTY OF GRADUATE STUDIES Department of Counselling Psychology  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA March, 1994 C) Mary Ann Sheller Carter, 1994  In presenting this thesis  in partial fulfilment  degree at the University of  of  the  requirements  for  an advanced  British Columbia, I agree that the Library shall make it  freely available for reference and study. 1 further agree that permission for extensive copying of  this thesis for  department  or  by  his  or  scholarly purposes her  may be granted by the head of my  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allovi'ed without my written permission.  (Signature)  Department of  ^ ^ O g N f^eWX <vx^ Y a - ^ c V A nrj^  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  b^\'v:)^\^H  11  ABSTRACT This study  (N = 123) contrasted a group of  sexually abused children in treatment, aged 6 to 12 years, with two comparable groups—other non-sexually abused children in treatment, and non-abused, nontreatment children—to determine whether differences in Rosebush Picture selection could be demonstrated.  All  children underwent evaluation procedures that included completion of a 12 picture selection (the Rosebush Picture Sort) and the Culture-Free Self-Esteem Inventory-2 (CFSEI-2).  A demographic questionnaire for  each child was completed.  Results showed no  differences in picture selection or self-esteem scores of children at the beginning or end-treatment stage. Sexually abused children's Rosebush Picture Sort (RPS) selections at mid-treatment were different than both comparison groups. The self-esteem scores of the two clinical groups in mid-treatment were significantly lower than the scores of the non-abused, non-treatment group, but did not differ from one another.  The  findings indicate that sexually abused children identify Rosebush Picture Sort pictures differently than other children.  It is possible that sexual abuse  affects the internal working model of children differently than other types of abuse and trauma.  Ill  TABLE OF CONTENTS TITLE PAGE ABSTRACT TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ACKNOWLEDGEMENTS CHAPTER 1:  INTRODUCTION  Page i ii iii vi vii viii 1  Nature of the Problem 1 Previous Research Problems 3 Child Sexual Abuse Assessment Obstacles . . . . 4 Definition of Terms 7 Justification of Study 10 Impact of Sexual Abuse on Personality . . . 10 The Conceptual Model 12 Evaluation of Sexual Abuse Effects from a .... Projective Perspective 14 Using Projective Measurement with Children. . 15 Summary 18 CHAPTER 2:  LITERATURE REVIEW  21  Sexual Abuse Impact Findings 22 Sexual Abuse Assessment and Interview . . . . 28 Comparative Studies of Sexually, Physically, NonAbused Children 30 Methodology and Data Analysis Questions . . . 41 Projective Techniques in Assessment 43 Projective Testing of Children 46 The Use of Drawings as Projective Techniques . 49 Drawings as a Diagnostic Technique with Sexually Abused Children 53 Limitations of Drawings as a Diagnostic Tool in Assessment 55 Rosebush Research . 56 Crandall/Allan Rosebush Study . 56 Bowden Rosebush Study . . . 57 Carter/Allan/Boldt Rosebush Study . . . . . 59 Rosebush Q-Sort Pictures . . . 60 Results 61 Limitation of Rosebush Studies . . . 66 Hypotheses 67 CHAPTER 3:  METHODOLOGY  Method Design of the Study Child Participants  70 70 70 71  IV  Therapists 73 Procedures 74 Instruments 76 Demographic Questionnaire . . . 77 Rosebush Picture Sort (RPS) 78 Culture-Free Self-Esteem Inventory (CFSEI-2)80 Data Analysis 83 Summary 88 Null Hypotheses 89 CHAPTER 4:  RESULTS  .  91  Phase One 91 Sample Characteristics 91 Reason for Referral 92 Gender 96 Age 98 Parent 98 Ethnic Heritage 101 Geographic Area 103 Gender Multivariate Analysis 105 Treatment Group Analysis 105 Phase Two 106 Hypothesis 1 107 Rosebush Picture Sort Beginning-Treatment 107 Rosebush Picture Sort Mid-Treatment . . . 108 Rosebush Picture Sort End-Treatment . . . 113 Hypothesis 2 114 Self-Esteem Beginning-Treatment 115 Self-Esteera Mid-Treatment . . . 117 Self-Esteem End-Treatment 117 Informal Findings 118 Culture-Free Self-Esteem Inventory-2 (CFSEI-2) Informal Findings 123 Summary of Findings 123 CHAPTER FIVE:  DISCUSSION AND CONCLUSIONS  127  Discussion Phase One Sample Characteristics . . . . . . Reason for Referral . . . . . . Gender Age Parent Ethnic Heritage . . . . . . Geographic Area Phase One Data Analysis Treatment Group Analysis . Phase Two Data Analysis . Hypothesis 1 Rosebush Picture Sort Beginning-Treatment Rosebush Picture Sort Mid-Treatment . . .  127 127 127 128 128 129 130 131 132 132 132 132 133 136  V  Rosebush Picture Sort End-Treatment . . . Hypothesis 2 . Culture-Free Self-Esteem Inventory . . . . Self-Esteem Beginning-Treatment Self-Esteem Mid-Treatment Self-Esteem End-Treatment Limitations of Study Conclusions Future Research  139 141 141 142 143 144 145 148 149  REFERENCES  152  APPENDICES  168  Appendix A. Preliminary Study Bar Graph . . . 169 Appendix B. Research Approval Forms 171 B-1 UBC Certificate of Approval . . . 172 B-2 B.C. Superintendent of Ministry of Social Services Letter of Approval 173 B-3 B.C. Manager Child and Youth Mental Health Services Letter of Approval 174 Appendix C. Therapist Letters and Permission Forms 175 C-1 Introductory Letter 176 C-2 Therapist Permission Letter . . . 177 C-3 Therapist Instruction Letter . . 178 C-4 Parent Permission Letter . . . . 180 Appendix D. Instrumentation Used In Study . 181 D-1 Demographic Questionnaire . . . . 182 Descriptions of Abuse Categories 183 Attachment Classifications . . . 185 D-2 Rosebush Sort Instructions . . . 186 Rosebush Picture Answer Sheet . . 187 Rosebush Pictures 188 D-3 Culture-Free Self-Esteem Inventory 190 Appendix E. Rosebush Picture Sort Statistics 192 E-1 Phase One Mean & Standard Deviation Scores 193 E-2 Phase Two Mean & Standard Deviation Scores 194 E-3 Chi-Square Beginning-Treatment . 196 E-4 Chi-Square Mid-Treatment . . . . 198 E-5 Chi-Square End-Treatment . . . . 200 Appendix F. Child Participant Comments . . . 202 Appendix G. Therapist Comments . 213 G-1 Beginning-Treatment Comments . . 2 1 4 G-2 Mid-Treatment Comments 215 G-3 End-Treatment Comments 217 G-4 Therapist CFSEI-2 Comments . . . 219  VI  LIST OF TABLES Page  Table 1  Abuse Classification Group 1 & Group 2  Table 2  Demographic Variable - Gender  97  Table 3  Demographic Variable - Age  99  Table 4  Demographic Variable - Parent  Table 5  Demographic Variable - Ethnic Heritage  . 102  Table 6  Demographic Variable - Geographic Area  . 104  Table 7  Results of Discriminant Function Analysis, Rosebush Picture Sort Variables Ill Self-Esteem Descriptive Statistics/ANOVA 116  Table 8  . 93  100  V l l  LIST OF FIGURES Page Figure 1  Child Participant Selection Diagram . . .  71  Figure 2  Phase One Data Analysis Design  84  Figure 3  Phase Two Data Analysis Design  Figure 4  All Groups Discriminant Function Scatterplot 110  Figure 5  Self-Esteem Lie Frequency Graph . . . . . 115  Figure 6  Rosebush Comment Frequency Graph  .....  86  . . . . 119  Vlll  ACKNOWLEDGEMENTS I would like to extend my sincere appreciation to Dr. John Allan, for providing his consistent support, supervision, and encouragement throughout the duration of this study. I also wish to extend my gratitude to my committee members. Dr. Beth Haverkamp and Dr. Walter Boldt, for providing me with judicious editorial and statistical advice. The academic challenges and interactions I have experienced with the Counselling Department faculty and students have broadened my professional and personal growth in unexpected and serendipitous ways. I give my thanks to the government officials of the province who approved this study. I give my heartfelt thanks to the psychologists and therapists throughout the province who enthusiastically participated in this study, by not only collecting data from their clients, but by offering additional information that proved invaluable in the data analyses. And equally importantly, I send a warm message of thanks to the 130 children whose personal contributions to this study may help provide clinicians with more effective assessment and treatment techniques. Numerous friends and colleagues have listened patiently at various stages of this study, and have offered constructive advice, often woven with humour. My appreciation is especially extended to Ms. Marie Jose Dhase, Dr. Elizabeth Huntsman, Ms. Catherine Malysheff, and Dr. Joyce Ternes. I would like to assure them my professional and personal pursuits extend beyond the rosebush garden. Family members have shown interest in the research topic and support of my academic endeavors. I am grateful to my children. Holly, Steven and Emma, for their on-going love, respect, and reassurance. And I owe a special debt of gratitude to my husband, Ian, who has stood beside me sharing the irreplaceable gifts of marriage, acceptance, goodness, integrity, and commitment throughout my personal and academic journey.  CHAPTER 1  INTRODUCTION  Nature of the Problem  "Child sexual abuse is a violation of a child's body, mind, and spirit" (Bagley & King, 1991, p.34). The short and long term effects of sexual abuse on a child's body, cognitive ability, and sense of identity can no longer be under estimated. In the past ten years, the problem of identifying children who have been sexually abused, and assessing the type and degree of that abuse, has become a major challenge for social service agencies, the criminal justice system, and society as a whole (Berliner & Conte, 1993; Kendal1-Tackett, Williams & Finkelhor, 1993; Melton & Limber, 1989; Myers, 1993; Rogers, 1990; Sgroi, 1989).  The societal consequences of early  sexual trauma are far reaching.  Children who have been  sexually abused engage, as adults, in a wide variety of anti-social and self-destructive behaviours, such as prostitution, sexual and physical assault, and drug abuse (Finkelhor & Browne, 1985; MacFarlane & Krebs, 1986).  Victims of sexual abuse inevitably carry the  scars of their abuse with them into adulthood, and many  2 victims suffer from sexually related interpersonal problems for the rest of their lives (Conte, 1985). Instances of childhood sexual abuse are difficult to substantiate since the act typically occurs in the privacy of the home or in a secluded environment purposely chosen by the abuser to ensure anonymity. Existing methods of assessing the sexually abused have not proved effective, yet a detailed and comprehensive psychological assessment is critically important since the adversarial legal system demands objective and corroborated information in order to determine whether or not a crime has been perpetrated.  A definite need  exists for research which will clearly identify specific underlying effects of child sexual abuse, and which will direct professionals towards a more effective and efficient means of providing a comprehensive sexual abuse assessment (Bagley & King, 1991; McGovern, 1993; Robin, 1993; Steller, 1992; Waterman & Lusk, 1993).  This study has as its purpose  the development and application of an instrument which will be able to distinguish sexually abused children from other non-abused children, or from children who may have been abused, but non-sexually. The remainder of this chapter outlines previous research problems, including the difficulties in providing accurate sexual abuse assessment; definitions  3 of terms; the justification for this study;  the  study's conceptual model; and projective test dynamics. A summary is provided.  Previous Research Problems  In an attempt to identify sexual abuse syndromes, earlier sexual abuse assessment studies have utilized children's drawings, as well as questionnaires and check lists that request adults or children to identify various behaviours and attitudes characteristic of the sexually abused child.  Although these studies pursued  valid questions, many of the selected measurements have been standardized on non-abused samples and the tests have yielded mixed results (Robin, 1993; Waterman & Lusk, 1993).  The empirical studies have not been  replicated, and retrospective reports have been found to have low reliability, being shaped by external and internal factors (Brewin, Andrews & Gotlib, 1993). Further compounding the problem, it has been shown that there are no identifiable socio-economic variables that distinguish sexually abused children and their families from non-abused children and their families (Farber & Egeland, 1987; Finkelhor, 1993; Friedrich, Grambsch, Broughton, Kuiper, & Beilke, 1991).  4 The measurements used in sexual abuse assessment studies primarily concentrate on an analysis of the drawings of children, adult perceptions of the child's behaviours and attitudes, and the child's own selfperceptions.  The findings suggest the absence of any  specific sexual abuse syndrome or pattern, as well as the absence of a single traumatizing process. No single symptom of sexual abuse has been found, although low self-esteem is one of the most frequently cited characteristics of the sexually abused child (KendallTackett et al., 1993; Wolfe & Gentile, 1992).  Child Sexual Abuse Assessment Obstacles  Typically, child sexual abuse assessment procedures and techniques are similar to the general procedures and techniques used in child psychological assessments.  The assessment protocols that  professionals are encouraged to use with sexually abused children (American Professional Society on the Abuse of Children, 1990; College of Psychologists of British Columbia, 1990; Steller, 1992; Wakefield & Underwager, 1989) fail to take into account, however, the possible differing emotional, psychological, and sexual development of the sexually abused child.  The  major drawback of these assessment protocols is that  5 they require the child to respond in the very manner that the experience of sexual abuse has inhibited. For example, the child may have been threatened by the offender if she shares any sexual abuse information. Researchers agree that sexually abused children are less likely to disclose their abuse than non-abused children are to invent misleading disclosures.  At the  same time, however, sexually abused children are highly accurate in what they report, but are liable to make numerous errors of omission (Berliner, 1991; Friedrich, 1990; Smith, 1992; Steward, Bussey, Goodman, & Saywitz, 1993).  Regardless of these current facts, earlier  suggested assessment procedures (Steller, 1992; Wakefield & Underwager, 1989) continue to emphasize unsolicited verbal disclosures from these children. Assessment criteria of this type make it difficult for distressed, sexually abused children to come forward with an account of their abuse.  Children  become inarticulate and confused when experiencing conflicting emotions (Garbarino, Stott & Faculty of the Erikson Institute, 1989; Riordan & Verdel, 1991). Furthermore, the sexually abused child's sense of trust in adults has been violated by both the ageinappropriate sexual intimacy with the offender, and by the failure of the adult care-giver to protect the child from the abuse.  The shame and guilt felt by the  6 child tends to produce defensive withdrawal and a denial of the abusive experience (Campis, Hebden-Curtis & Demeso, 1993).  Children find various means to avoid  recalling the abusive events they experienced (Friedrich, 1990; MacFarlane & Krebs, 1986), for example, denying it happened to them; they also find ways to minimize their low self-esteem by responding in a socially acceptable manner and ignoring their internal feelings (Campis et al., 1993; Kendal1-Tackett et al., 1993). A further problem arises when court personnel, police officers, or social workers who interview sexually abused children do not understand the child's communication, and misinterpret or misread the message contained in the child's play or story (Baartman, 1992; Garbarino et al., 1989).  Sexually abused children  often relate abuse experiences in a manner that is not specific enough for adults to understand. Suggested interview protocols in sexual abuse assessments encourage the use of open ended questions. Elementary school-age children have often not developed the fairly high degree of coherent comprehension and expressive ability that this type of questioning requires (Friedrich, 1990).  A further disadvantage of  this kind of questioning is that it relies on stereotypical responses (Berliner & Conte, 1993).  7 Moreover, the adults who interview sexually abused children often lack the training required to understand and to interpret the latent meaning of the verbal communications given (Horner, Guyer & Kalter, 1993). As a result, adult judgments may be faulty and they may make decisions before sufficient information is available and alternative theories or interpretations have been considered (McGovern, 1993). Generally speaking, current child sexual abuse assessment protocols are inadequate for understanding sexually abused children, and do not provide a set of standardized procedures that can measure the effects of sexual abuse on the personality development of the child.  Definition of Terms  Dissociation.  Dissociation is a defensive ego  function, in which the mind fragments portions of the self in order to survive.  Often, dissociation results  in the separation of thinking and feeling, for example when a child reports a terrifying experience in a flat, unemotional, monotone voice.  External events that  threaten people, psychologically and physically, call for this function (Hartman & Burgess, 1989).  8 Ego Structure.  Ego structure is the internal process  that integrates the pressures of personal impulses and conscience with perceptions from external reality (Sugarman, 1992). Emotional Abuse.  Emotional abuse is the performance of  acts or omission of acts by those responsible for the care of a child.  The process is likely to undermine a  child's self-image, sense of worth and self confidence (Province of British Columbia, Ministry of Social Services, 1989). Internal Working Model.  The internal working model is  an individual's psychological organization; consisting of basic attachment needs, thoughts, feelings, memories, and the use of defense mechanisms.  The  schematic image of self develops from how one is responded to and how these actions are internally processed.  These conditions begin to operate at an  unconscious level and include more than one internal working model as one develops (Crittenden, 1992). Physical Abuse.  Physical abuse is direct and indirect  behaviour that physically harms a child (hitting, using force, spilling hot water on a child), as well as failure to intervene when others are aggressive against a child (Garbarino, Guttman & Seeley, 1986). Primary Process Thinking.  Primary process thinking is  a separate process that develops simultaneously with  9 logical and neutral thought and is conceptualized as a subtype of affect-laden cognition (Dubowski, 1990; Fischer & Pipp, 1984; Russ & Grossman-McKee, 1990).  It  often involves images, symbols, fantasy and metaphors (Allan, 1988; Allan & Bertoia, 1992). Projective Techniques.  Projective techniques are  assessment methods that are founded on the broadly based principle of projection; the techniques utilize a stimulus in order to tap into the unconscious aspects of the participant's personality (Rabin, 1986). Rationalization.  Rationalization is the process in  which an individual substitutes a socially acceptable explanation of their conduct in place of the real reason (Chaplin, 1975). Self-esteem.  An individual's perception of his/her own  worth (Battle, 1992). Sexual Abuse.  Sexual abuse is an act of sexually  exploiting another person (Cavanaugh-Johnson, 1992). The sexual contact/interaction between the two individuals is performed solely for the purpose of fulfilling the needs of the instigator.  Acts of sexual  abuse include sexual exposure, threatened sexual assault, unwanted sexual activity including kissing, fondling, and attempted or actual anal, oral, or vaginal penetration (Rogers, 1990).  10 Justification of Study Impact of Sexual Abuse on Personality Development  The impact of sexual abuse on the personality development of children can be seen in cognitive and emotional distortions, and low self-esteem (deYoung, 1992).  Children cannot logically consent to age-  inappropriate sexual activities, as they do not understand to what they are consenting and they lack freedom of choice.  Children are also unaware of the  social meaning of sexuality (Finkelhor, 1979).  There  is growing evidence that the experience of sexual abuse may alter children's fundamental beliefs about themselves in relationship to others, thus contributing to impaired interpersonal relationships and the development of personality disorders (Stovall & Craig, 1990).  Sexually abused children develop low self-  esteem as a result of offending adults providing them with messages that are contrary to their internal beliefs (Friedrich, 1990). Sexually abused school-age children typically lack a strong, clearly defined sense of self.  They often  feel shame because they believe they have violated accepted moral principles (Garsee & Schuster, 1992). In many instances, they act compulsively in order to block out thoughts of distressing events (Kaufman &  11 Wohl, 1992).  Sexually abused children may be more  perceptive of adult behaviour than other children, while being less perceptive of their own, or other children's behaviour (Crittenden & Ainsworth, 1989). Children who have experienced sexual abuse in the past, and are currently being sexually abused, are the most severely depressed of clinical and non-clinical samples (Wolfe, 1990).  Such a finding suggests that a  history of sexual abuse continues to affect children and augments the effect of current abuse.  Further  complications arise when personality changes related to the abuse do not become evident until after the individual has been involved with the criminal justice system (Terr, 1990).  The use of non-adaptive defense  mechanisms, such as denial, protect sexually abused children from immediate emotional upset, but these defense mechanisms may also have subsequent negative effects, leading to future psychopathology (Bollinger & Cramer, 1990; Livingston, Lawson & Jones, 1993; MacFarlane & Krebs, 1986; McElroy & McElroy, 1989; Schetky, 1990; Shapiro, Leifer, Martone & Kassen, 1990). Childhood sexual abuse is a complex phenomenon, which includes a wide reinge of factors, including historical, social, cultural, situational, familial, genetic, and environmental factors. The effects of the  12 criminal justice system on the victim must also be taken into account (Rogers, 1990).  Although society is  less tolerant of child sexual abuse than other types of abuse (Bagley & King, 1991), there is a conspicuous absence of trustworthy methods of identifying young sexual abuse victims. At present, the professional community is faced with the problem of obtaining a variety of valid, reliable, non-threatening clinical tools that will accurately reflect the range of factors that need to be considered in assessing and identifying the sexually abused child.  The Conceptual Model  A broad conceptual model is an important primary structure to have in order to evaluate the information generated in the testing situation.  With such a model  in place, one can develop theoretical constructs, as well as evaluate the significance of test scores and test content (Sugarman, 1991).  By adopting an  "unbiased" conceptual model, it becomes possible to connect what is currently known about the phenomena of sexual abuse with what is observed in the attitudes, feelings and behaviours of sexual abuse victims (Conte, 1985; Finkelhor, 1988),  This type of model makes it  easier to understand the interaction of the dynamics  13 involved in sexual abuse child assessments (Brewin et al., 1993; MacFarlane & Bulkley, 1982). Unfortunately, there is discouraging empirical research that examines children's own reports of how they have been affected by abuse (Hartman & Burgess, 1989).  Previous sexual abuse conceptual models have  been cognitively and behaviourally focused (Finkelhor & Browne, 1985; Friedrich, 1990; Hoier, Shawchuck, Pallotta, Freeman, Inderbitzen-Pisaruk, MacMillan, Malinosky-Rummell & Greene, 1992; Sgroi, Blick & Porter, 1985).  These models are effective in  identifying fear and anxiety in children, but neither do they address the ego structure of the child, nor do they offer methods of restoring the child's self-esteem (Berliner, 1991). In order to assess the impact of sexual abuse upon an individual, one needs to obtain access to the individual's subjective experience.  One way to do this  is to find a reliable, non-verbal, projective measure that has the potential to discriminate sexual abuse effects from other abuse effects.  14 Evaluation of Sexual Abuse Effects from a Projective Perspective  It has been shown that anxious, traumatized children have difficulty verbally communicating a reliable expression of their thoughts and feelings (Garbarino et al., 1989).  When children consciously or  unconsciously use defense mechanisms, they express their feelings as if they belong to someone else, minimizing and under reporting the extent of their sexual abuse experiences.  Although children may  verbally deny any ownership of the experience, or the feelings associated with it, they often communicate the suggestion of abuse through their use of symbols or metaphors.  Through understanding the latent content of  the communication, one gains access to understanding the individual's internal working model. An alternative way for the clinician to gain insight into children's internal working models is by stimulating the child's emotions and fantasies, and the internal images and feelings which make up the child's internal world, but which the child cannot articulate (Ammann, 1991; Bretherton, 1990).  Images presented to  the child may trigger similar (though not necessarily identical) reactions to the reactions caused by the original stimuli. Mental imagery can be viewed as the  15 hypothetical connecting link between the processing of conscious or unconscious information and physiological change (Anmiann, 1991; Crittenden, 1989).  It can thus  play a causative and reactive role. An alternative assessment procedure involves obtaining from children, through their art, representative images of their traumatic sexual experiences, thereby gaining access to their unexpressed thoughts, feelings and reactions (Burgess, McCausland & Wolbert, 1981).  This method releases the  child from the pressure of verbal communication and helps to circumvent stereotyped defenses (Cornman, 1988).  The child's drawings become the unconscious  projection of emotional aspects of the personality, providing cues that are lacking in other measures (DiLeo, 1983; Jung, 1976; Kaufman & Wohl, 1992). Incorporating children's drawings into a non-verbal projective assessment procedure may prove to be an effective means of precisely discriminating sexually abused children from other types of abused children.  Using Projective Measurement with Children  Although clinicians regularly use projective techniques in child assessment, they seldom undertake  16 research analysis of the accumulated results (Marshall, 1993).  This may be due to a number of factors: the  interpretation of projective tests requires extensive training; children's reactions to the stimuli may reflect response contamination unwittingly caused by clinician suggestions; and cultural picture bias influences children's responses (French, 1993). Nevertheless, projective technique research is beginning to isolate specific characteristics that can be explored, while eliminating other confounding factors (Singer, 1981).  As Singer (1981) suggests,  formal projective technique research needs to identify the underlying psychological process that will be measured and to design the instruments accordingly. Projective measures for use with children include Kinetic Family Drawings (Burns & Kaufman, 1970), Make A Picture Story (Shneidman, 1952), Children's Apperception Test (Bellak, 1986), Goodenough-Harris Drawing Test (Harris, 1963), and (Buck, 1970).  House-Tree-Person  These measures complement the overall  child assessment, but are of limited value in providing reliable sexual abuse diagnostic descriptors.  These  projective stimuli measures require verbal and visualmotor responses from children. Three projective studies using the image of a rosebush were undertaken during 1984 to 1992 (Allan &  17 Crandall, 1986; Bowden, 1991; Carter, Allan & Boldt, 1992; Crandall, 1984) and have shown great promise in the development and application of a non-verbal projective technique for abused children.  Rosebush  pictures drawn by abused and non-abused children were sorted by adults and other children.  The findings of  the sorts showed statistical levels of significance, dependent upon the focus of the individual study.  For  example, counsellors discriminated between rosebush pictures drawn by coping and non-coping children (Allan & Crandall, 1986);  Bowden (1991) found abused children  sorted rosebush pictures differently than non-abused children.  Following these findings, the next step in  the development of this non-verbal projective technique is to determine if there are specific rosebush pictures, or rosebush picture sets, selected by a child sexual abuse sample. A standardized self-esteem inventory is included with this study to examine correlations of children's self-esteem level with diagnostic category.  The inventory consists of 60  yes/no questions which address common thoughts and emotions reported by sexual abuse victims.  18 Summary  Current procedures for assessing sexually abused children are based on cognitive and behavioural models, and require conscious, unsolicited verbal responses from the children.  These models do not satisfactorily  discriminate sexually abused children from children who have experienced other types of abuse.  Studies have  shown that sexually abused children often have lower self-esteem and exhibit more sexualized behavior than other abused children, but this information alone is not considered reliable enough by the legal system to be used as a basis for discriminating sexually abused children from other children, even when the information is combined with verbal report.  There is an urgent  need for a broader based test model, one which incorporates unconscious elements, given the limitations of current sexual abuse assessment procedures. The purpose of this study is to develop a nonverbal projective measure that will investigate and analyze the correlation between specific picture choices and children's diagnostic categories.  It is  assumed that the internal working models of elementary school age children who have experienced sexual abuse will be different than those of children who have  19 experienced other kinds of abuse/ or no known abuse. The personality dimension of self-esteem has been selected for inclusion in this study, as studies have shown that low self-esteem is a dominant personality trait in sexually abused individuals of all ages (Bagley & King, 1991; Briere, 1992; Briere & Runtz, 1991; Browne & Finkelhor, 1986; Conte & Schuerman, 1988; Faller, 1990; Gold, 1986; Kaufman & Wohl, 1992; Porter, Blick, Sgroi, 1985; Tong, Oates & McDowell, 1987; Wolfe, 1990).  As Farber and Egeland (1987)  state, assessing differences in abused and non-abused children does not identify these differences as causes of abuse. There are several reasons why this study differs from the majority of studies on sexual abuse. These are:  (a) the study requests abused children to  evaluate other children's drawings, rather than completing their own drawings; (b) the focus of the pictures is directed toward the visual and auditory senses of the participants with minimal verbal interaction with an examiner required; (c) parents will not be required to fill out questionnaires about themselves or their children; (d) the selected or designed instruments were chosen, after considerable analysis, for their ability to provide a valid measurement of significant concepts that are known to  20 affect sexually abused children; and (e) a design is used which incorporates findings from previous studies, and provides the opportunity for uncomplicated replication in future studies.  21 CHAPTER 2  LITERATURE REVIEW  This chapter includes the following topics: a synopsis of studies undertaken during the past ten years which investigate the impact of childhood sexual abuse; a critique of studies addressing sexual abuse assessment protocols and interviewing techniques; a review of physical, emotional, and sexual abuse studies; the use of projective techniques with children; and an analysis of the use of drawings as a diagnostic tool.  In conclusion, the background studies  of Rosebush drawing research and the hypotheses of this study are presented.  The criteria for selection of the  studies cited in this chapter are that the work is (a) current or a forerunner in addressing the topics; (b) representative of the issues of concern; (c) frequently cited as the rationale for later studies; (d) representative of a psychodynamic theoretical approach; and (e) one which uses children's drawings as indicators of suspected sexual abuse.  22 Sexual Abuse Impact Findings  This section will review the development of the internal working model, the long and short term impact of abusive conditions on the internal working model and the identified problems in sexual abuse studies that need to be addressed in the future. There are numerous longitudinal studies identifying how early relationships affect one's internal and external perceptual development (Ainsworth, 1991; Bowlby, 1988; Bretherton, 1991; Crittenden, 1989; Egeland & Sroufe, 1981; Hinde & Stevenson-Hinde, 1991; Hopkins, 1991; Kobak & Sceery, 1988; Main, Kaplan, & Cassidy, 1988; Schneider-Rosen, Braunwald, Carlson & Cicchetti, 1988; Waters & Deane, 1988).  As summarized in the following paragraph,  Steele (1983) addresses significant points that relate to early developmental processes of the internal working model that abused infants experience. Infants are unable to develop stable feelings of security when the parenting they receive is inconsistent in quality and quantity.  Infants develop  primitive representations of the parent as an unreliable comforter, and this internal attitude leads to future difficulties in their relationships with others and their feelings about themselves.  When this  23 negative process occurs, infant egos or self-identities become overwhelmed with the physical and emotional pain of neglect and injury.  Infant ego structures become  distorted and disintegrated, with the result that the infant withdraws muscularly and vocally. internal sensations are not validated.  The infant's  This  invalidation results in internal sensations becoming less important in decision making.  The infant learns  to respond only to external cues, disregarding internal cues.  Inadequate care-giving during infant development  may lead to low self-esteem characteristics later in life.  These children, in turn, may be more vulnerable  and susceptible to abusive circumstances as they continually seek secure attachments.  As they mature,  the sexually abused are covertly and overtly silenced by the offender, with the result that they expend their psychic energy maintaining this status for their personal safety, while at the same time depleting growth in their normal developmental processes (Hartman & Burgess, 1989). Studies identifying the long term impact of sexual victimization in childhood include findings of symptoms of post traumatic stress disorder, impaired selfesteem, depression, self-blame for negative events, feelings of helplessness, personal "emptiness," and psychological maltreatment (Briere & Runtz, 1991; Gold,  24 1986; Raczek, 1992; Schetky, 1990).  Sexual abuse  victims suffering the greatest trauma as adults report experiencing a greater frequency and duration of abuse, multiple perpetrators, penetration, physically forced sexual contact, perpetrator substantially older than victim, bizarre abuse features, feeling responsible for abuse, powerless, betrayed, and stigmatized (Anderson, Martin, Mullen, Romans, Herbison, 1993; Briere, 1992). Studies using the Child Behavior Checklist (Achenbach & Edelbrock, 1983), suggest that sexually abused children exhibit significantly more sexualized and aggressive behaviours than non-abused children (Friedrich, Urquiza, Beilke, 1986; Tong et al., 1987). Tong et al.(1987) found that sexually abused girls had lower self-esteem ratings on the Piers-Harris Selfesteem Inventory (Piers, 1990) than the non-sexually abused, but that there was no difference in the scores of the two groups of boys. The most current review of the impact of sexual abuse (Kendal1-Tackett et al., 1993) found that not all children who have been sexually abused exhibit internal or external symptoms.  However, sexually abused  children have more clinical symptoms than non-abused children, but not as many as other clinical child populations.  The study identified the absence of any  specific sexual abuse syndrome and concluded that there  25 is no single traumatizing process which occurs. The review implied that sexual abuse symptomatology is developmentally specific.  The studies showed the most  common symptoms for school age, sexually abused children are fear, neurotic and general mental illness, aggression, nightmares, school problems, hyperactivity, and regressive behaviour. Friedrich (1993) found that identification of pathological behaviours in sexually abused boys did not occur until the mid-treatment stage.  Although the  characteristic of low self-esteem is not significant in some study findings, it is hypothesized that this fact may be a result of sexually abused children initially minimizing their internal feelings in their selfreports.  Although there are no identifiable  demographic characteristics to exclude certain children from sexual abuse, a higher percentage of girls, children with step-fathers, children with poor parenting, preadolescence, children living without a natural parent, children having an impaired mother, and children witnessing family conflict are common risk factors in the sexually abused population (Finkelhor, 1993). Sexual abuse is a multidimensional, serious problem for children, regardless of long term effects (Browne & Finkelhor, 1986).  The need to establish a  26 conceptual foundation for sexual abuse standardized assessment instruments is very clear (Browne & Finkelhor, 1986; Conte, 1985; Friedrich, 1991). Studies of the effects of sexual abuse are usually based on global self-reports, anecdotal and retrospective clinical reports, lists of clinical problems, and previously established standardized tests validated for non-sexual abuse concerns.  The diagnosis  of sexual abuse is often dependent upon clinicians formulating opinions based on lists of symptoms or on court decisions (Melton, & Limber, 1989; Myers, 1993). Theories regarding why sexually abused children differ from non-sexually abused children are in the initial stages of formation.  For example,  developmental theory suggests that children's cognitive understanding of sexual abuse may determine the emergence and persistence of sexual behaviours (Friedrich, 1991).  Cognitive behaviourists point out  numerous reasons why the cognitive approach is sensible: (a) there is a specific nature of stimulus events that initiate responses in a predictable way; (b) there are features of classical conditioning in sexual abuse; and (c) the optional types of responses elicited from a cross-section of individuals may be controlled by the same stimulus, such as a threat to survival (Hoier et al., 1992).  Psychodynamic theory  27 includes understanding how children's cognitive/ emotional, and other internal processes react to the aversive properties of sexual abuse, e.g., children may sexually act out in an ego defensive manner because they identify with the offender (Friedrich, 1991).  In  reviewing adult retrospective reports, Brewin et al. (1993) recognize that both cognitive behavioural and psychoanalytic theories propose that it is advantageous for adult victims to find thematic links between their early and current experiences, so that they may reappraise their current sources of difficulties.  It  is evident in all theories that the impact of sexual abuse on individuals is debilitating and needs to be rectified. In sum, the immediate and long term impact of childhood sexual abuse on the internal working model is devastating to some, but not to all, individuals. There are no standardized measurements to identify sexually abused children.  Parental reports are the  preferred measurements in use. Child self-reports have proven to be unreliable due to the child's minimization of the abusive events experienced.  The majority of all  measurements are based on cognitive and behavioural theoretical foundations, omitting the psychodynamic component.  The following section will review studies  of sexual abuse assessment and interview protocols.  28 Sexual Abuse Assessment and Interview  The sexual abuse assessment and interview format used with children is influenced by a multitude of factors, including educational background and/or personal bias of investigators, selection of assessment measurements, and, most importantly, the welfare of the children involved.  The following section will outline  previously and currently identified issues in these areas. During the 1980's. North American communities scrambled to develop objective interview and assessment guidelines for police, social workers, psychologists, lawyers and counsellors to use with children who were allegedly sexually abused.  The limitations of personal  agendas, subjective bias, lack of sound methodological research, and inadequate understanding of sexual abuse dynamics soon became apparent (McGovern, 1991; Robin, 1993; Steller, 1992; Wakefield & Underwager, 1989). Child sexual abuse assessment and interview protocol (American Professional Society on the Abuse of Children, 1990; College of Psychologists of British Columbia, 1990) were based on early studies of nonsexual ly abused children's memory responses.  The  studies took place in laboratory settings. The protocols were developed with minimal understanding of  29 the impact of sexual abuse on the internal working model of children and of the effect of factors impacting upon children following a disclosure, e.g., delay of court process, suspected false disclosure, removal from home, family responses. Berliner (1988) provided an excellent discussion of the confounding factors involved in child sexual abuse assessments, recognizing that there is no universally accepted procedure for interviewing children or assessing the validity of their statements. In a follow-up study of current assessment and interview approaches, Berliner and Conte (1993) analyzed reported techniques for significant empirical findings and conceptual integrity.  They concluded from  their review that the inclusion of professional opinion is vitally important in sexual abuse cases. Additionally, if professional opinion is formed from a broad range of criteria, the judgment is more likely to be objective. Current studies suggest that assessments include ratings of children's global adjustment and include a rating of their adjustment that is specific to sexual abuse (deYoung, 1992; Shapiro et al., 1990; Steward et al., 1993; Wolfe & Gentile, 1992).  Sexual abuse  assessment checklists are currently being developed, but there are no studies that are quantitative in  30 nature and capable of generating a statistical diagnosis of sexual abuse.  The use of projective  techniques may be of value in addressing this problem. Although there are significant differences between sexually abused children and non-sexually abused children, the question of whether or not these differences are the result of sexual abuse or other coexisting factors is still of concern (Waterman & Lusk, 1993).  Comparative Studies of Sexually, Physically, and NonAbused Children  This section reviews ten representative studies that attempt to identify specific traits, behaviours, and internal processes that discriminate sexually abused children from other children.  These studies  illustrate the wide range of data collection techniques that are employed: retrospective clinical chart reviews, parent questionnaires, child questionnaires, child observations and the use of the projective techniques of sand play analysis, analysis of verbal responses to projective pictures, and analysis of children's graphic representations.  Research and  31 methodological limitations of these reviews will be discussed in the Methodology and Data Analysis Problem section. Obtaining 58 participant records from a psychiatric hospital, Deblinger, McLeer, Atkins, Ralphe and Foa (1989) investigated differences in posttraumatic stress descriptors in sexually abused, physically abused, and non-abused children, ages 3 to 13.  Using former in-patient chart notations, they  found that sexually abused children exhibit significantly higher rates of inappropriate sexual behaviours and had a higher number of post-traumatic stress disorder symptoms than either the physically abused or non-abused groups.  The sexually and  physically abused children showed more avoidant and dissociative symptoms than the non-abused children. Kolko, Moser, and Weldy (1988) compared emotional indicators of 103 sexually abused, physically abused, and non-abused child psychiatric patients.  The  children's average age was 9 years 9 months. The examiners used parent ratings of home behaviours and hospital chart behavioural frequency counts to collect the data.  There were few significant correlations  between the identified home and hospital symptoms of the groups.  32 None of the symptoms that discriminated the sexually abused from the non-abused children were found to significantly differentiate the physically and nonabused children.  The differences in the  psychopathology profiles of the sexually abused children included exhibiting greater degrees of sexual behaviour, mistrust, anxiety and fear.  Although these  findings provided empirical validation of behavioural and emotional symptoms of sexual abuse, the explanation of the mechanism or process which contributed to this finding was not identified.  The examiners stressed the  need for the development of specialized sexual abuse diagnostic and therapeutic techniques. Exploring the relationship of the severity of emotional distress to chronological age in sexually abused girls, Gomes-Schwartz, Horowitz and Sauzier (1985) requested parents of sexually abused pre-school, school-age, and adolescent girls to complete questionnaires regarding their children's behaviours and attitudes.  The findings indicated that there was  more clinically significant psychopathology in schoolage girls than the other two age groups.  This finding  lends support to the theory that elementary school-age children develop an increased emotional sensitivity (Westen, Klepser, Ruffins, Silverman, Lifton & Bockamp, 1991).  Although school-age children were seen as more  33 socially skilled than pre-schoolers, they exhibited more internalized anxieties of fear and impulsivity. School-age children displayed more angry, destructive behaviours.  The examiners indicated that other  psychological problems may appear at later stages of development, and that the sexually abused may have some predisposing, unidentified factors that make them more vulnerable to abuse.  They stressed the importance of  understanding the emotional harm that the sexual abuse experience has caused the individual. Comparisons of emotional and behavioural difficulties in 6 to 12-year-old girls (N = 258) were investigated by Mannarino, Cohen and Gregor (1989). The participants were divided into 3 groups; sexually abused, clinic controls, and normal controls.  The data  collection included the Child Behavior Checklist, the Piers-Harris Personality Inventory, the Child Depression Inventory and the State-Trait Anxiety Questionnaire. Once again, the sexually abused and clinic participants had more pathological symptoms than the normal control group on the Child Behavior Checklist. However, on the Piers-Harris and on the Child Depression Inventory self-report measures, there were no discrepancies between the three groups. The examiners hypothesized that the lack of significant  34 findings may be due to the inadequacy of the instruments in tapping into the emotional problems experienced by sexually abused children.  The State-  Trait Anxiety Scale analysis indicated that the sexually abused girls were more anxious than the other two groups. It is interesting to note that Mannarino et al. (1989) found that the factors of the familial nature of the abuse, the type of abuse, the number of episodes of abuse, and whether or not force was used, did not significantly correlate with the self or parentreported symptomatology. Although the White, Halpin, Strom and Santilli (1988) study focused on 2 to 6-year-old sexually abused children, it is included in this review as it addressed important issues to be considered.  The study compared  the behavioural characteristics of sexually abused children with those of a group of neglected and/or physically abused children and non-referred children, using a developmental questionnaire including additional items that are characteristic behaviours of sexually abused children. Contrary to previous findings, there were no developmental delays in any of the boy groups. There were developmental delays in the girl groups. This study found that the sexually abused boys were  35 significantly more interested in intimate body parts and behaviours of others than the girls.  The sexually  abused boys exhibited lower self-esteem than the controls, whereas there were no differences in the self-esteem ratings of the sexually abused girls and the controls.  The investigators suggested that future  studies need to include gender analysis of the data. The possibility of false positive and false negative disclosure in the respective reports in groups of sexually abused and other children needs to be taken into consideration.  An example of this is seen in the  previously mentioned studies.  The positive self-  esteem, self-report scores of sexually abused children may be due to the children's conscious or unconscious projection of their ideal self. In an innovative projective study by Harper (1991), the differential effects of sexual abuse and physical abuse were identified through sand table scene analysis.  Harper made the assiimption that children  play out their existing conflicts in a direct or symbolic manner.  The 40 participants were children  between the ages of 3 to 10 years of age.  The three  groups consisted of sexually abused, physically abused, and sexually/physically abused children.  The children  were asked to make a "picture" of a world of their own  36 choice in the sand tray with sand tray items. They each did this on four separate, one-hour occasions. The themes of the sexually abused children included the need for nurturance and protection.  The  sexually abused group did not use fantasy and were reluctant to provide a verbal explanation of their scene.  Their worlds were described as closed,  aggressive and rigid in comparison with the other two groups.  Unfortunately, the small sample size limited a  formal statistical analysis. Livingston (1987) compared sexually abused and physically abused children, mean age 9 years 7 months. The children were given a structured diagnostic interview at the time of hospital admission.  Of the  100 completing the interview, 28 qualified for the study.  The diagnoses were made by a child  psychiatrist, based upon the interviews.  There were  several significant differences between the two groups. Conduct disorders were less common in the sexually abused, but the sexually abused children had more major depressive disorders, somatic complaints and gender identity problems than the physically abused children. It is interesting to note that historical information on both groups of children included reported symptoms of attentional deficit disorders and oppositional  37 disorders at an early age, preceding either type of abuse report. Stovall and Craig (1990) investigated mental representations of self and others in a sample of 60 school-age girls, comparing scores of physically abused, sexually abused, and non-abused (distressed) girls on the WISC-R, the Thematic Apperception Test (TAT) and the Piers-Harris. From a psychodynamic perspective, the investigators assumed that unresolved traumatic experiences create chronically disturbed behaviour and arrested character development. The TATs were scored with the Internalized Object Relations Scale by two trained graduate students. The scale is used to differentiate between perceptions of self and others.  The statistical findings from this  scale indicated that the two abused groups significantly differed in comparison to the non-abused children. Comparisons of the Object Relations Scale and the Piers-Harris indicated the Object Relations Scale is more sensitive in assessing unconscious perceptions of self and others than the Piers-Harris Scale. The physically and sexually abused girls had negative unconscious perceptions of self and others and positive conscious perceptions of self and others.  The non-  38 abused girls had congruent conscious and unconscious perceptions of self and others. Although mental representations between physically and sexually abused children were not significantly different, there was a significant difference in comparison with non-abused children.  The findings  support the hypothesis that the difference is related to the actual abuse.  The overall results suggested the  importance of therapists going beyond child selfreports and conscious interactions during diagnostic and assessment procedures with sexually abused children. Yates, Beutler and Crago (1985) studied the relationship between drawings by victims of child incest and drawings by children with other diagnoses. The sample was obtained from a child psychiatry clinic. Several studies have linked a child's expression of their internal reality to art works produced, and have suggested that further research be done in this area (Harris, 1989; Klepsch & Logie, 1982; Wohl & Kaufman, 1985).  This study sought to identify categories that  separate the sexually abused children's drawings from the non-sexually abused children's drawings.  The  researchers selected 15 psychological dimensions with which to evaluate the drawings, e.g., quality of projection, level of anxiety, and confusion between  39 love and anger.  The investigators attempted to  determine which dimensions were more etiologically significant for sexual abuse identification.  They  hypothesized that the drawings done by incest victims were causally related to both hypo and hypersexualization and that the other differences between the abused and non-abused might include other qualities such as projection.  The two psychologist  raters of the drawings were purposely unaware of the children's diagnoses and rated the drawings on the basis of subjective impressions to the 15 dimension models.  Each decision was substantiated with an  example. In this study, 35 girls between the ages of 3.5 and 17 years, were used as participants. A comparison group (n = 17) was matched for age and socioeconomic background to the children who had experienced incest (n = 18). The diagnoses of the sexually abused girls were based on court identification.  An analysis of the  t tests indicated only two significant discrepancies between these 2 groups:  incest victims had  significantly less control over their impulses and tended to rid their conscious thought of painful memories. An F test showed significant group differences (p < .05). Incest victims were found to be more variable  40 in exaggerating or minimizing the sexual features of drawings; they were also less mature, and had difficulties directing a sexual instinct into an acceptable form. Although the study contains numerous methodological difficulties, the findings showed that female children who have experienced incest responded with an impaired ability to control their impulses in comparison with the other group.  It appears that the  internal working models of these girls were affected by the sexual abuse. Another study undertaken by Hibbard and Hartman (1990) compared 134 human figure drawings drawn by sexually abused and non-abused children, ages 5 to 8years old.  The drawings were scored using the Koppitz  Emotional Indicators list, a list of 30 items that are indicative of an area of behavioural or emotional difficulty, e.g., tiny figures, transparencies.  No  statistically significant differences between categorized items were observed, although there appeared to be trends. The main conclusion was that the drawings of sexually abused children demonstrated anxiety more often than the drawings of comparison children.  There were no gender, race, or socioeconomic  differences between the two groups.  41 In sum, the comparison studies show sexually abused children exhibit higher rates of sexual behaviours, anxiety, low self-esteem and major depressive disorders than other children.  Sexually  abused children exhibit inadequate impulse control, have impoverished fantasy and limited expressive verbal skills.  Unfortunately, the studies are difficult to  replicate due to the wide variety of projective measurements and non-projective measurements that are employed.  There are no instruments designed that  measure the psychological impact of sexual abuse on the internal working model of children.  Methodology and Data Analysis Questions  Although these previous studies are often cited in research, their limitations are seldom mentioned. Extremely small samples, poor inter-rater reliability, correlating unstandardized personality characteristics with drawings, neglecting to mention that control group members may have been sexually abused, and the omission of age and gender variability are oversights in previous studies.  Early studies have a lack of  consistent and standardized definitions of child abuse and often have no control groups. Differences that  42 were found were often assumed to be the consequences of the suspected abuse (Pearce, 1984). Previous research on the assessment and identification of sexually abused children has not led to a thorough understanding of the internal working model of sexual abuse dynamics for the child.  Yet to  be determined are the effects of sexual abuse on the covert functioning of the child. Both qualitative and quantitative researchers are beginning to address the question of how sexual abuse affects an individual's internal personal script and cognitive bias.  The authors of research discussed in  this review used different terms in describing internal perceptions of self (Finkelhor & Browne, 1985; Friedrich, 1990; Sgroi, 1989; Summit, 1983) and applied a provisional model in referring to the unconscious and conscious workings of the sexually abused person. Ultimately, a more specific and adaptable system is required. There are certain items in sexually abused children's drawings that provide strong indicators that their internal perceptions are directly related to the sexual abuse they experienced (Kaufman & Wohl, 1992), e.g., the presence of clouds in human figure drawings is more prevalent in pictures drawn by sexually abused children than non-abused children.  Some research  43 studies provide minimally corroborative measurements, but the samples are small and the case studies appear aborted. manner.  They are often presented in a provisional Consequently, the basic fundamentals of robust  quantitative research need to be supplemented with rigorous qualitative inquiry.  The dynamics surrounding  sexual abuse issues are of definite public concern, and the underlying issues are of national importance both in theoretical and practical terms.  Projective Techniques in Assessment  The use of projective techniques in the formal assessment of adults and children is a controversial subject.  The controversy surrounds the question of  whether there is (a) a necessity to understand the process underlying the observed behavior; (b) a relationship of the projective conclusions to the participant's total personality; and  (c) the validity  of the projective conclusions (Gittleman-Klein, 1986; Zubin, Eron, & Schumer, 1965). asked include:  Questions commonly  is similar information available from  other sources, has a reliable measurement tool of the identified psychological process been developed, are  44 the test results any better than chance, and have the parameters that influence projective techniques and their interaction been accounted for in the analysis (Meehl, 1965; Singer, 1981; Zubin et al., 1965)?  There  are parallels between projective testing results and other testing results.  For example, the results do not  always generalize to the social environment and many times the test administrator is more concerned with the study content than the experimental design (GittlemanKlein, 1986; Masling, 1965; Murstein, 1965). The use of projective tools in a thorough assessment of the personality incorporates consideration of the client's intellectual functioning, affect organization, object relations, and defense mechanisms, with the aim of respecting and understanding each client (Jaffee, 1990; Sugarman, 1992). The rationale underlying the use of projective techniques includes the following assumptions: (a) all behaviour is an expression of personality; (b) a stimulus response is brought about by a set of internal inferences that one formulates judgments about; (c) projective tests tap the durable essence of personality equally in different individuals; (d) the individuals taking the test provide material they can not or will not otherwise volunteer; (e) the more similar the  45 stimulus to the client, the greater the degree of projective identification; (f) the ambiguous stimulus is unimportant in comparison to the value of the response it elicits; and (g) projective techniques tap into various layers of the personality (Korner, 1965; Murstein, 1965; Zubin et al., 1965).  The possibility  of projective test results providing a means of identifying latent pathology has also been suggested (Korner, 1965). There is a need to get beyond simplistic descriptions of symptoms in understanding the total personality.  Reporting on a battery of tests becomes  useless if the findings provide insufficient data to answer the diagnostic question (Berliner & Conte, 1993).  There has been so much emphasis on the  statistical verification of testing, but respect for the nature of the unconscious has been neglected (Jaffee, 1990). Sugarman (1991).  This statement is supported by Sugarman (1991) suggests that by  adapting a broad theory based approach to an assessment, it is possible to expand the sources of data that emerge from the testing situation. Current studies using projective test coding are obtaining significant qualitative and quantitative data (Exner, 1991; Sugarman, 1992; Westen, 1991).  For  example, Westen (1991) has developed a five point  46 coding process which can be used to analyze correlations between sexually abused female reactions on Thematic Apperception Test cards and their scores on the Wechsler Adult Intelligence Scale-R,  Picture  Arrangement subtest. As a complement to current sexual abuse assessment protocols, the use of projective testing may prove to provide the much needed, sensitive, non-threatening instrument that distinguishes sexually abused children from other children.  Projective Testing of Children  This section outlines the benefits of projective testing of children, reviews projective testing studies, and presents the types of projective tests often used in child assessment. The use of projective techniques in assessment provides the clinician with a broader view of children's psychological functioning.  When using  projective techniques with children, it is necessary to consider the child's developmental level, degree of emotional and social maturity, and the relative fluidity of their ego boundaries (Rabin, 1986).  For  47 example, sexually abused children may be of superior intelligence while their emotional development is frozen at an earlier stage of development. Child studies using projective techniques have moved from the original psychoanalytic base to a more objective picture of personality development (Haworth, 1986; Rabin & Doneson, 1986; Westen et al., 1991).  For  example, factor analytic studies of the CAT, multivariate analysis of the TAT, and interrater reliability of the CBCL, indicate that personality issues such as character integrity and an understanding of social causality can be identified in elementary school age children through an objective analysis of projective techniques (Haworth, 1986; Westen et al., 1991). Projective techniques are assumed to be sensitive to the unconscious aspects of behavioural responses and to connect with ones' primary process thinking (Rabin, 1986).  Westen et al.(1991) found that certain  dimensions of object relations develop during children's elementary school years, e.g., their capacity for emotional investment increases.  Accepting  this finding, it is clear that the effects of sexual abuse on the internal working model of elementary school age children may be even more devastating to this particular age group.  48 Projective technique studies with children provide objective findings regarding the interaction of their primary process thinking, coping devices, and affectladen play (Levine & Levine, 1986; Rabin, 1986; Russ & Grossman-McKee, 1990).  In comparison to adults, their  responses may be a reflection of their actual reality, rather than their fantasy world (Rabin, 1986; Terr, 1990). Projective techniques are used in assessment to obtain specific levels of diagnostic information, which are then interpreted with other forms of data.  The  techniques are only as meaningful as the extent to which they are viewed with a deeper understanding of children's life circumstances.  One is cautioned that  children's verbal responses to projective tests may be influenced by their anxiety and other defenses, such as emotional numbness, lack of self-understanding, misinterpretation of reality in the testing situation, or their general verbal fluency (Bollinger & Cramer, 1990; Rabin, 1986; Westen et al., 1991).  McGrew and  Teglasi (1990) found that emotionally disturbed boys had much more difficulty than the comparison group in telling a complete, logically constructed story using the TAT cards.  Their stories were rambling and their  story characters lacked feeling.  This is a similar  type of response to that of sexually abused children.  49 Therapists need to incorporate alternative child assessment procedures to allow children an expressive means other than a verbal response.  Projective  techniques offer this alternative to children.  The  following section will review projective studies comparing drawings of sexually abused children with drawings of non-abused and other abused children.  The Use of Drawings as Projective Techniques  Using drawings as a projective technique with children provides more information than is obtained from an objective verbal measurement (Klepsch & Logie, 1982; Wohl & Kaufman, 1988).  Words are unable to  convey all that a drawing represents, and feelings may be altered through the filter of language (Harmer, 1986).  Symbols in a drawing bridge the gap between the  internal world of the child and that of the child's external reality (DiLeo, 1983).  An objective  measurement is likely to reveal the kind of attitudes the child perceives the examiner desiring, whereas the use of a projective technique makes portions of the internal self visible (Burns & Kaufman, 1970; Klepsch & Logie, 1982).  Verbal objective measurement provides a  picture of the way a child would like to be, while  50 projective measurement provides a view of the way a child truly is. According to Machover (1953), stereotyped defenses are less easy to apply to graphomotor projective responses than to verbal projective responses.  Drawings show the cognitive  stage, developmental level, flexibility, and overall personality integration of the child (Rubin, Schachter & Ragins, 1983).  Drawings become the graphic  representation of the child's internal world (Wohl & Kaufman, 1988). A study by Lewis and Livson (1980) compared differences in cognitive development, personality traits and drawings.  Children between the ages of 5  and 11 years (N = 72) were administered the WechslerIntelligence Scale for Children (WISC) or the StanfordBinet and the Goodenough-Harris Drawing Test. A written description of the children's behaviour during testing was recorded immediately after the test administration. The comparison of I.Q. scores on the WISC and Goodenough-Harris found that boys and girls who appeared more success-oriented obtained higher intelligence test scores than drawing scores. Boys who were shy and dependent and girls who were withdrawn, sombre and irritable received higher scores on the human figure drawings than the intelligence tests.  51 Lewis and Livson (1980) speculated that this discrepancy may have been due to social conditions experienced by the children in the two testing situations, which affect their different personality styles either positively or negatively. The study did not identify differences in the verbal and performance WISC scores of the two groups. Findings such as these show that the use of nonverbal stimuli, such as drawings, allows the individual to enter the unconscious perception of their self-concept. Often times, the affective implications of drawings are ignored in assessment (Harmer, 1986). Children are more comfortable with drawings than with direct inquiry.  Drawings are a means of safely  ameliorating the defensive affect of traumatized children, while obtaining valuable information at the same time (Lewis & Livson, 1980; Pynoos & Eth, 1986). The value of human figure drawings as a reliable diagnostic technique is beginning to be recognized. Acton and Moretti (1993) did a meta-analysis of empirical drawing research from 1949 to 1992. The analysis corrected for observed deficiencies of past reviews.  The results indicated that a large number of  drawing features show potential as measures of the constructs of anger, hostility, anxiety and thought disorder.  52 Rubin et al. (1983) analyzed drawings completed by children between 4 to 12 years of age.  There were 10  participants in each age group and gender.  Each  participant completed 4 human figure drawings on two separate occasions, one week apart.  The findings  indicated that human figure drawings reflected major changes in intellectual, behavioural and psychodynamic functioning.  The investigators concluded from these  findings that the reliability of using human figure drawings in clinical work may not be as significant as the sensitivity to change within the child that the drawings reveal (Rubin et al., 1983). A study done in 1973 (Prytula & Thompson, 1973) found no supportive results of correlations between the size of human figure drawings and the self-esteem levels of 10 to 13-year-old children.  Several years  later, a study of stress in hospitalized children compared human figure drawings, physiological and behavioural indices prior to, and following, an injection (Stumer, Rothbaum, Visintainer & Wolfer, 1980).  The findings showed significant emotional  indicators and graphic expressions of stress-laden themes in the pre-post drawings of the children.  There  were correlations between the pulse ratings and coping behaviours, but no correlations of the ratings and behaviours with the drawings, suggesting that acute  53 anxiety taps into an inner psychic process that is internally processed in a different way than the recorded behavioural responses.  The possibility of  anxiety affecting the internal working model of children differently than the recorded physiological and behavioural responses may be similar to the effect of sexual abuse on children. The significance of understanding the inner working model of the sexually abused child was recently addressed by Burgess and Hartman (1993).  They used a  series of event drawings with sexually abused children. The findings suggested that children initially process and organize information at a sensory level, and then use a higher cognitive process to label the information.  It follows that using drawings with  sexually abused children provides the assessor with greater understanding of children's internal working models and how they have been affected by the sexual abuse.  Drawings as a Diagnostic Technique with Sexually Abused Children  Studies have found it beneficial to use drawings as a diagnostic tool with the sexually abused child.  54 Drawings are a means of taking the pressure off the child to verbalize, while, at the same time, gaining access to the child's unexpressed thoughts and feelings.  The trauma of sexual abuse often interferes  with the child's ability to concentrate and attend to stressful questions.  The sexually abused child's  drawing identifies the impact of the abuse upon the child, allowing for an understanding of the child's ego functioning (Burgess et al., 1981; Kaufman & Wohl, 1992;  Miller & Veltkamp, 1989; Riordan & Verdel,  1991).  Drawings can be a permanent record for use in  treatment or by the court. They can be discussed with the child more than once (Allan, 1988; Miller, Veltkamp St Janson, 1987). Studies have shown that sexually abused children below the age of 7 years are more likely to include genitalia in their human figure drawings than nonabused children (Hibbard, Roghmann, & Hoekelman, 1987), and the inclusion of genitalia in elementary school children's drawings occurs very infrequently.  Sexually  abused, elementary school children's human figure drawings scored lower than their chronological age and demonstrated more signs of anxiety than human figure drawings of non-abused children (Hibbard & Hartman, 1990; Jones, 1989).  There were no gender, racial or  socioeconomic status differences in the sexually abused  55 and the non-abused sample (Hibbard & Hartman, 1990). Sexually abused (incest) children's family drawings were significantly different than other children's family drawings (Hackbarth, Murphy & McQuary, 1991). Although the above studies show significant promise for clinical use, the studies have limitations that need to be addressed.  Limitations of Drawings as a Diagnostic Tool in Assessment  The use of drawings in the assessment of sexually abused children presents some limitations.  The  drawings may have been completed by children to satisfy the adult requesting them; drawings taken out of clinical context do not provide enough information to qualify as a complete assessment; some children may not like to draw; children's perceptions of their drawings may not have been adequately explored; diagnosis cannot be made on the basis of one drawing; generalizations about drawings are possible in terms of principles only, not specific symbols in drawings; it is difficult to define widespread variables for scoring drawings; and finally, the chain of inference leading from children's drawing to assessment reports may be too  56 complex and subjective to be reliable or valid (Hagood, 1992; Miller & Veltkamp, 1989; Tritell, 1988). An alternative to the human figure and kinetic family drawings is being developed.  Rosebush Research  A method of inner personal evaluation was developed by Violet Oaklander (1978).  As originally  described by Stevens (1971), this technique, intended as an aid to children who were processing the emotions of painful issues, requested that children use the familiar metaphor of a rosebush as an artistic medium. This technique encouraged children to visualize themselves as a rosebush, draw detailed pictures of themselves as a rosebush, and then respond to questioning about the feelings of the rosebush.  Crandall/Allan Rosebush Study Building on this approach, Allan and Crandall (1986) developed and administered the Rosebush Visualization Technique (RVT) individually to 20 tenyear-old children (10 "coping" and 10 "non-coping"), and then asked them to "draw a picture of their rosebush."  Three independent counselling  57 psychologists, trained in projective assessment, then sorted the 20 drawings into "coping" and "non-coping" categories.  A significant difference (p < .05) was  found between the two groups.  Interestingly, after  each child's words and phrases were included with their drawing, the raters ability to discriminate between the two groups improved to the .01 level of significance. As the background of each child's drawing was examined, it was discovered and clinically noted that the rosebushes of coping, sexually abused, physically abused, and emotionally neglected children differed significantly.  Coping children drew colourful  rosebushes with abundant flowers; sexually abused children placed protective fences around many of the rosebushes; the physically abused children tended to include large numbers of thorns on the rosebush, and the drawings of the emotionally neglected children were extremely impoverished.  Clearly, the RVT appeared to  offer the possibility of providing a differential diagnosis.  Bowden Rosebush Study In 1991, Bowden (1991) continued the rosebush research by presenting previously drawn rosebushes to 42 abused and non-abused children.  Bowden's assumption  was that if a traumatic past experience was  58 internalized by the child, it was likely that this similarity would be recognized through rosebushes drawn by children who had experienced similar trauma. Bowden's subjective analysis indicated that nontraumatized children selected light, colourful, balanced designs filled with positive life qualities, while the abused and neglected children selected pictures depicting isolation, lacking midpoint horizon lines, and containing dark backgrounds. Each child was asked to evaluate and sort the rosebushes into seven Q-Sort categories from "most like" to "least like" themselves. Bowden found that abused and non-abused children globally discriminated in selecting the rosebush drawings of other abused and non-abused children {£ < .05). However, no significant correlation was found to exist within the specific participant diagnosis of physical abuse, sexual abuse, emotional abuse or non-abuse children, and drawings done by children with similar diagnoses. For clarification on this matter, Bowden suggested that an empirical approach be used to effectively discriminate which specific drawings differentiated the non-abused, sexually abused, physically abused, and emotionally neglected groups of children.  The more  important result of this study, however, was the fact that the children did not randomly sort the drawings.  59 Carter/Allan/Boldt Rosebush Study To further explore the possibility of rosebush drawing selection offering a differential diagnosis. Carter et al. (1992) examined two questions:  (a) to  what extent do sexually abused children select pictures completed by similarly diagnosed children as being most like themselves and reject pictures completed by children with other diagnoses as least like themselves? and (b) which specific pictures within each set of pictures are chosen as being more significant to one diagnosis than another? Participants for this study included 11 children, ages 7 to 12, in individual treatment with one therapist; 3 males, mean age 9 years 3 months; and 8 females, mean age 9 years 6 months.  Two males and 7  females were diagnosed as sexually abused.  One male  and 1 female were diagnosed as having an adjustment reaction disorder.  The sample Caucasian families were  of mixed socioeconomic and marital status. The individual children were previously given the RVT.  They visualized their own rosebush environment,  drew a symbolic representation of it, and responded to a series of questions as if they were the rosebush; e.g., "Tell me about your flowers." Sexually abused children often believe others can tell that they have been abused (Sgroi, 1982).  In  60 turn, the children become anxious in interpersonal relationships, feel like social outcasts, and develop physiological and somatic reactions to their internal feelings (Schetky, 1990).  To account for the anxiety  level of the sample, the Revised Children's Manifest Anxiety Scale (RCMAS) (Reynolds & Richmond, 1990) was individually administered to the 11 children during a therapeutic session.  The RCMAS is a standardized,  objective measurement scale of anxiety that children experience, including physiological anxiety, worry, over-sensitivity, social concerns, and concentration. The RCMAS includes a Lie scale.  Rosebush Q-Sort Pictures Q methodology was used as an objective analysis technique to obtain subjective points of view from this small group of individuals.  Because the participant  conducts the sort, this methodology avoids interrater reliability complications.  Twenty-four rosebush  drawings were selected for the Q-Sort from an accumulated collection of drawings used in the previous rosebush studies.  Six representative rosebush drawings  were chosen to be part of each set:  emotional neglect,  pictures 1-6; non-abuse, pictures 7-12; physical abuse, pictures 13-18; sexual abuse, pictures 19-24.  A forced  choice Q-Sort (McKeown & Thomas, 1988) of the 24  61 pictures was individually administered to the children. The children were asked to place the pictures in order from "most like themselves" to "least like themselves" in the following pattern: Frequency: Q-Scores:  1 2 2 3 4 4 3 2 2 1 10  9 8 7 6 5 4 3 2 1  E v a l u a t i o n C r i t e r i o n : "Most L i k e . . . " . . . . " L e a s t  Like..."  Results A data analysis was undertaken using one repeated measures factor design (Glass & Hopkins, 1984).  The  analysis showed the participants grouped into 3 types: Type 1 - 3  sexually abused females, 1 sexually abused  male; Type 2 - 3  sexually abused females, 1 sexually  abused male, 1 adjustment reaction disorder female and male; Type 3 - 1  sexually abused female.  Type 3 was  discarded from the data analysis as the client was extremely atypical of the total sample and not of importance in the analysis (McKeown & Thomas, 1988). Type 1 Z-scores (1.0 or greater than 1.0) indicated 4 non-abuse drawings were selected as most similar to Type 1 children (see Appendix A ) . Two physical abuse and 1 sexual abuse drawings were least similar to them.  Type 2 Z-scores indicated 2 non-  abuse, 2 sexual abuse and 1 physical abuse drawings were chosen as most similar to Type 2 children (see  62 Appendix A).  One emotional neglect and 2 physical  abuse drawings were least similar to Type 2 children. Data analysis of the 24 rosebush pictures was undertaken using one repeated measures factor design (Glass & Hopkins, 1984), hypothesizing that there would be no significant pictures within each set. The tentative conclusion appeared to be an over all significant difference between the pictures at the .01 level.  Emotional Neglect Set Fl (df 1,10) 6.35; Non-  abuse Set - F2 (df 1/10) 4.77; Physical Abuse Set - F3 (df 1,10) 21.32; Sexual Abuse Set - F4 (df 1,10) 4.11, The RCMAS total t-scores and subtest scores on Type 1 and Type 2 sexually abused children indicated there was no significant difference in levels of anxiety between the 2 Types at the .05 level of significance. Type 1 sexually abused children selected pictures from the non-abuse set as being most like themselves. The remaining two pictures in the non-abuse set were their next choices as being most like themselves. Type 1 children chose two physical abuse and one sexual abuse pictures as being most unlike themselves. Following this choice, a sexual abuse and emotional neglect picture were chosen. In summary. Type 1 sexually abused children chose non-abuse pictures as being most like themselves.  The  63 majority of sexual abuse pictures were chosen by Type 1 children as being least like themselves.  Reasons for  these choices include their possible denial of reactions to the abuse, their psychological reintegration since the abuse was reported, or an asyet unidentified cause. According to the therapist. Type 1 children had a positive attitude towards themselves; the children made good eye-contact, exhibited a sense of humour, were clean and well-groomed, played creatively, and were able to directly address their personal concerns. They chose non-abuse rosebush pictures as being most similar to themselves.  They stated they "liked" these pictures  more than the other pictures.  These children chose  physical and sexual abuse rosebush pictures as being least similar to themselves.  The emotional neglect  rosebush pictures were not significant. Type 2 children had negative attitudes towards themselves.  According to the therapist, they were  unmotivated during play, focused on themes of war, deprivation and abandonment. problem solving skills.  They did not demonstrate  There appeared to be more  ambivalence in Type 2 children's discrimination between sets of pictures.  For example, pictures from three  different sets were selected as being strongly similar to Type 2 children.  They chose sexual abuse pictures.  64 physical abuse pictures, and non-abuse pictures as being most like themselves.  The rationale underlying  this discrepancy may be due to several components, e.g., degree of reaction to abuse, length or type of treatment, low self-esteem, poor attachment patterns, or an internal attitude of identifying themselves as a victim. On the RCMAS, there was no significant difference in levels of anxiety between the two groups at the .05 level of significance. There were 7 significant pictures within the picture sets for Type 1 and Type 2 Z-scores.  The Type  2 Z-score for emotional neglect picture 6 was negative. This score indicated a denial of feelings associated with emotional loneliness. 6 was not significant.  Type 1 Z-score for picture  Type 1 Z-scores were  significantly positive for the non-abused pictures 7 and 8 in comparison to Type 2 Z-scores being in the negative range.  It is evident that Type 1 children  identified the pictures drawn by non-abused children as being similar to themselves. Physical abuse pictures 15 and 18 and sexual abuse pictures 20 and 21 had significant positive Z-scores for Type 2. These same picture Z-Scores were either close to zero or negative for Type 1.  Type 2 children  identified with pictures drawn by physically abused and  65 sexually abused children as being similar to themselves. It seems that the pictures were selected by all groups from their overall impressions of the pictures, and according to how they corresponded to their inner self-perceptions.  According to the children's verbal  report, the majority of their choices were made because they felt they "liked" the pictures and not because of specific items in the pictures. There did not appear to be a gender or single/two parent family status discrepancy between Type 1 and Type 2 participants.  Type 1 children had been in  therapy slightly longer than Type 2 children.  The mean  number of children's therapy sessions were 19.3 and 16.5, for Type 1 and Type 2, respectively. In summary. Type 1 children were positive, with healthy ego structures enabling them to handle tension with less use of denial as a defense mechanism.  They  were cheerful in their home, school, and therapeutic environment.  The effects of their sexual abuse  appeared to be externally and internally resolved at this stage of their development. Type 2 sexually abused children were described as negative and aggressive in their home, school, and therapeutic environment.  Their ego development  66 appeared thwarted and the necessity for this group to continue therapeutic treatment was apparent. Continuing the development of the RVT as a tool to distinguish sexually abused children from other children was strengthened by the Carter et al. (1992) research.  The need to find a set of discriminating  rosebush pictures appeared to be the next step in this process.  Limitations of Rosebush Studies  The participants in the previous Rosebush studies were obtained from a metropolitan area on the west coast of British Columbia.  Future Rosebush studies  need to obtain a larger number of participants from a broader geographic area.  This will increase the  strength and generalization of the findings. Q-Sort methodology was advantageous in analyzing the data in the previous Rosebush studies. The direction of the current study is to develop a projective assessment tool.  Because Q-Sort Rosebush  selections required a long period of time for children to complete, it will be beneficial during assessment procedures to provide children with a smaller number of discriminating pictures, requiring an alternative type of selection response.  67 The preceding Rosebush data analyses showed significant differences in the picture selections of abused children and non-abused children.  Furthermore,  differences in Rosebush picture selections were shown between sexually abused children having positive attitudes and sexually abused children having negative attitudes.  The possibility of Rosebush picture  selections being related to the stage of treatment, level of self-esteem, and the type of abuse, needs to be considered.  Hypotheses  In this subsection, the hypotheses on which this study is based are presented.  Each hypothesis is  prefaced by a brief statement of rationale.  Studies have found that sexually abused children exhibit more overt sexual behaviour and include more sexual images in their drawings than other abused children.  Therefore, it is possible that acts of  sexual abuse affect the internal working model of children differently than other abusive acts. Accessing this internal discrepancy through visual stimuli provides children with a non-threatening.  68 unconscious operation to which they can easily respond. Sexually abused children have developed strong defense mechanisms that protect them from ego threatening circumstances.  Once sexually abused children in  treatment begin to feel safe, and develop trusting relationships, their internal working model is more accessible.  Hypothesis 1.  Sexually abused children at different  stages of treatment will select Rosebush Pictures that are significantly different than pictures chosen by children with other abuse diagnoses, or children with no known abuse.  Low self-esteem is a significant trait of sexually abused children.  The possibility of children having  low self-esteem prior to being sexually abused must be considered.  Feelings of low self-esteem may create  thought patterns and behaviours that make children susceptible to any kind of attention from others, and, in turn, more vulnerable to being sexually, physically, and emotionally abused.  It is proposed that the self-  esteem of children is affected differently by sexual abuse than other types of abuse.  The differences  within children between the effects of sexual abuse and other types of abuse does not change during treatment.  69 Hypothesis 2.  Self-Esteem scores of sexually abused  children at different stages of treatment will be significantly different than the self-esteem scores of other abused or non-abused children.  70 CHAPTER 3  METHODOLOGY This chapter begins with the design of the study, a description of the participant sample, the therapists and the procedures that were undertaken.  This is  followed by a description of the instruments, followed by the data analysis relevant to each hypothesis.  Method Design of the Study  The rationale of this design is based on the tacit assumption that there is some element in Rosebush drawings that triggers a response from sexually abused children that is different than responses from other children. This study investigates the relationships between scores on Rosebush picture selections and child diagnostic categories, e.g., do sexually abused children choose a certain Rosebush Picture or set of Rosebush Pictures more often than children who were not sexually abused?  Included in the investigation are the  child participant's stage of treatment and level of self-esteem.  The dependent variables are the child  participants' scores on the set of 12 Rosebush  71 pictures.  The independent variables are the diagnostic  referral groups, stages of treatment (see Figure 1 ) , and level of self-esteem.  Mental Health Centres  Private Therapists  Private Psychologists  Child Participants  Treatment for Sexual Abuse  Beginning  Mid  Others in Treatment  End  Beginning  Mid  Non-Abuse Non-Treatment  End  One Group Only  Figure 1 - Child Participant Selection Diagram  Child Participants  Following approval from the University of British Columbia Behavioural Sciences Ethical Review Committee for Research involving Human Subjects, the Ministry of  72 Social Services, and the Ministry of Child and Youth Mental Health Services (see Appendices B-1, B-2, B-3), 66 sexually abused children and 46 other children in treatment were recruited from the case loads of therapists and psychologists in private practice, private agencies, mental health and social service offices of a western province. A descriptive introductory letter (see Appendix C1) was sent to the therapists and psychologists, followed by a telephone call or personal contact by the principal investigator to obtain therapist or psychologist consent (Appendix C-2).  Participating  therapists and psychologists were then sent specific research directions (see Appendix C-3) and asked to obtain written consent from parents of participating children between the ages of 6 years 0 months to 11 years 11 months and of normal intelligence (see Appendix C-4).  This age range was selected for  inclusion of participants in the study because the actual detection of sexual victimization is less probable in school-age children than younger children (Campis et al., 1993; Conte, 1991);  school-age  children are more likely to maintain the secrecy of abuse as they recognize the social implications (Anderson et al., 1993); school-age children have feelings of guilt and shame associated with the sexual  73 abuse they have experienced; and school-age children do not depend on the behavioural reenactment of the sexual abuse as the only means they use of mastering the trauma; if behavioural reenactment is used as a healing technique by this age group, it is less detectable as the children often play outside, or in an unsupervised environment.  Any child of normal intelligence, within  the defined age-range, and in treatment, qualified for inclusion in the study. The therapist was requested to place the signed Parental Consent Form with the child participant's clinical record.  Of the 130 child participant returns,  7 were excluded due to falling outside of age requirements: months;  2 returns were under age 6 years 0  5 returns were over age 11 years 11 months.  Eleven non-abused, non-treatment child participants were obtained from the families and friends of three psychologists and one nurse clinician.  Therapists  Provincial mental health representatives gave permission for the staff at 8 mental health facilities to be contacted, providing a mental health pool of 15 therapists or psychologists to administer the materials to the child participants (see Appendix B-3).  In  74 addition, 12 therapists and 4 psychologists in private practice or community services agreed to participate. Of the 31 therapists or psychologists administering the instruments to child participants, 4 were male and 27 were female.  There were 8 registered psychologists  having clinical, counselling and developmental backgrounds.  Of the remaining 23 therapists, 15 had  backgrounds at the masters level in art, counselling psychology, education, nursing, or social work.  Six  therapists held bachelor degrees in education, nursing, psychology, or social work.  The remaining two  therapists were a registered art therapist and a licensed practical nurse.  The therapists and  psychologists will be referred to as "therapists" in the remainder of the study. Eighteen of the therapists collected data in a large metropolitan area.  The remainder of the  therapists collected data in the major provincial zone locations:  interior (7), valley (4), northern part of  Island (2).  Procedures  Data were collected over an 8 month period.  Each  therapist was provided with a number coded set of materials for each child participant once the therapist  75 agreed to participate in this study.  The set of  materials included (a) a letter of instructions for the therapist (see Appendix C-3), (b) a demographic questionnaire (see Appendix D-1) (c) a Rosebush Picture Sort answer sheet and administration instructions (see Appendix D-2), (d) a CFSEI-2 question/answer sheet (see Appendix D-3), (e) a Therapist Permission Sheet (see Appendix C-2), (f) 2 Parent Permission Forms (see Appendix C-4) and (g) 1 cerloxed set of Rosebush Pictures (see Appendix D-2). The therapist completed the demographic questionnaires and returned them with the Rosebush Picture Sort (RPS) forms and Culture-Free Self-Esteem Inventory (CFSEI-2) answer sheets to the principal investigator (see Appendix D-1, D-2, D-3). Each individual set of data was given a numeric code to provide anonymity to the children and their therapist. The principal investigator identified the beginning stage of treatment as 1 to 3 sessions (Carter & Allan, 1992; Horvath & Greenberg, 1989; Mannarino et al., 1989).  The mid and end stages of treatment were  identified by the child's therapist and noted on the demographic questionnaire: "# sessions with therapist...; beginning rx...middle rx...end rx,..". Distinguishing the mid and end stages of treatment was an individual decision made by each therapist, based on  76 their knowledge of the child participant and the therapeutic process. If the returned data were incomplete, the principal investigator contacted the therapist and requested the missing information.  Once each set was  complete, the principal investigator assigned the child participant returns to one of three diagnostic categories, as the therapist noted on the demographic questionnaire item "reason for referral":  sexual  abuse, other treatment or non-treatment (see Figure 1). The non-treatment, non-abuse child participant group (N = 11) was used as a comparison for all the three treatment stages. Following the receipt of the data, the principal investigator provided each therapist with the selfesteem ratings of their number-coded participants. No other response information was released to the therapists.  Instruments  Three instruments were used to collect information:  a demographic questionnaire; the Rosebush  Picture Sort (RPS); Culture Free Self-Esteem InventorySecond Edition (CFSEI-2; Battle, 1992).  77 Demographic Questionnaire The demographic questionnaire (see Appendix D-1) provided information on the following: (a) therapist gender and educational level; (b) geographic location; (c) child gender; (d) child age; (e) child ethnic heritage; (f) child residence; (g) reason for referral; (h) degree of suspected abuse; (i) type of suspected abuse; (j) number of treatment sessions; (k) stage of treatment; (1) type of parental attachment; and (m) general comments.  The reason for referral, type of  suspected abuse, number of sessions and stage of treatment data were used as criteria to assign participants to groups (see Figure 1). The items on the demographic questionnaire were selected to obtain specific information with which to clarify potential correlations between sexual abuse and other factors; for example, whether there is a higher number of sexually abused children of single parent families in comparison to sexually abused children living in two parent families.  The demographic items  of interest were selected and compared using chi-square tests to determine if there were significant differences in demographic variables across the three treatment groups.  If differences were found, these  variables might need to be included in the final data analyses.  78 The type of parental attachment information was not under investigation in this study. The parental attachment information was collected for use in future research.  Rosebush Picture Sort (RPS) A set of 13 Rosebush pictures (see Appendix D-2) was selected for use in the study.  The original  pictures were drawn by sexually abused, physically abused, emotionally neglected, and non-abused children. The 13 pictures were chosen based on previous Rosebush research (Bowden, 1991; Carter et al., 1992), in which these 13 Rosebush Pictures discriminated between groups of abused and non-abused children, and sexually abused children with high and low self-concepts.  Four expert  child abuse therapists reviewed the 13 Rosebush Pictures for potentially distressing or threatening visual cues.  They approved this set of Rosebush  Pictures for research with abused children. The selection of discriminating pictures noted above was based on results derived from previous Q-Sort methodology findings (Bowden, 1991; Carter et al., 1992).  The Q-Sort method allowed children to  conceptualize and rank order a large set of Rosebush Pictures (N = 24). For the present study, a Likert response format was substituted for the Q-sort method  79 to investigate the strength of a child's identification with each picture.  The children identified how they  felt about each picture in this format, rather than ranking the pictures against one another.  Unlike the  Q-Sort method, the Likert Scale can be administered rapidly, facilitating inclusion of the Rosebush Picture Sort Likert Scale in an assessment package.  In  research or assessment with children, the length of time for test completion is of major importance as a lengthy assessment tool quickly becomes boring to children and may, therefore, influence their true responses.  Although the Likert Method is usually a  five point anchored scale, a shorter, even numbered four point scale was chosen to prevent participants from continuously selecting the midpoint of the scale (Dawis, 1987).  Further, using four choices provided  clearer information in relation to the strength of the hypotheses.  For example, either children did or did  not identify with each Rosebush Picture; no neutral choice was allowed. The Rosebush Picture Sort (RPS) was presented individually to each child by their therapist.  The  therapist was requested not to interact with the child or give the child any prompting during this test administration (see Appendix D-2).  The child rated  80 each picture on a four point Likert scale from "not at all like me" = score of 4...to "very much like me" = score of 1. The first picture was a test sample, drawn by a non-abused child.  The remaining 12 Rosebush Pictures  were arranged in the following rotating order:  picture  drawn by (a) non-abused, (b) physically abused, (c) sexually abused, and (d) emotionally neglected child. This picture order (a,b,c,d) was repeated three times, totalling 12 test choices.  This picture order was  decided upon to prevent children from making a set response choice or from using a sequencing bias, e.g., making a repeated negative response to four pictures in sequence that were drawn by a non-abused child.  And in  turn, the alternating picture order was selected to provide optimum conditions in which to tap into their internal working model (feelings, memories, defenses).  Culture-Free Self-Esteem Inventory - Second Edition fCFSEI-2) The Culture-Free Self-Esteem Inventory (see Appendix D-3; Battle, 1992) was individually administered to each child by their therapist or psychologist to obtain a measure of the child's selfesteem level.  Form A of the Culture-Free Self-Esteem  Inventory (CFSEI-2) is a 60 item, yes-no inventory.  81 intended to measure an individual's perception of selfworth.  The CFSEI-2, Form A, was standardized on  children, grades 1 through 6. The Culture-Free Self-Esteem Inventory (CFSEI-2) was chosen for inclusion in this study as it uses items that are often of specific concern to sexually abused children.  For example, items such as "I often feel  ashamed of myself," "I usually fail when I try to do important things," and "I spend a lot of time daydreaming," are common attitudes expressed by, or observed in, sexually abused children. includes 4 self-esteem subscales: Parental and Academic.  The CFSEI-2  General, Social,  For the purposes of this  investigation, only total scores on the CFSEI-2 were used.  The Lie scale is scored separately.  Lie subtest  research measured individual defensive responses to self-esteem items, e.g., stating they never told a lie. There were no gender differences and the majority of the elementary school-age sample showed a lack of defensiveness. A self-esteem inventory is considered reliable if participants respond to the stimulus items in a consistent manner across time.  The test-retest  correlations of CFSEI, Form A, ranged from .81 to .89 (n < .01) based on a sample of 198 elementary school children tested with a 48 hour interval between  82 testings (Battle, 1991).  Thus, children in grades 3  through 6 responded to CFSEI, Form A in a statistically reliable manner.  Further reliability of this inventory  was established when 33 children from the original 198 children were retested two years later.  The test-  retest correlation for these participants was .74 (p < .01) (Battle, 1991). Construct validity of the CFSEI, Form A, was established by selecting the 60 most discriminating items from a pool of 150. Concurrent validity of the CFSEI, Form A, was established by comparing these scale results with other established self-esteem measurements; for example, the Coopersmith Self-Esteem Inventory for Children.  The correlations between the  two tests were significant for all grade levels, supporting the validity of this instrument (Battle, 1991). Therefore, the valid, reliable Culture-Free SelfEsteem Inventory, Second Edition (CFSEI-2) was selected for inclusion in this Western Canadian study, since the items address specific concerns relevant to sexually abused children.  The inventory was standardized on  Canadian children and covers the age range defined in this study.  83 Data Analysis  To examine the hypotheses of this study, a one-way multivariate analysis of variance (MANOVA) was used to determine if there were significant overall differences in mean scores on the Rosebush Picture Sort (RPS) among three study groups (see Figure 2).  The multivariate  analysis was selected to evaluate simultaneously the effects of the independent variables/diagnostic groups (3) on the dependent variables/pictures (12). Conceptually, the MANOVA is an extension of the univariate analysis (ANOVA) technique.  The MANOVA and  ANOVA are suitable for testing the research hypotheses that there are no mean differences among the three study groups.  The MANOVA investigates if the mean  differences among groups on the combined dependent variables are larger than what is expected by chance. Tabachnick and Fidell (1989) outline several advantages of using MANOVA instead of ANOVA:  (a) by  measuring several dependent variables at once, instead of singly, the chance of discovering what it is that changes as a result of different variables or their interactions is improved; (b) when there are several dependent variables, MANOVA protects against an inflated Type I error by avoiding the use of multiple  84  a.  ROSEBUSH PICTURE SORT SCORES  MANOVA (2 X 12) Sexual Abuse/Other X Pictures  b.  DEMOGRAPHIC CHI SQUARE ANALYSES  BEGINNING STAGE  GENDER  Figure 2,  MANOVA (2 x 12) Gender X Pictures  MIDSTAGE  AGE  PARENT  END STAGE  ETHNIC  Phase One Data Analysis Design  AREA  85 tests of correlated dependent variables; and (c) MANOVA may reveal differences not shown in ANOVA when dependent variables are considered in combinations and may be more powerful than separate ANOVAs. An a priori decision was made that if the MANOVA results showed there were significant mean differences among the study groups, discriminant function analysis (DISCRIM) would be performed (see Figure 3).  The  primary goal of the discriminant function was to find the dimensions along which the groups differed and to find the classification functions that predicted group membership (Tabachnick & Fidell, 1989).  Discriminant  function analysis (DISCRIM) asked if there was some combination of variables that reliably separated the groups.  The purpose was to predict group membership  from a set of predictor variables; e.g., can predictions of group membership for a group of sexually abused children in treatment, a group of other children in treatment and a group of non-abused, non-treatment children be made reliably from a set of Rosebush Picture Sort scores? Thus, multivariate statistics provided appropriate analyses when there were many independent variables and dependent variables.  The intercorrelations among these  variables can be compared in one omnibus analysis.  86  ROSEBUSH PICTURE SORT SCORES a.  MANOVA(3X12) Beg. X Picts  MANOVA(3X12) End X Picts  MAN0VA(3X12) Mid X Picts  /  SA  OT  NON  SA  OT  NON  ^  ^  SA  NON  OT  b. DISCRIM{12X3) Picts X Beg.  DISCRIM(12X3) Picts X Mid.  DISCRIM(12X3) Picts X End. \  ^  SA  NON  OT  NON  c. Chi-Square Analysis of Pictures by Stage of Treatment - Secondary Analysis  SELF-ESTEEM INVENTORY SCORES d. ANOVA (3X1) Beg. X Est.  SA  OT  Figure 3.  *SA OT NON  NON  ANOVA (3X1) Mid X Est.  SA  OT  NON  Phase Two Data Analysis Design  = Sexual Abuse Treatment = Other Treatment = No Sexual Abuse, No Treatment  ANOVA (3X1) End X Est.  SA  OT  NON  87 MANOVA tests whether the mean differences among the three study groups' Rosebush Picture Selections were likely to occur by chance or reflect a significant difference.  The DISCRIM predicted group membership  through classification of picture choices. MANOVA and DISCRIM assumptions that must be met were similar to those of ANOVA: normal distribution of observations, homogeneity of variance and independence of observations. Chi-square analysis was used to compare the four Likert scale categories of response scores among the 3 groups within each treatment stage (see Figure 3). That is, the Rosebush Picture scores were treated as categoric response variables. A series of one-way analyses of variance (ANOVA) was carried out to compare standardized mean selfesteem scores on the three groups within the Beginning, Mid and End-treatment stage (see Figure 3). In sum, the data analyses were performed in two phases (see Figures 2 & 3).  The first phase used a  MANOVA to investigate gender differences in the sexual abuse group picture selections (N = 66) and a MANOVA to investigate overall treatment group differences (N = 111) in picture selections.  The three groups in the  three treatment stages were compared with respect to the demographic variables of gender, age, parent.  88 ethnic heritage, and area, using chi-square tests of independence.  The demographic factors were compared  within each treatment stage to see if any of these variables might be influencing the picture (predictor variables) selection findings.  If differences arose,  it would be necessary to include the demographic factors as variables in the subsequent analyses. The second phase of data analysis included MANOVA and discriminant function analyses with stage of treatment and diagnostic group picture selection comparisons.  A secondary analysis was completed using  chi-square analyses of the individual pictures in each stage of treatment.  The self-esteem score means of  diagnostic groups in relation to stage of treatment were investigated using ANOVAs.  Summary This study investigated the association of Rosebush Picture selection and child diagnosis in relation to participants' stage of treatment and level of self-esteem, using the instruments of the Rosebush Picture Sort (RPS), the Culture-Free Self-Esteem Inventory (CFSEI-2), and a demographic questionnaire. The data were analyzed using MANOVA, DISCRIM, ANOVA, and Chi-Square tests.  89 The participants in the study were elementary school-age children of normal intelligence.  The  participants were obtained from the case loads of psychologists and therapists.  These professionals  administered and completed the design instruments with the cooperation of the child participants.  Null Hypotheses Null Hypothesis 1: It is hypothesized that the Rosebush Picture Sort (RPS) selection of sexually abused children at different stages of treatment will not differ from those of other children in similar stages of treatment and will not differ from non-abused, non-treatment children. A multivariate and discriminant function analysis of the Rosebush Picture Sort (RPS) selection will show no significant differences among children at the three different treatment levels.  Null Hypothesis 2: It is hypothesized that the self-esteem mean scores of sexually abused children at different stages of treatment will not differ from those of other children at similar stages of treatment and will not  90 differ from the mean scores of non-abused, nontreatment children. An analysis of variance of the Culture-Free SelfEsteem Inventory-2 (CFSEI-2) will show no significant differences among children at the three different treatment levels.  91 CHAPTER 4  RESULTS The results of the data analysis will be formally reported in this chapter.  Informal information, e.g.,  remarks from participants and therapists will also be reported.  Phase One includes chi-square tests of  independence of the demographic characteristics of the sample, a multivariate analysis of Rosebush Picture Sort (RPS) selection by gender, and a multivariate analysis of Rosebush Picture Sort (RPS) selection by sexually abused children and other children in treatment (see Figure 2).  Phase Two pertains to the  multivariate and discriminant function analyses of the Rosebush Picture (RPS) selection according to stage of treatment and group, chi-square tests of independence of the pictures and groups, an analysis of variance of self-esteem scores with respect to treatment stage and groups (see Figure 3), concluding with an informal data analysis.  Phase One  Sample Characteristics This study investigated seven demographic variables using chi-square to test for distinguishing  92 characteristics of the sample, examining the three stages of treatment for any relationship between the three groups (sexual abuse, other treatment, no abuse/no treatment) and the following variables: reason for referral, stage of treatment, child gender, age, parent status, ethnic heritage, and geographic area. Reason for Referral The primary reason for referral listed for Group 1 was sexual abuse.  The type of sexual abuse ranged from  pornography to violent rape. According to demographic information from therapists, 8% of Group 1 participants experienced mild sexual abuse (see Table 1; Appendix D1).  Moderate sexual abuse, including fondling and  masturbation (see Appendix D-1), was listed for 62% of Group 1 participants.  Anal, rectal and/or vaginal  penetration were experienced by 30% of Group 1 participants.  Of the 66 sexual abuse participants, 17  children, or 26% of Group 1, were identified as not experiencing emotional or physical abuse.  The  therapists identified the remaining 74% of the sexual abuse group as experiencing emotional and/or physical abuse (see Appendix D-1).  The total number of cases of  emotional and physical abuse does not equal the number of sexually abused children, since  some children were  93  Table 1  Abuse Classification Information on Group 1 and Group 2 Group 1 Sexual Abuse FEMALE Degree  MALE  Abuse Type Sexl Emot  Phys  Abuse Type Sexl Emot  Phys  Mild Moderate Severe  3 26 14  13 12 8  14 9 1  2 15 6  4 7 4  8 6 0  Total  43  33  14  23  15  14  Chi-Square, 4 df = 23. 34 0001  Chi--Square, 4 df = 12.28 E = .015  n=  Group 2 Other Treatment FEMALE Degree Mild Moderate Severe Total  MALE -  Abuse Type Emot  Phys  Abuse Type Emot  Phys  6 7 0  2 2 2  12 6 6  8 6 0  13  6  24  14  Chi-Square, 2 df = 4.86 = .088  Chi--Square, 2 df = 4.48 E = .11  94 identified as experiencing more than one type of abuse. Within the three treatment stages, sexual offenders included both parents (1), father (2), mother (1), step-parent (4), grandmother (1), brother (2), uncle (1)/ babysitter (2), male adolescent (2), foster parents (l), family friend (1), mentally handicapped adult (2), or stranger (1). The frequency of abuse ranged from one event to an unknown number of sexual events.  The time period of  abuse ranged from infancy to the present day.  The  number of offenders per victim ranged from one offender to many. Although the above information was volunteered by some of the therapists and, therefore, may not be a complete representation of a pattern for the entire sample, it is similar to other sexual abuse epidemiological studies (Finkelhor, 1993).  For  example, the offender is usually related to, or known by the victim; abuse events range over a short or long period of time; and the children involved are adversely affected by the many different kinds of abusive circumstances. Therapists listed the following reasons for the referrals of Group 2 (n = 46) other children in treatment:  physical abuse, emotional neglect or abuse,  dysfunctional parenting, oppositional or bizarre  95 behaviour, inappropriate sexual behaviour, emotional swings, threats of suicide, poor social skills, depression, anxiety, aggression towards peers, school suspension, enuresis, nightmares, encopresis, learning disabilities, low self-esteem, sibling rivalry, somatic complaints, arson, sleep problems, and phobias. As previously noted (Kendal1-Tackett et al., 1993), many of these behaviours are often observed in sexually abused children.  The therapists did not identify  sexual abuse as a type of suspected abuse on the demographic questionnaires of Group 2 children, but this does not mean that sexual abuse had not occurred. According to the demographic information provided by the therapists, 28% of the Group 2 referrals have not experienced sexual, physical or emotional abuse. The remaining 72% of Group 2 are identified as experiencing physical and/or emotional abuse (see Table 1).  The total number of cases of physical and  emotional abuse does not equal the actual number of physically and emotionally abused children because some of the children are identified as experiencing both physical and emotional abuse. As shown in Table 1, Group 2 female and male children are identified as experiencing more emotional abuse than physical abuse. The largest number of data returns were assigned to the Mid-Treatment stage (see Table 2).  The members  96 of this stage represent 60% of the total study population.  In this Mid-Treatment stage, 54% were  diagnosed as sexually abused, and 31% were in the other treatment group.  The 11 Group 3 no abuse, no treatment  children represent 15% of the Mid-Treatment stage.  The  Beginning-Treatment and End-Treatment stage returns were similar to one another:  37% and 36%,  respectively. Group 1 sexual abuse; 31% and 33%, respectively. Group 2 other treatment; 31% for Group 3 in both treatment stages. Gender Female participants account for 55% of the study population (see Table 2).  This female/male ratio is  lower than other sexual abuse population data (Finkelhor, 1993).  The percentage of female  participants in the Beginning-Treatment stage is 63%, and 67% in the End-Treatment stage, but only 51% in the Mid-Treatment Stage (see Table 2).  In the Mid-  Treatment stage. Group 2 other treatment, there is a significant difference among groups with respect to gender ratio (chi-square = 9.09; p = .01); Group 2 other treatment has a larger representation of males than females.  97 Table 2  Demographic Variable - Gender  BEGINNING-TREATMENT Group* Female Male  1  2  8 5  6 5  Total 13 11 Chi-Square 2 df = 0.79  Total 8 3  22 13  11 35 p = .67  MID-TREATMENT Group* Female Male  1  2  3  Total  24 16  6 17  8 3  38 36  Total 40 23 Chi-Square 2 df = 9.09  11 74 E == .01  END-TREATMENT Group* Female Male  1  2  3  Total  11 2  5 7  8 3  24 12  Total 13 12 Chi-Square 2 df = 5,44  11 36 E = .07  *Group 1, sexual abuse; Group 2, other treatment; Group 3, no abuse, no treatment.  98 Age One-way analysis of variance showed no differences in the mean ages of the three groups in each stage of treatment (see Table 3).  The age range of the entire  sample spans 6 years 0 months to 11 years 11 months, with the mean age 9 years 5 months.  The Group 1  (sexual abuse) mean age is 9 years 2 months; the Group 2 (other treatment) mean age is 9 years 6 months; and the Group 3 (no abuse, no treatment) mean age is 9 years 7 months (see Table 3).  The mean age of the  total group of females (n = 68) is 9 years 2 months, ranging from 6.0 years to 11 years 11 months, and the mean age of the total group of males (n = 55) is 9 years 7 months, with the age range from 6 years 1 month to 11 years 11 months. Parent The chi-square test for the parent type was not significant in any treatment stage (see Table 4).  The  children reside with two birth parents in 29% of the families in the total sample:  Group 1 has 29%; Group 2  has 24%; and Group 3 has 36%.  In other two parent  families of the total sample, one of the parents is a step-parent.  In the total sample, 11% of the two  parent families include a step-parent: Group 2 has 15%; and Group 3 has 9%.  Group 1 has 9%;  Close to 50% of  the total sample are living in single parent families.  99 Table 3 Demographic Variable - Age BEGINNING-TREATMENT Group*  N  Mean  Std. Deviation  1 2 3  13 11 11  9.49 9.30 9.71  1.68 1.62 1.68  Total 35 9.50 F (df 2, 32) = .17; £ = .85  1.66  MID-TREATMENT Group*  N  Mean  Std. Deviation  1 2 3  40 23 11  8.80 9.25 9.71  1.66 1.51 1.68  Total F (df ^,  74 9.25 71) = 1.56; p = .23  1.62  END-TREATMENT Group*  N  1 2 3  13 12 11  Mean 9.15 10.17 9.71  Total 36 9.68 F (df 2, 33) = 1.16; p = .32  *Group 1-Sexual Abuse; Group 2-Other Treatment; Group 3-No Abuse, No Treatment.  Std. Deviation 1.79 1.52 1.68 1.66  100 Table 4 Demographic Variable - Parent BEGINNING-TREATMENT Group* 1-parent 2-parent Step  1  2  9 4 0  4 4 3  Total 13 11 Chi-Square 4 df = 3.13  Total 6 4 1  19 12 4  11 E = .54  35  MID-TREATMENT Group* 1-parent 2-parent Step Foster  Total  1 17 11 6 6  12 5 3 3  Total 40 23 Chi-Square 6 df = 1.56  6 4 1 0  35 20 10 9  11 E = .96  74  END-TREATMENT Group'" 1-parent 2-parent Step Foster  Total 7 4 0 2  6 2 1 3  Total 13 12 Chi-Square 6 df = 1.78  6 4 1 0 11 n = -94  * Group 1 - Sexual Abuse; Group 2 - Other Treatment; Group 3 - No Abuse, No Treatment.  19 10 2 5 36  101 In Group 1 and Group 2, 49% of the children are living in single parent families.  Group 3 has 56% of the  children living in single parent families. None of the Beginning-Treatment children reside in foster parent families.  There are 12% of Mid-Treatment children and  14% of End-Treatment children residing in foster parent families. Ethnic Heritage The chi-square test for ethnic heritage was not undertaken at the Beginning-Treatment and End-Treatment stage due to numerous empty cells (see Table 5).  The  chi-square test undertaken at Mid-Treatment was not significant.  Caucasians are the largest ethnic group  represented in all three treatment stages and sample groups.  The total sample is 90% Caucasian.  Group 1  sexual abuse and Group 2 other treatment have an 89% Caucasian representation. Group 3 is 100% Caucasian.  The no abuse, no treatment Aboriginal children  represent 7% of the total sample.  There are 3  Aboriginal children and 1 Latin American child in the Beginning-Treatment stage.  The Mid-Treatment stage  includes 7 Aboriginal children.  The End-Treatment  stage includes 1 Aboriginal child, 1 Asian child, and 1 East Indian child.  There are no identified children of  African or Inuit origin.  102 Table 5 Demographic Variable - Ethnic Heritage BEGINNING-TREATMENT Group*  1  2  Caucasian Aboriginal Latin Amer.  10 3 0  10 0 1  11 0 0  31 3 1  Total  13  11  11  35**  Total  MID-TREATMENT Total  Group* Caucasian Aboriginal  35 5  21 2  Total 40 23 Chi-Square 2 df = .30  11 0  67 7  11 E = .86  74  END-TREATMENT Total  Group* Caucasian Aboriginal Indian Asian  12 0 0 1  10 1 1 0  11 0 0 0  33 1 1 1  Total  13  12  11  36**  *Group 1 - sexual abuse; Group 2 - other treatment; Group 3 - no abuse, no treatment. **Chi-Square not computed due to niomerous empty cells,  103 Geographic Area Chi-square test results were not reported due to empty cells in each treatment stage (see Table 6). Combining the metropolitan resident participants from the three treatment stages accounts for 57% of the total study population.  Close to 47% of the sexually  abused children reside in a large metropolitan area. The remaining 53% of the sexually abused children reside in smaller metropolitan areas.  Of the children  in treatment for other reasons, 41% reside in a large metropolitan community.  The remaining 59% of this  group live in smaller communities.  The entire no  abuse, no treatment group resides in a large metropolitan area.  Although 57% of the study  population resides in a large metropolitan area, 15 other areas of the province are represented by the sexual abuse group and other treatment group, covering all treatment stages. To summarize the demographic data, the largest number of returns that were received fell into the Caucasian, Group 1 sexually abused female of the MidTreatment category.  Many of these children reside in a  large metropolitan area, with a single parent family. ANOVA and chi-square tests of independence were used to compare gender and parent constellation in demographic treatment groups.  The tests indicated no significant  104 Table 6  Demographic Variable - Geographic Area BEGINNING-TREATMENT Group* Interior Lower Main Eraser Val Total  Total  1  2  4 9 0  0 7 4  0 11 0  4 27 4  13  11  11  35  MID-TREATMENT Group*  1  Total  Interior Van.Island Lower Main Eraser Val  16 6 16 2  1 3 5 14  0 0 11 0  17 9 32 16  Total  40  23  11  74  END-TREATMENT Total  Group* Interior Lower Main Eraser Val Total  6 6 1  0 7 5  0 11 0  6 24 6  13  12  11  36  *Group 1 - sexual abuse; Group 2 - other treatment; Group 3 - no abuse, no treatment.  105 age differences within Beginning, Mid or End-Treatment groups.  There was an unequal ratio of female/male in  the Mid-Treatment stage, with more males than females in Group 2 other treatment.  Gender Multivariate Analysis A one-way MANOVA, with the Rosebush Picture Sort (RPS) scores as the dependent variables and gender as the independent variable, was performed to check for the presence of gender effect on the sexual abuse group.  The MANOVA test for Homogeneity of Dispersion  indicates the assumptions have been met:  F (df 78,  668)= 1.04, E = .387. The analysis shows no gender effect in the scores of 43 female and 23 male sexually abused children (N = 66, F (df 12, 53) = 1.50, n =.153)(see Appendix E-1). Given this finding, no follow-up gender analysis was undertaken.  Treatment Group Analysis A one-way MANOVA was undertaken comparing the mean scores of Rosebush Picture Sort (RPS) selections of Group 1 sexually abused children (n = 66) and Group 2 children in treatment for other reasons (n = 45). The data were not sorted according to stage of treatment in this first phase of analysis.  106 The MANOVA test for Homogeneity of Dispersion indicates the assumptions have been met:  F (df 78,  284) = 1.12, E = .216. There were no significant differences found between the picture score means of Group 1 sexually abused and Group 2 other children in treatment; F (df 12, 98) = 1.17, E =.312 (see Appendix E-1). The next one-way MANOVA was undertaken after the data were sorted into groups by stage of treatment.  Phase Two  The Phase Two analyses were divided into seven sections:  (a) a MANOVA analysis of the Rosebush  Picture Sort (RPS) scores at the Beginning-Treatment stage, (b) a MANOVA analysis of the RPS scores at the Mid-Treatment stage, (c) a MANOVA analysis of the RPS scores at the End-Treatment stage, (d) chi-square tests of independence for each Rosebush Picture within the three study groups, (e) an ANOVA of the self-esteem scores at the Beginning-Treatment stage, (f) an ANOVA of the self-esteem scores at the Mid-Treatment stage, and (g) an ANOVA of the self-esteem scores at the EndTreatment stage.  107 Hypothesis 1 The Rosebush Picture Sort (RPS) selection of sexually abused children at different stages of treatment will not differ from those of other children in similar stages of treatment and will not differ from non-abused/ non-treatment children. Of the 123 RPS returns, it appears only 4 male children (3%) made set choice responses.  The responses  were all negative: 3 males chose "This is not at all like me" for all responses or all but one of their responses; 1 male chose "This is not like me" for all but one of his responses.  Three of the males were in  Group 1, sexual abuse and 1 male was in Group 2, other treatment.  This Group 2 male was appropriately  classified according to the discriminant function analysis.  Their ages ranged from 7 years 0 months to  11 years 11 months.  Rosebush Picture Sort (RPS) Beginning-Treatment Stage: The MANOVA of the RPS scores of Group 1 sexual abuse, (n = 13), Group 2 other treatment, (n = 11), and Group 3 no abuse, no treatment, (n = 11), indicated no significant differences among the three groups at the Beginning-Treatment stage.  The null hypothesis is not  rejected, with F (df 24, 42) = 1.45, p =.142 (see Appendix E-2).  Therefore, a discriminant analysis was  108 not done.  A chi-square analysis was done following  this finding (see Appendix E-3). At the Beginning-Treatment stage. Picture 10 showed the only statistically significant trend in the chi-square analysis: E» =.009 (see Appendices D-2, E-3). Picture 10 is dark and has thorn-studded stalks filling the centre of the picture.  Children in treatment.  Groups 1 and 2, tended to identify with Picture 10 more than children not in treatment. Group 3.  Rosebush Picture Sort (RPS) Mid-Treatment Stage: The MANOVA of the RPS scores at the Mid-Treatment stage showed significant differences among the three groups (N = 73), F (df 24, 118) = 2.03, p =.007 (see Appendix E-2). A discriminant function analysis was completed using the twelve Rosebush picture variables as predictors of membership in three groups:  Group 1  sexually abused children in treatment; Group 2 other children in treatment; and Group 3 non-abused, nontreatment children.  For the 73 child participants'  scores, evaluation of assumptions of linearity, variance-covariance matrices revealed no threat to the validity of multivariate analysis: F (df 78, 6142) = 1.06, E = .341.  109 One discriminant function was calculated, with a combined chi-square (df 24) = 44.55, p = .007. After removal of the first function, there were no further associations between groups and predictors, chi-square (df 11) = 11.44, p = .41. The first discriminant function accounted for 77.57 % of the between-group variability.  The first discriminant function separated  the sexually abused group of children and other two groups from each other.  The sexually abused group had  a 75.0% predicted group membership.  The other children  in treatment group had a 59.1% predicted group membership and the non-abused, non-treatment group had a 72,7% predicted group membership. The standardized weights of correlations between picture predictor discriminant functions suggest that the best predictors for distinguishing between the sexually abused and the other two groups were Pictures 6, F (df 2, 70) = 4.25, s. = -02, and Picture 8, F (df 2, 70) = 4.02, n = -02. Picture 6 has a dark sky border, black lines for clouds, and thorny bushes. Picture 8 contains a barren, single flower face with "closed eyes" on a stalk. These pictures, plus Pictures 2 and 9, showed large standardized discriminant function coefficients (see Appendix D-2, Table 7): Picture 2 = -.565; Picture 6 = .627; Picture 8 = .923; and Picture 9 = -.523.  Standardized weights less  110  6.0  •  4.0 •  2.0 •  3  3 3  Second Discriminant Fxinction  1 1 2 3 1 1 1 3 2__ 1 11 1 3 2Ci. 11^ 132 111 O 2 221 1 112-111 1 J2E23 1 2 2 ^ 1 2 2 2 1 2  °* -2.0 •  3 1  3 2 3  •4.0 •  •6.0 •  OUT X XOUT  -6.0  -4.0  -2.0  2.0  4.0  6.0  First Discriminant Function  Ficmre 4 - All Groups Discriminant Function Scatterplot © Centroid A Centroid XS Centroid  Group 1 Group 2 Group 3  Sexual Abuse Other Treatment No Abuse, No Treatment  X -X OUT  Ill Table 7 Results of Discriminant Function Analysis RPS Variables  Correlations* Picture #  Univariate  Coefficients**  F (df 2, 70)  1  -.04  .96  -.27  2  -.16  .38  -.57  3  .03  .34  -.15  4  -.01  .94  -.31  5  -.00  .90  .00  6  .46  .02  .63  7  .08  .31  .25  8  .41  .02  .92  9  -.26  .18  -.52  10  .26  .20  ,30  11  -.07  .81  -.16  12  .27  .19  .05  Variance %  77.57  Eigenvalue  .67  (fi < .05)  *Correlations of Predictor Variables with Discriminant Function **Standardized Discriminant Function Coefficients  112 than .50 were not interpreted.  Picture 2 has strong  dark colours, and although there is "life" in the picture, the overall impression is rather stark. Picture 9 contains a fence with a "Keep Out" sign. In reviewing the discriminant analysis of the individual groups in the Mid-Treatment stage, within Group 1 sexual abuse, 18% or 7 children selected pictures more closely related to Group 2 other treatment choices. Two of Group 1 children identified more closely with the no treatment Group 3 picture choices.  There were 6 participants or 27% of Group 2  other treatment misclassified.  Four of the children  chose pictures that were similar to those chosen by Group 3 and 2 of the children chose pictures similar to the selection of Group 1 sexual abuse.  Three children  in Group 3 selected pictures that were similar to the selection of Group 2 other treatment children.  Only 1  child in Group 3 selected pictures that were similar to those selected by Group 1 sexual abuse. In summary, there are significant differences between the sexually abused mean discriminant scores and corresponding scores of other children in this study.  The hypothesis that the RPS mean discriminant  scores of sexually abused children at Mid-Treatment will not differ from the scores of other children at similar stages of treatment or no abuse, no treatment  113 children was rejected. The mean discriminant scores of sexually abused children at Mid-Treatment differ from scores of other children in treatment or no abuse, no treatment children. The chi-square analyses of individual pictures at Mid-Treatment showed Picture 8, a single, "sleeping" flower, as the only statistically significant trend; p = 0.02 (see Appendices D-2; E-4). Group 1 sexual abuse, either strongly accepted (choosing option 1) or strongly rejected (choosing option 4) this picture. Group 3 mainly rejected Picture 8.  RSP End-Treatment Stage: At this stage of treatment, there was no significant difference among the three groups with respect to picture selection (n = 36, F (df 24, 44) = 1.48, n = .129) (see Appendix E-2).  Discriminant  analysis was not undertaken following these findings. Chi-square analyses of this stage of treatment was undertaken. At the End-Treatment stage. Picture 5 showed a statistically significant trend; n =.098 (see Appendices D-2; E-5). Picture 5 shows the base of a tree with roots, a bird with a worm in its beak and insects (see Appendix D-2).  Groups 1 and 2 strongly  identified with this picture, while Group 3 mildly  114 identified with the picture.  It is possible that  Picture 5 tapped into positive internal feelings that the children in the End-Treatment stage had recently developed about themselves.  Hypothesis 2  The self-esteem mean scores of sexually abused children at different stages of treatment will not differ from those of other children at similar stages of treatment and will not differ from the mean scores of non-abused, non-treatment children. Participants with Lie scores one standard deviation below the mean (score below 40) were not included in the analyses.  The Lie subtest measures  defensiveness (Battle, 1992).  Defensive participants  do not accept valid, but socially unacceptable characteristics within themselves. The BeginningTreatment stage Lie scores of 2 sexual abuse participants and one other treatment participant were below 40 (see Figure 5).  The Mid-Treatment stage  indicated 5 sexual abuse participants and 2 other treatment participants with Lie scores below 40. The End-Treatment stage score of one sexual abuse participant was below 40, but there were no Lie scores below 40 of other children in treatment.  None of the  115 non-abuse, non-treatment group self-esteem Lie scores were below 40.  Number of Partici- 3 pants with scores below 40 0  I  BeginningStage  MidStage  GROUP 1  sexual abuse  GROUP 2  other treatment =  Figure 5.  EndStage  =  Self-Esteem Lie Score Frequency Graph  Self-Esteem Beginning-Treatment Stage: The ANOVA of the Self-Esteem scores of Group 1 sexual abuse, Beginning-Treatment stage; Group 2 other treatment, Beginning-Treatment stage; and Group 3 no abuse, no treatment, indicated no significant differences among the three groups (see Table 8);  F  (df 2, 28) = 1.51, E = .24, No further analysis was done with the Beginning-Treatment stage following this finding.  116 Table 8 Self-Esteem Descriptive Statistics and One-way ANOVA BEGINNING-TREATMENT Variable  N  Group 1 Group 2 Group 3 ANOVA:  Mean  Std.Dev.  11* 48.73 10* 48.40 10** 55.40  11.35 11.32 7.47  Min. Max. 30 27 37  65 62 64  F (df 2, 28) = 1.51; p = .239  MID-TREATMENT Variable  N  Group 1 Group 2 Group 3 ANOVA:  F  Mean  Std.Dev.  35* 47.11 21* 44.81 10** 55.40  10.48 9.68 7.47  Min. Max, 28 27 37  65 61 64  (df 2, 63) = 4.02; p = .023  END-TREATMENT Variable  N  Group 1 Group 2 Group 3 ANOVA:  12* 12 10**  Mean  Std.Dev.  53.17 48.67 55.40  7.71 9.46 7.47  Min. Max. 38 38 37  64 62 64  F (df 2,31) = 1.91; p = .17  *Participants with Lie scores that are one standard deviation below the mean are not included. **The total number of Self-Esteem scores in Group 3 is 10, not 11, because one subject refused to answer the questions.  117 Self-Esteem Mid-Treatment Stage: The ANOVA of the Self-Esteem scores of the three groups in Mid-Treatment indicated a statistically significant difference among the groups (see Table 8); F (df 2,63) = 4.02, £ = .023. Comparisons of the means reveal that the self-esteem scores of Group 3 no abuse, no treatment, are significantly higher than the mean scores of the other two groups. Mean levels of selfesteem scores in Group 1 and Group 2 do not differ from one another.  Self-Esteem End-Treatment Stage: The ANOVA of the Self-Esteem scores of the three groups at End-Treatment stage show no statistically significant difference (see Table 8); F (df 2, 31) = 1.91, p =.17.  Further analysis was not undertaken  following these results. In summary, the Self-Esteem scores of children in all three groups did not differ significantly at the Beginning or End-Treatment stages.  The mean scores of  Group 3 no abuse, no treatment, differed significantly from the sexually abused group and the other treatment group at the Mid-Treatment stage.  The Self-Esteem  scores of Group 1 sexual abuse, and Group 2 other treatment, did not differ significantly from one another at Mid-Treatment.  118 Informal Findings Even though the data was not statistically analyzed, the following information has been included since the conceptual impact of the information is relevant to the study hypotheses.  The comments were  recorded by the child's therapist. It is noteworthy that there were only ten comments on the RPS from Group 1 sexual abuse, at BeginningTreatment, while Group 2 other treatment, made 70 comments at this stage (see Figure 6; Appendix F).  The  self-imposed silence on the part of the sexually abused children did not seem to interfere in their completion of the RPS but may indicate their hesitancy to verbalize personal feelings. The RPS statements from Group 2 other treatment children, at Beginning-Treatment stage were much more expressive and imaginative than those of Group 1. These statements were full of symbolic meaning, mixed with the child's sense of reality: RPS #1: "I'm always caught in the middle of things." RPS #2: "I'd like to live in a big house with a big yard." RPS #3: "I'm always getting cut off from things I like." RPS #4: "I like rainbows and I try to find the charm at the end." RPS #5: "I'm not used to birds getting for their young. I have to get everything for myself." RPS #6: "I don't like rose buds because when you touch them it hurts."  119 75  £.... .•-  Number of Comments  BeginningStage  Midstage  GROUP 1  sexual abuse  GROUP 2  other treatment = ^ ^  Figure 6.  EndStage  = Hm  Rosebush Comment Frequency Graph  RPS #7: "The leaves falling remind me of me because I'm so clumsy and break things a lot." RPS #8: "I wouldn't want to be a flower because if anybody picks me then I'd be...it would hurt." RPS #9: "It reminds me of my Mom. She doesn't put up 'Keep Out' signs when she has a problem. She talks to me." RPS #10: "I don't feel like looking at dead flowers." RPS #11: "The sun is smiling. The squirrel is going to run up the tree." RPS #12: "I like being small. If there's a big bush, I like to crawl under it" (see Appendices D-2, F ) . There were an equal number of therapist comments for Group 1 and Group 2, Beginning-Treatment stage (see  120 Appendix G-1).  The therapists made mainly positive  comments regarding both Group's reactions to the RPS, e.g., "...very much liked and responded to RPS," "...very thoughtful."  None of the therapists responded  negatively about the administration of this technique. At Mid-Treatment, Group 2 other treatment children made twice as many spontaneous statements (n = 33) about the RPS pictures as Group 1 sexual abuse children (n = 16)(see Figure 6).  Mid-Treatment Stage Group 1  statements contained a richer amount of information, much of which is representative of problems that sexually abused children face.  For example, a feeling  of being stigmatized was expressed by a sexually abused male in his response to RPS Picture 1, "Sometimes I feel like the only one among those who are different from me" {see Appendices D-2; F ) . The need to maintain self-protection and the lack of trust felt towards adults was shown by a male's response to RPS Picture 2, "This has a fence around me:  it's like me."  Gender  conflict was expressed by a 7 year 11 month old male to RPS Picture 8, "This is like me...oh no, it's a girl; it's not at all like me."  RPS Picture 6 possibly  tapped into early memories of a young female who responded with, "When I was a baby." According to therapist responses, the majority of Group 1 and Group 2 Mid-Treatment participants enjoyed  121 making the RPS selections (see Appendix G-2).  In Group  2, a 10 year 2 month old female chose to draw her own picture of a rosebush upon completion of the RPS. Most children in both groups were reflective and cooperative. reluctant.  A few children were confused and In Group 2 other treatment, an 11 year 7  month old male thought the pictures were effeminate and said, "People will think I'm a fag." On reviewing the verbal comments from the MidTreatment children. Group 2 other treatment continued to be more expressive than Group 1 sexual abuse (see Appendix F). End-Treatment, Group 2 other treatment participants made more than four times as many statements (n = 44) about the pictures than Group 1 sexual abuse participants (n = 10) (see Figure 5). Group 2 statements expressed a range of feelings and demonstrated the children's integration of symbolism and reality.  Statements included the following (see  Appendices D-2; F): RPS #1: "The trees stand for everyone and I'm in the middle." RPS #2: "The roses are for girls, not boys." RPS #3: "Looks like someone's crying." RPS #4: "It's snowing. I like rainbows... kind of Christmassy and I like Christmas." RPS #5: "I like insects, so I can bug my mom." RPS #6: "The roses are red, like my rosebush would be." RPS #7: "The leaves are all falling off. Makes you feel fenced in. Spikes make you feel unwanted...like a grave yard."  122 RPS #8: "Whatever that thing is (pointing to flower), I'm not that small." RPS #9: "Makes you feel like you're not wanted." RPS #10: "Too dark, don't feel welcome. Looks like a mean plant with mouths on it." RPS #11: "Like smiling sun. If I was one of the flowers, it would feel good." RPS #12: "It's like somebody talking out loud to a bunch of people.  I wish I had a friend."  As in the previous two treatment stages. Group 1 statements were mainly descriptive and limited in imagination.  A Group 1, 10 year 11 month old sexually  abused female, wanted to know if the shape in RPS Picture 5 was a hand.  One child was more expressive at  this stage of treatment.  The young Group 1 male  responded to RPS Picture 4, "Volcano tree and storm are like me,"  Difficulty with self-control and power  issues appear to continue to affect this sexually abused young male, although his therapist perceived him to be at the End-Treatment stage.  He responded to RPS  Picture 8, "I hate 'Keep Out' because I always go in," and to RPS Picture 11 with, "It's light and that's not like me." The therapist of a Group 2 male wrote, "Initially quite ambivalent in doing test.  Once he realized that  it was not threatening to his 'artistic' abilities, e.g., he did not have to do any drawing himself, he relaxed a lot more" (see Appendix G-3).  The note from  this therapist supports the non-threatening nature of the RPS for children.  123 Culture-Free Self-Esteem Inventorv-2 (CFSEI-2) Informal Findings The children's unsolicited comments at all three treatment stages on the CFSEI-2 were mainly concerned with their undecided responses to the yes/no questions. The therapists of Group 1 participants responded with comments about the children more often than the therapists of Group 2 participants (see Appendix G-4) in all three treatment stages.  Some children had  difficulty with double negative questions and others felt that the Inventory had too many items. However, most of the children became involved in this exercise and it helped the therapist to communicate with them about their problems, e.g., "...said, 'It was good to get that out of me';" "had problems of acknowledging her sadness and disappointment, but the questionnaire gave opportunity to reflect and recognize changes she's making...;" "the test gave us the opportunity to talk more about how she's changed over the three years I've known her."  Summary of Findings The study sample consisted of 123 female and male elementary school-age children from 16 provincial communities.  The largest number of participants in the  three treatment stages dwell in a large metropolitan  124 area and are of Caucasian origin.  The largest number  of participants assigned to Group 1, 2 and 3 live in single parent families. The mean age of participants in the BeginningTreatment stage is 9 years 5 months.  The mean age of  participants in the Mid-Treatment stage is 9 years 3 months, and the mean age of participants in the EndTreatment stage is 9 years 7 months.  The mean age and  age range of the participants is similar to other studies that report a predominance of prepubertal abuse (Kendal1-Tackett et al., 1993). made up 55% of the total sample.  Female participants The majority of Group  1 participants experienced moderate sexual abuse.  The  majority of Group 2 participants experienced physical and emotional abuse, but none experienced reported sexual abuse.  Group 3 participants were not in  treatment and had not experienced known abusive circumstances. The Phase One studies indicated that there were no gender differences in the Rosebush Picture selection scores, and that there were no significant differences found when comparing the means of the two treatment groups.  There were no significant demographic  differences that might influence the Phase Two analyses.  125 In the Phase Two analyses, MANOVA and DISCRIM differences were found in the Mid-Treatment stage when the three study groups were analyzed, but not in the Beginning or End-Treatment stage.  Group 1 picture  selection patterns for Pictures 2, 6, 8, and 9 differed from Groups 2 and 3 picture selection patterns. As a cross check on the MANOVA and DISCRIM analyses, the three study groups were compared on the basis of response frequencies, using chi-square tests, at the stage of treatment.  The chi-square results,  however, must be viewed with extreme caution due to a number of cells with very small or zero frequencies. There were three pictures that discriminated between "'' groups.  Beginning-Treatment stage Group 1 and Group 2  strongly identified with Group 3 did not.  Picture 10, p =.009, while  Group 3 did not identify with Mid-  Treatment stage Picture 8, p =.02, while 43% (n = 17) of Group 1 did.  End-Treatment stage Groups 1 and 2  strongly identified with Picture 5, p =.098, while Group 3 did not. There were Self-Esteem score differences in the Mid-Treatment stage but not in the Beginning or EndTreatment stage Self-Esteem scores.  Group 3 no abuse,  no treatment, had significantly higher Self-Esteem scores than the other two groups.  126 In the three stages of treatment. Group 1, sexually abused children made significantly fewer, unsolicited Rosebush picture comments than Group 2 other children in treatment.  127 CHAPTER 5  DISCUSSION AND CONCLUSIONS The general findings of this study offer a contribution to the literature on assessment procedures that discriminate sexually abused children from other children in treatment and from non-abused children, through the use of a non-threatening, non-verbal technique.  This chapter examines the trends and  significance of the demographic findings and the implications of the significant differences between the three groups in picture selections and self-esteem levels in relation to stage of treatment.  The  limitations of this study are also examined.  The  chapter concludes with suggestions for future Rosebush Picture Sort research.  DISCUSSION  Phase One Sample Characteristics  Reason for Referral The reasons for referral of the sexual abuse and other treatment sample represent a broad range of characteristic sexual abuse behaviours and other types of child pathological behaviours that warrant clinical  128 intervention.  This broad range of referrals permits a  greater generalizability of the study findings.  Gender As evidenced in Phase One, there are no gender differences in the Rosebush Picture Sort (RPS) profiles of female and male participants.  This is similar to  other study findings (Hibbard et al., 1987; Reinhart, 1987).  In Mid-Treatment stage. Group 2 other treatment  children, there is a higher percentage of male than female participants.  This shift in gender distribution  suggests that male participants in Group 2 other treatment have not yet disclosed their abuse.  This  reversed gender finding lends support to previously cited current and retrospective findings that indicate sexually abused, school-age males are reluctant to report their abuse and often do not report their abuse until mid-treatment or adulthood (Campis et al., 1993; Friedrich, 1993; McGrew & Teglasi, 1990; Reinhart, 1987).  Age The age range of children in this study spans 6 years 0 months to 11 years 11 months.  Children, ages 6  to 7 years of age and 10 years of age, are cited as being the most vulnerable to sexual abuse (Finkelhor,  129 1993).  Other findings show that sexually abused  children between the ages of 6 years 0 months to 11 years 11 months are less likely to make a purposeful verbal disclosure as (a) the children perceive that they are responsible for the abuse; (b) they are aware of the potential consequences of the disclosure; and (c) their defense mechanisms may have removed the abusive experience from their conscious awareness (Green, 1993; Livingston et al., 1993).  The risk of  sexual abuse for this age group, combined with the fact that it is less likely to disclose abuse, demonstrates the importance of using this age group in the development of a non-verbal, non-threatening technique. The results of this study can be used for the development of a technique by which sexually abused children in this age group can be identified without placing the children in a personally frightening or verbally threatening position.  Parent Although Finkelhor (1993) has stated that living in single parent families is a sexual abuse risk factor for children, the single parent situation does not appear to be a discrepancy in the three groups of this study sample.  The results in this study are  130 strengthened by the finding that all three Groups have similar single or two parent circumstances. The Beginning and End-Treatment Group 1 sexual abuse have no step-parent category.  It is conjectured  that this circumstance may be related to a step-parent being the sexual offender, and that the step-parent was removed from the home at the time of disclosure, prior to therapy; or else the offender's identity came to light during treatment, and the step-parent left the home.  This conjecture is supported by demographic  findings showing no foster parent category in the Beginning-Treatment stage.  In the Mid and End-  Treatment stage, 16% of Groups 1 and 2 children are in foster placement.  Ethnic Heritage Caucasians were the largest ethnic group represented in all three treatment stages and sample groups (see Table 5).  As reported in current  epidemiological studies (Finkelhor, 1993; Reinhart, 1987), race and ethnicity have not been found to be sexual abuse risk factors. The relatively small percentages of ethnic minorities in this study sample limit any interpretation of the ethnic similarities of picture selections within the treatment groups.  131 Geographic Area Although the majority of the study population resides in a large metropolitan area, 16 areas of the province are represented by the sexual abuse and other treatment groups, covering all treatment stages. These 16 areas span the four major regions of the province. Therefore, the participants could be representative of the province as a whole, giving a broader range of generalization to the study findings.  It appears that  the sexual abuse of children is prevalent in all communities and is not necessarily related to rural isolation or crowded urban demographic factors. In summary of the demographic findings, there were no significant differences in the study groups gender preference of pictures, age, or area of residence.  The  study findings of reversed gender ratio in MidTreatment Group 2, ethnic heritage limitations, and the type of parental status in treatment stages, provide possibilities for further Rosebush Picture Sort investigations.  132 Phase One Data Analysis  Treatment Group Analysis There were no differences found when comparing the entire sample of Rosebush Picture Sort scores of the two treatment groups.  Previous studies (Bowden, 1991;  Carter et al., 1992) did find significant differences in Rosebush picture selections, comparing abused and non-abused children or sexually abused children with high and low self-concepts.  These studies used Q-Sort  methodology (ranking each picture against another picture) in contrast to Likert Methodology (rating identification with each individual picture) used in this study.  Noting this difference in the two studies,  it will be beneficial to carefully address the rationale underlying the methodology used in future Rosebush Picture Sort research.  Phase Two  Hypothesis 1 The Rosebush Picture Sort (RPS) selection of sexually abused children at different stages of treatment will not differ from those of other children in similar stages of treatment, and will not differ from non-abused, non-treatment children.  133 Rosebush Picture Sort (RPS) Beginning-Treatment Stage: According to the MANOVA analysis, there were no statistically significant differences in the picture selections of Group 1, Group 2 or Group 3 at BeginningTreatment . The chi-square (df 6, p = .009) analysis of Picture 10 (see Appendices D-2; E-3) showed differences between the three groups of children at the BeginningTreatment stage.  Group 2 identified with Picture 10  (73%), Group 1 somewhat identified with Picture 10 (46%) and Group 3 somewhat rejected Picture 10 (46%). This picture selection preference is similar to the Bowden (1991) study, in which the abused children identified with this picture and other pictures containing dark, gloomy backgrounds.  This finding may  indicate that the sexually abused, and others, in the early stages of treatment closely identify with the darkness and the thorny stalks that they see in the picture, finding a correspondence with their own sense of emotional chaos. The fact that there are limited significant differences in picture selections at the BeginningTreatment stage may be a result of the interaction of children's primary process thinking, coping devices, developmental level, and affect-laden decision making skills.  As suggested in previously cited studies  134 (Levine & Levine, 1986; Rabin, 1986; Russ & GrossmanMcKee, 1990), children at the Beginning-Treatment stage are hiding behind their defenses and do not feel safe enough to uncover or disclose painful or anxiety producing material; e.g., therapists reported that two sexually abused participants were threatened with death if they disclosed their abuse. As reported in other studies, children do not necessary disclose their sexual abuse during the initial interview, but do so at a later date (Conte & Schuerman, 1988; Friedrich, 1993; Herzog, Staley, Carmody, Robbins & vanderKolk, 1993). The RPS assessment technique appears clinically valuable in providing children with a projective, nonthreatening, non-verbal visual medium, but seems unable to tap into their internal working model without upsetting their conscious cognitive defenses.  Comments  on the RPS that were made by several of the sexually abused children at the Beginning-Treatment stage provided useful information for the therapist to incorporate into the assessment and treatment process; e.g., RPS Picture 1: trees," RPS Picture 9:  "It's crammed up in a bunch of "I don't go in places where it  says 'Keep OutI'" (see Appendices D-2; F ) . However, the majority of comments made by sexually abused children at Beginning-Treatment were abrupt and limited: e.g., "too black," "bird," "little girl."  135 It appears that sexually abused children have difficulty expressing themselves verbally at BeginningTreatment.  This assertion is supported by a comment  from a therapist, "Child stated he didn't enjoy the Culture-Free Self-Esteem Inventory (CFSEI-2) as much as the RPS, perhaps due to the fact that this child is uncomfortable with verbal strategies" (see Appendix G1). As previously cited (Harper, 1991; Kaufman & Wohl, 1992; Mannarino et al., 1989), it is possible that the internal working models of sexually abused children are affected differently than the internal working models of children suffering from other types of abuse and trauma.  In turn, sexually abused children create and  employ alternative defense mechanisms to protect themselves.  For example, most decision making of  sexually abused children may be based on their own external environmental cues, to the exclusion of their internal feelings.  They may have responded to the  pictures in the manner that they thought the therapist would approve of, rather than recognizing or acknowledging how they truly felt about the picture. Children initially process and organize information at a sensory level and then use higher cognitive processing abilities to label the information (Burgess & Hartman, 1993; Stumer et al., 1980).  Acute  136 anxiety may tap into an inner psychic process that is internally processed in sexually abused children in a different way than their observable behavioural responses, similar to the findings of Stumer et al. (1980).  This type of defensive response to external  stimuli (Rosebush Pictures), may have originated in the child's early stages of development (Steele, 1983).  Rosebush Picture Sort (RPS) Mid-Treatment Stage: Pictures 2, 6, 8, and 9 discriminated between Group 1 sexual abuse children. Group 2 other children in treatment and Group 3 no abuse, no treatment children (see Appendix D-2).  These same pictures  discriminated between abused/non-abused and high/low self-esteem children in the earlier Rosebush Picture studies (Bowden, 1991; Carter et al., 1992), e.g., sexually abused children with low self-esteem strongly identified with Picture 6.  According to the  discriminant analysis, 75.0% of Group 1 sexual abuse participants were correctly classified.  In viewing the  All-Groups Scatterplot (see Figure 4 ) , Group 1 sexual abuse, falls to the left side of the plot; Group 2 other treatment, center around the middle; and Group 3 no abuse, no treatment, are entirely on the right side. The chi-square (df 6, p .02) analysis of Picture 8 (see Appendices D-2; E-4) showed a significant  137 difference between the two groups of children in treatment and the children in Group 3, no abuse, no treatment.  Ninety percent of Group 3 did not identify  with Picture 8.  Close to 43% of Group 1 sexual abuse  and 32% of Group 2, other treatment, did identify with Picture 8.  This picture shows an isolated, "sleeping"-  faced figure.  The children's level of identification  with the picture may be a symbolic representation of the way they feel about themselves.  The Mid-Treatment  children in Group 1, who did not identify with Picture 8, appear to be representative of Group 1 sexual abuse Picture 8 selection in all treatment stages. Of those in the Mid-Treatment stage Group 1 sexual abuse who were misclassified, the males (n = 7) outnumber the females (n = 2). Five males chose pictures more similar to the RPS selection of Group 2, and two males chose pictures more similar to the RPS selection of Group 3.  Both females chose pictures more  similar to those of the Group 2 selection.  The  demographics of the misclassified subset are typical of the total sample in geographic area, age range, and parent status categories.  However, 22% of the subset  are Aboriginal, in comparison to the 6% Aboriginal in the total Mid-Treatment stage. A possible explanation for the misclassified, sexually abused males and females choosing pictures  138 more similar to those of Group 2 is because (a) their internal working model remains defended, and although the therapist knows they have been sexually abused and believes they are at Mid-Treatment stage, their defenses block the visual stimuli from tapping into their internal feelings regarding sexual abuse; (b) other abusive circumstances are less threatening for them to deal with and, in turn, serve as their focus and defense barriers against the sexual abuse issues; (c) they become dissociated when presented with any task they feel is threatening to their sense of self; or (d) their picture choices were merely chance selection. In Mid-Treatment Group 2 other treatment, one male participant and one female participant made a selection of Rosebush Pictures similar to Group 1 sexual abuse. It is possible that these children may be withholding sexual abuse information.  It is suggested that  although these Group 2 participants have not verbalized sexual abuse events, their internal working models (psychological organization of thoughts, feelings, memories, and defenses) unconsciously identify Rosebush Pictures that are similar to those chosen by Group 1 sexual abuse participants. The two male participants in Group 2 other treatment, and two male participants in Group 1 sexual  139 abuse, made selections similar to Group 3 no treatment, no abuse.  Their ideal self image may be the part they  are presenting in their picture selection, similar to responses in previously cited findings (Livingston, 1987; Stovall & Craig, 1990). Within Group 3 no abuse, no treatment, the RPS selection by one male and one female participant is more similar to those of the Group 2 other treatment, sample.  One female in Group 3 selected pictures  similar to the selection of the Group 1 sexual abuse sample.  It is possible that these three children have  suffered some type of abuse that has not yet been disclosed.  Rosebush Picture Sort (RPS) End-Treatment Stage: There were no significant differences in the RPS scores of the three groups at End-Treatment stage. The chi-square analysis showed a noticeable trend in Picture 5 (see Appendices D-2; E-5), with the sexually abused children and other children in treatment identifying with this picture more than did children not in treatment.  The internal working models of Group  1 and Group 2 participants may be more sensitive to acknowledging positive qualities within themselves as a result of therapy and, in turn, identifying with the  140 positive aspects of Picture 5 more strongly than Group 3 no treatment, no abuse children. A young sexually abused female told her therapist she did not like "m" birds [RPS Picture 2], or birds drawn like the letter "m".  Children who have suffered  anal penetration often draw this shape in their pictures (J. A. B, Allan, personal communication, Dec. 14, 1989).  This 8 year old child was sexually abused  by adults during her first two years of life. Her statement provides the therapist with an important area to investigate during treatment. In summary, it appears that the format of the Rosebush Picture Sort was highly acceptable to the child participants and their therapists.  Only 3% of  the sample made set choice responses. Three of the males who made set choice responses on the RPS were in the Mid-Treatment stage; two are from Group 1 sexual abuse, and one is from Group 2 other treatment.  The fourth male is from Beginning-  Treatment stage. Group 1 sexual abuse.  According to  the discriminant analysis, the four male RPS selections fit their Group placement, and although they made set choice responses (choosing option 4 for over 90% of the pictures), they were correctly classified within their diagnostic group.  141 The need for a projective technique that is sensitive to the effects of sexual abuse when compared with the effects of other types of abuse becomes more critical as the understanding of the short term and long term impact of sexual abuse on the personality development of the child is understood.  The Rosebush  Picture Sort provides a non-threatening, visual stimulus, allowing access to primary process thinking within the individual.  The findings and information  obtained in this study appear to support the use of the Rosebush Picture Sort as a diagnostic tool that addresses the need for a differential, psychodynamic, assessment technique.  Hypothesis 2 The self-esteem mean-scores of sexually abused children at different stages of treatment will not differ from those of other children at similar stages of treatment and will not differ from the mean-scores of non-abused, non-treatment children.  Culture-Free Self-Esteem Inventorv-2 (CFSEI-2) Findings  Twelve percent of Group 1 sexual abuse CFSEI-2 scores were deleted from the analysis due to low Lie  142 scoreso  Lie scores below 40 were seen in only 6% of  Group 2 other treatment.  This defense mechanism is  twice as significant in the sexually abused sample than in the other children in treatment sample.  None of  Group 3 had Lie scores below 40.  Self-Esteem Beginning-Treatment Stage: According to the ANOVA, there are no differences in the self-esteem levels of the sexually abused, other treatment, or non-abused children at BeginningTreatment, as measured by the CFSEI-2. This finding supports other study findings that have shown that low self-esteem is not always a distinguishing characteristic of sexually abused children (KendallTackett et al., 1993).  However, some studies analyze  participant data following hospital in-take assessment or use previous hospital records for retrospective research.  The study findings are based on data  obtained either at the beginning of treatment, or at a stage in treatment that is unknown (Cosentino et al., 1993; Gomes-Schwartz et al., 1985; Gold, 1986; Mannarino et al., 1989; Miller et al., 1987; Stovall & Craig, 1990).  This type of data collection leads one  to question the reliability of the composite selfesteem findings in the literature.  143 Self-Esteem Mid-Treatment Stage: The ANOVA of the self-esteem means reveal that the participants in Group 3 no treatment, no abuse, are significantly different than the other two groups. It appears that sexually abused children and other children in treatment have lower self-esteem than children not in treatment, according to their responses on Culture-Free Self-Esteem Inventory-2 (CFSEI-2), but the scores of the two groups of children in treatment do not significantly differentiate from one another. There is a possibility that the normative school population on which the CFSEI-2 was standardized contained unidentified, sexually abused children and unidentified, other children in need of treatment or at different stages of treatment.  This possibility could  affect the strength of the test item selection in regard to discriminating the self-esteem levels of sexually abused children from other children, as shown in the Rosebush Picture Sort findings. As stated earlier, many of the CFSEI-2 items reflect problems that sexually abused children experience.  The fact that the two treatment groups do  not differentiate when completing a verbal exercise, but do differentiate when completing a non-verbal exercise, strengthens support for the diagnostic value of the differential, non-verbal RPS technique.  144 Self-Esteem End-Treatment Stage: There were no differences in the self-esteem levels of the three Groups at the End-Treatment stage. It is postulated that this situation may represent the consequences of the therapeutic healing that the sexually abused and other children in treatment have experienced.  In summary, the strengths of this study lie in: (a) the randomly selected population sample; (b) the application of a non-threatening, non-verbal measurement technique to the three sample groups at different stages of treatment;  (c) the collection of  sensitive demographic data from the child's therapist, rather than imposing this task on child participants or their parents; and (d) the comparison of participants' self-esteem at different stages of treatment.  The  demographic information about the three groups of participants indicates that the samples were comparable with respect to range of abuse, therapist background, child participant age range, parental status, ethnic origin, and residential area.  This information enables  comparisons with other studies to be made and the replication of this study in other settings. The study findings show no differences in Rosebush Picture Sort selections or Culture-Free Self-Esteem  145 scores of no abuse, no treatment children, sexually abused children and other children at the Beginning or End-Treatment stage.  Sexually abused children's RPS  selections at the Mid-Treatment stage are different than both comparison groups.  These findings indicate  that sexually abused children identify RPS pictures differently than other children.  The findings further  suggest that sexual abuse may affect the internal working models of children differently than other types of abuse and trauma. The CFSEI-2 scores of Group 1 and Group 2 in MidTreatment differ from the no abuse, no treatment Group 3, but not from one another.  It is possible to reason  that the verbal-auditory response mode of the CFSEI-2 instrument does not tap into the internal dynamics of the sexually abused child, although the Inventory is sensitive enough to discriminate the self-esteem levels of coping and non-coping children.  Limitations of Study The demographic data was obtained from the caseloads of 31 therapists or psychologists.  It is  possible that the therapists or psychologists made diagnostic errors and mislabelled the stage of treatment, although according to the statistical analyses, the misclassified children represent only 12%  146 (n = 9) of the Mid-Treatment study sample.  The size of  the no abuse, no treatment sample was small, particularly in comparison with the Mid-Treatment groups.  This small size influences the strength of the  findings, particularly in the Mid-Treatment stage.  The  Beginning-Treatment stage was defined as one to three sessions with the assumption that the children would not have developed a strong relationship with the therapist during that time. incorrect.  This assumption might be  Because previous study findings of these  pictures differentiated the abused and non-abused children, it is possible that the non-differentiating Rosebush Pictures in this study may hold greater significance with other participant samples or when presented to participants in a different format.  The  type of measurement of Rosebush Picture selection needs to be considered.  For example, following the  identification of significant Rosebush Pictures through Q-Sort measurement, the use of an interval measurement and weighted Likert scales with these same pictures would strengthen the technique. To continue the development of this non-verbal, diagnostic, classification model, future RPS studies will benefit from using cross-validation procedures with data from this study, including an identified sample as a fourth comparison group.  147 The study sample is not representative of all children in therapy, but rather is representative of sexually abused children and other children in outpatient treatment and children whose parents are in health care professions in a Western Canadian province. Finally, this RPS study is a preliminary work in an on-going research project.  At this point in time,  children's selection of a rosebush drawing or the selection of specific rosebush pictures cannot be considered as indicators of sexual abuse status.  148 CONCLUSIONS  This study (N = 123) contrasted a group of sexually abused children in treatment, aged 6 to 12 years, with two comparable groups—other non-sexually abused children in treatment, and non-abused, nontreatment children—to determine whether differences in Rosebush Picture selection could be demonstrated.  All  children underwent evaluation procedures that included completion of a 12 picture projective measure (the Rosebush Picture Sort) and the Culture-Free Self-Esteem Inventory-2.  A demographic questionnaire for each  child was completed. Results showed no differences in picture selection or self-esteem scores of children at the Beginning and End-Treatment stage.  Sexually abused children's  Rosebush Picture Sort selections at Mid-Treatment were different than both comparison groups.  The self-esteem  scores of the two clinical groups in Mid-Treatment were significantly lower than the scores of the non-abused, non-treatment group, but did not differ from one another.  The findings indicated that at Mid-Treatment,  sexually abused children identify Rosebush Picture Sort Pictures differently than other children.  The findings  of this study further indicated that sexually abused children, in all stages of treatment, provided few  149 unsolicited verbal expressions of their personal opinions but they had no difficulty in responding to the RPS Pictures, It is possible that sexual abuse affects the internal working models of children differently than other types of abuse and trauma. In using the RPS findings, it is important to remember that the Rosebush Picture Sort findings are based on the interactions of the picture set, not a single picture within the set, and if a child selects pictures similarly to Group 1, Group 2 or Group 3, the picture selection alone is not sufficient information for a clinical diagnosis.  Future Research It is hoped that future Rosebush Picture Sort research will identify a number of Rosebush Pictures that discriminate between sexually abused children and other children.  Using a larger representative sample  of three study groups, including minority groups, will also be of value in the continuing development of this classification technique.  It is suggested that future  demographic information include parental status and whether the abuse is intra-familial or extra-familial, as these variables may influence the degree of impact  150 of sexual abuse on children.  The intra-familial or  extra-familial variable may influence the participant's Rosebush Picture Sort selection. Accidental detection of sexual victimization of school-age children is less probable than in other age groups.  It is important to have a non-threatening,  diagnostic technique that will facilitate the possible detection of sexual abuse in this age population, to aid in the protection of children. Learning from the stage of treatment findings in this study, it will be helpful in future sexual abuse studies to define and include the stage of treatment of the participants.  Identifying Rosebush Pictures that  tap into gender identity issues for this age group will be of diagnostic and therapeutic value.  As reported in  these findings, it would be interesting to compare histories of language development in sexually abused children and other children.  The legal and social  service system personnel need to be informed as to the possible limitations of verbal expression in sexually abused children, particularly when regarding children experiencing other kinds of abusive circumstances. Future research may need to investigate the formulation of sexually abused children's defense mechanisms when compared with the development of defense mechanisms of other children.  151 Future research in assessment of sexually abused children needs to continue to explore the development of a non-verbal projective measurement as a way of tapping into childrens' unconscious perception of self in relationship to their conscious perception of self. Learning to identify discrepancies between these two perceptions will be of great value to the clinical findings in the assessment process.  152 REFERENCES Acton, B. & Moretti, M. (1993, May). Hope for human drawings: A new look at drawing features. Poster paper presented at the Canadian Psychological Association Annual Convention, Montreal, Que. Achenbach, T. & Edelbrock, C. (1984). Child Behavior Checklist. Burlington, VT: University of VT. Ainsworth, M. (1991). Attachments and other affectional bonds across the life cycle. In C. Parkes, J. Stevenson-Hinde & P. Marris (Eds.), Attachment across the life cycle (pp. 33-51). NY: Tavistock/Routledge. Allan, J. (1988). Inscapes of the child's world: Counseling in schools and clinics. Dallas: Spring. Allan, J. & Bertoia, J. (1992). Written paths to healing: Education and Jungian child counselling. Dallas: Spring. Allan, J. & Crandall, J. (1986). The rosebush: A visualization strategy for possible identification of child abuse. School Guidance & Counselling, 21, 44-51. American Professional Society on the Abuse of Children (1990). Guidelines for psychological evaluation of suspected sexual abuse in young children (APSAC Task Force). Chicago: Lucy Berliner, Ammann, R. (1991). Healing and transformation in sandplay. LaSalle, IL.: Open Court Publishers, Anderson, J,, Martin, J., Mullen, P,, Romans, S. & Herbison, P, (1993), Prevalence of childhood sexual abuse experiences in a community sample of women, Journal of American Academy of Child/Adolescent Psychiatry. 32, 911-919, Bagley, C. & King, K. (1991). search for healing, NY:  Child sexual abuse: The Tavistock/ Routledge,  Baartman, E,M, (1992), The credibility of children as witnesses and the social denial of the incestuous abuse of children. In F. Losel, D, Bender, & T. Bliesener (Eds,), Psychology and law (pp, 345351). NY: Walter de Gruyter,  153 Battle, J. (1991). Self-esteem research: relevant findings. Edmonton, ALTA.: of Alberta Printing Services.  A summary of University  Battle, J. (1992). Culture-Free SEI: Self-Esteem Inventories for Children and Adults; Second Edition. Austin, TX: Pro-Ed, Inc.. Bellak, L. (1986). The TAT, CAT and SAT in clinical use (4th ed.). NY: Grune & Stratton. Berliner, L. (1988). Deciding whether a child has been sexually abused. In E.B. Nicholson & J. Bulkley (Eds.), Sexual abuse allegations in custody and visitation cases: A resource book for judges and court personnel (pp. 102-123). DC: American Bar Association. Berliner, L. (1991). Therapy with victimized children and their families. In J. Briere (Ed.), Treating victims of child sexual abuse (pp. 29-46). Toronto: Jossey-Bass. Berliner, L. & Conte, J. (1993). Sexual abuse evaluations: Conceptual and empirical obstacles. Child Abuse & Neglect. 17, 111-125. Bowden, S. (1991). Rosebush Q-Sort. Unpublished master's thesis. University of British Columbia, Vancouver, BC. Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. NY: Basic Books, Inc. Bretherton, I. (1990). Open communication and internal working models: Their role in the development of attachment relationships. In R.A. Thompson (Ed.), Socioemotional development (pp. 57-113). Omaha: University of Nebraska Press. Bretherton, I. (1991). The roots and growing points of attachment theory. In C. Parkes, J. StevensonHinde & P. Harris (Eds.), Attachment across the life cycle (pp. 9-32). NY: Tavistock/ Routledge. Brewin, C , Andrews, B. & Gotlib, I. (1993). Psychopathology and early experience: A reappraisal of reprospective reports. Psychological Bulletin. 113. 82-98.  154 Briere, J.N. (1992). Child abuse trauma: Theory and treating of the lasting effects. Newbury Park, CA: Sage. Briere, J. & Runtz, M. (1991). The long-term effects of sexual abuse: A review and synthesis. In J. Briere, (Ed.), Treating victims of child sexual abuse (pp. 3-14). Toronto: Jossey-Bass. Browne, A. & Finkelhor, D. (1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99, 66-77. Buck, J.N. (1970). The House-Tree-Person Technigue, Revised Manual. Los Angeles: Western Psychological Services. Burgess, A. & Hartman, C. (1993). Children's drawings. Child Abuse & Neglect. 17, 161-168. Burgess, A., McCausland, M., & Wolbert, W. (1981). Children's drawings as indicators of sexual trauma. Perspectives in Psychiatric Care. 2, 50-58. Burns, R.C. & Kaufman, S. (1970). Kinetic Family Drawings: (K-F-D). NY: Brunner/Mazel. Campis, L.B., Hebden-Curtis, J. & Demaso, D.R. (1993). Developmental differences in detection and disclosure of sexual abuse. Journal of American Academy of Child/Adolescent Psychiatry. 32, 920924. Carter, M.A. & Allan, J.A.B. (1992). Counselling issues in the therapeutic alliance with sexually abused children: Counsellor role, psychological development, and resistances. Guidance & Counselling. 7, 34-44. Carter, M.A., Allan, J.A.B., & Boldt, W.B. (1992). Projective assessment of child sexual abuse. British Journal of Projective Psychology, 37, 50-60. Cavanaugh-Johnson, T. (1992, November). Treating sexually intrusive children. Paper presented at workshop of Act II, Douglas College & B.C. Institute of Justice, Coquitlam, BC. Chaplin, J. (1975). Dictionary of psychology. York: Dell Publishing.  New  155 College of Psychologists of British Columbia (1990)» Guidelines for the conducting of psychological assessments in cases involving sexual abuse(CPBC Task Force). Vancouver, BC.: R. Colby, J. Ternes, M.A. Carter, M. Elterman & P. Wilensky. Conte, J. R. (1985). The effects of sexual victimization on children: A critique and suggestions for future research. Victimology: The International Journal, 10, 110-130. Conte, J.R. (1991). The therapist in child sexual abuse: The context of helping. In J. Briere (Ed.), Treating victims of sexual abuse (pp. 87-98), Toronto: Jossey-Bass. Conte, J. & Schuerman, J. (1988). The effects of sexual abuse on children: A multidimensional view. In G. Wyatt & G. Powell (Eds.), Lasting effects of child sexual abuse (pp. 157-171). Newbury Park, CA: Sage. Cornman, B.J. (1988). Impact of childhood cancer on the family. Unpublished doctoral dissertation. University of Washington, Seattle, Crandall, J. (1984). Projective technigues with children: Assessment through guided imagery, drawing, and post-drawing inguiry. Unpublished master's thesis. University of British Columbia, Vancouver, BC. Crittenden, P.M. (1989, September). Internal representational models of attachment relationships. Paper presented at the meeting of the World Association of Infant Psychiatry and Allied Disciplines, Lugano, Switzerland. Crittenden, P.M. (1992). Treatment of anxious attachment in infancy and early childhood. Unpublished manuscript. Crittenden, P.M. & Ainsworth, M.D.S. (1989). Child maltreatment and attachment theory. In D. Cicchetti and V. Carlson (Eds.), Child maltreatment: Theory and research on the causes and conseguences of child abuse and neglect (pp, 432-463). Cambridge: Cambridge University Press. Dawis, R.V. (1987). Scale construction. Counseling Psychology, 34, 481-489.  Journal of  156 Deblinger, E., Mcleer, S.V., Atkins, M.S., Ralphe, D. & Foa, E. (1989). Post-traumatic stress in sexually abused, physically abused, and nonabused children. Child Abuse & Neglect, 13, 403-408. deYoung, M. (1992). Credibility assessment during sexual abuse evaluation. In W.O'Donohue and J.H. Geer (Eds.) Sexual Abuse of Children, Vol. 2 (pp. 256-282). Hillsdale, NJ: Erlbaum. DiLeo, J.H. (1983). Interpreting children's drawings. New York: Brunner/Mazel. Bollinger, S. & Cramer, P. (1990). Children's defensive responses and emotional upset following a disaster: A projective assessment. Journal of Personality Assessment, 54, 116-127. Dubowski, J. (1990). Art versus language: Separate development during childhood. In C. Case & T. Dalley (Eds.), Working with children in art therapy (pp. 7-22). NY: Tavistock/Routledge. Egeland, B. & Sroufe, A. (1981). Developmental sequelae of maltreatment. In R. Rizley & D. Cicchetti (Eds.), Developmental perspectives on child maltreatment (pp. 77-92). San Francisco: Jossey-Bass. Exner, J. (1991). The Rorshach: A comprehensive system. Interpretation. 2nd Ed., Vol. 2.. NY: Wiley. Faller, K. (1990). maltreatment.  Understanding child sexual Newbury Park, CA.: Sage.  Farber, E.A. & Egeland, B.E. (1987). Invulnerability among abused and neglected children. In E.J. Anthony & B.J. Cohler (Eds.), The invulnerable child (pp. 253-288). NY: The Guilford Press. Finkelhor, D. (1979). What's wrong with sex between adults and children? American Journal of Orthopsychiatry, 49, 692-697. Finkelhor, D. (1988). The trauma of child sexual abuse: Two models. In G. Wyatt & G. Powell (Eds.), Lasting effects of child sexual abuse (pp. 61-82). Newbury Park, CA: Sage.  157 Finkelhor, D. (1993). Epidemiological factors in the clinical identification of child sexual abuse. Child Abuse & Neglect. 17. 67-70. Finkelhor, D. & Browne, A. (1985). The traumatic impact of child sexual abuse: A conceptualization. American Journal of Orthopsychiatry. 55, 530-540. Fischer, K. & Pipp, S. (1984). Development of the structures of conscious thought. In K, Bowers & D. Meichenbaum (Eds.), The unconscious reconsidered (pp. 88-148). NY: Wiley. French, L. (1993). Adapting projective tests for minority children. Psychological Reports, 72. 1518. Friedrich, W.N. (1990). Psychotherapy of sexually abused children and their families. NY: W.W. Norton. Friedrich, W.N. (1991). Sexual behavior in sexually abused children. In J. Briere (Ed.), Treating victims of child sexual abuse (pp. 15-28). Toronto: Jossey-Bass. Friedrich, W.N. (1993, May). An integrated perspective on working with sexually abused children and their families. Paper presented at the meeting of the National Children's Mental Health Conference, Seattle, WA. Friedrich, W.N., Grambsch, P., Broughton, D., Kuiper, J. & Beilke, R.L. (1991). Normative sexual behavior in children. Pediatrics, 88, 456-464. Friedrich, W., Urquiza, A. & Beilke, R. (1986). Behavior problems in sexually abused young children. Journal of Pediatric Psychology, 11. 47-57. Garbarino, J., Guttman, E., & Seeley, J. (1987). The psychologically battered child. San Francisco: Jossey-Bass. Garbarino, J. Stott, F.M. & Faculty of the Erikson Institute (1989), What children can tell us. San Franscisco: Jossey-Bass. Garsee, J.W. & Schuster, C.S. (1992). Moral development. In C.S.Schuster & S.S. Ashburn  158 (Eds.)/ The process of human development: A holistic life-span approach (pp. 467-489). Philadelphia: J.P. Lippincott & Co. Gittleman-Klein, R. (1986). Questioning the clinical usefulness of projective tests for children. Developmental and Behavioral Pediatrics. 7, 378-382. Glass, G.V. & Hopkins, K.D. (1984). Statistical methods in education and psychology, second edition. Englewood Cliffs, NJ: Prentice Hall. Gold, E. (1986). Long-term effects of sexual victimization in childhood: An attributional approach. Journal of Consulting and Clinical Psychology, 54, 471-475. Gomes-Schwartz, B., Horowitz, J.M., & Sauzier, M. (1985). Severity of emotional distress among sexually abused preschool, school-age, and adolescent children. Hospital and Community Psychiatry, 36, 503-508. Green, A. (1993). Child sexual abuse: Immediate and long-term effects and intervention. Journal of American Academy Child/Adolescent Psychiatry, 32, 890-902. Hackbarth, S.G., Murphy, H.D. & McQuary, J.P. (1991). Identifying sexually abused children by using kinetic family drawing. Elementary School Guidance & Counselling, 25, 255-260. Hagood, M. (1992). Diagnosis or dilemma: Drawings of sexually abused children. British Journal of Projective Psychology, 37, 22-33. Harmer, E.F.(1986). Graphic techniques with children and adolescents. In A.I. Rabin (Ed.), Projective techniques for adolescents and children (pp. 239263). NY: Springer. Harper, J. (1991). Children's play: The differential effects of intrafamilial physical and sexual abuse. Child Abuse & Neglect, 15, 89-98. Harris, D. (1963). Goodenough-Harris Drawing Test. NY: Harcourt. Brace & World. Harris, P. (1989). Children and emotion. Blackwell Ltd.  NY:  Basil  159 Hartman, C.P. & Burgess, A.W. (1989), Sexual abuse of children: Causes and consequences. In D, Cicchetti & V. Carlson (Eds.)/ Child maltreatment: Theory and research on the causes and consequences of child abuse and neglect (pp. 95-128). Cambridge: Cambridge University Press. Haworth, M. (1986). Children's apperception test. Rabin, A.I. (Ed.), Projective techniques for adolescents and children (pp. 37-72). NY: Singer.  In  Herzog, D.B., Staley, J.E., Carmody, S., Robbins, W.M. & vanderKolk, B.A. (1993). Journal of American Academy Child/Adolescent Psychiatry, 32, 962-966. Hibbard, R.A. & Hartmann, G.L. (1990). Emotional indicators in human figure drawing of sexually victimized and nonabused children. Journal of Clinical Psychology, 46, 211-219. Hibbard, R.A., Roghmann, K. & Hoekelman, R.A. (1987). Genitalia in children's drawings: An association with sexual abuse. Pediatrics, 79, 129-137. Hinde, R. & Stevenson-Hinde, J. (1991), Perspectives on attachment. In C. Parkes, J. Stevenson-Hinde & P. Harris (Eds.), Attachment across the life cycle (pp. 52-65). NY: Tavistock/Routledge. Hoier, T., Shawchuck, C. Pallotta, G., Freeman, T., Inderbitzen-Pisaruk, H., MacMillan, V., MalinoskyRummell, R. & Greene, A. (1992). The impact of sexual abuse: A cognitive-behavioral model. In W. 0'Donohue and J.H. Geer (Eds.), Sexual abuse of children. Vol. 2 (pp. 100-141). Hillsdale, N.J.: Erlbaum. Hopkins, J. (1991). Failure of the holding relationship: Some effects of physical rejection on the child's attachment and inner experience. In C. Parkes, J. Stevenson-Hinde & P. Harris (Eds.), Attachment across the life cycle (pp. 187-198). NY: Tavistock/Routledge. Horner, T.H., Guyer, H.J. & Kalter, N.M. (1993). Clinical expertise and the assessment of child sexual abuse. Journal of American Academy Child/Adolescent Psychiatry. 32, 925-933.  160 Horvath, A.O. & Greenberg, L.S. (1989). Development and validation of the working alliance inventory. Journal of Counseling Psychology, 36, 223-233. Jaffe, L. (1990). The empirical foundations of psychoanalytic approaches to psychological testing. Journal of Personality Assessment, 55, 746-755. Jones, L. (1989). Hearts wish. Early Child Development and Care, 42, 175-182. Jung, C.G. (1976). Publishing.  Man and his symbols.  NY:  Dell  Kaufman, B. & Wohl A. (1992). Casualties of childhood: A developmental perspective on sexual abuse using projective drawings. NY: Brunner/Mazel. Kendall-Tackett, K., Williams, L. & Finkelhor, D. (1993). Impact of sexual abuse on children: A review and synthesis of recent empirical studies. Psychological Bulletin. 113, 164-180. Klepsch, M. & Logie, L. (1982). Children draw and tell. NY: Brunner/Mazel. Kobak, R.R. & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation, and representations of self and others. Child Development, 59, 135-146. Kolko, D.J., Moser, J.T. & Weldy, S.R. (1988). Behavioral/emotional indicators of sexual abuse in child psychiatric inpatients: A controlled comparison with physical abuse. Child Abuse & Neglect, 12, 529-541. Korner, A. (1965). Theoretical considerations concerning the scope and limitations of projective techniques. In B. Murstein (Ed.), Handbook of projective technigues (pp. 23-34). NY: Basic Books, Inc. Levine, J, & Levine, R. (1986), Projective techniques with children in the clinic setting. In A, Rabin (Ed,), Projective technigues for adolescents and children (pp. 323-337). NY: Springer. Lewis, H.P, & Livson, N. (1980). Cognitive development, personality and drawing: Their  161 interrelationships in a replicated longitudinal study. Studies in Art Education, 22, 8-11. Livingston, R. (1987). Sexually and physically abused children. Journal of the American Academy of Child/Adolescent Psychiatry, 26, 413-415. Livingston, R., Lawson, L. & Jones, J.G. (1993). Predictors of self-reported psychopathology in children abused repeatedly by a parent. Journal of American Academy Child/Adolescent Psychiatry, 32, 948-953. MacFarlane, K. & Bulkley, J. (1982). Treating child sexual abuse: An overview of current program models. In Social work and child sexual abuse (pp. 69-91). Binghamtom, NY: The Haworth Press. MacFarlane, K. & Krebs, S. (1986). Techniques for interviewing and evidence gathering. In K. MacFarlane, J. Waterman, S. Conerly, L. Damon, M. Durfee, & S. Long (Eds.), Sexual abuse of young children (pp. 67-100). NY: The Guilford Press. McElroy, L.P. & McElroy, R.A. (1989). Psychoanalytically oriented psychotherapy with sexually abused children. Journal of Mental Health Counselling, 11, 244-258. McGovern, K. (1993). Was there really child sexual abuse or is there another explanation? Child & Youth Services, 15, 115-127. McGrew, M. & Teglasi, H. (1990). Formal characteristics of Thematic Apperception Test stories as indices of emotional disturbance in children. Journal of Personality Assessment, 54, 639-655. McKeown, B. & Thomas, D. (1988). Newbury Park, CA: Sage.  Q-Methodology.  Machover, K. (1953). Human figure drawings of children. Journal of Projective Techniques, 17, 85-91. Main, M., Kaplan, N. & Cassidy, J. (1988). Security in infancy, childhood and adulthood: A move to the level of representation. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research (pp. 66-106). Monographs of the  162 society for research in child development, serial no. 209, 50. Mannarino, A.P., Cohen, J.A. & Gregor, M. (1989). Emotional and behavioral difficulties in sexually abused girls. Journal of Interpersonal Violence. 4, 437-467. Marshall, W. (1993). Anachronistic obstacles to effective training in research: The dissertation, the thesis, and the comprehensive examination. Canadian Psychology, 34, 176-182. Masling, J. (1965). The influence of situational and interpersonal variables in projective testing. In B. Murstein (Ed.), Handbook of projective techniques (pp. 117-130). NY: Basic Books, Inc. Meehl, P. (1965). Structured and projective tests: Some common problems in validation. In B. Murstein (Ed.), Handbook of projective techniques (pp. 83-88). NY: Basic Books, Inc. Melton, G.B. & Limber, S. (1989). Psychologists' involvement in cases of child maltreatment. American Psychologist. 44. 1225-1233. Miller, T. & Veltkamp, L. (1989). Assessment of child sexual abuse: Clinical use of fables. Child Psychiatry and Human Development. 20, 123-133. Miller, T., Veltkamp, L., Janson, D. (1987). Projective measures in the clinical evaluation of sexually abused children. Child Psychiatry and Human Development. 18. 47-57. Murstein, B. (1965). Assumptions, adaptation level, and projective techniques. In B. Murstein (Ed.), Handbook of projective techniques (pp. 49-67). NY: Basic Books, Inc. Myers, J. (1993). Expert testimony regarding child sexual abuse. Child Abuse & Neqlect. 17. 175-185. Oaklander, V. (1978). Windows to our children. UT: Real People Press.  Moab,  Pearce, J.W. (1984). Characteristics of abused children: Research findings. Canada's Mental Health. 32, 2-6.  163 Piers, E.V. (1990). Pier-Harris Children's Self-Concept Scale. Los Angeles: Western Psychological Services. Porter, F.S., Blick, L. & Sgroi, S. (1985). Treatment of the sexually abused child. In S. Sgroi (Ed.), Handbook of clinical intervention in child sexual abuse (pp. 109-145). Toronto: Lexington Books. Province of British Columbia, Ministry of Social Services (1989). Protecting our children ISBNO7726-7176-1. Prytula, R.E. & Thompson, N.D. (1973). Analysis of emotional indicators in human figure drawings as related to self-esteem. Perceptual and Motor Skills. 37, 795-802. Pynoos, R.S. & Eth, S. (1986). Witness to violence: The child interview. Journal of the American Academy of Child Psychiatry, 25, 306-319. Rabin, A. (1986). Concerning projective techniques. In A. Rabin (Ed.), Proiective techniques for adolescents and children (pp. 3-13). NY: Springer. Rabin, A. & Doneson, S.L. (1986). Miscellaneous projective techniques. In A.I. Rabin (Ed.), Projective techniques for adolescents and children (pp. 306-323). NY: Singer. Raczek, S. (1992). Childhood abuse and personality disorders. Journal of Personality Disorders, 6, 109-116. Reinhart, M. (1987). Sexually abused boys. Abuse & Neglect, 11. 229-235.  Child  Reynolds, C.R. & Richmond, B.O. (1990). Revised Children's Manifest Anxiety Scale TRCMASI. Los Angeles: Western Psychological Services. Riordan, R.J. & Vardel, P. (1991), Evidence of sexual abuse in children's art products. The School Counsellor. 39, 116-121. Robin, M. (1993). Beyond validation interviews: An assessment approach to evaluating sexual abuse allegations. Child & Youth Services. 15, 93-114.  164 Rogers, R.G. (1990). Reaching for solutions: The report of the special advisor to the Ministry of National Health and Welfare on child sexual abuse in Canada. Ottawa: Minister of National Health and Welfare. Rubin, J., Schachter, J. & Ragins, N. (1983). Intraindividual variability in human figure drawings: A developmental study. American Journal of Orthopsychiatry. 53, 654-667. Russ, S. & Grossman-McKee, A. (1990). Affective expression in children's fantasy play, primary process thinking on the Rorschach, and divergent thinking. Journal of Personality Assessment. 54. 756-771. Schetky, D.H. (1990). A review of the literature on the long-term effects of childhood sexual abuse. In R.P. Kluft (Ed.), Incest related syndromes of adult psvchopathology (pp. 35-54). DC: American Psychiatric Press. Schneider-Rosen, K., Braunwald, K., Carlson, V. & Cicchetti, P. (1988). Current perspectives in attachment theory: Illustrations from the study of maltreated infants. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research (pp. 194 -210). Monographs of the society for research in child development, serial no. 209, 50. Sgroi, S.M. (1989). Vulnerable populations: Vol. 2. Toronto: Lexington Books. Sgroi, S., Blick, L. & Porter, F. (1985). A conceptual framework for child sexual abuse. In S. M. Sgroi (Ed.), Handbook of clinical intervention in child sexual abuse (pp. 9-37). Toronto: Lexington Books. Shapiro, J., Leifer, M., Martone M., & Kassem, L. (1990). Multimethod assessment of depression in sexually abused girls. Journal of Personality Assessment, 55, 234-248. Shneidman, E.S. (1952). Projective Techniques Monographs: Manual for the Make A Picture Story Method. The Society for Personality Assessment, Inc.  165 Singer, J. (1981). Research applications of projective methods. In A.I. Rabin (Ed.), Assessment with projective techniques: A concise introduction (pp. 297-331). NY: Springer. Smith, G. (1992). The unbearable traumatogenic past: Child sexual abuse. In V.P. Varma (Ed.), The secret life of vulnerable children (pp. 130-156). NY: Routledge. Steele, B.F. (1983). The effect of abuse and neglect on psychological development. In J.D. Call, E, Galenson, & R.L. Tyson (Eds.), Frontiers of infant psychiatry (pp. 42-51). NY: Basic Books. Steller, M. (1992). Child witnesses in sexual abuse cases: Psychological implications of legal procedures. In F. Losel, D. Bender, T. Bliesener (Eds.), Psychology and law (pp. 360-364). NY: Walter de Gruyter. Stevens, J. (1971). Awareness: Exploring. experimenting, experiencing. Moab, UT: People Press.  Real  Steward, M., Bussey, K., Goodman, G., Saywitz, K, (1993). Implications for developmental research for interviewing children- Child Abuse & Neglect, 17, 25-37. Stovall, G. & Craig, R.J. (1990). Mental representation of physically and sexually abused latency-aged females. Children Abuse & Neglect. 14/ 233-242. S t u m e r , R.A., Rothbaum, F., Visantainer, M. & Wolfer,J. (1980). The effects of stress on children's human figure drawings. Journal of Clinical Psychology. 36, 324-331. Sugarman, A. (1991). Where's the beef? Putting personality back into personality assessment. Journal of Personality Assessment, 56, 130-144. Sugarman, A. (1992). The psychoanalytic approach to inference making during diagnostic testing. British Journal of Projective Psychology. 37, 3449. Sununit, R. (1983). The child sexual abuse accommodation syndrome. Child Abuse & Neglect, 7, 179-193.  166 Tabachnick, B.G. & Fidell, L.S. (1989). Using multivariate statistics, second edition. NY: Harper Collins. Terr, L. (1990). Row.  Too scared to cry.  NY:  Harper &  Tong, L., Oates, K. & McDowell, M. (1987). Personality development following sexual abuse. Child Abuse & Neglect, 11, 371-383. Tritell, S.A. (1988). Diagnostic and therapeutic uses of drawings of children who have been sexually abused. Unpublished doctoral dissertation, Wisconsin School of Professional Psychology, Milwaukee, WI. Wakefield, H. & Underwager, R.C. (1989, April). Technigues for interviewing children in sexual abuse cases. Paper presented at the Fifth Annual Symposium in Forensic Psychology, San Diego, CA. Waterman, J. & Lusk, R. (1993). Psychological testing in evaluation of child sexual abuse. Child Abuse & Neglect. 17, 145-159. Waters, E. & Deane, K. (1988). Defining and assessing individual differences in attachment relationships: Q-Methodology and the organization of behavior in infancy and early childhood. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research (pp. 194 -210). Monographs of the society for research in child development, serial no. 209, 50. Westen, D. (1991). Clinical assessment of object relations using the TAT. Journal of Personality Assessment, 56, 56-74. Westen, D., Klepser, J., Ruffins, S., Silverman, M., Lifton, N., & Bockamp, J. (1991). Object relations in childhood and adolescence: The development of working representations. Journal of Consulting and Clinical Psychology, 59, 400409. White, S., Halpin, B.M., Strom, G., Santelli, W. (1988). Behavioral comparisons of young sexually abused, neglected and nonreferral children. Journal of Clinical Child Psychology, 17, 53-61.  167 Wohl, A. & Kaufman, B. (1985). Silent screams and hidden cries. NY: Brunner/Mazel. Wolfe, D. (1990). Child abuse: Implications for child development and psvchopathology. Newbury Park, CA: Sage. Wolfe, V.V. & Gentile, C. (1992). Psychological assessment of sexually abused children. In W.0'Donohue and J.H. Greer (Eds.), The sexual abuse of children: Theory, research, and therapy (pp. 143-187). Hillsdale, N.J.: Erlbaum. Yates, A., Beutler, L.E. & Crago, M. (1985). Drawings lay child victims of incest. Child Abuse and Neglect, 9, 183-189. Zubin, J., Eron, L. & Schumer, F. (1965). The challenge of projective techniques. In J. Zubin, L. Eron & F. Schumer (Eds.), An experimental approach to proiective techniques (pp. 1-49). NY: John Wiley & Sons, Inc.  168  APPENDICES  169  APPENDIX A - PRELIMINARY STUDY BAR GRAPH  170 Appendix A  BAR GRAPH FROM PRELIMINARY STUDY Bar Graph Pictures Picture #7 Picture #18 Picture #21 Picture #1 Picture #8 Picture #15 See Bowden (1991) See Bowden (1991) See Bowden (1991) See Bowden (1991) Picture #20 Picture #6  correspond with  Rosebush Pictures RPS RPS RPS RPS RPS RPS RPS RPS RPS RPS RPS RPS  #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12  2 score 2.5 2.0 1,5 1.0 0.5 0 -0.5 -1.0 -1.5 -2.0-  m 0  TYPE 1 TYPE 2  -2.5picturef 1_ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 emotional neglect non-abuse physical abuse sexual abuse FIGURE 5 .  BAR  GRAPH OF PICTURE SCORES FOR TYPE 1 AND TYPE 2  171  APPENDIX B - RESEARCH APPROVAL FORMS  172 Appendix B-1 B92-267  The University of British Columbia Office of Research Services  BEHAVIOURAL SCIENCES SCREENING COMMITTEE FOR RESEARCH AND OTHER STUDIES INVOLVING HUMAN SUBJECTS C E R T I F I C A T E  of  A P P R O V A L  INVESTIGATOR:  Allan, J.  UBC DEPT:  Counselling Psychology  INSTITUTION:  Community agencies;office of therapists  TITLE:  Rosebush picture sort  NUMBER:  B92-267  CO-INVEST:  Carter, M.A.  APPROVED:  I s  The protocol describing the above-named project has been reviewed by the Committee and the experimental procedures were found to be acceptable on ethical grounds for research involving human subjects.  Director, Research Services and Acting Chairman  THIS CERTIFICATE OF APPROVAL IS VALID FOR THREE YEARS FROM THE ABOVE APPROVAL DATE PROVIDED THERE IS NO CHANGE IN THE EXPERIMENTAL PROCEDURES  Ministry of Social Services  Province of British Columbia &  Superintendent of Family and Child Service Parliament Buildings Victoria British ColumHi'j  173 Appendix B-2  July 6, 1992  Ms. Mary Ann Carter, M.A. Registered Psychologist Collingwood Professional Building 216 - 3540 West 41st Avenue Vancouver, British Columbia V6N 3E6  \  •  •'. '^^  '  :  Dear Ms. Carter: Thank you for your recent letter outlining your proposed research. I am pleased to provide you with my consent to my wards being interviewed and tested for your research purposes, with the following conditions applied: 1.  That confidentiality of my wards will be maintained and nothing will identify them personally, or as wards of the Superintendent.  2.  That interviewing will be conducted in a safe and supportive environment.  3.  That interviewers will obtain any necessary parental or guardian consents for those children who are in the care of the Superintendent by agreement only; i.e. those for whom the Superintendent has custody only, not guardianship.  I wish you every success with your research and your doctoral work and will be most interested to hear of the results. Yours sincerely.  aj^\  royce Rigaux, RSW Superintendent of Family and Child Service  %^  • S^ W • • tK**^  «'*  British Columbia  iviii n o i l y  uommuniiy ana i-amiiy Health 1515 Blanshard Street Victoria British Columbia V8W .^f^°-  KJi  Health  174 Appendix B-3  November 17, 1992 Mary Ann Carter, M.A, Registered Psychologist Department of Counselling Psychology Faculty of Education 5780 Toronto Road Vancouver, British Columbia V6T1L2 Dear Ms. Carter:  I have reviewed your research project. I am satisfied that this research will not compromise client confidentiality and will be undertaken with written parent/child consent. Client access to services will not be compromised by virtue of their agreement or refusal to participate in the research project. It is my understanding that this research has received approval from Fraser Valley/North Shore Regional Mental Health Office, and that this approval has being conveyed by the Regional Office to all Mental Health Centres involved in the study. In recognition of the time to be spent by our therapists in gathering the data, I would appreciate the results being made available to staff in some verbal (symposium) or written form, eventually. All the best in completion of your research project. Yours sincerely.  Clem Meunier, Ph.D. Manager Child and Youth Mental Health Sen/ices cc:  Mr. Paul Charron Director Adult Services  Mr. David Brown Child and Youth Coordinator Fraser Valley/North Shore Region  Ms. Christine Kline Regional Director Fraser Valley/North Shore Regional Office Attachment D;\DATA\MEUNIER\ROSEBUSH.Itr  175  APPENDIX C - THERAPIST LETTERS AND PERMISSION FORMS  T H E U N I V E R S I T Y OF B R I T I S H  COLUMBIA 176 Appendix C-1  Department of Counselling Psychology Faculty of Education 5780 Toronto Road Vancouver, B.C. Canada V6T 1L2 Tel: (604) 822-5259 Fax:(604)822-2328  Date  Name T i t l e or Firm S t r e e t Address C i t y , B.C. Postal Code Dear  ,  As a psychologist or child abuse therapist in the Province of British Columbia, you are aware of the unexpected ordeal encountered by children following their disclosure of child abuse. I am involved in a doctoral study program at the University of British Columbia under the supervision of Dr. John A.B. Allan, ''(822-4625). We are developing a standardized, nonverbal paradigm (Rosebush Picture Sort) to discriminate types of abused children through their selection of rosebush pictures. As a psychologist or therapist, your participation in this study is of extreme importance. All therapists in private practice and health related eitployroent who assess or treat six to twelve year old abused children, are being asked to participate in this research. The picture sort has been reviewed and approved by some psychologists and therapists in the province who treat abused children. Your cooperation in administering the picture sort, the Culture-Free Self-Esteem Inventory and completing the individual questionnaire will be greatly appreciated. The time required for each child will be one hour.' Results from the questionnaires and picture sort will become part of child abuse research at the University. You can be assured the child's and your responses on the forms will remain anonymous, i.e. identified only by a code number. You are welcome to withdraw from this research at any time and such a withdrawal will not jeopardize your involvement in future university research. I will call you within the next few days to find out if you are willing to participate in this research project. Yours truly.  Mary Ann Carter, M.A. Registered Psychologist  177 Appendix C-2  THERAPIST PARTICIPATION PERMISSION  Dear Psychologist / Therapist: Please sign and return this agreement with the enclosed materials.  I [do] [do notl agree to participate in the data collection for the Rosebush Picture Sort research. I will obtain written permission from parents or guardians for their children to participate in this research. This permission will be kept with the client's record in my place of eiiployment. The Rosebush Picture Sort and the Self-esteem Inventory will be individually administered in a safe and supportive environment to clients who are six to twelve years old and of normal intelligence. The coded materials will be returned to the investigators, maintaining the confidentiality of the client at all times.  Name  Date  THE U N I V E R S I T Y OF B R I T I S H  COLUMBIA 178 Appendix C-3  Department of Counselling Psychology Faculty of Education 5780 Toronto Road Vancouver, B.C. Canada V6T 1L2 Tel: (604) 822-5259 Fax:(604)822-2328  Dear Thank you for agreeing to participate in the research project we discussed on the telephone on . I have enclosed set (s) of materials. Each set contains a demographic sheet, a Rosebush Picture Protocol and a Culture-Free Self-esteem Inventory. The Picture Sort Cards are included and to be returned when you have completed your testing. Please sign and return the Therapist Participation sheet with the above forms. Two Participation Permission forms are included and to be signed by the parent or guardian before the test materials are administered to their child. Make sure the parent or guardian has signed one form for your record and kept one for their records. Directions for administering the Rosebush Picture Sort can be found inside the front cover of the Rosebush Picture Sort. If you wish further information on the use of Rosebush drawings in therapy, see Allan, J. (1988). Inscapes of the child's world. Dallas, Spring Publications. PLEASE ADMINISTER THE MATERIALS IN THE FOLLOWING ORDER: Once the child has established rapport with you and is relaxed, administer the Rosebush Picture Sort. Towards the end of the session or following the picture sort, administer the Culture-Free Self-Esteem Inventory. Culture-Free Self-Esteem Inventory Complete this yes/no form following the Rosebush Picture Sort. It will take no longer than 15 minutes. Read the directions out loud to the child. Read the sentences out loud to the child marking their yes/no answers on the protocol. Upon completion, please check the form to make sure all questions are answered. Note any unusual behaviors or comments the child makes. Make as few comments to the child as possible, other than reading the self-esteem statements. If the child asks you what a word means, provide a simple synonym. Please check each form to make sure all the sentences are answered.  179 Appendix C-3 Complete the demographic information sheet on each child. Please administer the materials to the child (ren) within the next two weeks and return the forms to me. If, for some unforeseen reason* you are no longer able to participate, I would appreciate you retuirning the blank protocols and Card sort to me. Tnahk you very much for participating in this study, Y<SUrs truly,  Mary Ann Carter, M.A. Registered Psychologist  THE U N I V E R S I T Y  OF B R I T I S H  COLUMBIA ISO Appendix C-4  Department of Counselling Psycbology Faculty of Education 5780 Toronto Road Vancouver, B.C. Canada V6T 1L2 Tel: (604) 822-5259 Fax:(604)822-2328  PARTICIPATION PERMISSION  Dear Parent / Guardian: We are from the University of British Columbia, Department of Counselling Psychology doing research which involves children sorting nonthreatening pictures and conpleting a paper/pencil questionnaire . The project is under the supervision of Professor John Allan (822-4625). We request your permission to allow your child to participate in the research. We will have no contact with your child or know their identity. The child's therapist will administer the questionnaire and pictures, recording the responses on a number coded page. With your permission, the therapist will complete one page of demographic information about your child including your child's age, gender, ethnic heritage, family grouping, geographic location, stage of treatment, and diagnostic considerations. The therapist will return the three numbered pages to us. Should this research project be acceptable to you, please circle the appropriate word below and return this form to your child's therapist. Please understand that denying consent will not jeopardize any treatment that your child receives. A copy of this consent form has been included for you to file with your own records. I [do] [do not] consent for my child to participate in this research.  Name  Date  181  APPENDIX D - INSTRUMENTATION USED IN STUDY  182 Please fill in all requested information and check choices. A p p e n d i x DATE  THERAPIST  male  female  highest degree earned  D-1  major  GEOGRAPHIC SITE Lower Mainland  Fraser Valley  Interior  Northern B.C.  Vane. Island  Other Specify  CHILD » male  female  age- yr./mo.  It sessions with therapist  Please check if child is in assessment or treatment and indicate what stage:  beginning rx  assessment  end rx  middle rx  ETHNIC HERITAGE Aboriginal (Native)  African  Asian  Caucasian  Indian  Inuit  CHILD RESIDENCE 2 parent  1 parent  foster par.  step par.  other (specify)  REASON FOR REFERRAL:  From the background and your clinical impression, please check the degree of severity of suspected abuse that you believe the child has experienced: [check only one on each row, i.e. for emotional, check either severe, moderate or mild.] THERAPIST'S OPINION REGARDING DEGREE OF SUSPECTED ABUSE ^EMOTIONAL  PHYSICAL  SEXUAL  (threat of death)  (broken bones)  (forced penetration)  (sporadic neglect)  (bruising)  (fondling)  (verbal put downs)  (consequential spanking)  (porno, pictures)  SEVERE  MODERATE  MILD  *see attached sheet for further description of abuse categories  CHECK TYPE OF ATTACHMENT CHILD HAS WITH PARENT/GOARDIAN (check only one)  •secure  anxious/ambivalent  anx ious/avo idant  anxious/disorganized  *see attached sheet for description of attachment categories  - Participation Permission Sheet signed by parent/guardian and on file. yes  no  COMMENTS REGARDING CHILD'S ATTITUDE BEFORE, DURING AND AFTER ROSEBUSH PICTURE SORT AND SELF-ESTEEM INVENTORY: e.g. sleepy, in angry mood, silly (use back of page i f you need more space).  DESCRIPTIONS OF ABUSE CATEGORIES  133 Appendix D-1  SPECTRUM OF EMOTIONALLY ABUSIVE BEHAVIORS MILD EMOTIONALLY ABUSIVE BEHAVIORS CRITICISM - The adult often negatively refers to the child when speaking to the child or to others about the child. SELF-CARE - The adult does not provide the child with clean, mended, appropriate seasonal clothing or self-care guidelines, e.g. routine toothbrushing, bathing. AVOIDANCE - The adult makes limited eye contact with the child and has little conversation with him or her. hWDERATE RMOTIONALLY ABUSIVE BEHAVIORS SPORADIC NEGLECT - The adult is unreliable in meeting the overt needs of the child, e.g. attends very few school functions, shows little interest in child's activities. INTIMIDATION - The adult threatens to "put the child in a foster home", "out on the street" etc. DENOUNCING - The adult constantly belittles any accomplishments of the child. SEVERE EMOTIONALLY ABUSIVE BEHAVIORS THREATS - The adult threatens to kill or maim the child. ABANDONMENT - The adult leaves the child unsupervised for long periods of tine. CASTIGATION - Any interaction between the adult and the child is peppered with strong, undermining statements from the adult that consistently attack the ego defenses of the child. SPECTRUM OF PHYSICALLY ABUSIVE BEHAVIORS MILD PHYSICALLY ABUSIVE BEHAVIORS SPANKING - The adult uses their hand or an object to hit the child with once or twice, on the child's clothed or unclothed body or buttocks. RESTRICTION - The adult physically prevents the child from moving by pinning their body or body parts to a surface or forcefully holding the the child, against their will. SLAPPING, PINCHING, EAR PULLING - The adult slaps the child on the face or head, pinches the child's body or limbs or leads the child by pulling the ear.  MODERATE PHYSICALLY ABUSIVE BEHAVIORS  184 Appendix D-1 BRUISING - The adult hits the child severely enough to leave bruise marks on the child's skin. HAIR LOSS - The adult pulls the child's hair with enough force to remove clumps of hair, possibly causing scalp abrasions. CONFINEMENT - The adult ties or chains the child to a stationary object. SEVERE PHYSICALLY ABUSIVE BEHAVIORS [the situation is severe enough to require hospitalization] BROKEN BONES - The adult intentionally  \  breaks bone(s) of the child.  BURNING - The adult intentionally burns the skin of the child by using cigarettes, hot liquids or heating elertents. STARVATION - The adult denies the child access to any nutrition.  SPECTRUM OF SEXUALLY ABUSIVE BEHAVIORS (Sgroi, Blick, Porter, 1985) MILD SEXUALLY ABUSIVE BEHAVIORS NUDITY OR DISROBING - The adult parades nude around the house in front of all or some of family members. The adult may disrobe in front of the child. OBSERVATION OF THE CHILD - The adult surreptitiously or overtly watches the child undress, bathe, excrete, urinate. KISSING - The adult kisses the child in a lingering and intimate way.  MODERATE  SEXUALLY ABUSIVE BEHAVIORS  GENITAL EXPOSURE - The adult exposes his or her genitals to the child. FONDLING - The adult fondles the child's breasts, abdomen, genital area, inner thighs or buttocks. MASTURBATION - The adult masturbates while the child observes ,the adult observes the child masturbating; the adult and child^observe each other while masturbating themselves or each other. SEVERE SEXUALLY ABUSIVE BEHAVIORS FELLATIO OR CUNNILINGUS - The adult has oral-genital contact with the child. DIGITAL PENETRATION OF THE ANUS OR THE VAGINA. be inserted. PENILE PENETRATION OF THE ANUS OR THE VAGINA.  Inanimate objects may also  ATTACHMENT CLASSIFICATIONS  185 Appendix D-1  SECDRE ATTACHMENT - The children show genuine pleasure in being with adults, making good eye contact, smiling and conversing. Children are able to express their positive and negative feelings openly with adults. Children seem secure in their own identity. These children use and respond to reason and are eager to learn. They seem to have a vital, lively energy.  ANXIOUS/ AMBIVftLHIT ATTACHMENT - Children have ambivalent feelings towards adults, sometimes showing sadness or fear of the adult and other times being joyous when they are together. They show a lack of trust in adults and will have little verbal communication with them. Although they seem to comply with the wishes of the adult and at time almost "take care of them," there is an underlying chronic anxiety that is evident in their play, drawings, peer relationships and verbal expression. They may seem eager to please, obedient and coitpliant to adult requests. They seem to day dream a lot and are shy and reserved.  ANXIOUS/ AVOIDANT ATTACHMENT - Children avoid their parent in a neutral and nonconfrontational manner by playing at friend's homes, staying late at school and spending a lot of time in their rooms or watching T.V.. Children do not play with their parent, show little shyness with strangers and show no distress when separated from their parent. They may have a seemingly excessive sense of self-reliance and put an enphasis on their independence of any need for help or support. Though these children seem detached and aloof, professionals often sense a core of anger.  ANXIOUS/ DISORGANIZED ATTACHMENT - Children show contradictory behavior patterns. They may appear to be in a good mood while playing but will compulsively and unexpectedly strike out and hit or make derogatory remarks. These children have undirected, misdirected, inconplete and interrupted movements and expressions during play. After long periods of contented play, they may burst into tears or express anger without any logical rationale. These children will not seek out parental comfort or protection when in fearful or threatening situations. They often appear dishevelled, incoherent at times, unkempt and unsocialized. They are restless and fidgety.  186 Appendix D-2  ROSEBUSH PICTURE SORT DO NOT GIVE THE CHILD ANY PROMPTING REGARDING THEIR CHOICE OF PICTURES At the beginning of the session, once rapport is established with the child, and you feel the child is relaxed, say, "NOW, I AM GOING TO SHOW YOU A PICTURE."  [show child the sample picture]  "BELOW THIS PICTURE ARE FOUR STATEMENTS:  [read statements out  loud] "THIS IS VERY MUCH LIKE ME, THIS IS LIKE ME, THIS IS NOT LIKE ME AND THIS IS NOT AT ALL LIKE ME.  WHICH ONE OF THESE  STATEMENTS MATCHES THE WAY YOU FEEL ABOUT THIS PICTURE?"  Once  child has made a selection, thank them and go to picture one.  If the child has difficulty making a choice on the sample, encourage their participation saying, "THINK OF YOURSELF AS A FLOWERING BUSH...(pause)...WHAT KIND OF PICTURE WOULD YOU BE LIKE?  HOW DO YOUR FEELINGS ABOUT YOUR PICTURE COMPARE WITH YOUR  FEELINGS ABOUT THIS ONE?" Follow with "HERE ARE OTHER PICTURES I WANT YOU TO LOOK AT VERY CAREFULLY AND DECIDE HOW YOU FEEL ABOUT EACH ONE.  TAKE YOUR TIME  AND MARK YOUR ANSWER THE SAME WAY YOU DID FOR THE FIRST PICTURE" [point to answered response]. Sit quietly while the child completes the task, noting the child's behavior and verbal comments. Once the child has finished, check the protocol to make sure all of the answers are completed. Thank them for completing the task. If the child absolutely refuses to cooperate, ask the child if he or she is willing to complete this activity in your next meeting. If the child continues to be uncooperative, note what the child finds objectionable about the task.  187 ROSEBUSH PICTURE SORT ANSWER SHEET Appendix D-2 ( s i z e 8 1/2 X 11) ROSSBOSH PICTORB SORT OPTIONS  So  C.  D.  u  A„  8.  C.  0.  2.  k.  3.  C.  0.  3.  A.  3.  C.  0.  4>  ft.  S.  c.  D.  5-  A-  So  c.  D.  6,  A,  B.  c.  0.  7c  A.  B.  c.  D.  8.  A.  B.  c.  D.  9,  A,  B.______^  c.  0.  10.  a.  B.  c.  D.  Il-  A.  B.  c<  D,  ls.  Ik,  B.  —  D.  Therapist Commancs. C3:  t  RPS SAMPLE PAGE ( a l l p i c t u r e p a g e s l a m i n a t e d , 8 1/2 x 11)  CO  1^ S. A.  THIS I S VERS MUCH I-tKE ME.  B.  SIfIS IS LIKS HS.  C.  THIS IS aOT LIKE MS.  D.  THIS I S HOT AT AX-L tIECE HE,  188  n R05EBU5H PICTURES  Appendix  ^"2  6.  8.  n.^ r>vij;) J"  i^^  •/;;'i^^f^l?P^*w -^Z  11  12.  Appenaix t»-d  Culture-Free Self-Esteem Inventories  lisli  Second Edition  FORMA \ \ ^  Date.  -Age.  ^  Date of Birth  School / Class Total  S  A  P  Directions Please mark each statement in the following way: If the statement describes how you usually feel, make a check mark (V) in the "yes" column. If the statement does not describe how you usually feel, make a check mark (V) in the "no" column. Check only one column (either yes or no) for each of the 60 statements. This is not a test, and there are no right or wrong answers.  ©1992,1981 by PRaED, Inc. 1 2 3 4 5 94 93 92 91  Additional copies of this form (#5187) are available from PRO-ED, 8700 Shoal Creek Boulevard, Austin, TX 78758 USA 512/451-3246  ^V-^H  1. I spend a lot of time daydreaming 2. Boys and girls like to play with me 3. I like to spend most of my time alone 4. I am satisfied with my school work 5. I have lots of fun with my mother 6. My parents never get angry at me 7. I wish I were younger 8. I have only a few friends 9. I usually quit when my school work is too hard 10. I have lots of fun with my father 11. I am happy most of the time 12. I am never shy 13. I have very little trust in myself 14. Most boys and girls play games better than I do 15. I like being a boy / I like being a girl 16. I am doing as well in school as I would like to 17. I have lots of fun with both of my parents 18. I usually fail when I try to do important things 19. I have never taken anything that did not belong to me 20. I often feel ashamed of myself 21. Boys and girls usually choose me to be the leader 22. I usually can take care of myself 23. I am a failure at school 24. I find it hard to make up my mind and stick to it 25. My parents make me feel that I am not good enough 26. I never get angry 27. I often feel that I am no good at all 28. I have many friends about my own age 29. Most boys and girls are smarter than I am 30. Most boys and girls are better than I am 31. My parents dislike me because I am not good enough 32. I like everyone I know 33. Children pick on me very often 34. I like to play with children younger than I am 35. I like to be called on by my teacher to answer questions 36. I would change many things about myself if I could 37. There are many times when I would like to run away from home 38. I am as happy as most boys and girls 39. I can do things as well as other boys and girls 40. I often feel like quitting school 41. I worry a lot 42. My parents understand how I feel 43. When I have something to say, I usually say it 44. I never worry about anything 45. I am as nice looking as most boys and girls 46. Other boys and girls are mean to me 47. I know myself very well.... 48. I am doing the best school work that I can 49. People can depend on me to keep my promises 50. My parents think I am a failure 51. I always tell the truth • 52. I need more friends 53. I always know what to say to people 54. My teacher feels that I am not good enough 55. My parents love me 56. I never do anything wrong 57. Most boys and girls are stronger than I am 58. 1 am proud of my school work 59. I often get upset at home 60. I am never unhappy  191 App^di3^,J)-3 D D D D D D D D D D O O D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D O D D D D D D D D D D D D D D D D a D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D • • • •  • %  192  APPENDIX E - ROSEBUSH PICTURE SORT STATISTICS  193 Appendix E-1  PHASE ONE MEAN & STANDARD DEVIATION SCORES* SEXUAL ABUSE/OTHER CHILDREN IN TREATMENT  SEXUAL ABUSE FEMALE/MALE  PICTURE  GROUP  MEAN  S.D.  GROUP  MEAN  S,D,  1  S,A. Other  2.83 2.69  1.05 .99  Female Male  2.88 2.70  1,05 1,06  2  S.A. Other  2.52 2.44  .92 1.06  Female Male  2.35 2.87  ,87 ,92  3  S.A. Other  2.70 2.56  1.10 1.08  Female Male  2.58 2.78  1,07 1.17  4  S.A. Other  2.41 2.29  1.11 1.06  Female Male  2,40 2.48  1.16 1.08  5  S.A. Other  2.33 2.36  1.17 1.19  Female Male  2.35 2.44  1.21 1,12  6  S.A. Other  1.97 2.33  1.04 1.09  Female Male  1.79 2.35  .91 1.19  7  S.A. Other  2.64 2.73  1.09 1.03  Female Male  2.51 2.74  1.06 1.18  8  S.A. Other  2.77 3.09  1.24 .97  Female Male  2.56 3,09  1.28 1.08  9  S.A. Other  2.58 2.60  1.15 1.18  Female Male  2.61 2.44  1,16 1.20  10  S.A. Other  2.18 2.31  1.04 1.15  Female Male  2.02 2,44  1.01 1.08  other  2.09 1,87  1.11 1.01  Female Male  1,95 2,22  1.05 1.17  S.A. Other  2.18 2.53  1.05 1.12  Female Male  1,93 2,70  .96 1,11  11 12 *  Score 1 - "very much like me" , 2 - "like me", 3 - "not like me", 4 - "not at all like me'  194 Appendix E-2  PHASE TWO RPS MEAN AND STANDARD DEVIATION SCORES* TREATMENT STAGES BEGINNING  MID  GROUP  MEAN  S.D.  MEAN  S.D.  MEAN  S.D.  1  1 2 3  2.85 2.46 2.73  .99 1.04 .79  2.80 2.73 2.73  1.11 1.03 .79  3.00 2.92 2.73  1.0 .79 .79  2  1 2 3  2.92 2.46 1.91  .86 1.04 1.04  2.38 2.36 1.91  .90 1.18 1.04  2.62 2.42 1.91  .96 .90 1.04  3  1 2 3  2.69 2.91 3.0  1.03 .94 1.0  2.70 2.41 3.0  1.11 1.14 1.0  2.46 2.75 3.0  1.20 1.06 1.0  4  1 2 3  2.39 2.09 2.27  1.26 .94 1.01  2.35 2.41 2.27  1.10 1.05 1.01  2.54 2.17 2.27  1,13 1.12 1,01  5  1 2 3  2.31 2.18 2.55  1.17 1.08 .82  2.48 2.36 2.55  1.18 1.22 .82  2.15 2.25 2.55  1,28 1.29 .82  6  1 2 3  2.54 2.0 2.91  1.13 1.0 1.04  1.88 2.32 2.91  1.02 1.21 1,04  2.0 2.5 2.91  1.08 .91 1,04  7  1 2 3  2.62 2.36 2.64  1.26 1.03 1.21  2.68 3.09 2.64  1.10 .97 1,21  2.31 2.67 2.64  .95 .99 1.21  8  1 2 3  2.77 2.82 3.36  1.36 1.08 .67  2.50 3.14 3.36  1.22 .99 .67  3.08 3.42 3.36  1.12 .79 .67  9  1 2 3  2.15 2.91 2.0  1.07 1.04 .63  2.70 2.46 2.0  1.16 1.22 .63  2.46 2.75 2.0  1.27 1.14 .63  PICTURE  END  Score 1 = "very much like me", 2 = "like me", 3 = "not like me", 4="not at all like me".  Append:i x E-2  10  11  12  1 2 3  2.23 1.73 2.64  1 2 3  2.15 1.73 1.73  1.21  1 2 3  2.62 2.36 2.82  1.33 1.12  .93  1.10 .81 .79 .79  .98  2.13 2.55 2.64  1.04 1.18  1.95 1.96 1.73  1.09 1.13  2.18 2.46 2.82  1.04 1.14  .81  .79  .98  2.31 2.50 2.64  1.25  2.23 2.08 1.73  1.01 1.17  1.92 2.67 2.82  1.07  .91 .81  .79 .76 .98  196 Appendix E-3 Total  BEGIN. GROUP Picture 14 4 5  4 3 4  Total 3 13 Chi-Square, 6 d f = 3 . 2 0  11 P  1 2 0  1 2 3  4 2 2  13 11 11  8 35 =0.78  •Picture 25 4 3  4 3 2  Total 7 12 Chi-Square, 6 d f = 8 . 2 3  9 P  0 2 5  1 2 3  4 2 1  13 11 11  7 35 =0.22  -Picture 33 5 2  5 2 4  Total 3 10 Chi-Square, 6 d f = 4 . 6 6  11 P  1 2 3  2 0 1  -Picture 41 2 3  1 2 3  4 5 4  Total 12 6 Chi-Square, 6 df = 4.78  13 P  5 4 3  3 4 4  13 11 11  11 35 =0.59 3 0 1  13 11 11  4 35 0.57  -Picture 1 2 3  5 2 7  1 4 2  Total 8 14 Chi-Square, 6 ^ = 9 . 7 7  7 P  -Picture 61 2 3  4 4 0  3 4 3  4 2 3  Total 8 10 Chi-Square, 6 d f = 4 . 3 4  9 P  3 4 1  3 1 2  13 11 11  6 35 =0.13 3 1 4  13 11 11  8 35 =0.63  197 Appendix E-3 BEGIN. GROUP  1  2  3  4  Total  4 2 2  1 5 4  4 2 1  4 2 4  13 11 11  Total 8 10 Chi-Square, 6 df = 5.98  7 P  Picture1  7  -•  1 2 3  1  fi—  .  1 2 3  4 2 0  1 1 1  2 5 5  3  12 P  5 3 7  2 3 2  Total 7 15 Chi-Square, 6 df = 7.94  7 P  Total 6 Chi-Square, 6 df = 6.12 Picture,  35 10 = 0.43 6 3 5  13 11 11  14 35 = 0.41  Q  4 1 2  1 2 3  2 4 0  13 11 11  35 6 = 0.24  Picture. 1 D —. 4 7 0  2 1 6  7 2 3  Total 11 9 Chi-Square, 6 df = 17.11  12 P  1 2 3  1  13 11 11  35 3 = 0.0089  11-.  1 2 3  4 4 4  1 2 2  Total 15 12 Chi-Square, 6 df = 5.87  5 P  Picture1 1 2 3  0 1 2  5 5 5  3 0 0  13 11 11  35 3 = 0.44  1 9-.  4 3 1  Total 8 Chi-Square, 6 df = 3.77  2 3 3  2 3 4  8  9 P  5 2 3  13 11 11  35 10 = 0. 71  198 Appendix E-4 MID GROUP •Picture 11 2 3  Total 8 6 5  11 7 4  Total 10 19 Chi-Square, 6 d f = 5 . 2 5  22 P  -Picture 21 2 3  7 3 0  15 8 3  14 2 2  Total 18 26 Chi-Square, 6 df = 10.28  18 P  -Picture 31 2 3  7 6 5  8 6 2  12 5 4  Total 15 16 Chi-Square, 6 d f = 2 . 6 6  21 P  -Picture 41 2 3  8 6 1  12 7 3  9 6 4  Total 19 22 Chi-Square, 6 d f = 1 . 4 2  19 P  -Picture 51 2 3  11 5 3  10 6 7  8 3 2  Total 18 23 Chi-Square, 6 d f = 8 . 2 0  13 P  ---Picture 61 2 3  11 7 0  19 7 1  11 7 3  6 2 3  Total 27 21 Chi-Square, 6 d f = 9 . 5 8  11 P  14 6 2  40 22 11  22 73 =0.51 4 6 1  40 22 11  11 73 =0.11 12 5 4  40 22 11  21 73 =0.85 8 4 1  40 22 11  13 73 =0.96 11 6 2  40 22 11  19 73 =0.22 4 6 4  40 22 11  14 73 =0.14  199 Appendix E-4 MID GROUP  TOTAL  Picture, 7__. 6 1 2 1 3 2  14 6 4  7 5 1  24 Total 9 Chi--Square, 6 df = 3.34  13 P  Pictv , ire 1 2 3  40 22 11  27 73 = 0.77  Q  12 1 0  5 6 1  11 4 5  13  12  20 P  10 4 7  8 5 2  17 21 Total Chi--Square, 6 df = 10.64  15 P  Total  chi--Square, 6 df = 13.42 Pictv 1 2 3  13 10 4  12 11 5  40 22 11  73 28 = 0.037  , Q_ .  8 7 2  Picti 1 ire 1 2 3  14 6 0  40 22 11  20 73 = 0. 10  1n .  14 5 0  12 7 6  9 3 3  25 Total 19 Chi--Square, 6 df = 9.36  15 P  5 7 2  40 22 11  14 73 = 0.15  1 11—.  12 4 4  4 4 2  34 Total 20 Chi--Square, 6 df = 3.69  10 P  1 2 3  Picttixr€L 1 2 3  18 11 5  6 3 0  40 22 11  9 73 = 3.69  1 9_.  11 8 3  9 3 4  Total 19 22 Chi--Square, 6 df = 7.60  16 P  13 5 1  7 6 3  40 22 11  16 73 = 0.56  200 Appendix E-5 END GROUP  3  4  Total  3 1 5  4 8 4  5 2 2  13 12 11  9  16 P  6 7 3  3 2 2  Total 7 16 Chi-Square, 6 d f = 7 . 6 2  7 P  Picture 1 1 2 3  1 1 1 0  Total 2 Chi-Square, 6 d f = 7 . 6 0 Picture 2 1 2 3  Picture 3 1 2 3  1 1 5  4 2 1  2  2 2 2  4 5 4  6  13 P  6 4 3  1 2 4  Total 9 13 Chi-Square, 6 d f = 5 . 5 7  7 P  Total 7 Chi-Square, 6 d f = 2 . 1 7 Picture 4 1 2 3  Picture 5 1 2 3  2 4 3  2 2 7  2 2 2  Total 11 11 Chi-Square, 6 df = 10.71  6 P  Picture 6 1 2 3  6 5 0  5 1 1  5 6 3  1 3 3  Total 7 14 Chi-Square, 6 d f = 7 . 4 4  7 P  9 36 =0.27 3 2 1  13 12 11  6 36 =0.27 3 3 4  13 12 11  10 36 =0.90 4 2 1  13 12 11  7 36 =0.47 3 3 2  13 12 11  8 36 = 0.098 2 2 4  13 12 11  8 36 =0.28  201 Appendix E-5 END GROUP  TOTAL  Picture 7 — • 3 1 2 1 3 2 .  4 5 4  5 3 1  Total 6 13 Chi--Square, 6 df = 5.36  9 P  Pictvir€J 1 2 3  1 3 4  13 12 11  8 36 = 0.50  Q  2 0 0  1 2 1  4 3 5  4  12 P  3 3 7  2 3 2  Total 8 13 Chi--Square, 6 df = 7.85  7 P  2 Total Chi--Square, 6 df = 5.06 Q_  6 7 5  13 12 11  18 36 = 0.54  _.  1 2 3  4 2 2  4 4 0  13 12 11  8 36 = 0,25  in —.  1 2 3  2 3 6  3 6 3  7 11 Total Chi--Square, 6 df = 10.14  12 P  5 2 0  3 1 2  13 12 11  36 6 = 0.12  T 1  6 3 4  2 2 2  Total 13 13 Chi--Square, 6 df = 3.55  6 P  1 2 3  1 2 3  3 5 5  4 2 1  Chi-Square, 6 df = 6.32  6 3 3  3 4 4  2 2 0  13 12 11  4 36 = 0.74 0 3 3 = 0.39  13 12 11  202  APPENDIX F - CHILD PARTICIPANT COMMENTS  203 Appendix F  CHILDREN'S COMMENTS ABOUT PICTURES General Comments Group 3 - No Abuse, No Treatment 10.3 female - "Some I knew it was like me...Lots of background & grass, not just one thing. I like having things around me and having lots of colour." Picture 1 Group 3 - No Abuse, No Treatment 11.3 male - "Some looked odd to me, something not quite right [#1,6,12], something missing," Beginning-Treatment Group 1 - Sexual Abuse 9.8 male -"It's crammed up in a bunch of trees." Group 2 - Other Treatment 11.11 female - "I really like trees." 10.7 female - "Too much forest." 11.10 male - "I'm always caught in the middle of things." 6.10 male - "I don't like grass." 9.11 male - "I like to play in forests." 7.7 male - "I couldn't go traveling in the forest." Mid-Treatment Group 1 - Sexual Abuse 8.2 female - "Too colourful." 6.1 female - "That's all alone, that's not like me. " 8.6 male - "Sometimes I feel like the only one among those who are different from me." Group 2 - Other Treatment 8.0 female - "I like trees and grass." 9.6 male - "Nobody to talk to." 10.2 female - "It's all trees. I'd be very lonely." End-Treatment Group 2 - Other Treatment 7.8 female - "Bush isn't fluffy." 11.7 male - "Looks very crowded in the picture." 11.11 male - "The trees stand for everyone and I'm in the middle."  204 Appendix F  10.4 male - "Makes you feel crowded in by all the trees." Picture 2 Beginning-Treatment Group 1 - Sexual Abuse 9.7 female - "Birds." Group 2 - Other Treatment 11.1 female - "It's bare., should have lots of pretty things around it." 10.7 female - "I like to play outside and watch T.V. in the house." 6.10 male - "I don't like smoke." 7.6 male - "The tree looks good...a good drawing." 9.11 male - "I'd like to live in a big house with a big yard." Mid-Treatment Group 1 - Sexual Abuse 6.1 female - "I like these birds." 7.9 female - "I like rain because it makes everything grow." 7.11 male - "This has a fence around me." It's like me." Group 2 - Other Treatment 8.0 female - "I like house, trees, birds, sunshine and flowers." 9.6 male - "Nobody to talk to." End-Treatment Group 2 - Other Treatment 7.8 female - "It's a skinny bush." 11.11 male - "The roses are for girls, not boys." 9.7 male - "There's a house." 10.4 male - "Nice feeling that you're finally home. That it's a nice day." Picture 3 Beginning-Treatment Group 2 - Other Treatment 11.10 male - "I'm always getting cut off from things I like." 11.1 female - "Seems so plain. House seems  205 Appendix F  disfigured. Seems sad." 10.7 female - "Don't know why [I don't like it]." 6.10 male - "I don't like watering plants." 7.6 male - "I wouldn't go outside near any flowers." 9.11 male - "I don't like to garden." 7.11 male - "No way." Mid-Treatment Group 1 - Sexual Abuse 8.2 female - "Too black." 7.11 male - "That's a lawn mower." Group 2 - Other Treatment 9.6 male - "Someone always looks after me." End-Treatment Group 2 - Other Treatment 7.8 female - "Bush has no leaves." 11.7 male - "Looks like someone's crying." 10.4 male - "Everything is white. It feels cold, like there's nothing in it." 11.11 male - "I don't live in a house like that." Picture 4 Group 3 - No Abuse, No Treatment 11.11 male - "Half cold/half warm. raining out.  Don't like  Beginning-Treatment Group 2 - Other Treatment 11.1 female - "I like rainbows, I like the colours." 10.7 female - "I like rainbows and I try to find the charm at the end." 6.10 male - "I like the rainbow." 9.11 male - "I don't like rainbows or the rain." 7.11 male - "The sides kind of flat...sort of." Mid-Treatment Group 1 - Sexual Abuse 7.5 female - "Nice rainbow." 7.9 female - "I like this picture." Group 2 - Other Treatment 8.0 female - "It gives feeling."  you  a  heart  warming  206 Appendix F  10.2 female - "This would grow lots because of purple rain. I really like this." 9.9 male - "Like me but I don't like it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "Volcano tree & storm are like me." Group 2 - Other Treatment 7.8 female - "Rainbows, nice but the bush doesn't have red roses." 11.11 male - "I like the water." 10.4 male - "It'snowing. I like rainbows... kind of Christmassy and I like Christmas." Picture 5 Beginning-Treatment Group 1 - Sexual Abuse 9-7 female - "Bird." 9-8 male - "I like birds." Group 2 - Other Treatment 11.10 male - "I'm not use to birds getting food for their young. I have to get everything for myself-" 11.1 female - "Like colours, bugs." 10.7 female - "I like the bugs and birds." 6.10 male - "I like the bugs." 9.11 male - "I like collecting bugs & like to watch birds." 7.6 male - "I don't like looking at insects." 7.4 male - "If I were a flowering bush." Mid-Treatment Group 2 - Other Treatment 8.0 female - "I like birds, ants, grass and trees." 9.9 male - "Like me but I don't like it." 10.2 female - "I really love it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "Like to collect bugs." 10.11 female - "Is that a hand?" Group 2 - Other Treatment 7.8 female - "I don't like bugs." 11.11 male - "Because of bugs and stuff."  207 Appendix F  9.7 male - "I like insects so I can bug my mom." 10.4 male - "It has birds and animals- I like the forest but not too many trees." Picture 6 Beginning-Treatment Group 1 - Sexual Abuse 9.7 female - "Flowers." Group 2 - Other Treatment 11.1 female - "They should have more colour." 10.7 female - "I like roses." 6.10 male - "I like roses and the birds and sky." 7.6 male - "The blue sky is sort of blackish." 9.11 male - "I don't like rose buds because when you touch them it hurts." Mid-Treatment Group 1 - Sexual Abuse 6.1 female - "When I was a baby." Group 2 - Other Treatment 8.3 female - "Lots of friends...no." End-Treatment Group 1 - Sexual Abuse 6.1 male - "Black clouds and black birds." Group 2 - Other Treatment 7.8 female - "The roses are red, like my rose bush would be." 11.11 male - "Roses, I don't like flowers." 10.4 male - "It's got the sun and flowers." Picture 7 Beginning-Treatment Group 2 - Other Treatment 11.10 male - "The leaves falling remind me of me because I'm so clumsy and break things a lot." 11.1 female - "Like how tree stands out." 10.7 female - "I don't like Fall because the trees are so bare." 6.10 male - "I like the colour black." 9.11 male - "I like climbing trees and I like windy days."  208 Appendix F  7.6 male - "The trees looks kinda weird." Mid-Treatment Group 2 - Other Treatment 8.3 female - "What's this one...losing things? Somebody always there." 10.2 female - "I would be right next to the trees. It's a beautiful one. I really love it." End-Treatment Group 1 - Sexual Abuse 6-1 male - "I like black and I like green leaves and green sun and blue clouds." Group 2 - Other Treatment 7.8 female - "I don't like trees without leaves." 11.11 male - "I like Fall." 9.7 male - "I like the fence so I can punch it." 10.4 male - "The leaves are all falling off. Makes you feel fenced in. Spikes make you feel unwanted-.like a grave yard." Picture 8 Group 3 - No Abuse, No Treatment 8.4 female - " One with rose was not like me. Sunny faces and people smiling not like me." Beginning Treatment Group 1 - Sexual Abuse 9.7 female - "Flowers." Group 2 - Other Treatment 11.10 male - "I do have someone to talk to when I want to." 11.1 female - "Would rather have lots of flowers & trees and not the open space." 6.10 male - "It's drawn nicely." 7.6 male - "I wouldn't want to be a flower because if anybody picks me then I'd be ...it would hurt." 9.11 male - "I don't like flowers." Mid-Treatment Group 1 - Sexual Abuse 7.11 male - "This is like me oh no, it's a girl, it's not at all like me."  209 Appendix F  Group 2 - Other Treatment 8.0 female - "It has clouds, blue sky and a bit of sunshine." 8.3 female - "All alone." 9.9 male - "Like me but I don't like it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "It's a happy face, flower and it's sunny." Group 2 - Other Treatment 7.8 female - "The stem is blue." 11.11 male - "Whatever that thing is (pointing to flower) I'm not that small." 9.7 male - "I've never seen a face on a rosebush." 10.4 male - "It's like you're caved in...the flower has no others to keep it company." Picture 9 Beginning-Treatment Group 1 - Sexual Abuse 9.7 female - "Cat." 9.8 male - "I don't go in places where it says 'Keep Out.'" Group 2 - Other Treatment 11.10 male - "It reminds me of my Mom. She doesn't put up 'Keep Out' signs when she has a problem. She talks to me." 11.1 female - "Looks harsh. Looks mean." 10.7 female - "'Keep Out' signs scare me because I think of there being weird people out there." 6.10 male - "I don't like cats. I can't see any words on the bottom sign." 7.6 male - "The people should be allowed to go in there if it isn't dangerous." 9.11 male - "I like cats." Mid-Treatment Group 1 - Sexual Abuse 8.4 female - "Don't like 'Keep Out' building." Group 2 - Other Treatment 8.0 female - "I like cats, flowers, sunshine and I also like the whole picture." 8.3 female - "Won't let me be in the group."  210 Appendix F  10.2 female - "That's pretty." 9.9 male - "Like me but I don't like it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "I hate 'Keep Out' because I always go in." Group 2 - Other Treatment 7.8 female - "I don't like to see 'Keep Out' signs. 11.11 male - "I like this because of the water and the cat." 9.7 male - "I would 'Keep Out' if I saw a sign like that." 10.4 male - "Makes you feel like you're not wanted." Picture 10 Beginning-Treatment Group 1 - Sexual Abuse 9.8 male - "It's too dark." Group 2 - Other Treatment 11.1 female - "Like prickles because they stand out. Sun and house also stand out." 6.10 male - "I like prickle bushes because they grow raspberries." 9.11 male - "I like this because I play with my friends alot at night." 10.7 female - "I like to play in the dark." 7.6 male - "I don't feel like looking at dead flowers." Mid-Treatment Group 2 - Other Treatment 8.3 female - "All alone at night." 10.2 female - "Everyone would say how beautiful I was right next to the house." 9.9 male - "Like me but I don't like it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "I like dark." Group 2 - Other Treatment 11.11 male - "Because of darkness I like black."  211 Appendix F  9.7 male - "I can draw a house just like that." 10.11 male - "Too dark, don't feel welcome. Looks like a mean plant with mouths on it." 7.8 female - "The house is flying." Picture 11 Beginning-Treatment Group 1 - Sexual Abuse 9.8 male - "I don't go outside much." 9.7 female - "Little girl." Group 2 - Other Treatment 11.10 male - "This person is playing with an animal and I always play with my cat." 11.1 female - "Like smiling sun. If I was one of the flowers, it would feel good." 9.11 male - "I like to climb trees and I like pets." 6.10 male - "I like rats because it's raining." 10.7 female - "I don't like picking our fences. If you touch a cat you may get a disease." 7.6 male - "The sun is smiling. The squirrel is going to run up the tree." Mid-Treatment Group 2 - Other 8.0 female 8.3 female 9.9 male -  Treatment - "The tiny cat is nice." - "People looking at other things." "Like me but I don't like it."  End-Treatment Group 1 - Sexual Abuse 6.1 male - "It's light and that's not like me." Group 2 - Other Treatment 7.8 female - "The sun's up, it's shiny, there's a nice blue sky." 10.11 male - "It's a nice day...it's just started." 9.7 male - "I have a cat." 11.11 male - "Everything looks happy." Picture 12 Beginning-Treatment Group 2 - Other Treatment -  212 Appendix F  11.10 male - "I don't like flowers." 11.1 female - "I feel like there should be something around me." 6.10 male - "I like the roses." 9.11 male - "I don't like flowers." 10.7 female - "I like being small. If there's a big bush, I like to crawl under it." 7.6 male - "The sun isn't whole." 7.11 male - "I guess" [chose positive response]. Mid-Treatment Group 2 - Other Treatment 8.0 female - "I just like flowers." 8.3 female - "No one to talk to." 10.2 female - "The most like me. I really love it." End-Treatment Group 1 - Sexual Abuse 6.1 male - "It's very little (the flower) and there's a yellow, good sun." Group 2 - Other Treatment 11.11 male - "Flower!" 9.7 male - "I don't like roses." 10.11 male - "It's like somebody talking out loud to a bunch of people. I wish I had a friend." 7.8 female - "Makes me feel good."  213  APPENDIX G - THERAPIST COMMENTS  214 Appendix G-1  THERAPIST COMMENTS Note: 2 therapists sent rosebush pictures that were drawn by the child following the research exercise. BEGINNING TREATMENT Group 1 - SEXUAL ABUSE 7.10 male - "Child responded very well to RPSidentified with many of the pictures- [cfsei-2] child stated it took too long, he didn't enjoy it as much as the r.b., perhaps due to the fact that this child is uncomfortable with verbal communication, responds more to art therapy strategies- child participated seriously- did not have adverse behaviour afterwards." 11.5 male - "Looked through pictures very quickly, almost without thought." 11.7 female - "Child seemed to answer question with thoughtfulness & seriousness - no evidence of negative behaviours." 7.7 male - "Difficulty focusing at first, agitated, wanting to move about physically." 10.9 female - "To most items with the exception of #8, very quick response." 11.2 female - "Before very comfortable, interested & curious - during very much liked & responded to RPS." Group 2 OTHER TREATMENT 7.11 female - "The child seemed to find many of the pictures looking happy & sad. 'this looks happy so this is like me' for at least 2 of her pictures." 7.11 male - "7,8,9 done fast- #10 short thoughtful pause." 7.11 male - "Attitude was 'that's over, it was easy.'" 10.6 female - "#6 slow to choose (positive), #7 fast to choose (negative)." 9.1 female - "Eager to please, fluctuations of certainty, was impulsive."  215 Appendix G-2  11.6 female - "Very thoughtful of #4." MID TREATMENT Group 1 - SEXUAL ABUSE 6.4 female - "She is looking to me for nurturing while she's in foster home - quite sure none of pictures were like her-" 8.9 female - "She often responded to the overall feel of the picture - that's a pretty one - that's like me a lot, or I don't like that one, that's not like me." (#1 & #5) 7.2 male - "Told therapist it was fun - pictures are nice, nice how they blend in the colours & they draw really nice, who drew the book?" 7.5 male - "Seemed to be balancing choices rather than choosing based on pictures." 8.6 male - "Very reflective." 7.3 female - "The basis of choice did not give me confidence - it seemed related to the 1's and 4's only and was made quickly." 9.4 male - "111 at ease." 6.4 female - "Liked rosebush - it was fun, very reflective & cooperative." 11.4 female - "Before RPS relaxed, cooperative, during involved, quick, definite responses." 10.1 female - "Child was somewhat reluctant to do RPS. stopped at #6, encouraged to go to #10, then said enough, however did go on to complete #11 & #12. Reluctant behavior is an overall pattern with the child but is changing over time." 8.9 female - "Appeared happy, cooperative, always eager to please, accommodate." 7.11 female - "Cooperated throughout." 8.1 male - "Hesitant to cooperate - after, was proud.  216 Appendix G-2  during sleepy & sick." 10.1 female - "Approached tests in a positive manner, enjoyed self during testing, was content & settled afterward." 7.2 male - "The child was very cooperative, the parents had spoken to him prior to him doing the exercise with me. He thought it was a test, so I explained that it wasn't, we were doing it to help other children by collecting data information from 'our' kids. He seemed relaxed & eager." 7.5 male - "This child approached task in a positive cooperative manner. My impression was that choices were made to balance the RPS answers rather than on reaction to pictures. However, there was no resistance to this task." 10.1 male - "Very keen & interested in testing. His attitude was positive. His approach was lively & confident. Afterwards painted picture of house." 7.3 female - "Child was nonchalant about doing RPS. Not visibly affected." 9.6 male - "This boy can hardly read & spent most of time looking at the words vs. pictures. Often I wondered if he knew what was written there. He is quite non-verbal." 7,3 female - "Child excited about completing research. In good happy mood. Found RPS too abstract for her." 7.3 female - "This child was willing to cooperate." 6.0 female - "This child was willing to cooperate." Group 2 - OTHER TREATMENT 11.10 female - "Mild tension, unsure worried before about ability to be successful - after relaxes. Pictures seemed very affirming for her, relaxed smiling." 11.7 male - "Cooperative seemed somewhat insulted by the RPS." 8.3 female - "Attitude o.k. Let's get this over with so  217 Appendix G-3  I can get to the good modelling clay [would much prefer to be a boy]." 8.7 male - "Eager to do activity." 9.9 male - "Liked happy pictures. He confused some of pictures he didn't like with 'not Like me.'" 8.1 male - "Rather guarded approach at first, but became emotionally engaged fairly quickly. Thoughtful & talkative immediately afterwards. Aggressive play followed testing." 8.8 female - "Anxious to perform task. Cheery disposition. Confused about some of pictures. Had trouble understanding them." 9.0 female - "Child cooperative throughout process. She completed all pictures in approx. 5 min." 11.7 male - "Child thought the pictures were 'effeminate' except for the last two. Said 'people will think I am a fag.'" 9.3 male - "Sometimes seemed as though choices were varied just for the sake of variation." 9.6 male - "He commented on most as he first saw them 'What's this supposed to mean?' By 1/2 way he would just make the statement." 10.2 female - "When finished inventory chose to draw her own picture of a rosebush." 6.6 male - "Spent little time on pictures." 9.0 female - "She really looked at the pictures & made very thoughtful selections. Was very enthusiastic about doing this." 11.4 male - "Thoughtful." END TREATMENT Group 1 - SEXUAL ABUSE 8.4 female - "Very clear on choices. Took time & pronounced answer."  218 Appendix G-3  8.11 female -  "Said she did not like the 'm' birds."  Group 2 - OTHER TREATMENT 11.3 female - "Very definite on Rosebush. Trouble yes/no on self esteem." 7.0 female - "Thoughtful & pensive. Discrepancies reflect conflicts in her situation." 9.11 female - "Bright, spontaneous, became quite serious & thoughtful." 11.1 male - "Initially quite ambivalent in doing test. Once he realized that it was not threatening to his 'artistic' abilities i.e. he did not have to do any drawing himself, he relaxed a lot more." 9-7 male - "Some confusion initially re: what part of picture he was to focus on but once given flowering bush example, he responded easily."  219 Appendix G-4  CFSEI THERAPIST COMMENTS BEGINNING TREATMENT Group 1 - SEXUAL ABUSE 7.0 male - "He didn't seem to enjoy it as much as Rosebush. He stated it took too long. Perhaps due to the fact that this child is uncomfortable with verbal commxinication, responds more to art therapy strategies." 7.1 female - "I don't believe she understood the double negative kinds of questions." 11.9 male - "He was able to concentrate & challenged the redundancy of some questions." 11.7 female - "Child seemed to answer questions with thoughtfulness and seriousness. No evidence of negative or reactive behaviours." 7.7 male - "Approached Inventory guarded. After testing, made up story about spending last evening with girl friend while her mother died & mother was buried Scime evening. Agitated after testing as well." Group 2 - OTHER TREATMENT 9.1 female - "Quite anxious, fluctuations of certainty versus impulsive, no thought put into response or maybe covering over incomprehension of questions. Painted picture after of happy family, but excluded her sister. Noted to be guarded about sharing angry feelings in family. Started to break out of this through art." MID TREATMENT Group 1 - SEXUAL ABUSE 8.8 female- "She commented that she wished she was a newborn again." 7.5 male - "His approach to the Self-Esteem Inventory was cooperative and thoughtful," 10.1 male - "Experienced anxiety over some of the items in the Self-Esteem Inventory. Would lay his head on  220 Appendix G-4  the table & think. The questions involving father were anxiety provoking for him but he worked it through. After the test, Steven readily engaged himself in physical play and painted a picture of a house." 7.3 female - "Thoughtful about questions on selfesteem. " 9.6 male - "Was far too long for him. At the end he said, 'Thank God.' Several I wondered if he comprehended. At about #50 he began playing with a kaleidoscope." 9.4 female - "A little ill at ease but answered readily." 11.4 female - "Somewhat guarded affect/tone flat. Except for #37 there are many times when I would like to run away from home... resounding YES- after exercise she went to clay. Built a large volcano. Talked about feelings of anger toward peers. Possible sexual overtones during amplification of the volcano symbol," 8.8 male - "Thoughtful about self-esteem but did not understand all the words." 10.1 female - "Seemed to like doing the self-esteem. Appeared to think about items. Asked for repetition & I think answered to the best of her ability." 8.1 male -  "Cooperative, proud."  11.2 female - "Was concerned with the purpose of the questions after the test." Group 2 - OTHER TREATMENT 9.9 male - "Troubled with questions of self identity." 8.7 male - "Liked questions. A little silly and hyper became serious. At one point said 'it was good to be able to get that out of me.'" 9.3 male - "I would question some of his answers to the Inventory, as he has developed many defense in response to experience." 11.7 male - "Seemed self-aware during Inventory."  221 Appendix G-4  11.6 female - "Responded with quite a bit more confidence than she anticipated. Had problems of acknowledging her sadness & disappointment but the questionnaire gave opportunity to reflect & recognize changes she's making, sense of hope." 7.3 female - "She seemed very self-aware during the Inventory." END TREATMENT Group 1 SEXUAL ABUSE 8.2 female - "Gave thoughtful answers. Did not become bored or withdrawn with the exception of showing slight discomfort when a word was used that she did not understand, i.e., ashamed., when the word was explained to her she regained her comfort level." 7.10 female - "Seemed a little uncomfortable at first but within minutes was telling me about herself without prompting." 9.10 female - "Became very engaged while doing the self-esteem. Her feelings about friends led into a new area of discussion, which was extremely open for this child." 6.1 male him."  "Self-esteem questions too difficult for  Group 2 OTHER TREATMENT 11.3 female - "Unable to answer many of the absolutes as yes or no." 9.6 female - "Became quite serious & thoughtful during process & occasionally asked for clarification. The test gave us opportunity to talk more about how she's changed over the 3 yrs. I've known her. This is her second time with me at M.H. and a good experience." 9.8 male - "Calm & assured in responding to self-esteem Inventory."  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053989/manifest

Comment

Related Items